The AM Forum

Just got this off of ebay. Its a little rough. The mod iron is an Inca N-21, modulators are a pair of 811As. So far I can tell it does need a pair of 811s (one had a bad filament, the other is questionable), new filter caps (one is a GE pyranol cap marked 8 muf 1000 volts that looks ok, the other is unmarked and rusty and looks like it seeped out around the insulators at some point in its life), and some wiring fixed. I drew out a schematic, my guess is the thing wasn't built for much power, the HV transformer is an 800 volt center tapped unit, the only marking I can find on it is TX-4247. No clue what the second filter cap is or what the filter chokes are, both are Thordarson chokes I know. Also, I have no idea whats up with the mica cap across the primary of the mod iron, and my guess is the variable resistor must be the bleeder (its a huge wirewound variable one). I'm looking to use it with the 814 RF deck that I'm gathering parts to build, I'm wondering if I could power the 814 off of the same power supply? I'll post pictures later. And if you are wondering what is up with the 5R4's filament transformer circuit, that transformer is actually a 6.3 volt unit and the pot must've been used to drop the primary voltage down so that it only puts out 5 volts.

Are you even going to get 700 volts out of that thing?
I doubt it.
811's like 1200+ volts on the plates.
320 ma on peaks with a high vacuum rectifier is going to suck, solid state or MV rectifiers would be better with less voltage drop, I would scrap the entire thing.

For a modulator, you want a stiff 1200 to 1500 volt supply.

You should get around 150 watts of audio as it stands, don't know what your plans are.  The cap across the mod. transformer is probably intended to cut high audio frequencies.  Good luck.

73,
Chris

Your schematic is also missing a center tap on the secondary of the driver transformer.  This should be grounded if running the 811As at zero bias, or connect to a negative bias supply if running higher plate voltage.

The biggest thing I wanted out of it was the plate and mod iron. What I might end up doing is solid state the power supply and replace the 811s with 809s, which is what I originally wanted to do anyway was a pair of 809s modulating an 814 until I found this thing. I was only looking to build a 75 to 100 watt transmitter and that's all the more audio power I need to ask from this. And I don't remember seeing a center tap on the driver, which that may not be staying in there anyway.

Here's the pictures.

I'd like to be able to use the existing plate iron and mod iron, but maybe change the tubes out to 809s. I'd also like to use the modulator plate supply for the final too (yes I know, probably not the best solution but it's far better than running the crap I have been for the last several years). I'm looking to get about 1000 volts from it, which means the current filter has to go. I'm thinking about going with a full wave solid state rectifier and maybe a cap input filter, good idea or not? As it is right now that would give me a little over 1100 volts I believe, but the transformer has 3 taps on the primary, 105, 115, and 125 volts, our line voltage here is about 117 to 119, so I think if I put it on the highest tap then that should drop the voltage down some into the 1000s. Also thinking about getting rid of the existing driver transformer and putting a backwards audio amp transformer in there and then driving it with an amp I already have that has mic inputs on it. First thing I want to do is strip that power supply section off of there and see how much room I have to work with. Going to a solid state rectifier means I can also get that mess with the pot and 6.3 volt transformer out of there and that should lighten the thing up some, as it sits it weighs like 77 pounds and is a pain to move around, I'd prefer to have a little less weight since I imagine it will get moved a lot. Also want to get rid of that cap on the mod iron, I looked and its a .002 mfd mica.

Bias can also be applied if needed through the centre tap of the 811a filament transformer using a zener diode etc. A small voltage above ground there appears as a negative voltage at the grids of the tubes. Try and locate any specs on the plate transformer and the Inca mod iron. Their ratings might limit what you want to do with this.

Looks like the modulator has possibilities. My Clegg Zeus modulator/PS only puts 900v on a pair of 811as. They work fine of course. Good luck with it.

Al VE3AJM

I'm still trying to figure out who made the plate iron, the only markings on it are TX 4247 on the top under the terminals and then the terminals are marked 0, 105, 115, and 125 on the primary side and 800, 0, 800 on the secondary, so as it is right now the thing probably does only make about 700 volts, if not less than that. What I really want to know is what that iron can handle. Should I keep the choke input filter, or could I switch it out to a cap input? Either way I'm going to probably solid state the supply. Also looking for specs on the Inca N-21, I could at least put a voltage across the windings and try to figure out what the turns ratios are, but that doesn't help with current ratings, how much audio it can handle, etc. I might keep the 811s too, if I switched to 809s they would definitely help lessen the load on the iron, but the 811s wouldn't be pushing it as hard either if I was only asking 100 watts or so from them.

