The AM Forum

NOTE:  To maintain continuity, here is the beginning thread of this project:

http://amfone.net/Amforum/index.php?topic=35443.0




Hola,

I'm wondering how much power can be expected out of a 6146B in class C when it is piss beat?

The data sheet shows with 600 volts, it will do 62 watts carrier out.  I assume they mean 248 watts peak with 100% modulation.

http://frank.pocnet.net/sheets/079/6/6146B.pdf


Has anyone tested 6146Bs really hard to see what they will put out maximum?  What maximum voltage and current parameters have you used?

I usually find the upper limits and then back the rig off to a conservative level.  The maximum power out is usually MUCH higher than the published datasheet. However, with four tubes, I cannot do better than about 70w carrier and 300 watts peak output per tube.  And this is after trying different plate voltages, screen currents, drive levels, loading , etc. Even to the point of abuse.

What is the limiting factor here, the filament emission or is it the plate voltage?  I am considering switching to 6DQ5 or 6LF6 sweep tubes which have twice the fil emission, and a higher plate voltage rating.


My goal with four tubes was to see at least 300 watts carrier and 1500w pep output. (X5)    Is this do-able with four 6146Bs with lots of air - without destroying them?  


T

Reference thread for more info about project:
http://amfone.net/Amforum/index.php?topic=35443.0

What about 6293? It's a rugged 6146A.

With a big Crown M600 as a modulator, and a negative cycle attenuator, I could get 800 watts PEP out of a single 4D32 with a B+ of about 700. A 4D32 is roughly equivalent to a pair of 6146A's except the huge cathode has enormous peak current emission; much more then a 6146.

Jim
Wd5JKO

Pete,

Interesting on the 6293.  The datasheet showed its pulse ratings at 3KV maximum plate voltage!  I wonder if that is the pulse rating on regular 6146Bs?   I was worried about flashing them over, since I have a series PDM modulator that can be variaced up to 2500V.   I was thinking of keeping the carrier at 250 watts and having big headroom with 2KV peak voltage.  Maybe the plate dissipation will handle it.

Jim, I will take a look at the 4D32.  


The good news:  I forgot that there are old, used tubes in the rig for testing.  I just replaced the four with NOS 6146Bs. With 1600V,  I get an easy 325 watts of carrier and JUST kiss 1500 watts pep.  HA!  I'll bet the old tubes had a dud in there.   Now I need to know how high I can run the plate voltage for peaks and still maintain proper plate dissipation on carrier conditions.   For example, 275 watts of carrier and 1800 w pep would be a great goal. (X 6.5)  Then back the plate voltage off for comfort.

T

All the rigs that ran a pair did 100 watts, with 700 volts on the plate.
700 volts and 200 ma was typical on phone I think.

With four, you should get 200 watts at 700 volts and 400 ma.

No reason to push them harder than that, just use a bigger tube.

My 3x 4D32 rig runs 1250 volts and 300 ma, and does 300 watts carrier out.
Plate dissipation is the limiting factor there.
The 4D32 is rated at 600 volts on the plate with plate modulation, 700 for CW, but they work fine at much higher voltages, I would assume the 6146 does also.

If you can not get 400 ma of plate current at 700 volts, there is a problem.
If you get 400 ma at 700 volts and only get 70 watts out, you have an output tank circuit problem.

Brett,

Three people have recommended the 4D32 so far.  They are even cheaper than the 6146Bs.   I would have probably used them if I had looked into it.   Too bad the socket pin-out is different than the 6146B....  ;)  It wud need some rewiring.

Well, with the new tubes, I am seeing much more than what the datasheet says, so I am now happy. When I can only duplicate the sheet, something is usually wrong.

I'll see how this works out and go from there.  Frank/GFZ  said that maybe the PDM filter won't handle any more power without saturating, so I might be at the limit now and should be satisfied.

T

So what is the power in, power out?

You have 140 watts of plate dissipation (ICAS), 1600 volts on the plates, so you are only 1000 volts over the maximum ratings, what is the plate current at 325 watts out?

When I mention 1600V, it can be cornfusing since most of us think in terms of a standard plate modulated rig.  

This is a series modulator, so the PDM modulator tubes have the bulk of the voltage across them during resting carrier conditions. I have about 700 volts across the RF finals -  and during full modulation the modulator puts the additional 900 volts across the RF finals creating a 1600 V peak.  

Normally, with a conventional p-p modulator and iron, the same voltages would occur at ~ 120% positive modulation.

At 325 watts carrier,  700 volts, I am drawing about 575 mA.  This is 400 watts input.   325w/400 w = 81% efficiency.  This is about 75 watts of plate dissipation. The tubes are rated at 140 watts. (35w X4)    So well within ratings. As long as the tubes handle the HV, dissipation is not an issue.

I have actually run the HV up to 2000 V and set the PDM carrier back down to 300 watts to see upwards of 150%+ modulation. Still experimenting.  I think these tubes can take more plate voltage than we would expect, especially in pulse service, as long as the internal structures don't overheat from high average currents.

The tube plates show no signs of color and I have air blowing by them for good measure...

In the shakedown phase, my idea here is to push things to their limits until I start to break parts and then back off to a more conservative level. This way the rig becomes more reliable and the parts that are to break have already broken in testing, not on the air... :-)

T

 

Hey,

I just reread the RCA data sheet on the 6146B and noticed their comment that the screen must be modulated in AM Telephone service. WTF?

We seen screen taps on the PP primary side of modulation transformers.....for improved linearity....

Don't know if I ever saw class C rigs with the screen also modulated.

ever consider the 8122? I have a pile of them in matched pairs and even a matched quad.

r

Tom,

We did exactly that as part of new product development.   The recognized term for this is HALT  (Highly accelerated life testing).  Once you pushed the "corners of the design" and failures started to occur you would analyze the failure(s) and determine if it should be fixed.  In some cases design changes and or component changes occurred and in other case a given failure was not determined to be significant to the end user and things were left alone.  Once the first failure occurred and the analysis and fix took place you would continue to push it higher until the next failure occurred and just keep doing this until all the data was gathered.  Upon completion it provided a good demonstrated model of a given circuit.   

Joe, W3GMS         


http://en.wikipedia.org/wiki/Highly_accelerated_life_test   

Well, I think you are safe with running higher voltages.
The 4D32 is just like two 6146's stuck together, and I have been running them at 1250 volts on the plates, over 2500 volts with modulation, for years with no issue at all.

I do 1250 volts and 300 ma with three 4D32's, with a total of 150 watts plate dissipation.
375 watts in, 300 out, 80% efficient, 75 watts plate dissipation out of 150....

1/4 inch copper tubing plug in tank coil, no switches or taps.

I never figured it out before, but it seems I am running the tubes very light.
All the old rigs seemed to get 100 watts out of a pair of 6146's, or under 100 watts for a 4D32, so that is what I run them at, but I wonder why the old rigs ran things so light.

On my rig, if I load it up more, the plate current tends to go up, but the power output does not go up much.
I tried different tank coil values, but did not see much change....

The 2x 4x150a rig has a total of 500 watts of plate dissipation, but runs the same power input and output as the 4D32 rf deck. The 4x150 wants 1500 to 2000 volts on the plates, but I only give them 1250...

At 1250V and 300ma, they only get warm to the touch, and cooling is with a low noise muffin fan in the bottom of the chassis, not moving much air.

Same deal though, if I increase the current, power does not go up as much as the current...
 

Rob

That's the normal for plate modulating a tetrode amplifier.  There are several good articles in both the west coast (W6SAI) handbook and the east coast (ARRL) handbook of the 40s - 50's era.  

The most familiar approach is found in the ART13 for the 813 final.  Not only is there a secondary for the Plate but a secondary for the Screen.  Another approach is to just pick the correct screen dropping resistor from the plate supply - on top of the modulation secondary or to install a reactor in series with the dedicated screen supply which is called allowing the screen to "self-modulate."

Al

PS:  There is a fairly good article that describes plate modulating a tetrode

http://www.w8ji.com/amplitude_modulation.htm

Scroll down to the section "Plate modulated Tetrodes."

Brett said:

"On my rig, if I load it up more, the plate current tends to go up, but the power output does not go up much.
I tried different tank coil values, but did not see much change....

Same deal though, if I increase the current, power does not go up as much as the current..."


Brett,

Initially, I had the same problem with my 6146B rig.   It turned out that I needed MORE screen voltage and current. Once I reduced the value of the plate-to-screen dropping resistor, the tubes came to life and the power output jumped.  The combination of grid drive, loading, screen current, plate current, tank impedance, etc. is important and you will find a sweet spot where it all comes to life and the plate tuning and power output is robust and sharp.

Carefully put a VOM probe on the screen and see how much voltage is there under full drive. It should agree with the current as speced in the datasheet. Actually, my screen voltage is slightly higher than specs, but the current is right on the nose.  Also, run your grid current a little harder than specs to be sure your peak power is available under modulation.

Soft tubes are another possibility, but check the screen first. Also, measure the grid DC voltage when driven harder. Some circuits charge up with too much grid leak bias and make the tube harder and harder to drive. Output power can actually decrease with more drive cuz the charging effect adds bias at an increasing rate.  A bias steering diode might solve this on some circuits.

Once loaded correctly, more drive should increase plate current, show an increase in screen current and saturate the output power at some point. (class C)

Let me know your test results.

Joe, that is a cool write up - and your thoughts on shakedown runs are pretty much my own beliefs.  I don't want to blow out tubes or mod transformers, but I have no mercy for most any other components that may fail. We should be especially careful when pushing mod iron, but hopefully other cheaper components will fail first. Disc caps, resistors, arcing, instabilities and the like should be rooted out like the weeds they are... ;D

T

Everything is metered and adjustable on my rigs, so its easy to adjust things.
Turning the screen voltage up exceeds the screen current recomendations.

Grid drive is also above normal, with a very high bias voltage (fixed and grid leak).
Modulation is not a problem, at 300 watts carrier, if I crank things up, I can get close to 2000 watts pep, badly overmodulating.
So the tubes respond well to more voltage, but loading it harder seems to decrease the efficiency.

300 watts is a nice power level for 40 meters, so I do not worry about it much, if I got 400 watts out, I doubt anyone would notice any change.

And, I can not remember ever having a tube fail (in anything) for 30 years...

Brett,

Sounds like you have a good handle on it.  If you can do 2KW pep, the rig must be FB as it is.  Your three 4D32s rig is like mine on steroids.  I might have done three or four 4D32s if I'd thought of it.


BTW, do you have any pics you could post of your two homebrew superhet recievers?

T

Well, I should really look into picture web hosting.

Try this.

http://www.flickr.com/photos/tags/flicker/


klc

I never saw an AM xmtr where the screen wasn't modulated.

Pictures (some very old) here:

http://n2dts.smugmug.com/Ham-radio

Hi Brett -

Wow, amazing.   Youse gots the sickness as bad as me.... ;D

Nice combination of old buzzard and newer rigs. I like the receivers.

The "dream" lives on.

Tnx for posting them.

T

Darn, swept the Dual Quads rig.  I'm having a bad time with severe high frequency audio distortion above about 1700Hz.  The audio is perfect at 1 Hz, 100Hz, 1000 Hz.   Frank thinks it's the PDM filter design, so trying some different values.  

Blew out the MOSFET driver.  And I thought I had this in the bag.     Plate modulation with iron is so much easier... [sigh]

T

Ahhh... don't get discouraged!   ;)  PWM with tubes is challenging, but the results are better when you finally get it working.

It could be the PWM filter, but there are other possible causes such as insufficient bypassing in the cathode of the modulator.

Did you do the pulse width range test?  This test checks the pulse width range, and is done with carrier only.  You start with a carrier, and then adjust the DC offset going into the pulse width modulation generator.  As you decrease the "on" pulse, there may be a point where the waveform may ring or otherwise not stay square.   

