The AM Forum
March 29, 2024, 09:53:43 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: 1 2 [3] 4 ... 7   Go Down
  Print  
Author Topic: Looking for ideas for a ~25 watt plate modulated homebrew Pissweaker rig  (Read 62268 times)
0 Members and 1 Guest are viewing this topic.
KD1SH
Member

Offline Offline

Posts: 739



« Reply #50 on: September 06, 2020, 06:25:55 PM »

  Like mopeds - too fast for the side of the road but not fast enough for traffic - 100 watt rigs are awkward: too much to drive an amp but not enough to strap on their own.
 
Logged

"Gosh, Batman, I never knew there were no punctuation marks in alphabet soup!"
—Robin, in the 1960's Batman TV series.
KD1SH
Member

Offline Offline

Posts: 739



« Reply #51 on: September 06, 2020, 06:36:46 PM »

Hi, Larry,
   I remember you mentioning your Cheyenne the other day on 6m AM. I've heard that while they're very similar to the DX-60, they're not exactly the same. To me, it's a much nicer looking rig than the DX-60.
   Hope to see you at Vintage Radio next Saturday!

I just dug out my Cheyenne transmitter the other day, basically a DX-60 with a built in VFO. I need to come up with a power supply for it first. Bill, we were talking about this rig a couple weeks ago on 50.4, you had inspired words to say about the screen modulation which I have not done in decades. Also, thinking of modulating my "BC-3.75" mini hybrid mil rig, and thought of suppresser grid modulating a 6AG7 also. How much power did you get out of yours Mike?
Larry
Logged

"Gosh, Batman, I never knew there were no punctuation marks in alphabet soup!"
—Robin, in the 1960's Batman TV series.
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #52 on: September 06, 2020, 10:17:41 PM »

What about a 4-125 modulating a 4-65 Smiley


Or a 4-250 modding a 4-125?

Not a huge fan of paralleling tubes.  If you need more than one tube, you need a different tube.

--Shane
KD6VXI

Shane, thanks for your suggestions.  I agree it is preferable to use one tube instead of several in parallel for the RF output, for many reasons - stability, neutralization, output impedance and capacity especially on the higher HF bands......   But here we are paralleling the series modulator, which in this  configuration actually offers more benefits  than running a single tube.  Providing a lower impedance source via parallel tubes provides better capability for providing the necessary plate current on positive peaks, and reduces the overall dissipation by allowing less voltage drop in the modulator for the same output voltage and current, thus making the whole affair more efficient.  They are easy to balance through cathode degeneration, as well as easier to drive as lower audio level is required.

The Eimac series of tubes you suggest are fine players at both RF and as modulators, but they sing much better at significantly higher voltages than the sweep tubes and the Raytheon 4D32, and when putting them all together in series, it would require a much higher overall power supply voltage to produce similar power levels with the 4-65 and the bigger siblings.  This would make components more expensive, require more complex isolation, and probably a larger package for the same overall performance in the series-modulated architecture under consideration here.  And  it appears we are still focusing on a low-end power level for either stand-alone QRP AM, or conveniently driving a linear to full scrote while getting the lead (-er Iron) out of the signal path.  What's not to like?
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #53 on: September 06, 2020, 10:35:06 PM »

This is exactly how I like to design stuff. Come up with a possible design on paper and then try to improve on it. Keep trying different ideas until there are enuff reasons to stay with the latest one.  No harm trying stuff on paper.

From what I am seeing, there are several reasons to stick with the quad 6LF6s series modulating a 4D32. The sweep tubes are simply better suited for the job.  My first reason for this rig is as a driver for the various linears. It will require 10-30 watts carrier. The sweep tubes are best suited for this. Second, I want a pissweaker rig. I have enuff 100 watt class rigs, but no PW rigs. The sweep tube modulator is best for this purpose. The 3-500Z modulator would be more suited for higher voltage as Rick says. I think keeping it optimized in the 10-40 watt range is the best choice here.  We can design up a matching MOSFET driver suited to this lower range.

And finally, I get the impression there are a lot of guys who might like to build this rig once proven, working FB and there is a schematic and operating procedures for it. Whether as a stand-alone 35-40 watt AM rig or linear driver at 10-20 watts carrier, this will make a great RF source with transparent audio.  Making the modulator a 3-500Z would knock some guys out of the game due to cost and generally larger parts needed..  It would also ruin the pissweaker mystique by making it 125 watt capable.