I would stay choke input, solid state the supply, and change the mod tubes.
With zero bias triodes, I think they do not work as well at low voltages because you would need to bias them ON more to get it clean.
At 700 volts, you will have almost no resting current.

Fine iron for a 100 or 150 watt rig running both off the supply.
Does the choke have any rating?

A 4D32 would be a good tube to modulate, or a pair.
 

Go with the 809s for the tubes? Not sure what the ratings are for the chokes, I do have part numbers though. Both are Thordarsons, T13C30 and T44918. Maybe someone has a catalog out there that lists those. Also want to change the caps, one shows signs of a slight leak around the insulators, not sure what value it is. The other is what I said earlier, a GE pyranol cap marked at 8 muf 1000 volts. It's not leaking so might be good for now. The final I was looking to use was an 814, I might build a seperate 1000 volt plate supply for it and run it at about 90 watts. 809s are supposed to do 120 watts with zero bias at 700 volts, so I might be able to push the 814 to 1200 volts and get about 130 watts out of it. That mod iron should be able to handle the plate current of the 814, I would probably run it around 140 ma.

That Inca xfmr is probably 75-80 yrs old. hope it's still good.  For a modulator you can use a single choke supply, no need for two chokes.  The Thordarson 13C30 is a standard catalog model but the other one with the five digit part number is a special, no way to tell what its ratings are.  Same is true for the plate iron TX-4 digits, special.

I'll try to look up the ratings for the 13C30 later today.

Fred

Yea I am going to check the mod iron out later. Evidently I'm slightly dyslexic, that plate transformer says XT 4247, not TX. Then what I might do is just pull the leaky cap and the choke before it and just run the first half of the filter

Stripped everything off of the chassis this afternoon. The mod iron appears to be OK. Ended up with a large pile of parts that aren't going back on, like the rectifier socket and filament transformer, driver transformer, the pot used to control the filament voltage, and lots of wire. Once I get the chassis and front panel fixed up I'm going to remount the plate and mod iron, filter choke, modulator filament transformer, and start rewiring everything. Now the only problem is I don't have the RF deck even started yet so I'll have the modulator done but no way to test it.

Looking at your schematic, the fact that only a single 5R4 was used in the original set up HV supply, might indicate that the plate transformer is not rated for that much secondary current, and then again not. Same with the mod transformer. Not rated for much voltage and current. That was a strange/marginal arrangement. You could go cap input or choke input since the supply would only be used for the modulator. FW, FWB etc to get various voltages after solidstating the supply and adding some surge protection. And the primary has taps on it as well. In any case the VA rating of the transformer remains the same whatever it is.

The Inca modulation transformer data might be unobtainable as to voltage and the max dc current through the secondary winding. Theres almost nothing worthwhile online with any details about any Inca transformers. As you say, the impedance transformation can be derived on your own easily. If you're certain that the modulation transformer is good, I would protect it by insulating it above chassis ground, and/or you could use a modified heising arrangement where there would be no DC on the secondary winding, to be on the safe side.

Al VE3AJM

No problem.  It's easy to test the modulation deck by substituting the RF deck with a 5K - 7K power resistor. (or calculate whatever value  the RF deck impedance would be under full operation)    You can sample some audio off the power resistor(s)  for the scope with a HV divider or HV probe and run some tones through to see how it looks.


T
  

I was thinking of going with modified heising, finding a choke is the hard part. I want to run the final at 1100 to 1200 volts at 140 ma, so my calculations say I'll need at least a 35 henry choke.

Now that you have the iron stripped off, how about better quality photos of the connections on the modulation transformer?  How much does the modulation transformer weigh?   If the transformer is not cross-laminated why bother looking for a modulation reactor?  Use it as is.

So, 1) Better quality photos of the connection (use a better light source or turn on the flash so that Luma 635 has a better chance), 

2) How much does it weigh?

As far as the schematic is concerned, just start over with a proven design out of a handbook.