One of the jobs of the PWM filter is to be able to contain (store) the entire "off" time of the PWM waveform.  What can happen  is this: as you approach 0% duty cycle, the waveform can "collapse" before the modulator actually turns on again at the next PWM cycle.  The fix for this is to increase the inductance of the input inductor.  The input inductor's inductance should ideally be just big enough to allow the modulator to fully store the "off" pulse at 1% duty cycle.  Increasing the inductance beyond this point will both decrease the efficiency of the modulator somewhat, and also will contribute to mathematical integration of small on-time pulses.

Experimentation will definitely reveal what's going on.

Regards,

Steve

Maybe you need more resistors ;D

Never built a PDM rig so I can't add much of anything

I would think the problem has to be in the filter, no?
What else could change with frequency so much?

Brett,

It may be the filter- I dunno at this point.

After a lot of testing and experiments, I still cannot find the problem.  Even with a resistive load across the PDM modulator tubes, I get harmonic distortion.   (no RF involved)    Even when they are running right, I don't think I will be totally happy with tubes in the modulator.

So I've given up on the PDM 6LF6 sweep tubes and tore them out. I'm replacing them with a single SS 4KV  IGBT device.

http://ixapps.ixys.com/DataSheet/DS99385B(IXEL40N400)_.pdf

It will sit on a copper spreader and heatsink with air blowing by it.  This will be a near perfect switch for PDM and the 6146B finals will make the rig the best of both whirls.  At $45 a pop, you know I will be careful with these IGBTs.

I just ordered an inductance meter, so will see if the cores got ruined or the caps are still good in the filter. The rig sounded pretty good until I had a flashover and then went to hell. Though, it always had that high frequency distortion from the beginning.

My goal is to get this thing sounding pristine and will not give up on the project til it's working right...

Frank / GFZ has been helping a lot via emails and Steve /QIX  suggested using the IGBT. Thanks for their help.    Dennis/ W7TFO has offered to send me some used 6146s for free, so no worry about running out.

Fred - Mo bigger resistors might solve it - 200 watters - ya never know...  ;D
T

Seems like a fun project.
Trying things until you find something that works real well is the fun part.

I do not really follow it, but people have built PDM rigs using tubes, have they not?


I really dislike rigs that have blowers or fans, at least if you can hear them.
There is nothing like a silent rig if its close to the operating position.

I hear guys running amps with big blowers in the background, and it bothers ME hundreds of miles away!

Very, very few people have [successfully] built a tube PWM rig.

It's so different from building a transformer coupled modulator, and the concepts are vastly different.

The thing is, with transformer coupled modulators, there are degrees of working "correctly".  It can be pretty poor, and still work or if someone is very clever, has the right components and does everything perfectly, the rig will sound superb.  And then there's everything in between.

Not generally so with PWM.  It's either right or it's horrible - at least most of the time.  They're not THAT hard to get right, but the rules for making it so are hard and fast.  Tubes are harder than solid state in the PWM world for a number of reasons:  the high voltages involved; tubes are poor switches and need compensation; the impedances are higher so the filter design is more critical; heater voltage is required, necessitating a special transformer; etc. etc.

But, if you can get over the technical hurtles, tube PWM rigs work well.  I've built a number of them over the years and they sound every bit as good as the best out there.

I wrote up a fairly detailed account of the last tube PWM rig I built including a lot of theory about how things work, and work-arounds and corrections for various problems that most builders will encounter.  If anyone is interested, I'll post the link.

Regards,  Steve

Tom,

Back when the 4D32 was a rare tube I replaced it in my Viking One with 3 x 6146B's and switched to an outboard 811A modulator that has its own plate supply. The Viking plate supply and the modulator plate supply are connected to the same variac so I can roll them up and down to set the power. My 35 year old 3 x 6146B's running @750 volts can do ~135 watts carrier output without complaining and the 811's will modulate them close to 100% to produce a bit more than 500 watts PEP. Rolling the variac up so that the viking's plate supply is above 800 volts it's easy to reach over 150 watts carrier but I don't push it that hard.

Looking forward to hearing the tube PDM rig!

Rob W1AEX

Yep, they can be a challenge, alright.

In the past I've built three big tube PDM transmitters. All involved 4-1000As.  They eventually worked FB, but they were especially dangerous due to the floating supplies, filament transformers, 10KV power supply, floating input RF circuits at HV, etc..  To work on it required extraordinary care. After being away from it for a while, going back into it for a repair was a big risk.  So I eventually tore each one down and went back to conventional plate modulation, which is what I run now for the 4X1 rig.

A big tube PDM rig also needs a few specialized parts, like a custom wound PDM filter that can get very big if wound on an air core. We need to wind a custom fil transformer to handle full plate voltage. The damper diodes can be finicky - I eventually went to tube rectifiers for one big rig. It goes on and on.

The 6146B rig runs a more "reasonable" 1600 volts, which I can test without excessive fear, but still dangerous.   I suppose I would eventually find the problem with this particular tube PDM modulator, but again, I probably wouldn't be satisfied without linearization, compensation and all the other stuff needed for tube modulators. The time involved could be better spent elsewhere with optimizing an IGBT modulator.  A hybrid - one half SS and one half tube is the best of both whirls. I still want tubes for the ease of QSY on any band and the fun of tubes.  But 100% solid state? -  I've had several class E SS rigs and became bored with them... but certainly nice rigs in operation!

In the meantime I found a beautiful heatsink in the cellar that is 5X5X5" with 16 fins.... about 320 sq" that will fit into the 6LF6 space perfectly for the new IGBT. It will have a copper spreader and plenty of air, so should be a good start.


Now I need to come up wid a new name.  It's no longer a dual quads rig.   Maybe a Hybrid name of some kind....

More as the parts arrive...

T
 

I would love to see this !

Ok, here's a link to the article.

Keep in mind, this article is *old* - I wrote it back in 1986 or thereabouts.  There are parts available today that were not available back then.  But the basic theory is exactly the same.  So, if anyone wants to actually build a tube PWM transmitter based on this article, please contact me first about newer technologies.

The whole PWM generator and driver will be MUCH MUCH easier than the one shown in my article because I had to make everything - triangle wave generator, controls, comparator circuits, etc.  Now, all of this is available in ONE CHIP !!

Otherwise, the article is as true today as ever.

The article is here: http://www.classeradio.com/pdm_article.html (http://www.classeradio.com/pdm_article.html)

There is 1 transmitter on the air today that is based on these circuits.  Larry NE1S has a kW PWM rig.  I still have my complete PWM generator and driver - everything right up to the grid of the 4-1000 PWM modulator tube.  It didn't seem right to dismantle this PWM generator and driver - I used it for so many years.

Regards,  Steve

That really seems like doing it the hard way.

Well, that article is very old, and now there are much better components available, and much of the circuitry can be literally reduced to one component!  The whole PWM generator including the anti-aliasing filter fits handily 1 small PC board.

Other than that, what part(s) of the circuit do you think can be simplified?  This is not rhetorical - I'm very interested.

Regards,  Steve

I'll bet he means compared to standard plate modulation with a mod xfmr, tube PDM is a harder and more involved way to do the same thang.


This is a point when we compare a plate modulated rig that has been carefully optimized with negative feedback and running the modulators in AB1 to a well designed and working tube PDM rig.  The PDM rig, if set up correctly, should be a little cleaner on the test instruments, but probably not enough to notice to the average ham ear.

T

I meant from the standpoint of using a nice mod transformer and some zero bias triodes.

A bunch of chips, wave generators, filters, damping diodes, etc, plus a tube/tubes and the high voltage.

For sure, PWM is really, really far from any transformer coupled modulator - quite a different animal entirely.  And, it is somewhat more challenging to implement with tubes than a tube transformer coupled modulator for sure.  In solid state, it's not so difficult for many reasons.

On the plus side, the power supply will be smaller, the transmitter will be lighter and it will cost somewhat less to build it (because you don't need to buy a high quality mod transformer and reactor).  It it's done right, the audio results will be quite good, and of course it's more efficient.

Steve and Brett, I have been following this thread with great interest.  I have always been more comfortable building the traditional AM rig with plate modulation, and have avoided PDM due to the cost and complexity of getting involved with the PDM generator and the filter circuits.  But a medium power PDM rig using tubes for the switch and the RF section seems very interesting to me now.  I was somewhat disappointed when Tom pulled the tube switch and changed to the solid state switch, as I wanted to see a successful completion of tubes for the power stages and learn more about it from his experiences.  But I understand his reasoning, and continue to follow his progress. 

That said, I am very interested in pursuing a PDM rig using tubes for the RF and the switch.  I have a few questions, and apologize in advance for the apparent thread hijack. 

I have several RF decks that I might consider using.  Is it possible, or should I say, practical, to run the RF deck with the cathodes referenced to ground and the switch in the B+ side of the RF deck?  If not, I will build an RF deck that will be compatible with the cathode floating at HV to allow the switch between RF stage cathode and ground.

When you talk of a simple, small board where most of the PDM functionality is implemented in one chip, is that the same board that drives solid state mosfet switches, or a completely different board?   If it is the same board, is it easy to interface it to a tube switch, either triode or tetrode?  Is there any advantage in using either a triode or tetrode for the switch?  I have an adequate supply of NOS 6146s, 4D32s, 4-65s, 4-250s, 4-400s, 4x-250s, etc.  As an initial experiment, I might even like to try the 6AS7, 6080, or 6336 for a low power rig, just to get some experience.  Designing, building, and tuning the PDM filter sounds like fun, after reading the document you supplied in the link.

Brett, looking at the photos you linked was quite an inspiration and has influenced me very positively to get back into more home brewing projects.  Your receivers are a class act, very nice construction.  I especially like the way you mounted the IF transformers on sockets so that you can swap and troubleshoot.  I will do the same thing on mine.  Your modular approach to transmitter building, with variable supplies and extensive metering is the way I like to build.  Your attention to detal and organization produced equipment that you can be proud of.  I would like to see more photos of the 4x250 based modules, either modulator or RF stages.  Looking at your shack, it is mostly home brew, very little in the way of manufactured rigs.  Way cool!

  Tom,

   If the audio response  problem continues after using the SS PDM, look at the 6146 self modulated screen circuit. Best I can follow in your earlier post is you have a 25K screen dropping resistor, and four .001 uf capacitors in parallel bypassing the screens to the cathode cage. The reactance of .oo4uf at 5 Khz is about 8K. Perhaps the self modulated screen concept was getting phase delayed, and at reduced level as you swept above 1700 Hz?

Jim
Wd5JKO

I like building, at least simple stuff, the class E and pwm stuff seems a lot more complex and needs more fine tuning.

Since a lot of the pictures are old, they show some non home brew gear, but that stuff is long gone and the only piece of ham gear I have is a 32V3 that has no cabinet or audio driver transformer (gave them away), so its worthless, but makes a good exciter.
Modern rigs float in and out, none of it gets used much if at all, except the sdr receivers.

The idea of using octal sockets for the IF transformers (and spacing things out) was not a good idea, but helped with the prototype. I had to change the IF amp tubes to something with less gain as building it that way made things unstable with high gain tubes.
I only had PC board mount IF transformers, hard to chassis mount, but thay have a very broad response, being intended for car radios (AM).

The home brew receivers out perform anything else I have had (for AM).
Much less noise then anything else I have ever tried, even the sdr receivers.
 
There are pictures of the 4x150 and 4D32 rigs, top and bottom, plus the 4x150 when it only had one tube.
I plan on redoing both with better grid circuits and more space on the 4x150 deck, and I might add a 3rd tube....

I use a cordless hand drill, a dremel, files, a few hole punches, and spray paint from cans.
Its also odd that nothing I built had a design, I took a bit of this, a bit of that, and tried different things.
If it worked, I used it, if not, I re did it.
That method would be very expensive with class E or the PWM type stuff, since I do not have those sorts of parts in the junk box.
Lots of respect for the guys who design and build that sort of stuff though.