So, I will draw up a basic schematic for quad 6LF6 series modulators and a single 4D32 final and then get back with Frank and see what he can come up with for a simple MOSFET audio driver.  It looks like we will be grounding the grids of the 6LF6s and directly driving the screens with the MOSFET driver. No audio transformers anywhere. This will make it more stable and will simplify things further.  We might look into audio NFB (neg feedback) later on.

Tom, I agree on all counts!  I think together we have all shaken this in every conceivable direction and narrowed it down to a practical and fitting solution that addresses your original goals.

It will be above the pissweaker class enough that it will be usable on a regular basis, with ease of enabling the afterburner without switching to a different rig when a bug needs a good squashing.  It is practical in size, weight, and friendly regarding parts procurement. The isolation issues are trivial with the voltages present.  With careful layout and  ventilated enclosure, I would expect convection cooling would obviate any requirement for a noisy fan.  Four sweepers should run much cooler overall than just one, or two in parallel.

Even though you will use the hybrid approach to provide power and drive (Baby Blue and the FT), there is no reason why others could not clone the core design in a stand-alone package.  A DDS with an intermediate amplifier could drive a single 4D32 in grid drive configuration with ease, and and the power supply requirements are low enough in current demands that even a modest transformer with a voltage doubler could provide all the energy needed while coasting.  The other power supply components are small and inexpensive, so this design can be taken in many directions by others in the future!  What a contrast to a single 6AQ5 modulated oscillator!

Looking forward to seeing this come together with more of Franks's design as all your ideas unfold!
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
km6sn
Contributing
Member
*
Offline Offline

Posts: 91


« Reply #54 on: September 07, 2020, 05:57:57 AM »

OK, Ok True Heresy coming up here. Fasten your seatbelts.

If you goal is to use hollow-state and see fire in the hole, that is one thing.

However, if you goal is to make an economical clean pissweak rig to use barefoot or to drive a linear, here is another approach.

http://amfone.net/Amforum/index.php?topic=46102.new;topicseen#new

It turns out two 74ac08 gates in parallel are an ideal drive for 100 ohm twisted pair to drive a NCP81074
fet driver on 80,40, and 20. For the RF deck, a single Transphorm  TPSH34208 (TPSH34206 would be even better) in class D is all that is needed. Power is scalable by changing the power supply voltage for the modulator and Class D RF deck.

In the project shown, everything is on .1 inch boards with ground plane, or on bare copper dead bug  (live bug, in the case of the NCP81074).

Driving a 3-500 one-holer, I get broadcast-quality audio reports.

Steve WA1QIX suggested changing to an analog modulator, and I think I will do that. Thanks for the idea Steve.

The whole shebang will use less power than the 3-500 modulator filament! Less shack heat and audible noise.

OK, thanks for bearing up under this Truly Heretical post!

Regards,

Rod

Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #55 on: September 07, 2020, 10:07:42 AM »

Rick,

All good points.  I was making that recommendation based upon the voltage levels with Baby Blue, which iirc was 2kv or so max.  The 4 series tubes would have made a killer match up if he would have wanted to make it a 'companion'.

I've run 6LF6s at 1.2kv before, they really get up and go!  Great tube for sure.  The 6LF6, 8950 and 8908 where my all time favorite "back then".

Never really messed with a 4D32 but from everything I've read it's another killer.  The gubmint must have loved it for a reason, so....

One thing is for sure:  No matter what lineup Tom chooses, it's gonna be a nice rig!

--Shane
KD6VXI
--Shane
KD6VXI
Logged
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #56 on: September 07, 2020, 12:37:44 PM »

Tom, we have both seen these links before, we just did not remember....

If these do not give you the ultimate confidence in going with the quad of 6LF6 sweepers, nothing will.  Just strap in your Raytheon 4D32, then STRAP!

73, Rick

The thread:

http://amfone.net/Amforum/index.php?topic=4821.0

Article:

http://www.amwindow.org/tech/htm/wb9eckseriesmod.htm

Hi-Res Schematic:

http://www.amwindow.org/tech/pdf/eckfinal2.pdf
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #57 on: September 07, 2020, 12:55:23 PM »

Rick,

All good points.  I was making that recommendation based upon the voltage levels with Baby Blue, which iirc was 2kv or so max.  The 4 series tubes would have made a killer match up if he would have wanted to make it a 'companion'.