Al

I did a rotation of what I think is the power transformer.  At least I think, in spite of the blurry pic, I see a connection for 115 volts.  I tweaked the rotation and lighting which begs the question what about the other transformer?

Back to this transformer: Be aware that the center tap on this transformer is meant to be at or close to ground potential.  The only real choice for this transformer is center tapped to ground full wave circuit.

Depending on your goals, the _ABSOLUTE_ number 1 thing you can do get close to, if not achieve a really low distortion system is to eliminate the driver transformer.

Some are good - and the audio may be "acceptable".  I've personally have never owned one that is even remotely close to good.

The test?  Put a 400hZ triangle wave, modulated 100% and look to see how good it is when demodulated with a local, good quality detector.

I have achieved the good triangle wave "test" with tube transmitters with modulation transformers.  The driver was a direct coupled driver.  That is critical.  Also, this enables you to put negative feedback around the system (including the modulation transformer) and boy does THAT help a less than perfect mod transformer.

But it all depends on your goals for the project.  Anyway, it sounds like a fun and challenging en-devour !

Regards,  Steve

Yes that is the power transformer, there are 4 connections on the primary side, marked 0, 105, 115, and 125. the secondary is 800, 0, 800. I'm going to stick with the full wave choke input circuit, just solid stated. The pictures are poor because they were taken on a cell phone with no flash (best camera I have unfortunately).


The mod iron weighs around 20 pounds. As far as I can tell, the primary is a single winding with four taps. The secondary appears to be two seperate windings, each winding has 3 taps. Taps 1, 4, and 5 are one winding, and 2, 3, and 6 are the other. The transformer is potted, I had the top cover off this afternoon.

I've already done a good bit of cleaning up, the only things I am keeping are the mod iron and plate iron, filament iron for the modulators, one of the filter chokes (probably the Thordarson 13C30, if I can find the specs on it), and the bleeder resistor. The mica cap (it was a .002 mfd one) is gone, the 5R4 is gone, along with its filament supply, the second choke and cap are going away as well, along with the driver transformer. That driver is a mystery too, it is also an Inca piece, but the secondary side of it has 4 connections, two of them were tied together and grounded, I guess that's the center tap. I missed that when I drew the schematic out originally but noticed it yesterday when I tore everything down yesterday.


At the moment it turning out to be more of a learning experience at building something (or in this case, completely reengineering) so I am trying to keep it as simple as possible. I already have a decent amp that I can use as a driver, so I'm just going to get a small push pull output transformer and use that backwards connected to drive the grids of the 809s.

You do not need to sound like a broadcast station to have fun.
I thought the rule was 10 watts per pound for mod iron.
20 pounds would be 200 watts?

Depends on how its made though, the cvm-5 had cast iron end bells....

 

That's about what i figured, and i'm only asking about 100 from it. The case is just thin sheet metal with a removable top cover, so the weigh is mostly transformer plus the potting material. I don't really intend on sounding like a broadcast station, as long as its clear and intelligible sounding and can make 100% modulation, thats fine with me. I do intend on using a negative peak limiter, but more as a safety measure for the mod iron.

A good idea on the neg peak limiter.  Use the 3-diode neg peak limiter.  That is the best one, and guarantees a load on the mod xfmr at all times (well, unless the final fails, but that's another story!).

The ratings for the Thordarson T-13C30 choke;  swinging choke, 8Hy - 25Hy, 150ma - 0ma, 200ohms, 1600 RMS TEST Voltage.  Choke is no good for HV, it's a choke for voltages under 500VDC.  I would use a different choke.

Fred

Could the 105Vac tap on the power transformer be used to get a few more volts out of the secondary?

What could go wrong?

R...........

Why not?

That choke IS on the light side for 811's (320ma), but for less current you could insulate it from the chassis and run somewhat higher voltages.
It did? run at whatever voltage in the original setup...

You should be able to get a good approximation of the Inca mod transformer's ratings.

We need to know these parameters:

1) dimensions L W H
2) Weight (this is important)
3) DC resistance of the primary and secondary windings
4) Inductance of both of the above windings
5) Any other observable data that you think may be useful.

With this, it should at least be possible to come with an educated guess on the power rating.  The resistance/inductance will be a clue into the current rating.  The voltage rating is a bit more sketchy, but if it has lots of taps and/or split primaries and/or secondaries, that will also be valuable.