Brett, looking at the photos you linked was quite an inspiration and has influenced me very positively to get back into more home brewing projects.  Your receivers are a class act, very nice construction.  I especially like the way you mounted the IF transformers on sockets so that you can swap and troubleshoot.  I will do the same thing on mine.  Your modular approach to transmitter building, with variable supplies and extensive metering is the way I like to build.  Your attention to detal and organization produced equipment that you can be proud of.  I would like to see more photos of the 4x250 based modules, either modulator or RF stages.  Looking at your shack, it is mostly home brew, very little in the way of manufactured rigs.  Way cool!
[/quote]

Very impressive!

This is mostly in response to Rick's query.

If you are going to use a tube PW modulator device, it is *significantly* easier to float the RF amplifier's negative rail than it is to float the modulator's.  For the RF, you essentially have to bypass the cathode circuit to RF ground (and make that capacitance part of the last element if your PWM filter), and drive the grid with a dc (and audio) isolated feed.  The grid circuit tuning components need not float at all - only the DC grid components (grid leak resistor, RF choke and metering).

The existing PWM generator board I use for the class E rigs is a proven, easy to assemble circuit that is perfect for any PW modulator application.  The circuit is published, and if you want a board and parts, they will again be available when I do the next run of PC boards which is planned for later this month.

For the PWM switch tube, use the highest current tube available.  I've built a lot of PWM tube rigs, and have found tetrodes are somewhat easier to work with, but triodes are fine and also work.  My first PWM tube rig used a pair of 250th tubes modulated by a single 833A.

The PWM driver is the hardest part of it, but it's not THAT hard.  The driver must produce enough output to drive the grid as hard as it can be driven without damage, and the driving waveform should be reasonably unaffected by the grid load.  A source follower is the best way to accomplish this.

You really should have analog compensation in your PWM driver.  This is discussed somewhat at length in the PWM (PDM) document - a link to which is in one of the previous posts.  Analog compensation drives the tube harder as the On-time of the pulse increases, and drives the tube less hard as the on-time decreases.  This both linearizes the tube (tubes are not perfect switches), and works to allow the transmitter to be modulated to 100% negative cleanly.  Analog compensation is done in all of the tube PWM broadcast transmitters that were made by Harris, etc. for this same reason.

Anyway, a worthy project for sure.

Oh, one more thing - I've been asked a number of times if I started with a low power transmitter for the first PWM rig.  No.  It's almost the exact same amount of work to build a big transmitter as it is a smaller one, so what the heck!  Go for the big one if it's no more difficult  ;D

Steve, thanks for the detailed response to my questions.  I definitely would like to order the complete set of boards and parts for the PDM generator when your next production run is available. 

I have been studying both the  PDM document you linked, as well as the newer documents on the class E website.  If I understand correctly, the analog compensation is required only for the tube switch, but not for the mosfet switch.  It appears this circuitry is documented in figG.  It is not clear to me whether that part of the circuitry is based on a board, or just manually wired. 

In any case, I think I will build the transmitter in a modular fashion, so I can use the PDM generator with either a larger or smaller switch, filter, and RF deck.  None of my RF decks are easy to isolate the cathodes for high voltage, so I will just build a new deck, I have all the needed parts available without purchase.  I am confident that I can calculate values and construct the PDM filter portion of the transmitter, but I expect I will need a bit more guidance on interfacing the PDM generator to the switch tube.  We can address that once I get started building the generator.  (I am thinking maybe a pair of 4-400A RF tubes modulated by either a 304-TL or 833A switch tube).

I really appreciate the resource of knowledge and boards/parts you provide to the AM community for this type of equipment. 

Hi Jim -

Yes, 8K certainly seems a low impedance compared to the existing screen resistor. (I currently use 15K for the screen dropper)


As you suggested, when I test the rig with the new IGBT modulator, I will keep this in mind.  I could always reduce the four screen bypass caps to 250 pf each.

I should have some results before Friday if the parts come in. Right now I'm setting up the modulator heatsink.  Looks like I will be mounting the IGBTs on Sil-Pads, since the ones I have (pink)  are good for about 5KV or so.  I noticed that even though the IGBT pad itself is good for 4KV, the leads are only spaced 0.118" from the heatsink.  Sounds like an arc waiting to happen without the pads.   Never used IGBTs before and too expensive to screw them up.


BTW, Rick, your plans sound like a very cool PDM project - 440As X an 833A.   You will love PDM once it is running right.  I remember the first PDM rig I built in 1991. It was a pair of 4X1s X a pair PDM.  The filter coils were about 40" long and wound on 8" diameter PVC pipe - thousands of turns. The coils were hung from ropes and did pull-ups in the rack when I hit lows according to Frank/GFZ.   Tron/ HLR came over for a visit and just stared at the rig as it was on the air pouring on the soup.. He had a glazed look in his eyes and said, "That's the way to do it - I like it!"    ;)

T

Ok, sounds good.  The 833A will be MUCH easier to drive than a 304TL. A 304Th - maybe.

My biggest PWM rig was a pair of 450TLs modulated by a single 4PR1000.  That was quite the transmitter !!!  The PWM filter coil - I still have it - is an air wound coil that's around 30 inches long and 12 inches in diameter.  Wound with #22 magnet wire.  Worked great - and that's more or less the kind of coil you want.  It took me a few hours to wind it.

The 2nd coil can be wound in layers, and that's how I made mine.  I put waxed paper between layers to prevent arcing between them.

The first inductor of the PWM filter *must* be large enough to completely contain the PWM waveform at the smallest duty cycle - typically 1% or 2% on-time and 99% or 98% off time.  If the waveform collapses or otherwise begins to distort from a square wave, the PWM filter input coil is too small.  It should be JUST big enough to contain the entire waveform when the filter is properly terminated and no bigger.  Making the coil of a higher inductance than necessary will compromise the efficiency of the modulator somewhat.

Anyway, just a few pointers.  It'll be interesting to see how the project progresses.

Tom, you may need an active pull-down with your solid state modulator to maintain the waveform.

OK, So it appears to me that I must decide on the approximate power level of the RF final (Plate voltage and current) before calculating the filter component values.  One of the reasons I thought about starting smaller is the higher voltages involved on the plate side of the RF deck, typically double what you would normally run as a linear or conventional plate modulated final.  Not only is it necessary to float the filament side, it is also necessary to provide adequate spacing of the DC components in the plate circuit to avoid bypass and blocking capacitor breakdown and arcing. (I don't think the pi net components need to be oversized, as they will not see the higher DC potential.)  Extremely high B+ voltages are more dangerous, even in an enclosed rack.  So I still might start smaller with three or four 4D32s in the final, running lower voltage but maybe higher current.  If I do go with a pair of 4-400As, either the 833A or a 3-500Z should work fine as a switch tube.  Lots of options, just need to nail down the best compromise.  Thinking out loud, maybe its time to tear apart and rebuild a new RF deck from the dual 4-400A G2DAF linear that I built in the air force barracks in 1967! Although I never received any negative signal quality reports, it certainly does not meet today's IMD goals.

The voltages across the DC blocking cap will be the same as in a standard transformer coupled plate modulator.  The only difference is that the (dc) voltage is there all the time as opposed to only on peaks.

I ran a 10kV power supply with my rig - 4000V across the RF amplifier and 6000V (dc - filtered) across the modulator.  This allowed for 150% positive peaks.

Typically, a 2.5 to 1 ratio is used, with 1 being the DC voltage at carrier, and the 2.5 being the total power supply voltage.

A pair of 813s modulated by either an 833 or a some other tube makes a nice combination because the 813 is a relatively low voltage tube.  You'll need a 5000V power supply, and the rig will run 800 watts DC input (2000V at 400mA).  Not bad  - a good medium power rig and very practical in all respects.

I see.  I was thinking a 2-1 ratio, but that would just barely allow 100 percent peaks.  2.5 to 1 makes sense now.  

Based upon power supply components I have on hand, I was initially looking at keeping the total power supply voltage down to 3500 or 4000 volts.  I could go higher but that would mean using new rather than available parts.  (A bridge rectifier and a pair of oil caps in series would put me around 8000 volts, but I would rather not do that initially.  For safety, for this rig I prefer a dedicated internal power supply, rather than a shared external one.)  So maybe my initial build might still be best to go with a quad of 4D32s, and maybe two or more 4-65A tubes for the switch.  I have 6 NOS 4D32s and 10 NOS 4-65As.  That would allow me to build it on two decks and use convection cooling.  Power supply components on the bottom floor of the rack.  So many different options, but modular construction will allow me to scale up later.  

I am wondering if the switch driver would be considerably different for 4-65A switch tubes as compared with an 833 or 3-500, and is that driver part of the printed circuit board or hand wired?  I am sure these details will be worked out in time.  Lots to build during the cold winter months.

By the way, one of the things that I really like about the transmitters that Brett, N2DTS builds is the plug-in coils for the pi-net.  Easy to get just the right inductance for any band with no band-switching on the plate tank.  I will definitely build that way.  The ceramic pillar insulators and the metal plate holding two banana jacks allows for short or long coils, simple, neat, effective.  We can sure learn a lot from just looking at photos of other folks innovation!

Sounds good except for the 4-65As.  I believe the 4-65A would make a poor switch because it is a relatively high impedance tube.  The 4d32 would be better, or if you can find them cheaply (and I think there are a lot of these out there) the 6DQ5 - which is a very good switch !

I've had some experience with 4-65As in various types of service, and they definitely like a lot of voltage, and are a bit anemic in the current department.

Here is one such PDM transmitter using two 6883s (12V version of 6146).

http://www.lu8jb.com.ar/PWM_6883.htm

The switch tube grid is switched in turn by a HV bi-polar transistor with its base switched by the PDM signal of about 5V.

Phil - AC0OB

The audio in that particular implementation will have some distortion due to the lack of analog compensation, and in fact the drive may actually decrease somewhat with increases in the on-time (duty cycle) due to a resistor pull-up to the modulator tube grid (the opposite of what is required).

One thing about PWM, shorcuts are costly with respect to audio quality.  But this is amateur radio, not broadcasting, and there's a lot of stuff on the air that will sound a whole lot worse  ;D  (and some that will sound a whole lot better, too!).  Depends on how much work you want to put into it and what the final goal is.

Steve, can your pwm generator board simply plug into the 6883 rig linked to in the post above yours ?

It should work although a MOSFET is better than a bipolar (it would still work either way).

Beware: That 6883 circuit has fundamental weaknesses that will yield less than ideal results.

That 50 watt 6146 schematic has an interesting feature...

1) Notice he uses a single zener diode for fixed bias in series with the GRID leak resistor instead of placing it in the cathode. This way no power is robbed from the plate circuit.  I've not seen that done before in a class C stage to replace the protective bias fixed supply.  

Would there still be enough residual grid current to drop a voltage across the zener if RF drive is lost?


The builder obviously knew his way around transformers, as some of them look homebrew. Coming up with the perfect PDM filter inductors in compact form like that is not especially easy.


Maybe with  a few mods it would make a great little driver rig.


T

 

I think I have that same meter in my junk box.

And the chokes? look like the plastic spools radio shack sells wire on.
Just buy 4, glue them together and hook the wires together?
That is slick, no need to even unwind the wire...

I've used the zener / capacitor approach in my own PWM transmitters (in the modulator).  In fact, the tube PWM article shows an "adjustable zener" for such a purpose.

The zener fixed bias circuit in the RF amplifier technically isn't needed because typically the modulator tube is simply turned off, and there is no voltage across the RF amplifier at that point..... so no protective bias is needed.  Usually at least from a linearity standpoint, 100% grid leak bias has typically worked better, at least in my RF amplifiers.

But, if one did choose to use it, the circuit's ability to maintain the fixed bias is dependent on the leakage of the capacitor(s), and other components that might drain energy from the fixed bias storage capacitor.

Addendum: This is a WIP.

My feelings exactly.