I've run 6LF6s at 1.2kv before, they really get up and go!  Great tube for sure.  The 6LF6, 8950 and 8908 where my all time favorite "back then".

Never really messed with a 4D32 but from everything I've read it's another killer.  The gubmint must have loved it for a reason, so....

One thing is for sure:  No matter what lineup Tom chooses, it's gonna be a nice rig!

--Shane
KD6VXI

Shane, I agree those Eimac sibling tubes would sing for their supper with the Baby Blue 2 KV supply all out.  If Tom was just going for a full-strap rig, those would be good choices.  We kinda got off-track and escalated the ideas well beyond his original intent, and we are back in focus now.  Thanks again for your input.

Going the other way, with much lower voltages, one thing I would love to try would be a handful of 6336 dual triodes for series modulating another one or two.  The 6336 is like a pair of 5AS7/6080 triodes in the same envelope, with double the heater current and huge graphite plates.  Perhaps that would make a nice rig with a low impedance plate tank wound on toroids.   So many project ideas, so little time.
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #58 on: September 07, 2020, 02:23:57 PM »

I drew up a preliminary schematic for the four 6LF6s series modulating the 4D32.  Take a close look and see if you can add or change any parts values or find mistakes.  Especially check the floating HV area of the 4D32 and the returns to the cathode rather than ground.

I am not sure if the GFZ driver carrier adjustment will substitute for the cathode diodes or whether they will help with NPL action too.  The modulator 100K from plate to cathode will be dissipating some power... maybe 250K is better.  We will want to look into that closer.  Summer Breeze runs just one 4D32 and can do an easy 100 watts out at 600V. So I am thinking that a total voltage of 1500V will be more than enuff to dial in 40 watts of carrier out and have reserve for huge positive peaks if desired. I have a Variac on the HV, so no problem testing for the sweet spots.

It's really a simple circuit and has only a handful of voltages needed:  Two filaments, one HV, one fixed grid and the MOSFET driver voltage, whatever that works out to be. Amazing, really.

The 4D32 grid and screen meters will have to be floated in Plexiglas, though the cathode meter is at ground potential.  I would add a HV meter too making it a total of four meters.  The grid and screen current will have to be mentally subtracted to get a true plate current reading. The screens of the 6LF6s could also use a meter, so let's make that five meters.  The 6LF6 screen current data sheet shows 70 mA each. That is a lot if we multiply times four tubes. But I remember my quad 6LF6 PDM modulator hauling down 200 mA screen current to make big audio peaks. It didn't hurt anything. The MOSFET audio driver will be feeding the screens, so no big deal to haul down current.

The RF choke between the modulator and 4D32 cathode should be between 0.5 to 1 MH or so. Testing will show what is invisible to audio and still work right.

I added a 5KV safety cap at the RF input to back up the other coupling cap in case it fails.

I am lining up some parts as you can see.  It will definitely be four 6LF6s modulating a single 4D32. This has the potential to be my most hi-fi, cleanest TUBE AM rig, ever. With the coming simplified GFZ MOSFET audio driver we have all bases covered. I'm hoping others will build it up in the future too.

T

***  LOOK   FOR A NEWER UPDATED SCHEMATIC -   Some important changes on 9-8-2020, 9 posts down.  ***



* DSCF0010 (4).JPG (335.24 KB, 1280x960 - viewed 445 times.)

* DSCF0014 (4).JPG (330.08 KB, 1280x960 - viewed 444 times.)

* DSCF0012 (4).JPG (334.56 KB, 1280x960 - viewed 453 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #59 on: September 07, 2020, 02:29:42 PM »

More:


* DSCF0015 (4).JPG (323.76 KB, 1280x960 - viewed 367 times.)

* DSCF0009 (4).JPG (347.09 KB, 1280x960 - viewed 414 times.)

* DSCF0017 (2).JPG (325.26 KB, 1280x960 - viewed 373 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #60 on: September 07, 2020, 02:32:54 PM »

The pics are coming out better today:


Hollywood, Baby Blue and Summer Breeze all ready to fire up and do radio service...