From the picture, the Inca mod transformer looks reasonable for the job, but the numbers will tell all.

I had a rather long reply typed up and then went to post it and lost it.

Anyway,

The case is 5.5 inches high by 6.5 inches long by 5.5 inches wide. It weighs 20 pounds. I have no way to measure inductance. The resistance of the primary is 124 ohms across terminals 2 and 4, and 172 ohms across 1 and 5. The first winding of the secondary is 97 ohms across terminals 6 and 9, 24 ohms across 6 and 10, and 73 ohms across 9 and 10. The other winding of the secondary is 129 ohms across terminals 8 and 11, 35 ohms across 7 and 11, and 94 ohms across 7 and 8. The first picture is the primary side and the second picture is the secondary side.

The Inca is probably a 125 watt xfmr.  Most makers of the day made 125 watters as a standard catalog size.  Best to use the entire primary, forget the taps.  Put both secondaries in series and use the whole of both windings or move the B+ up one tap from the bottom end.  Only other way to connect the secondary is to put both windings in parallel but the step-down would probably be too much for most Class C finals.

You can study the xfmr until you're blue in the face but the bottom line is, it's the only mod xfmr you have and there are only a few ways to connect it up.

OTOH, if you have twenty other mod xfmrs to play with, keep experimenting.

Fred

it could overheat

a simple test for a xfmr is to excite it as planned but without any load and see how much current is required just to energize the primary ... you can use a small resistance in series and measure the voltage drop then with Ohms law can calculate what is known as the magnetizing current .... this is a simplification for what is going on but yields useful info .... for example this xfmr you are interested in draws 50mA at 105V applied and after a half hour or so gets barely warm then you are sure that the xfmr is basically ok and can test at higher voltages .... so say the current draw doubles to 100mA at 117V applied .... the apparent power in the xfmr has increased from 5.5W to nearly 12W and the transformer runs a bit warmer but not hot

so where is this helpful .... it is an observed fact that electromechanical devices have a correlation between internal temperature and expected life ... it is basically stated that a 10 degree rise in temperature results in a halving of life expectancy ... the onset or critical temperature is dependent on materials and methods of construction

please run the test .... hv iron is on the endangered species list

Ok, good info.  Based on the weight and size alone, it's over 200 watts.    For instance, the UTC S22 is almost identical with respect to size and weight.  The S22 is a 250 watt transformer.

When you look at the size of some mod transformers, that one should do real well at 200 watts or below.

Yes!

I once did an experiment with my pair of 813s plate modulated by a pair of 813s.

The existing mod transformer was a 60 pound old broadcast unit. It worked very FB and was hi-fi at whatever power level I wanted using 2200 volts. It was the only transformer in the chain and driven by the WA1GFZ MOSFET 11N90 audio driver into p-p 813s.

I wanted to see how an ART-13 mod transformer would work in its place, so I clip leaded it in. The ART-13 transformer weighs maybe 10 pounds? and can fit in the palm of your hand.  I was using a modulation choke in modified Heising, so the mod transformer carried only audio, no DC current.

I was amazed. That thing modulated a 300 watt carrier no problem. The low end was noticeably absent as expected due to the lack of mass. But it was reasonably clean above 200 Hz and the guys on the band thought it sounded like a DX-100  - not bad.

It continued to work FB for ten minutes... no problems.  This is a transformer that was built when men were men and transformers were designed to survive a war. (WWII)

The bottom line is if modified Heising is used and the transformer is of good quality, yours should hang in there.  Be sure to use spark gaps for protection and be VERY careful of unexpected audio surges.

T

I was kind of thinking of the typical 100 watt rig, or even the G76 mod iron.
70 watts carrier and plenty of modulation, running the rig 100 volts over normal, and the very small mod transformer never gave any problems.
I worked someone who ran one 200 volts over normal at 100 watts without issue.

Yea that transformer is pretty heavy, its actually slightly heavier than the plate transformer is. If I put a modulation choke on it, finding one with enough inductance and be capable of handle the amount of current for the final I want to run could be difficult. Running an 814 at around 1100 volts 140 ma is going to have a plate impedance around 7800 ohms. I remember reading before to calculate the minimum inductance for the choke by multiplying the final's plate impedance by .0045, so that would be 35 henries. I could either find a single choke that was 35 to 40 henries or so or use two 20 henry chokes in series. Either has to be able to handle at least 150 ma (and thats right on the edge) at at least 2500 volts.