After my next two HB projects are completed this is what I was going to use.

I have a bunch of 6DQ5's from an HT-40 restore so I thought I would use them:

I was thinking more protective bias is needed for a dumb mistake, no drive or cata-crapout short situation. Like when the modulator itself shorts or the MOSFET driving it shorts. I've had this happen several times and it will put full voltage on the RF final.  A shutdown circuit would be even better for the final, but more complex with floating circuits. A 33V 50 watt zener is cheap and simple insurance to cover most problems occurring outside the final.



T

That same web site has many PDM rigs shown. This one is solid state, 220 watts for 40M AM:

http://www.lu8jb.com.ar/PWM_460_40m_(2).htm

He has many of the schematics there...The PDM is based upon a NE555 and an LM311...Quite simple!

Jim
Wd5JKo

It should be pointed out that lu8jb uses a 75kHz sampling frequency whereas Steve's PDM's use about 110kHz so the PDM filters will have different component values.

75kHz was the standard sampling frequency for some of the Harris/Gates PDM rigs so it should be more than sufficient for Ham audio.

However, I prefer the higher sampling frequency of Steve's circuit since it should yield slightly smaller values of PDM filter components.

Phil - AC0OB

That's a good point about the switching frequency.  The main reason why such a low frequency was used with the tube PWM transmitters is because, for various reasons (that have already been covered here), it is very hard in tubes to maintain the PWM waveform over the entire range of duty cycles.  So, there is less mathematical integration of the waveform as a percentage of the entire cycle at lower frequencies than at higher frequencies.

The old Harris PWM rigs used 70kHz, and I used around 70kHz in all of my tube PWM rigs.  It is possible to change the PWM boards to put out a lower switching frequency by changing a couple of component values.

Yep, that was a typo. Harris' PDM generators generally used a 70kHz standard.

l8ujb uses about 75kHz which is 5X15kHz, 15 kHz being the highest audio frequency expected.

In one implementation, Harris/Gates used a 35kHz cutoff, 3 inductor PDM filter system (with appropriate caps). There was a feedback path, tapped right after the first inductor, that fed back to the PDM module for distortion correction.

My plan for the circuit was to see if it could operate at the approx. 110kHz frequency with the 6DQ5 sweep tubes, and if not, then change the component values at the RT and CT ports of the 25701 PDM chip.


Lots of fun stuff to play with.  :D

Phil - AC0OB

4 - 6883 RF, 2 -6JM6 PDM => 400 watts AM!

http://www.lu8jb.com.ar/PWM_6883%20X%204_(1).htm

No schematics, but good photos of what appears to be well thought out modular construction.

Edit: added a schematic of one of his QRP PDM rigs...uses the Ti494 PWM chip....

http://www.ti.com/lit/ds/symlink/tl494.pdf

Looks like something to play with and no high voltage! See attachment.

Jim
Wd5JKO

Steve

I added the red highlight to a quote from one of your earlier posts.

Why not do the following:

Between the output of the PDM filter and the modulated RF stage... add a Zener diode with its cathode connected to the modulator output. This will add a negative offset voltage (equal to the Zener breakdown voltage) to the modulated B+ that drives the RF stage. As a result, the PWM will not have to go all the way to zero duty cycle pulses to produce 100% negative modulation peaks.

Stu

Hi Stu,

I was wondering when we were going to hear from you !!!!  That is a good suggestion, and I had actually tried that idea way back when (wow, we're talking 1973 here!!).

Unfortunately, it didn't correct the problem because the problem is not necessarily resolution of small pulses (although that is important), but rather it is the issue of discharging the PWM filter (and associated stray capacitances) when the current falls to near 0.  There isn't enough current.

To solve the problem, I had come up with several experimental solutions:  An active pull down is one.  That definitely works, but it does affect the efficiency slightly.  Ok, I thought this next one was clever - use an inverse modulator that discharges increasing amounts of energy from the filter (as the on-pulse falls to 0), and recycles the energy back to the power supply.  Or the same sort of thing that takes a constant amount of energy from the filter and recycles it.  Never actually built this, but did come up with some designs that seemed reasonable.  You see this sort of thing (energy recycling) in many switching power supplies, and I've done it myself in transformer coupled PWM modulator/power supplies.

Ok - why, someone might ask, doesn't this same filter discharge problem occur with the class E rigs.  Same thing is happening - the voltage will eventually fall to 0 (or near 0) and there won't be any current to discharge the filter.....sounds right, yes?.... well, NO... but only because of what I call the "class E gift".  This is an odd side effect of MOSFETs, where, if you drive the MOSFET with RF, the MOSFET will produce a NEGATIVE VOLTAGE on the drain (in the absence of any other drain voltage), and at a reasonably significant current, and this discharges the PWM filter - even with 0V output from the modulator.  So that's why there is no active pull down on any of the solid state pulse width modulators that we use with the class E rigs.

I've got a Valiant around here that I am very tempted to modify with a PWM.  The original modulator isn't there anyway.  The hard work will be to float the negative lead of the RF amplifier, and I haven't looked into all that's involved there, but it can't be that hard.  Put a bridge around the power transformer (with a cap input power supply) which should yield around 2000V (or a little less).  Use a couple of 6DQ5s in the modulator and the publish the design.  And, fit all of it inside of the original Valiant case, of course!!

It'd be an interesting project for sure.

Steve, et al.

Yes... very interesting!


A great example of: "the Devil is in the details"

I'll move back into "listening mode" on this discussion.

Stu

Update:

Got the IGBTs for the modulator today and mounted them on the heatsink.  Hope to try some tests by tonight. Received the triangle audio generator too, so loaded for bear.

Steve, I'm gonna try it first without the pull down circuit to see how it looks and go from there. Say a prayer that two $45 devices don't go up in smoke... :o

A PDM Valiant would be a hoot.  What ever happened to your 500 watt PDM Ranger?


BTW, it's interesting that the LU8 built a quad of 6146s X  two 6DQ5 PDM rig. I didn't see a schematic with it. Maybe he's the "QIX or Tron South of the Border."  


T

MUCH BETTER, but a new problem.

I fired up the IGBTs and got the rig working.   At about 500 volts on the rig (200V on final and 300 on modulator) the 100 HZ triangle tone looked good.  The high end above 1700 HZ was much cleaner, but still not perfect above 5 KHZ.   I think the problem with response may be the screen bypasses, so will change them to 250pf each.


* But the big problem now seems to be that the IGBTs are going into a slow thermal runaway when the voltage is above 600 volts.  When the modulator is not in the circuit (cathode of the 6146Bs grounded)  the RF output is stable, even with 800V across the finals.   But when the modulator is working, when above 600V, the plate current  starts at about 400 ma and then after 45 seconds it creeps higher and higher until it pins the plate at 1A.  The IGBT collector waveform starts to get rounded at the top and shrinks down as the power increases in runaway.   Everything is stable when at 500 volts for five minutes at a time.  But go above 600V and it starts to creep.

I felt the big heat sink and it appears to get warm all over its surface quickly, so there is good thermal transfer.  The IGBT case does not seem hotter than the heatsink, which is about 320 sq ".   I tried it with a fan blowing on the heatsink, but it did not have a big effect on the minimum level when the runaway started. That seems strange, but may be related to the IGBT junction.   I am using a copper spreader too.

The pulse on the IGBT input looks very square with no spikes and stays stable thru out the runaway event.  The collector output has no spikes, but is slightly more rounded at the top  - but doesn't seem much more rounded  than other PDM pulses I've seen.  I see no evidence of parasitics.

I have a 30 ohm resistor is series with each gate. I shorted them out and it made no difference to the runaway problem. I also have a 320 ohm resistor from gate to emitter for each IGBT for stability.

I'm using two stacked pink Sil-Pads to mount each device.

Maybe I need resistors in the emitters for thermal stability?   The 6LF6, (140 watt total) tubes handled 2KV no problem, but the IGBTs are stuck at 600V and are rated at about 320 watts total.  It can't be dissipating THAT much power even though the pulse isn't  perfect at only 600V. . What gives?

T

Do you have a current schematic of the circuit you can post?

Phil - AC0OB

Nothing drawn up yet, but I will later, Phil.

I just realized that my input pulse AFTER the 30 ohm gate resistor looks rounded. It is square coming outa the gen.  I have it acting too analog right now.  I need to change some values. 

Also, I am using two stacked SIL-Pads.  Frank tells me the thermal resistance is too high with two and probably causing the junctions to get too hot.  

So I might just mount them on the edge of the copper with paste and do it right. They have built in insulated pads for 4KV, so might as well.  The leads are too close to the heatsink, thus my reason for the SilPads.

I'll get this yet.  The audio is starting to come alive and look decent for a change.

T

Well, don't listen too long.  You are very often the "voice of reason" in technical discussions!

  Tom,

   This issue sounds thermal for sure..

One thing I did once was to put the electrical isolation between the copper spreader and the heatsink, while mounting the power device directly to the copper spreader. This way I get the low thermal resistance between the transistor and the copper spreader, and the next thermal barrier between the copper, and the heatsink was also very good since the area insulated had a lot of square inches. I used a overhead transparency sheet for an insulator, cut with a scissors, and holes made with a hole punch. A Kapton sheet would have been better, but I used what I had around. Used insulated shoulder mounts from old TO-3 transistor mounting kits along with counter bored holes in the copper spreader.

This was in a DC power supply where I only needed to hold off 50 volts. I was not worried about capacitance to ground..YMMV.

For thermal grease, consider the microprocessor stuff, such as Artic Silver Five:
http://www.newegg.com/Product/Product.aspx?Item=N82E16835100007
keep in mind that it's electrically conductive whereas the white "Owl Sh_t" is not.

Jim
Wd5JKO

Tom, it does sound like a thermal runaway.  They are bipolar transistors (with an insulated gate), so thermal runaway is definitely a real thing.

The bigger question is: why are they getting SO hot?  Normally, PWMs don't get overly hot.  I modulated 6 6DQ5s in parallel with 2 LITTLE TO-247 package IGBTs, and the heat sinks barely broke temperature.  I used a pure square wave drive, and drove them to +12V. This is important.  They have to turn off all the way, too.

Something's amiss.  Might be a turn-off or rise/fall time problem - could be lots of things, but measurements with the scope should reveal all.

Found the runaway problem...

Turns out it was heatsinking.  I had the two IGBTs each sitting on two stacked Sil-Pads.  The junctions were probably on the verge of roasting.  

As a test I tried mounting just ONE IGBT directly on the 1" X 5" copper spreader with white paste, no Sil-Pads and no external heatsink.  Fired up the rig. It ran warm at 500V and warmer at 700V for a minute with no sign of thermal runaway. The direct paste connection did the job.


Reduced the input resistor values and believe the input waveform cleaned up somewhat.  The collector waveform is very good - just a slight sign of slope.

Plan to mount two IXDD414s  right next to each of the two IGBTs to give them a good kick in the pants at 12V.

I see the audio is clean up to about 5 KHZ.  The screen caps need some decreasing and I shud be there with 10KHZ audio limits - the design of the filter.   At this point I think the PDM filter is fine.

T

Update:

The original heatsink location was cramped and  had poor air circulation.  I cut a 5" hole underneath  the heatsink and added a muffin fan to blow air thru the fins UP.  WHAT A DIFFERENCE!   With just one IGBT installed, the heatsink gets luke warm running a 175 watt carrier at 750w pep.  

One of the IGBTs shorted after the runaway abuse before, so I ordered another.  But with the new airflow, ONE actually does the job.

I added a 2uf / 5KV cap right next to the damper diodes to dump the spikes.  The short leads are important unless the HV supply filter caps are located very close by.

I need to run some higher power tests yet, but waiting for the new IGBT to arrive.

It appears the rig needs to be loaded very heavily to see perfect audio highs. Loaded lightly and the highs slant like phase distortion.  I think the 1000 ohm PDM filter needs to go higher, like 1500 ohms or so. I shouldn't have to piss beat the 6146s so hard to satisfy the PDM filter.