* DSCF0013 (4).JPG (338.93 KB, 1280x960 - viewed 437 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #61 on: September 07, 2020, 03:41:52 PM »

Tom,

Did you see my posts referencing the links to the WB9ECK series-modulated rig?  (Reply #56 in this thread)  They should give you great confidence in your plans.

Your schematics look good.   Just a few comments....

I do not believe you are going to be able to set the carrier level using a diode string under the series modulator tubes.  Just ground the cathodes through the resistors,  The carrier set point will be a function of the amount of static voltage drop across the series modulators, and this is controlled by the DC bias component supplied to the 6LF6 screen grids via the MOSFET driver circuit.  Frank's design will likely include the components to be able to adjust the bias set point, and include the audio drive to the screens with the same circuit.  

I would also discuss  the T/R switching method with Frank, as you will probably be leaving the HVPS live during receive periods.  You show an isolated bias supply for the 4D32, and I assume that is cutoff bias, but you probably do not want the entire 2000 supply volts to appear across the 4D32 during receive. It may be best to have the series modulators also close to cutoff during receive, and have two resistors, one across the modulator, and one across the final, to share the 2000 volts, half across each part of the series circuit.  The resistor across the modulator will perform the NPL function, but the resistors will be large enough to be inconsequential with regard to dissipation, both in standby and transmit mode.  If the resistor across the series modulator is too low a resistance, you will not get close to the zero crossing on negative peaks.  It will also cause the 4D32 to endure the entire plate voltage supply during standby, probably not favorable.  Bottom line, you do not want excessive NPL current through that resistor.

I see you are metering screen current at the hot end of the screen dropping resistor.  If you do this, you might as well put the plate current meter up there too, and read each current individually.  An alternative is to leave the meter in the negative supply circuit as drawn, then move the screen current meter to the screen end of the dropping resistor to reduce the voltage on the meter assembly.  It would still need to be isolated, but the lower the exposed voltage, the better.

Let's see what Frank has in store for the driver and carrier set circuitry before proceeding further.
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #62 on: September 07, 2020, 04:08:40 PM »

Bery, bery good, Rick!

I made all the meters HV isolated so that there is now a separate plate current reading. The screen is measured after the dropping resistor.  I got rid of the cathode diodes. Might as well make a plexiglass panel to mount and house the meters. I did just that with the PDM rig.

Yes, I believe Frank will have the audio driver modulator carrier idle and audio on the same lead into the screens.

The 1 to 2KV HV is keyed in Baby Blue via a step start and second relay. It drops off during receive.  All of my HV supplies in the shack do this. It was more of a safety issue to me than anything else. I hate that scary 60 Hz hum sound when in receive.

Yes, that's a floating grid bias fixed cutoff supply.  

I have no RF final stage keying cuz when the modulator is cut off with that 50K cathode relay/resistor, all current stops to the final.  At least my 6AQ5 did it like that.

I'll talk to Frank about the two resistors across the final and modulator. Sounds logical.

** So essentially, the only cap to ground in the 4D32 cathode/ grid circuit is the 500 pF 5KV, correct?  All other leads go to the 4D32 cathode....

I made the changes on the schematic and will leave the old one up there in case there are some more comments today - then change it tomorrow with the updated one.

You are making this too easy. I usually have to suffer more to get to this point... :-)

T


* DSCF0016 (4).JPG (319.61 KB, 1280x960 - viewed 414 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #63 on: September 07, 2020, 04:56:02 PM »

Yes, the only cap to ground is the 500 pF, 5 KV bypass.  This will probably be small enough that it will not attenuate any highs, as you have a very low impedance source modulator. (I am not even sure that you need an RF choke between the cathode and the modulator plates, but it should not cause any problem either way.)  All other returns go back to the cathode, not ground, so you have true series modulation.  This is equivalent to high-level plate modulation as far as the final is concerned.

I believe the bypass capacitor from cathode to the 2.5 mH grid choke should go from cathode to the bottom of the grid leak resistor, not the junction between the choke and the resistor.  Your grid current meter must also be isolated, of course.

Since you are metering the plate voltage at the high side with an isolated meter, you might as well meter the screen current from the high side, and put the screen dropping resistor back where it was, keeping the potential on all the meter cases (except the grid current meter) all relative to the high voltage.  (I was looking a minimizing the HV on the isolated meters if you were going to measure plate voltage on the low side.)