As a low power driver with the Variac at 600V, it will do 50W / 300w pep  to drive my 4X1 linear amp.  As a stand alone rig at 1400V, I think 200 w carrier / 1000w pep is a reasonable goal without beating it.

Thanks to Frank / GFZ for his many helpful suggestions!

T

Below:

The new air system, heatsink holding the two driver  IX414s and two IGBTs on copper spreaders.

BTW, I added flexible copper braid to the tube plate cap connections.  The solid wire was too stiff and pulled a cap from the tube - the glue failed.   Now there is flexibility to thermal expansion and mechanical strain.


Notice the rig is evolving....  I added the legs stand to give the fan below plenty of room for air intake.  I found that  it needed a few inches of clearance or else the air eddy currents ruined the flow.

There was now room to mount the new 2uf / 5KV snubber cap below. (and who knows what else)


It's sounds absolutely great in the off-air monitor and has the normal PDM "shark fins."  It sweeps from 1 HZ to about 10KHZ with flawless audio.  The triangle tests look good at 10, 100, 1000 HZ, etc.    If I can just iron out some minor issues, it will be a FB rig.

T

Very tight building, lots of stuff in a small box.

After a lot of work, problems and modifications, the  the "Dual Quads" 6146B / 6LF6 PDM rig is working well.

Since my last post, I blew up five $45 IGBT PDM switchers one by one - they all died from thermal runaway at only 1100V.  I went back to the 6LF6 sweep tubes.  The rig will now handle  1800V at  300 w carrier out, 1500 w pep out - without one crap out so far, since the mods. I love tubes!

I ran out of modification room, so split the PDM switcher and RF deck into two separate cabinets.  Also, I discovered that the original PDM filter, wound on toroids, was 1/10 the inductance it should be. Space-out! That was one of the main problems that I finally figured out when I got an L/C meter... duhhhh.   Now I'm using my old 4-1000A air wound PDM filter coils for the time being. Overkill. They are thousands of turns on 5 gallon PVC pails.  I tapped them to give the perfect 12mH  / 16 mH  inductances for a 1,000 ohm RF load.

The audio sweep tests look flawless with 10, 100 and 1000 Hz triangle waves looking textbook. I added a fan to each set of quad tubes blowing air directly down on them. After a transmission, they are cool enough to lay fingers on without discomfort.  And I am really hitting them hard at 1500w pep.  Without air flow, my laser temp gun shows a bulb temp approaching 300 degrees  F.  With air, they are usually about 110 degrees F. The tubes are mounted on plates, so air can go underneath for the bottom seals too. I can't stress enough the importance of aggressive cooling, even for these smaller convection-style tubes like 6146Bs and so forth.  With great cooling, these tubes will last a long time, especially at maximum operating parameters that are based on no air.

I had it on 75M last night and the locals thought it sounded FB. Rob /AEX and Steve/ QIX both sent me recordings which I felt were decent sounding to my ear.

The rig will sweep from 1 Hz to 9 Khz before the PDM filter starts to roll off.  I will be controlling the audio bandwidth from low level.  I still need to do some analog compensation of the switch tubes. It will modulate down to about -94% and I have seen 160% positive when at 250 watts carrier, according to Steve's mod monitor file.

The RF finals run a hefty -110V grid bias, which is very heavily into class C.  The switch tubes are probably in the 90% efficiency area.  There is very little heat coming out of the cabinets after a long key down.  


All in all, I am very happy with this rig's performance and expect it will be a nice little medium power rig to fill in the gaps.  It's working so well that I'm designing up a much larger tube PDM rig  - but we'll have to see if my ambition holds. This project was a real b-buster ... ;D

Here's a bunch of pics showing the latest progress.

T

More shots. Check out the 4X1 PDM coils....


All HV points are protected by Plexiglas or the cabinets.


Notice the little castor wheels to allow easy servicing access.


The blue cabinet on the second shelf holds the 2200VDC  1.5A supply, (with a step start)

Notice the PDM generator in the silver box on the right - 2nd pic.

Station shots:  Fabio I, Fabio II, Mr. Clean, Rico Suave, Dr Love -

This RF deck has been thru a lot of changes in the last two months.

Notice that both the RF deck and PDM units have SO-239 jacks on the front panel for PDM in - PDM out access. This will allow any PDM filter to be added without tearing things up.  Eventually I will wind up smaller coils that fit inside one another.  They will reside with short leads between the two units.

Thanks to Frank / WA1GFZ for his help via email to help me solve the many issues I came across with this project!

T

Good looking transmitters Mr. T! Congrats on the "dual quads".. Looks great! I am wondering if the blue and white color scheme might just be your thang..

Oh yea.. Guess it is!

Heheheh - pretty funny!     Just imagine driving around in that Fabio car. I wonder how long it would be before you got dragged out and beat senseless?   


OK, now all eyes are on you. When is that new ??  pill mawl gonna hit the airwaves?


T

more power in oil?  amazing. 70W carrier..

Don't wear out all the 6146's. I just gave up all my 807's to help restore a working museum piece computer in the UK. 807's are used as row or column resets in that. It has many many 807's.

Wow, that sounds like it was a LOT of work.

Buckets?
Is there RF on the buckets?

I think that is crazy, bucket coils in the shack behind glass!

Could you not make something small on a toroid or something?

Those are PDM filter coils handling 170 KHz square wave pulses.  They were originally wound for my KW   4-1000A PDM rig.  They are being used temporarily cuz my small toroidal filter was wound at 1/10th the proper inductance by error.  Actually, a filter like I'm using doesn't get any better than that. Less inter-turn capacitance than a toroidal based filter means better filter isolation.   Some designs use an air core as the first inductor and toroids afterwards.

The plan is to try winding compact coils one inside the other, all air core.The filter will mount between the two cabinets with short coaxial leads.

T

My thoughts on those PDM coils too.  Shades of Tesla; with a one or two turn pulsed primary you could stand some serious wig on end.  ;D

Where's that classic charlie Chaplin pix?

Seriously, love those blue and white, plexiglass cabs too.  very nice.

What would happen if you stuck your head in a filter like that when modulating?

Very nice work Tom...........

Could you keep your coffee warm in the buckets?
Could you turn them into bug zappers?

Tom,

Its been great following both your progress and set backs.  Personally I am thrilled that you went back to the tube PDM and if you had not, you probably would have never found were the problems were.   So now you know the good, bad and ugly! 

So now try and get on the air since others would like to hear the rig that has put you through the hurdles of hell! 

Oh, it does feel good when the issues are concurred and the results speak for themselves.

Joe, GMS 

I appreciate the comments Joe and all!

You make a good point about following through.  One of my weaknesses is after getting a rig built, I have a low tolerance for the aggravation of crap-outs.  If I cannot get the rig running perfectly in say, a few days, I tend to want to tear it down and try a different approach.  After the initial PDM tube modulator problems and then the five IGBTs blowing up, I almost threw in the towel.  But often we are right at the point of success when we give up.  It paid to stay the course.

Now the next shakedown that I never finished before - Last night I had the 4-1000A X 4-1000As plate modulated rig on the air. (Fabio II)  At 4KV it flashed over four times and blew out my HV fuses.  Today I opened it up and found that the input tuning bandswitch was arcing over to ground. Melted the wafer.  I replaced it.  The arcing started again but this time the screen pin flashed 1/2" to ground. I thought I blew a 4X1 cuz the screen current was dead. Turns out that I forgot to solder back on the RF choke lead, so one screen was floating..

Gots it going again and this time I was running tone stress tests. Suddenly there was arcing on the scope under modulation. I thought I had blown the mod transformer... yikes!    Investigation showed that one of the Teflon SO-239s at the Bird wattmeter was shorting and was hot as a pistol.  Repaired that.

It's now hanging in there with no shutdowns.

I'm still left with some odd high frequency distortion to figger out.  It's all a matter of pushing the rig and discovering the weak links until nothing fails anymore under normal operating condix.  This is often the most difficult part of building a rig and takes lots of patience. I mean, all it takes is one part out of 100 to be arcing internally, out of sight. (in a scary high voltage environment) And each time it arcs it does damage to other expensive parts.  Try to find THAT problem and keep a sane attitude... ;D

**  I think the bottom line is that when we are actively building something, we see the progress every day. Satisfying.  But when troubleshooting, we can blow hours and hours and sometimes see nothing for our efforts. Frustrating.  When we finally lick a difficult problem, there is finally a payoff.

T

Just an update on the Dual Quads tube PDM rig...

The rig has been running very FB except that it would not modulate less than about -85% negative. As normal, the tube switch could not hold the pulse when the filter field collapsed as the pulse narrowed. This required some analog compensation to make the pulse more linear.

The biggest improvement was to drop the PDM switch frequency from 170 KHz to about  75 KHz.  This immediately improved the negative modulation to about -92%!   I then added a 2 ohm resistor to the source of the 11N90 MOSFET.  The 11N90 drives the four 6LF6 sweep tube switches. In addition, I added a 170 ohm resistor in parallel with a 1N4007 diode  in series with the 11N90 gate. (diode anode towards gate)   This improved the negative modulation capability approaching -95% or slightly better under voice modulation.  This is all we can hope for using a PDM tube rig without elaborate circuitry.  

With the PMC-300A SOFT negative peak limiter working, the extra -5% negative is masked and the rig sounds FB. I don't need the hard NPL clipper anymore.  The positive peaks are stellar - I've had them up in excess of 200%, depending on where the PDM carrier is set.  I usually run the pos peaks at about 140% max.

The rig sweeps and modulates as well as Fabio II, the 4-1000A rig, but is slightly cleaner and has more high end, due to the transformerless PDM tubes.   At 1700V, it is putting out about 350w carrier and 1500 w pep. I usually run it at about 250w carrier and 1400w pep.

It is buttoned up and except for modifications and maintenance, I consider this a finished rig, caw mawn.  Oh yeah - and drop the big PDM coils and build up a compact set. I'll wind them so they slip one inside the other.  And 6" short cables from filter to the rig.

My thanks to Frank / WA1GFZ for his email help getting me thru the hard parts.   These analog compensation suggestions were from him today.

This was a very difficult project for me and I almost gave up a few times. I couldn't be happier with the final results.

I'll be on with The Dual Quads and Fabio II over the coming weeks...  ;D

T

As a final post to show the results of this tube PDM project, here's a recording Jeff / W2NBC made today of this rig on 40M.  (150 mile path.)


I finally hooked up the CRL-PMC-300A modulation limiter and the Dominator II peak limiter.  I think the audio is getting dialed in for my voice.   My goal is large dynamic range with minimum processing. The limiters come in only on the peaks.  I have more density available if conditions get poor.   Since this recording I've added a slight touch  of compression (about 2-3 dB)  right after the mic preamp, for better voice control.


You are hearing a cool and quiet four 6146Bs in class C parallel,  pulse width modulated by four 6LF6s in switch class D parallel at 200 watts of carrier, about 1250 watts pep, 1600 V total across the finals and modulators, muffin fan air on all tubes.   (W1AEX followed by K1JJ):

T

Fantastic! As close to "in-person" as I've ever heard from your station. Lock it down.

TNX Steve.  Yep, it's getting better.  

To be critical, I hear a little undesirable mid bass in there.  (100 - 125 Hz)  I EQed it out today.  

This little 6146B PDM rig is working so well I am tempted to scale up a new one using 813s.    


I played around with the new Aphex noise gate. The PDM rig is silent, but when using the 4-1000A rig, if we reduce out about 3-5 dB of blower noise with a smooth release, it is reasonably transparent. Any more and the gating starts to become noticeable.  But 3-5 dB can sometimes make a difference when on the threshold of "OK"  or "Good"  background noise levels.


T

Tom,

  Any chance that rig could be on 14,330 this coming Monday night?  ;)

Jim
Wd5JKO

Sorry, Jim, but the rig was built for 160, 75 and 40M only. 


Looks like the noise gate for AM was a bust. Even a small amount of noise reduction was noticed and got negative comments. Same as last time. Why would it be any different this time?  Guess I'll let it be on AM.  On SSB, the noise gate works FB and seems tolerable by the masses.

I'm thinking of taking out the four 6146B finals and replacing them with four 4D32s.   That would double the output power to an easy 500w using small octal-type tubes.  Small, silent and cool, caw mawn.
 

T

Tom, if that is you on the MP3, that station really sounds great!

RF equipment using sweep tubes as intended!  - a twist back to normalcy for certain! Bet they last a long time.

Thanks, Pat.


I went ahead and ripped out the four 6146 finals.  The sockets for (4) 4D32s are being installed now. 

So the rig will soon be (4) 4D32s in RF class C .....pulse width modulated by (4) 6LF6 sweep tubes in class D.   1800V on the whole thing. Probably good for 500- 600 watts carrier. Amazing power for a compact, quiet, low heat summertime rig.

Should have it working once the tubes arrive by the end of the week.  Will take some pictures.

T

It's always easier when the final deck is built on a removable mounting plate.  I pulled out the 6146Bs  plate and almost have the 4D32s ready to go.  I figger four may make an efficient 500 watts of carrier. Maybe compare to a Johnson 500.

See the empty space in there in the second pic?  Tight fit - JUST fits.  (Those suppressors are non-inductive 50 ohm resistors - I measured them with an MFJ-259B at 1:1 at 30 Mhz.)


The 4D32 is an amazing looking tube, like a baby 813.  The  internal plate structure is about 70% of the 813's size. Wish I had done a little rig like this for the summertime long ago.

T

Easy to put in or take out for servicing.

These ceramic sockets are also used for the 8877 tube. Very strapping.  At 2KV total PDM voltage, they need to be well insulated..

T

Getting closer.  Bottom all wired and ready to install and do some testing.  

The grid leak and screen dropping resistors all needed different values.  (About 1/5 their old values)

This 4D32 mounting plate gets bolted to a piece of Plexiglas which insulates it from chassis ground. The whole plate floats at 2 KV under full PDM modulation.

T

  Tom,

   This is getting exciting!

Some random comments. I suspect your drive requirement will go up quite a bit, and that the added circuit capacitance will require you to make some changes. I don't remember if you neutralized the quad of 6146's, but you will likely have to do that to the 4D32's in order to make 40M behave. I wouldn't be surprised if you get 400 watts with only three tubes. One less tube might make the RF drive, and Rf capacitance issue more manageable.

Good Luck...

Jim
Wd5JKO

Thanks for the comments, Jim.

So you think the 4D32s are more apt to take off than the 6146s?   I loaded each grid with a lower swamping resistor this time - 4.7K in series with a .003 cap to ground. I hope this will help.  No neutralization.

Well, I just discovered the space-out of the year. I tried the filaments and nothing happened. The grid meter was negative.  I wired the sockets wrong!  The pins are off by one pin rotation.  So just stripped it down and starting over.  Isn't life grand?

T

Tom, a case in point. I once took a pair of 7591's in a Central Electronics 20A, and could make > 50 watts out on 160-40M. These were low capacitance tubes, like 0.2pf Cgp like a 6146. Later I tried a pair of EL34's, and I had > 60 watts on 160M, and instability on 80-40m. Removing one tube made 80m work fine, but 40m was still a no go. The circuit wasn't easy to add neutralization, so I skipped that step. The EL34 had over 1 pf Cgp, and that was the deal breaker.

The 4D32 according to Raytheon has 0.4pf Cgp, 30pf Cin, and 16 pf Cout.
The 6146B according to RCA has 0.22 pf Cgp, 13 pf Cin, and 8.5 pf Cout

Those grid swamping resistors are in effect all in parallel....How big is the driver? I suppose that method could work with a big triode too, but even if it did all that Cgp regeneration might skew the modulated envelope linearity a lot. Later on, playing around with the 20A, and a single EL34 in the final, I added a N/C circuit, and it was set correct when I had a nice trapezoid on 20M.

I bet two of the 4D32's will do what the quad of 6146's did. Might start with two, and work your way up to three...then ponder the trip to four...

73,
Jim
Wd5JKO

Very interesting about your experiences with those tubes, Jim.

The quad 6146Bs worked perfectly on 160 - 40M, so at least there is a good baseline to start with.

I didn't know that the small C g-p  could affect linearity via regen feedback.  I'll run some linearity tests when it's running and see. I could always add neutralization later. Right now I am using a broadband toroidal input to the grids for easy band switching.

Well, it should be a challenge to get working with all four tubes. I would have laid it out in a square, but got painted into a corner with this small box and layout. The tubes will have plenty of air though.

Should have it fired up tmw with some luck and will let ya know how it goes. Yes, 500-600 watts would be very cool for a little rig like this. 

T

Well, I fired it up and it actually puts out power.   There was a slight tuning instability, but not bad. It put out 700 watts pep with the modulator bypassed. I MIGHT need to add neutralization, but not sure yet.

I need to reduce the tank impedance to about 1/2 since the 4D32s are drawing more current.  These tubes certainly generate more heat.  My filament voltage drops to about 5.5 volts, so this needs to be corrected too. (6.3V is normal)   The screen dropping resistor needs to be reduced to 1/4. Screen current is very low, thus low power out.  This should be capable of  2KW+ pep output (audio peaks) when optimized.

More tmw.

T

Pics with QUAD 4D32s installed:

Did a lot of changes today.

1) Swamped the 4D32 grids each with 2.3K =  600 ohms total input impedance.   Now very stable but harder to drive.  I have an MRF-150 solid state amp dedicated as a driver, so no problem.

2) Replaced the tank coil with a heavy duty silver plated coil.  The old coil was underrated for the new job. I reduced the tank impedance for higher Q for the 4D32s.

3) Padded the screens from 200 pF to 400 Pf for each tube.  (Better stability)

4) Tapped the filament transformer to give a measured 6.2 V.  (Close enuff)

5) Replaced the 15K screen dropping resistor with a 5K, 30 watt unit.

6) Made shorter and more direct connections to the RF mounting plate.


Results:

RF tuning is rock solid with a peak in screen current, dip in plate current and maximum power coincidence.

Easy 500 watts of carrier on 75M.

Audio looks good and stable. No parasitics or other instabilities.


Problem:   (now solved - 6LF6s are OK for the job))

The four 6LF6 PDM modulator tubes cannot handle the peak power needed to fully modulate a 500 watt carrier. The rig will do 1800 watts pep at 1700 volts, but I don't dare raise the voltage higher. So, I have throttled the carrier down to about 350 watts which allows for excellent positive peaks.  In hindsight, four 4D32s in the final really need six 6LF6 modulators OR four 4D32 modulators.

The rig is running very FB now, and I will do some thinking about replacing the 6LF6s with 4D32 modulators.

This is like putting a bigger motor in a car and discovering that the infrastructure can't handle it anymore.

T

Check out the new beefy tank coil. I notice efficiency has gone up somewhat. Lots of air on the finals. I can touch the tube glass after a transmission. The transformers to the left are the floating fixed grid bias supply and the filament transformer - both wound with HV wire on 120V Variac cores - and insulated from chassis for 2KV.

How much are you getting out on six meters?

Nothing that I can measure, thank goodness.    Are you suggesting that I neutralize it?


BTW, I'm finding I can't trust my peak reading Bird wattmeter. At 500 watts of carrier, the REA software mod monitor shows 140% - 160% modulation.   The Bird does not show the equivalent audio power peaks. Probably too slow or something.  The scope shows heavy modulation too.  So guess I may be OK using the four 6LF6 modulators after all.  The 6LF6s do show some serious screen PDM current under heavy modulation.


The 4D32 finals show no plate color at 500 watts carrier.  They are very deep into class C.  The 6LF6 modulators show no color in class D switch, as expected.


T

I have a Bird 43 wattmeter.  The peak detector is very poor - as are virtually all power meters.  Stu AB2EZ in fact pointed this out to one of the manufacturers and apparently they agreed.  Stu can tell you more, but suffice to say, having designed a number of them, a good peak detector is not necessarily an easy thing to design or implement.

Nah. You're doing great.

Never trust any meters over 30.

Ok, I have not been following this thread at all, but you have sufficient modulator capacity.  I looked up some ratings on the 6LF6 and from what I could find, the tubes will do more than an amp of peak plate current, 1.4A to be exact (from the rating sheet I read anyway).   So, if this is derated a bit to - say - 1A, at 1700 volts total DC, that's a whopping 6800 watts.  It should be enough ----

But you may not be able to realize this power level (at least not with good linearity) unless there is reasonably significant analog compensation in the modulator (again, I have not been following the thread, so maybe you have analog compensation).   Tubes are poor switches and need help.  Analog compensation increases the drive level of the pulse train as the tube is required to produce more current.  So, longer ON pulses are also driven at a higher amplitude (this is positive modulation).  Shorter ON pulses are driven at a lower amplitude (this is negative modulation).  In fact, at 90% negative and above (in other words, getting close to 100% negative) the tube should be operating pretty close to an analog series modulator because the drive will be very low at this point.

So, real numbers... With 1700V, I would personally run the RF amplifier at around 570VDC.  Maybe that's where you're running it....for 500W out at, say, 75% efficiency that would be 667 watts in, which translates to 1170 mA of plate current.  I think 4  4D32s will do that all OK.  570V will leave a LOT of room for positive peaks.  Not quite 200% positive but a bit more than 175% positive, and that's still pretty good.

If the rig will properly modulate a triangle wave up to the voltage limit of the power supply, then all is well  8)

OK on the 6LF6 tube capabilities, Steve.   Yes, it looks like they will work fine.

Tmw I will measure the finals high voltage and see where we're at.  I set it up by feel and will run some more tests. 600V across the finals seems reasonable to give plenty of headroom.

That Bird peak reading wattmeter is really a fooler. Not only are the peaks wrong, but as the audio frequency is increased, the readings are lower. Whereas the REA shows full modulation at all audio freqs.

Frank and I discussed analog comp a couple months ago and I was able to make some improvements to get the negative peaks close to =95%.  I run the NPL to cover it after that.


If I were to increase the PDM carrier ratio somewhat, I'll bet I could see 600 watts carrier and still see 140% positive modulation.  Those 4D32s have so much emission it's scary. Same wid the 6LF6s.


T

Of course the real acid test with respect to linearity is the triangle wave.  What do your triangles look like up to at least 150% positive and down to 95% negative?

Analog compensation is necessary for the positive peaks as well as the negative peaks.  Tubes aren't great switches, so the idea is to keep the voltage drop in line as the amount of current increases.  Driving the tubes harder during positive peaks is just as important as driving them at a lower level as the negative modulation increases.

When I was doing tube PWM design, I actually never hooked anything to the rig except for a triangle wave generator until I could prove linear modulation up to the positive peak capability of the transmitter.  It was truly amazing all the places where the linearity could be compromised including the RF amplifier itself!!  Those non-linearity gremlins sure find a way to sneak in  ;)

Steve,

The last triangles I ran after optimization looked perfect up until saturation using the old 6146B finals. (140%)   I will run some tests today using the new 4D32s.

The thang that concerns me is that I am not seeing much more PEAK power using the 4D32 finals compared to the 6146Bs.  The flat-top point seems to occur at the same place.  At flat top, the 6LF6 start drawing big grid current and the 6LF6 plate current starts to dip. I am using regulated screens and grids for the 6LF6s.  It seems the 6LF6 plate voltage is dropping below the screen voltage, thus saturation.   Other than increasing the overall HV, what else can be done?  I have 2200 volts available, but hesitate to ramp it above 1700v.