I also prefer to switch and step-start my HV supplies for my AM rigs, but for my 4-400 linear on sideband, I leave the HV on and use cutoff bias during receive.  Too much relay clatter on sideband exchanges.....

Back to the resistors.... You will want the NPL resistor across the series modulator, but I would leave it out for the initial testing, then add it and compare performance under modulation.  Since you will kill the HV on receive, no need to shunt a resistor across the 4D32 final.  If the HV was always there, then it becomes an issue causing all the stress to Mr. Raytheon.

I wonder how fast your power supply bleeds down when going from transmit to receive?  If it is fast enough, then you probably do not even need a relay in the minus side of the series modulator.   If it does bleed down very slowly, then we really need to look at the overall picture to determine what components see the HV until it is gone.  But if you stay within the capability of the 4D32 to handle all the voltage, and have the NPL resistor across the sweepers, then there should be no issue.

I am not familiar with the grid circuits you have used earlier, for example the dual quads.  Did you intend to have an "EL" network for the grid drive?  The two .001 uF  (1000 pF) caps before and after the EL network just function as DC blockers, and have insufficient reactance to be active as part of the matching network.

You have me thinking of building one too!  No parts to buy, my junkbox is full (except for the LF6 toobs).  I wonder how it would drive my 8877 homebrew linear?  It has a 4000 volt 2 amp CCS power supply with variac. The plate transformer is over 250 pounds, from a Temco 5 KW S\shortwave broadcast transmitter. Filter is a 290 uF oil cap at 5 KV.  That certainly deserves a glitch resistor, ya?

Please keep me informed as you and Frank work on the input and DC set circuitry.  He does MAGIC with the sand state FETs.  I will be curious to see how he wraps the ribbons on it!



Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #64 on: September 07, 2020, 06:33:50 PM »

OK on all, Rick.  I did the changes.

I am using the double grid DC blockers just to keep the HV off the RF input.  I will still float the input components as if they were hot in case, just to be cautious.

The HV supply dies slowly from a 200K bleeder. It works FB with the other rigs, so not worried.

This is a good start for today.  I hope you jump in and build one too.  That will undoubtedly accelerate our overall progress.  I will start looking around for parts and chassis materials.

Great driver for an 8877 linear. I have one here too that will be used from time to time. I also have a pair of 4X1s in linear and Mr. Ugly.  All need an AM driver that the FT-1000D could not fulfill.

The stand-alone 40 watt PW rig version shud be fun too.  I think we will be limited to 30-40 watts carrier due to the 6LF6 heat. But that's just FB for me cuz the peaks will be huge and clean. The NPL will be very important to be set up without splatter, but I know it can be done. I plan to put a stiff breeze on them. A muffin fan blowing air directly down, centered, is the best method I've found. The 4D32 will get some stray ventilation but will not really need much as a loafer.

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #65 on: September 07, 2020, 09:57:46 PM »

Rick,

I forgot to comment on Bob/ECK's rig.  I contacted him a month ago about his rig. He said:    "As time went on and I was using my series mod rig I discovered some instability issues.  I added parasitic suppressors in a couple of different places."   And,    "If I had it to do over again I’d use an 807W series modulating an 807W."

So I am prepared to add some parasitic suppressors.  I'm not sure about the 807W comment though.

Frank is making good headway with his MOSFET audio driver design.    He observed from the data sheet that at 0 volts on G1 and 160 on the screen it is hauling 1400 ma. He would be surprised if it needed more than 80 volts peak on the screens which is a good thang.  

Also, he found it necessary for the audio driver to have a negative supply. There is distortion at the negative peak without it.  So a +- supply will be required and the elimination of the diodes.
And he is working on adding negative feedback.  NFB will be the icing on the cake.

Looking back at some of my earlier schematics, I found that I had  .001 bypass caps from the top AND bottom of the grid leak resistor to ground. Is there any reason for me to pull out that top bypass on the rigs or is it OK?  I figgered that the 2.5 mH choke was enuff of impedance blocking.