A friend said the following concerning the 6FL6s in PDM service:

"The 6LF6 can deliver 1.4 amps peak but that is running a low duty cycle. PDM service is very different. Say your saturation voltage is 100 volts. At 1 amp peak current the tube dissipation is 100 watts peak. Say you are running .5 amps peak per tube with 100 volts saturation voltage the dissipation is now 50 watts which is a lot closer to the tube ratings. You should be able to get away with 4 tubes if you blow plenty of air across them.. Time will tell to see if the emission hangs in. One 6DQ5 would not fully modulate two 6146s (equal to 1 4D32) so maybe the 6LF6 has more poop."


So maybe I am running out of headroom with the four 6LF6s, I dunno.   I can run 500 watts carrier now, but the peak power is the same as before.  I will run some more tests today.

T

That's interesting about the 6DQ5.  I had a Viking II, out of which I removed the analog modulator, put a bridge around the power supply (with a cap input filter), and used a 6DQ5 as a pulse width modulator.  The 6DQ5 *did* modulate the pair of 6146 tubes without issue.  That was many years ago, so I cannot tell you the parameters under which I ran the RF amplifier.  I might have used a pair of 6DQ5s, but I don't think I did - doesn't ring a bell.  Probably didn't own 2 6DQ5s at the time!

Another rig from back then used a single 4-400 pulse width modulator to modulate an 813.  5kV total supply voltage; 2000V @ 200mA on the final.  The 4-400 is *not* a good pulse width modulator.  It is a low current, high impedance tube, but with coaxing it did the job nicely.

Also designed a transmitter using a single 4PR-1000 to modulate a pair of 450TLs.  The modulator was not particularly bigger than the RF amplifier there either.  That transmitter had a 10kV power supply, and was designed to modulate up to 150% positive.  This was a proof of concept for a broadcast transmitter I was designing way back when (another story for another day).  At 500mA of plate current, that transmitter was absolutely linear up to 150% positive.

Your modulator is pretty hefty, and I would expect it to be able to modulate the power levels you're expecting.

I do know that with tetrode modulators, I ran as low a screen voltage as possible, with as much grid #1 drive as possible and then used analog compensation to make it happen at high levels of positive modulation.

The thing is, with PWM, the tubes are run at peak current and peak dissipation values that are higher than the average values, and by a fair amount at times.  With the analog compensation operating, the grid #1 current on modulation peaks (positive) was considerably higher than the maximum average grid current from the data sheet, but the actual average grid current (and grid dissipation) was not, and the modulator tubes didn't seem to show any ill effects.

Unfortunately, with PWM and tubes, high voltage is your friend with respect to power out, efficiency and modulation capability, which is why I used 10kV on that 450TL rig and 5kV on the 813 rig (and about 2200V on the Viking II, as I recall).  But high voltage is NOT your friend with respect to construction complexity and all of that..

It's a non-trivial project, but you are keeping at it, and eventually will get it !  I can honestly say "I feel your pain" having going through this iteration myself !!!!

Interesting.

You used 2200 V  on the Viking rig? That's encouraging, cuz I'd like to hit that level too.    I've been creeping up the HV on my  4D32s and 6LF6s and now up to 2KV.  Maybe I'll go to my Variac limit of 2200V.  It certainly runs well at these levels.  I was just testing at 500 watts out and seeing voice peaks of 140%, so it is working.  With a nudge of the carrier ratio, I am seeing 600w but slightly lower % modulation.

I'm not sure where the breakdown weak link might be...  if the tubes can handle 2200V, I can make the surrounding infrastructure do it too.

It's really a fun rig to play with - being small and compact, all tubes, efficient and a real strapper considering the modest tubes used.
 
T

More testing with the triangle wave at 100 Hz, 1Kz...

The triangle and linearity looks very good at 1600V up to 160%. This is at 300 watts carrier out.    But when I ramp it to 2000V and run 500 watts carrier, I cannot get above about 110% positive without the triangle going non linear and the peak gets rounded. There is a power saturation limitation in there somewhere.  

When I push the rig to this non-linear point, I see the 6LF6 grid current start to soar and the 6LF6 plate current start to dip down. Normally they are both constant.   The maximum peak power capability is not much different than when using the old 6146Bs in the final.  

Today I will disconnect the PDM filter and put a simulation power resistor load across the 6LF6s to conquer and divide. We'll see if the peak saturation is related to the PDM filter, RF finals or the 6LF6 PDM stage itself.  

The rig presently runs beautifully at 250 watts at 160% positive as is, but then again, I could have kept it as a 6146B rig and done the same.  So I gots to get this solved even if it means putting in four 4D32 PDM modulators.  I wanna see that peak audio power soar unlimited like an eagle, caw mawn.


More later.

T

Ok, well don't change ANYTHING until you know what is actually happening  :D

I would first put a scope on the output of the PWM filter.  Look at that waveform.  Does it round off there?  If so, the problem's in the modulator.  If not, obviously the RF amp is at issue.  How LOW does the waveform at the output of the filter get (this would be the positive peaks) before you see the waveform flattop?

Then, I would look at the PWM waveform at the INPUT of the PWM filter.  You may need to construct a special H.V. probe to do this.  I used a resistive voltage divider (with non-inductive resistors), and a "gimmick" capacitor across the series resistor to make the probe's frequency response flat to at least 50 megacycles or so - all of this looking into a standard 10x scope probe.  It's important that the probe not change the waveform in any way.

So, after all of that business about the probe, what does the waveform look like at the input to the PWM filter?  How LOW does it go when the tube is turned on.  Does the LOW voltage (tube on) value change with loading and/or with pulse width.  If you put the triangle wave into the modulator while observing the output, do you observe a variation in the ON and OFF voltages of the PWM waveform (measured at the input to the filter)?  Ideally, you should not, except for whatever the analog compensation is doing. In the absolutely ideal TUBE pwm world, the ON voltage at the input of the PWM filter might even be a bit lower (tube turning on harder) during positive excursions.  Usually, you cannot get this, but the ON voltage not RISING during hard turn ons is very important to the generation of positive peaks.

These are important tests, and will reveal much.

Regards,  Steve

Excellent advice, Steve, tnx.

Good results so far....

Tests showed that the output of the PDM filter was clearly showing PDM modulator flat topping, so the problem is definitely in the modulator as expected.

I increased the 6Lf6 screen voltage from 85V to a higher level - and reduced the 6LF6 grid voltage from +1.5 to ground.  This resulted in more 6LF6 plate current, though hopefully the four  can handle it. (500 ma PDM, meter pinned)

The good news is that the peak power has now risen from about 1500w to about 3000w before flat topping.  I plan to put a bigger 6LF6 plate meter (1A) in there and try again.  


The good sign is I am now getting a clean, straight,  brickwall flat top of the positive peaks. Before it was a squirrelly rounding look.  I feel the modulators are now saturating hard, as they should, at maximum peak power.


Before this modification, I unplugged two of the 4D32s.  Using the limited modulator, I found that I could get out the same RF power as four tubes. So, the final is OK. It just needed a modulator that could keep up with it.

I may add a screen meter to the 6LF6s too.  The problem is that I don't know what currents are reasonable for the PDM grid/screen and plate set of 6LF6s.   The RF 4D32s are drawing 1A  RF carrier current and at the same time the 6LF6s are drawing 500 ma of PDM current.  The grid is at about 5ma now that the screen voltage has been increased. Who knows what the screen current is now.


More tests. Looking better - just don't want to blow the modulators until I meter them better.

** Update:  At 300 watts carrier, using a negative peak limiter, I am able to get the REA mod monitor to flash at 190% voice modulation.  Now THAT is a breakthrough!


T

After a long test into the dummyload at 400 watts carrier and 180% voice modulation at 1900V, the 6LF6 modulators show no color and there is very little heat being blown out by the fan. This is at 500 ma of modulator screen current and about 1A of plate current..

The 6LF6 grid draws only 5ma, which seems reasonable. Drawing more grid current doesn't seem to help anything.   The screen current needs to be measured, but I plan to set the screen voltage at the minimum needed to do the job.

So looks like the problem was that more modulator saturation was needed via more screen voltage, thus more plate current.   All other parameters are basically where they were before.

T

Check out the 6LF6 tube under "Average Pulse conditions"

http://tubedata.milbert.com/sheets/084/6/6LF6.pdf

I see no screen dissipation  power rating.

In my rig, for the screen I am measuring 500 ma at 97V. (125ma per tube)  This is what I need to get decent saturation performance. (48 screen watts, 12 watts per tube)

They are talking about 275V screen maximum  at 40 to 70 ma.    275V at 40 ma = 11 watts per tube.


My current at  125 ma per tube seems excessive. But the power in watts is only 12 watts.
Is this reasonable or are the 6LF6s going to see a short life due to screen burnout?

The grid and plate currents are well within specs.

T

Sounds like very good progress indeed!

So, what are your operating parameters and what is PDM current a measure of?

For instance, in PWM transmitters I measure the following:

Total power supply current (that may be your PDM current)
Total power supply voltage (that's probably your 1900 V)
Load current (this is the current as measured at the output of the PWM filter and before any loads)
Load voltage (this is the voltage measured from the output of the PWM filter to the load common point, which may be the H.V. supply or ground, depending on the topology. In solid state, the RF amplifier common point is usually ground, but in tubes it's usually the H.V. supply)

So, what are the operating parameters at carrier (no modulation)?  Be very interested to see the numbers (and hear the rig on the air, too!).

Been kind of off the air for the past number of days, as I have a very academically oriented job interview coming up (in other words, I'm going to be interviewed by much younger, recently graduated computer science majors who are going to ask me about the finer points of polymorphism,  multiple inheritance, virtual classes, and all other theoretical aspects of object oriented programming because that's all they know at the moment) even though I've been doing it for decades!   A friend of mine who is a very brilliant software engineer was recently submarined in a similar manner by a really irrelevant interview question related to generating Fibonacci numbers???!!!??? (and this was relevant to anything?? - NOT)  Suffice to say, I've pulled out my books and reference manuals and have been going over some of the finer points of OO design  :P

OK Steve -

Well, at this point everything is running FB at my targeted power, so getting into the specific parameters would be a waste of time at this point.  I do plan to finalize everything and post a schematic later on. A few guys have requested one.

All I need to know now is if the 500 ma modulator screen current (125 ma /  11 watts per tube) is excessive and will burn out the 6LF6s. Please re-read my last post and give me your opinion.

Once I get a few more things buttoned down, I think this rig is a wrap... ;D

A good luck at the interview.  Fibb numbers,  1,1, 2,  3 , 5, 8, 13, 21, etc  are simply the sum of the last two digits. Fibb ratios, 0.618, 1.618, etc.  The mystics try to make a big deal out of its predictive value in the markets. BS.

T

More info for the future builder's archives:

Getting ready to button it up.  Still need to wind a smaller PDM filter later on.

After  optimizing everything, the 6LF6 screens are at 110V, 500 ma = 55 w / 4 = 13.75 w per tube.   Not sure how to reduce this voltage, unless I went to bigger tubes that cud handle it like the 4D32s.  If I go below this  screen voltage I see premature flat topping.  I am told  that 7 watts per tube is the recommended dissipation  number, so I am above it. .
 
That screen current is there even if I reduce the plate voltage and run PW, of course.
  
 
Man, it’s really playing now.  At 1600V I get 300 watts carrier out with 200% positive modulation!    At 2200V FULL STRAP, I get 600 w at about 170% positive.   Nice and clean wid the triangle tests, all the way to maximum audio %.   Swept audio response is like a Flex... perfect, no transformers.  (1 Hz to 10 KHz, the PDM filter limit)
 
I will probably run it at  300w-400w most of the time.
 
It's really is a neat way to go for a summertime rig.    
 
T

The 4D32 is a great tube for the money.
Tough, I run three at 1250 volts on the plate, get 300 watts carrier out, with no color and they do not seem to get hot at all.
1200 to 1500 watts pep, do it all day long for years.

Hi Tom,

I was interested in the parameters mostly because I was trying to calculate the efficiency of the modulator.  The total supply current SHOULD be related to the ratio of the total power supply voltage and the voltage supplied to the load.  So, in theory a 50% duty cycle waveform should result in half the power supply voltage at twice the power supply current at the output of the PWM filter if the modulator were 100% efficient - which it is not (and none are).  So, more current is added to the supply side and that additional power is used up in the form of heat in the modulator.

Ok, to solve your screen dissipation problem use analog compensation on the modulator screens (I had to do this with the 4-400 PWM modulator screen grid for the same reason).  In this way, the average power dissipated by the screens will be a whole lot lower and it will be averaged out over the modulating waveform.

You provide a higher screen voltage when it's needed (under high positive peaks), and a lower screen voltage when it's not needed.  In fact, it will probably help to further linearize the modulator.  The quiescent (no modulation) modulator screen voltage can be reduced and keep the dissipation at a safe level.

Use a source follower MOSFET preceded by a MOSFET voltage amplifier, etc..  make the DC level and audio gain of the screen analog compensation independent of each other so you can get everything adjusted to a fine degree.  Such a system should work FB!

Yo Steve -

That sounds like a great idea to lower the screen dissipation and provide some analog compensation.

Could you draw up a circuit using 11N90 MOSFETs and post it here?  If so, I will build it up and give it a go.  There are a few others who have contacted me regarding schematics for this project, so it could be a benefit for them too.

Thanks.

T

After adding the four 4D32s, the rig started having key-up instability problems on 40M.  I figgered it was time to add a little vacuum variable neutralizing cap to get the rf final neutralized.

I sampled some RF off the 4D32 plate and fed it back thru the 30 pF 10KV cap -  and into the bottom of the broadband input toroid, which then feeds the grids. It worked like a charm and I was able to "cold neutralize" it by tuning out most of the feedthru signal from grid to plate.

The rig keys very cleanly now on all bands, though it needs slightly move RF drive since neutralization is really negative feedback.  I can now take out some of the grid swamping that kept it stable before.

Neutralization is always a good policy, especially when using multiple tubes.

T

Look for the tiny vac variable mounted on a Plexiglas board with a tuning knob, to the left of the tubes.  I have been experimenting using three final tubes instead of four.  

Seems like the latest JJ project is almost a rap!  I am sure its been very rewarding and as always, I like to follow along with what your doing.  Everyone learns from tracking your progress.   

As I have said before, its nice when a problem is found a solution is developed rather than totally switching gears.   

Joe, GMS 

Tom

The plate bus-to-ground voltage is AM modulated RF.

Capacitive coupling between the plate bus and the grid bus (due to each tube's plate-to-grid capacitance) puts AM modulated RF between the grid bus and the cathode bus.

So controlling this via neutralization (or other means) becomes even more important in an AM modulated transmitter.

It is also important that the input (grid drive) circuitry, including the grid-to-cathode, etc. capacitance of the tube(s), not make the grid bus-to-cathode bus impedance too high...and definitely not inductive.

The is why, in transmitters (for example, a Ranger or a Valiant) with a tuned circuit on the input side of the tubes (i.e. on the output of the preceding driver stage), one must use a relatively small value of coupling capacitance (typically 2 or 3 times the combined grid input capacitance of all of the parallel tubes) between the input tuned circuit and the grid bus.

Stu

Joe,
Yep, we think it's a wrap and then another problem or inspiration rears its head.  Jeff / W2NBC hit home when he said it pays to refine - refine - refine.   It took more patience than I had in the past to continue working on the same rig. It was always "NEXT!"

This time I built Fabio II, got it working well and then came back six months later and spent another 3 months refining it. Same with this Dual Quads rig... finished it and later came back and made some major improvements. That's the key... make it a work in progress and it will result in a very nice product in the end.



Stu:  Interesting on RF neutralization reducing the plate voltage from undesirably modulating the grid. I never heard that before.


Your comment about isolating RF driver stages better...  I am using an MRF-150 solid state amp to drive a broadband toriodal transformer into the grids of the 4D32s.  Do I still need to take some isolation precautions using smaller capacitors?  Due to the isolation from HV requirements, I now use 500 pF doorknobs to couple both transformer leads to the grids and the common floating metal plate.

T

Tom

I think that since your RF driver presents an impedance from the grid bus to the cathode bus that is less than the impedance of the sum of the input capacitances of the tubes... you don't have to use a smaller input coupling capacitor.

Also, the neutralization will greatly reduce the amount of modulated RF coupled from the output to the grids.

Somehow, in the case where the voltage between the grid bus-to-cathode bus (i.e. the input voltage) is provided by a driver having an output tuned circuit with a high impedance, the use of a relatively small coupling capacitor keeps the impedance from grid bus-to-cathode bus from getting too high. Too high an impedance will result in too large a fraction of the plate bus-to-ground RF voltage appearing between the grid bus and the cathode bus. I.e. there is a voltage divider formed by: the impedance of the tubes' plate-to-grid capacitances (adding in parallel), in series with the grid bus-to-cathode bus impedance.

I'm not entirely sure why this works as well as it does. I think that it may be a non-linear effect that involves the diode between the grid and the cathode.

In any event, using a 33pF coupling capacitor between the input tuned circuit and the grid of the amplifier tube... instead of a 330pF coupling capacitor... fixed an oscillation problem in a 2-tube transmitter that I built recently. That transmitter employs a 6CL6 crystal oscillator, with a tuned output, driving a plate modulated 6550.

That little transmitter works great, with 300V B+ and 8W of RF output at carrier. The modulation is very linear. I use a modulator with a negative peak limiter. Positive peaks easily exceed 125%.

Stu

Sounds good, Stu.  

I'm finding the neutralization required more than just a cold neutralization alignment.  When I decreased drive on 40M I saw some fuzziness on the carrier.  By fine tuning the neut cap trial and error, I was able to eliminate this last remaining instability. A few extra pF was required.

The 4D32 broadband input works FB, but it is not always 1:1 across all bands. So I made up an unbal to unbal T-match tuner that takes the solid state amp and matches it to the broadband input. This tuner previously had a smaller bandswitch that smoked, so I added a bigger one and more strapping coils. 

The 4D32 rig is putting out the same power on all bands now and is sounding sweet. Really pleased with the rig.  I still think the 6LF6 modulators are limiting the quad 4D32s somewhat, but not by much.

T

The input tuner - still needs labels. Bandswitched 160 - 40M:

Here's a fitting epitaph to complete this project.  

I added a CRL SEP-400 multi-band compressor followed by a CRL PMC-300A  independent negative and positive peak limiter to the PDM rig. These are 1982 vintage processors.    

Here's a short recording made by Jeff/ W2NBC on 75M today.   For once, I am very satisfied with what I am hearing and plan to call it a wrap.  The processing was the final touch needed to this transparent PDM audio platform.   Just what I've been trying to achieve.  (Assuming what I hear on my own playback computer speakers is accurate.)

T

There's nothing like a professional quality W2NBC air check to reveal the truth. Sounds real clean Tom. I always favor a nice push in the presence rise area that adds clarity and you've got that just right in my book. Sounds very natural and well balanced. So... now that it's perfect, once it has an official "Vu" name we probably will never hear it again since you always tear your creations apart once the challenge is gone!

Rob W1AEX

Hola Rob -

Tnx for the "smooth" report. Glad it's not "tubby."   ;D


Yes, Jeff / W2NBC has helped immensely with setting up my audio chain and dialing it in.  As he says, it can become a hobby within a hobby.   Those who have advanced to the point of setting up low level audio easily can forget how difficult it can be for those of us who still struggle.  There are so many techniques required that we should probably have some more tutorials posted.

Though, it depends on the person and many are happy with a D-104 and that's it. Whatever makes us happy.

One thing I notice:  When I listen to TV and FM announcers, (and many good sounding commercial AM stations) I hear very muted sss's and ch's. It's like these consonants are barely there and are VERY compressed.  In contrast, if you listen to most hams and my own recording, you will hear vary hard ss's and ch's.  What's that all about?   Is it my own imperfect teeth structure - while announcers are selected for perfection?    My own sss's are not spitting and sound clean, but are still very harsh in contrast.

Maybe they are hammering on the D-esser. I have one here but try not to use it.

Or maybe this is the difference between tailoring our audio for communications in marginal conditions vs:  rock-crushing broadcash signals for hi-hi FB entertainment.


T

If you listen closely to current AM broadcast, the audio processing model is to cram as much spectral energy into NRSC standards and poke a hole in your speakers while driving to work.. The bottom line $$ is to be HEARD. The result is some pretty grungy sounding stations, even the flagships from NYC:

All were recorded on an SP-600 off the diode load at 13 kHz- The first is CBS 880, followed by 660 WFAN, and then a familiar voice.. There is no comparison  ;)

Hi-Fi isn't on your AM dial any more. It's found on your HAM dial!

I guess the only way I can explain AM broadcast audio on the male voice, is that it always seems to have a metallic note superimposed on every word, as if all energy is directed to the tweeter. 

Here in the Albany area we have WGY AM/FM. It's a 50K station. The FM side sounds a little better but has the same note. I never heard anyone sound like that in person and have only heard two hams stations sound like that.


Your super smoooooooooooooooooth on your audio Tom 8) 8) 8) 8) 8) 8)

You get six COOLS, My max rating!

Jeff and Tom,

I listen to WCBS 880 in the mornings and now and then throughout the day if I want to catch one full loop of the news. When Pat Carroll does the station ID during her part of the news loop she hits the "C" in WCBS really hard and it comes searing through in such a way that it sounds a lot like the sibilant sound that a station makes when you are tuned off frequency slightly. I've often thought that the station engineer must cringe every time she does that. He probably has to lift his headphones off his ears to let the blood drip out as well.

Rob W1AEX

TV audio pretty much has a ton of compression.

Most of the FMs out of NYC are using a gizmo to make the highs have more fuzz.

Eddie Van Headache would be proud.  

The worst thing someone can do IMHO is to boost bass when they don't have the pipes in the first place.

If you don't have the pipes start smoking or embrace your inner Mickey Mouse.

Sorry, life is nought fayr.   

Thanks for the comments, guys.  Jeff, that comparison to the AM broadcash stations using the heavy processing is quite an ear opener....  ;D   I definitely prefer a more mellow sound for ham radio.

Well, after a lot of time trying various audio chains and running lots of tests on and off the air, I've finally settled on a simpler lash up.

I tried running a 528E > 31 band EQ >  CRL SEP 400 multiband compressor > CRL PMA 300A limiter >  PDM generator or solid state audio driver.  This chain worked very well, but was a handful to keep optimized and I found it stunted the  positive peaks. This was probably due to all the EQ/ phase changes in the chain.  I currently prefer higher positive peaks vs: symmetrical audio.  There was also a lot of background fan noise and a "busy" processed sound that comes with compressors.  I also diddled with a Berhinger 9024 digital multiband processor. It worked "OK," but gawd, the menus are a nightmare... give me knobs.

Bottom line is I recently replaced the SEP 400 and PMA 300 with a simple multiband limiter called the Dominator II. It is an extremely fast, transparent, multi-band peak limiter.   Now I can easily dial in moderate density in a flash, keep high positive peaks - and I like the cleaner natural sound. And most importantly, I do not need a negative peak limiter. With -95% negative, my voice naturally produces + 120-140%  positive peaks, depending on frequency. So I can run reasonable audio levels and be clean without banging on an NPL.  

The chain now is:   528E preamp > 31 band EQ  > Dominator II multi-band peak limiter  >  MOSFET audio driver or PDM generator

I like the natural sound very much now and feel like I have full control with less knobs. In addition, the tube PDM rigs takes exactly the same audio settings as the 4-1000A plate modulated rig. This has never occurred before and makes life much easier to switch rigs.

Just wanted to update the overall progress. As Jeff says, audio can be a hobby within a hobby.   ;)

T