Another thing I noticed on Bob/ECK's rig:  He used a resistor from the grid to the cathode for grid bias.  And then a blocking cap that fed the input RF circuitry. All of his input circuitry was at ground potential. Is that something we should be doing without a fixed cutoff bias supply?  I seem to remember that with the series modulator there was never any problems with damaging currents when the RF drive was taken off. This was on my 6AQ5 rig. It wud be so much simpler without floating the input RF circuitry.  Or, maybe the input L/C can already be connected to ground potential because it has an RF path to the cathode thru the 500 pF @ 5KV cap...

BTW, I found that I needed an 1 mH inductor and small capacitor between the plate of the series modulator and cathode of the RF tube due to instability. Maybe that is what Bob was referring to. I think it was more to keep the RF out of the audio. He said he didn't keep any documentation. This will all be determined in testing which can be a fun challenge.


T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w8khk
Member

Offline Offline

Posts: 1201


This ham got his ticket the old fashioned way.


WWW
« Reply #66 on: September 08, 2020, 02:24:07 AM »

You bring up many different points here, so I am not sure I will hit all of them in one response.....

Bob mentioned instability, and an 807W.  The 807W has a form factor similar to the 6146, not so tall, shorter leads, less lead inductance, thus easier to tame parasitics.  His comments probably were not related as much to the modulator.  Typical bypassing and parasitic suppression techniques apply no matter what tube we are using.  I think the 4D32 will be fine so long as input and output circuits are isolated/shielded, top and bottom of chassis is probably sufficient.

Biasing and protection gets interesting, so many options.  Perhaps a landline discussion would save some time, I have so many irons in the fire right now.  But to start, we can use fixed and grid leak in combination, grid leak by itself, clamp tube, and another option is to have a method to detect that drive is available and just use the proper amount of grid leak bias, simplifying the circuit and build significantly. 

A simple diode detector that enables or disables the TR switching and HV if there is no drive would allow you to use just grid leak, tied back to just the 4D32 cathode.  The grid leak bias is produced by grid current through the leak resistor and a capacitor is preferred to retain that charge between RF cycles to bias the grid deep into Class-C for big positive modulation peaks.  RF chokes and bypass capacitors may be added for metering the grid current, but the DC path for generating the grid bias must be tied back to the cathode of the final, not the chassis ground.  The bias is developed across the grid leak resistor, and the capacitor in parallel with this resistor maintains the bias.  A clamp could use the same filament transformer for the 4D32, but I think there are simpler ways of protecting the final without resorting to a clamp tube.  Perhaps just sampling the grid current with a small relay would allow a fail-safe to prevent HV application in absence of RF drive.  I would lean toward eliminating the fixed bias supply, and simplifying the grid current metering wiring as much as possible.  I am not sure about the need for screen current metering either.  I was thinking a relay on the screen to ground the screen voltage to the cathode in the event no drive is present, you then protect the final in absence of drive, and you discharge the supply when switching to receive via the grounded screen dropping resistor.  But let's see what Frank suggests for T/R issues before we add redundant functions.

According to his schematic, Bob just had a grid leak resistor from grid to cathode, no capacitor.

The input RF circuit need not be floating.  The coupling capacitor from the grid tuning circuit will address the difference in DC and modulation, but have a low enough reactance at the operating frequency to pass the RF.  If you wish to protect the driver, I would suggest an RF choke in parallel to the input, like the safety choke at the pi net output, instead of another series capacitor.  The Input network can be an L, Pi, link coupled tank, whatever works for you, capacitively coupled to the final grid.  The 4D32 is a very stable tube, even without neutralization, but you must make sure that the additional wiring for the grid current metering is properly filtered and bypassed such that you do not provide an unwanted feedback link causing parasitics.  Metering is always desirable, but the circuitry sometimes "gets in the way."

The need for some negative bias on the modulator does not surprise me.  Bob's design used a negative bias on the 12AU7 cathode follower that drives the 6LF6 series modulator.  I indicated this in my original simplified modulator block diagram with the driver tube.

It's late, I will try to get back and have a look at progress tomorrow.
Logged

Rick / W8KHK  ex WB2HKX, WB4GNR
"Both politicians and diapers need to be changed often and for the same reason.”   Ronald Reagan

My smart?phone voicetext screws up homophones, but they are crystal clear from my 75 meter plate-modulated AM transmitter
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #67 on: September 08, 2020, 02:45:54 PM »

Hi Rick -

OK on all.

I updated the schematic based on all of our recent ideas. It is simplified and improved. I still need to make a decision about the RF drive failsafe if there is no fixed cut-off bias. A simple RF detected relay on the input may work. Your screen voltage idea seems interesting.

Having meters for all functions is important. I have many times tried to get away without a grid or screen meter but ended up adding one in later when there was no room... :-)   Hopefully these HV insulated meters will work without adding in lead length in the wrong places. It appears they will all have bypassing.  All meters will be safely mounted in an enclosed Plexiglass support.

WE saw Frank's MOSFET audio driver and it looks simple enuff with only two MOSFETS. He was able to get NFB included. It is comprised of the input and output stages of his most famous GFZ audio driver board except it is single ended for our purpose. Once finished maybe John/JSW can make a professional PC Board for it.   We need to decide on the audio drive needed for the 6LF6s to choose a critical resistor value and proper power supply level so it will probably be one of the last things we build.

Time now for slowly accumulating parts that are missing. I need to buy some alum sheet metal and RF tank parts. I'm really running low this time... Wink

T

** UPDATED SCHEMATIC BELOW on 9-8-2020 **

EDIT: Schematic correction: The screen meter "+" should go to the "+" side of the plate meter, directly to the B+.  Right now the plate meter reads the plate current and screen current combined.



* DSCF0001 (8).JPG (326.41 KB, 1280x960 - viewed 468 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
w9jsw
Two shots of Whisky
Contributing
Member
*
Offline Offline

Posts: 592



« Reply #68 on: September 08, 2020, 05:32:36 PM »

I'm your huckleberry....
Logged
KD1SH
Member

Offline Offline

Posts: 739



« Reply #69 on: September 09, 2020, 09:15:53 AM »

Since this thread has been blessed with a notable lack of dumb questions, I feel I've just got to step in and fill that void. Why is the center tap of the 4D32's heater transformer tied to the cathode? The 4D32's heater is isolated from the cathode internally, so other than getting hot - as the heater should do - why does it need to get involved in the action at all? I've been trying to find specs on the 4D32's heater/cathode max voltage, but haven't seen it anywhere.
Logged

"Gosh, Batman, I never knew there were no punctuation marks in alphabet soup!"
—Robin, in the 1960's Batman TV series.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #70 on: September 09, 2020, 11:23:44 AM »

Since this thread has been blessed with a notable lack of dumb questions, I feel I've just got to step in and fill that void. Why is the center tap of the 4D32's heater transformer tied to the cathode? The 4D32's heater is isolated from the cathode internally, so other than getting hot - as the heater should do - why does it need to get involved in the action at all? I've been trying to find specs on the 4D32's heater/cathode max voltage, but haven't seen it anywhere.

Hi Bill,

The heater filament to cathode  voltage rating is usually only a few hundred volts on many tubes and by connecting the fil xfmr CT to the cathode of indirectly-heated tubes keeps the fil to cathode potential to near zero. It is not a problem with directly heated cathodes like a 3-500Z or other power grid tubes, but when a cathode depends on a separate heater, it is a good practice. In our case we are playing with 1500 volts or so that will be a swinging dick of destruction...

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
KD1SH
Member

Offline Offline

Posts: 739



« Reply #71 on: September 09, 2020, 12:03:20 PM »

Ah, yes, the infamous S.D.D.  Figured that was the idea, but I wasn't sure exactly what the allowable cathode/heater potential was. I was thinking for some reason that it would be a lot higher. But then, the physical spacing between heater and cathode must necessarily be small. On the 807 it's only something like 130 volts. Now, of course, we shift the burden to the heater transformer; probably hard to find anything with isolation greater than 2500 V these days.

Since this thread has been blessed with a notable lack of dumb questions, I feel I've just got to step in and fill that void. Why is the center tap of the 4D32's heater transformer tied to the cathode? The 4D32's heater is isolated from the cathode internally, so other than getting hot - as the heater should do - why does it need to get involved in the action at all? I've been trying to find specs on the 4D32's heater/cathode max voltage, but haven't seen it anywhere.

Hi Bill,

The filament to cathode  voltage rating is usually only a few hundred volts on many tubes and by connecting the fil xfmr CT to the cathode of indirectly-heated tubes keeps the fil to cathode potential to near zero. It is not a problem with directly heated cathodes like a 3-500Z or other power grid tubes, but when a cathode depends on a separate heater, it is a good practice. In our case we are playing with 1500 volts or so that will be a swinging dick of destruction...

T
Logged

"Gosh, Batman, I never knew there were no punctuation marks in alphabet soup!"
—Robin, in the 1960's Batman TV series.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #72 on: September 09, 2020, 02:06:04 PM »

Ah, yes, the infamous S.D.D.  Figured that was the idea, but I wasn't sure exactly what the allowable cathode/heater potential was. I was thinking for some reason that it would be a lot higher. But then, the physical spacing between heater and cathode must necessarily be small. On the 807 it's only something like 130 volts. Now, of course, we shift the burden to the heater transformer; probably hard to find anything with isolation greater than 2500 V these days.


I've been in that boat of looking for a HV insulated filament xfmr many times.  A great solution is to take the core of a Variac and use the existing turns for the 120 V primary winding. Take off the arm leaving a donut hole to wind wire on.  Then wind  the appropriate number of HV insulated turns to come up with a 6.3V secondary with a center tap. Around 20 turns or so, depending on the exisiting primary turns should do it. Trial and error under load until you get 6.3 VAC.  The secondary wire can be HV insulated wire or regular wire wound onto a Kapston? or Mylar equivalent sheet that covers the 120V winding. I have also used a few layers of black electrical tape... or both the tape and insulated wire.  The wire needs to handle the fil current; in this case about 4A for the 4D32 fil.

I've done this for 4X1 PDM rigs and other applications that needed 10++ KV insulation.  Same goes for fixed grid voltage transformers that need to float due to series modulation.  And even screen supplies.  Series modulation thru the cathode means some extra work with meters, supplies and floating some parts at times.  But in the end that RF final acts just like a conventional class C final but without mod iron... predictable, efficient and easy to tune up.

T

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #73 on: September 09, 2020, 09:55:12 PM »

Planning and digging around for parts.

I like to do a little every day on various projects, even if just an hour each.  I was able to dig out some key parts from the cellar stash.

Yeah, I know, the vacuum cap is overkill, but that's all I had. It's 700 pF and 15 KV. It will be seeing 1500 - 2KV on peaks.  I found a nice roller inductor and a 2100 pF loading cap. I need to get two turns counters.

You will see the stripped down Variac ready for its 6.3V filament  HV insulated winding for the 4D32.  I also found a nice 10A 6.3V fil xfmr for the four 6LF6s which requires 8 Amps.

Gonna order some 1/8" aluminum sheet stock off eBay. Good deals there and cheap shipping. I want to make a chassis for this rig so I can hide the fil transformers and any power supply stuff underneath.   I am not sure if I will build a HV power supply into the rig yet. But with a 48" X 24" piece of aluminum sheet coming, I can build just about anything I want.

Would  like to fit the (newly designed for this rig) GFZ simplified MOSFET audio driver and +- power supply underneath too. The top level should have mainly RF stuff and tubes with viewing windows.  I'm not sure yet how to mount the four meters for HV protection, but it will be safe with a Plexiglas front cover so there is no way to touch the meter or hardware.

Hopefully by plowing into the project a little, I am apt to do more each day. It all adds up over time.  Not a bad start, huh?

T



* DSCF0010.JPG (320.74 KB, 1280x960 - viewed 375 times.)

* DSCF0013.JPG (324.29 KB, 1280x960 - viewed 343 times.)

* DSCF0006.JPG (319.55 KB, 1280x960 - viewed 410 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #74 on: September 09, 2020, 09:57:58 PM »

The parts need a good cleaning.   The 4D32 is a little too far away from the vacuum C1 plate capacitor for a short lead.   I want it in front for the viewing window. Must figger that out.

This used to be my workbench.  As you can see the rigs took over the workspace.  I have to roll in a work table when needed. The other side of the room has more tables with HB rigs and a 6' rack amplifier -  and the main operating position.  Everything works and each rig is ready to get on the air with a few switches...


* DSCF0012.JPG (326.5 KB, 1280x960 - viewed 364 times.)

* DSCF0011.JPG (319.93 KB, 1280x960 - viewed 374 times.)

* DSCF0008.JPG (322.48 KB, 1280x960 - viewed 402 times.)
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
Pages: 1 2 [3] 4 ... 7   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.07 seconds with 19 queries.