DC through the transformer is not going to hurt anything, its voltage spikes you have to watch out for.

1100 volts does not seem excessive.

I have had two mod transformers fail, both my fault, one was a 30K1 mod transformer when I did full power 20 hz testing, before I knew better.
The 2nd one shorted when it had voltage only on one side of it (rf side) because the modulator was not turned on.

Never had one fail in normal use at below max ratings.

So proper transmit/receive sequencing might be a better idea? Would it be better to key the high voltage on the final and modulators (it appears that the transmitter this modulator came from was designed to do that with the external relay connection on the primary side of the plate iron) or to leave the high voltage on and bias the tubes on/off to go from transmit to receive (using a resistor on the center tap of the filament transformer, and probably fixed bias on the final or a cathode resistor there too).

Both ways work.
The voltage can go way up with no load if you leave it on, and can cause noise on RX if you have any current flow.

Hi Shelby,

Take a look through pages 4 to 9 of this thread to get an idea of the problems and solutions to keying up and un-keying a big plate modulated rig.

http://amfone.net/Amforum/index.php?topic=33816.75

Yes, a sequencer is a great idea.  I sequence/ key the high voltage transformer's 240VAC primary and the cathode (fil xfmr center taps of the RF and modulator tubes)  of my 4-1000A plated modulated rig.... as well as the screens, RF drive, antenna, and a few other things.

I also use Hall effect fault / shutdown circuitry that monitors current levels on the tube's various elements.

Another good idea discussed in this thread is putting a power resistor across the mod iron output that is open when transmitting and during un-key it loads the modulation output to quench any arcs. (relay controlled)   Whatever it takes to have the rig come up quietly with no arcs and then un-keys without arcing, audio pops on the air, etc. Smooth as glass.

Below is the sequencer I built that uses variable delays for key up and independent decays for un-key. It all still works FB.

W1VD / Jay makes a nice, simple sequencer board that could probably do the job for you.
This will help preserve your mod iron.  Getting a handle on keying transients is part of the shakedown process of building a big rig.

T


Below: The second picture is the simpler pre-assembled four stage sequencer board that you can purchase.

This has been mentioned before but I can’t find a full explanation or how to
find the mod transformer frequency response without damaging it.
Anyone care to explain?
Best regards
Ted / KC9LKE

I'll take a crack at it ....

to really understand what is going on in a transformer you need an accurate at least third order model ...higher orders are more accurate .... the problem is that the vendor of the named transformer is not around to help provide this information ...so we will need to set up some tests to help determine some of these parameters

this is not easy but is very interesting and will help knowledge of electronics grow...

if I could schematic capture, I would put in the model now...

now break it down to the low freq , mid freq, and high freq models and note how which parameters change in importance...

all of the models I have seen do not account for when the transformer is driven past ratings into the non-linear regions ... this is not easy and will challenge your reasoning ...

it would be easy just to give up and do something else .... just how curious are you ?

The modified Heising my very well buy you nothing other than more weight in the modulator !   That transformer is MORE than likely an Amateur modulation transformer which is designed for DC.  If you're not exceeding the DC current rating, it's should be all OK.  That's a pretty good size mod transformer.

Sequencing - THIS IS IMPORTANT !!!!!!!     You need a 2 stage sequencer for this rig.

Keying - my suggestion having built literally hundreds of transmitters, tube and solid state? -> Key the H.V. supply on and off.  Safe, reliable, no guesswork.  Use a step start and your rectifiers and other components will not be subject to any current spikes.

Here is a sequencer I use at Rattlesnake Island.  Also is used by Wayne WA1SSJ and a few others.

(http://www.classeradio.com/sequencer.jpg)

Does the adjustable hold and delay just change the amount of time that particular sequence stays in that state? Does this need to be fine tuned in any particular way?

I do not sequence anything other then the step start.
Most ham transmitters had no complex sequencing.

Its not a bad idea, but it adds complexity and failure points.

Yes, that's it.  The RX hold time determines how long, once you go from tx to rx, the system will actually delay before releasing the antenna change over relay, shut down the RF drive (or whatever you have controlled by the RX side of the sequence).

The TX delay is how long, when you from from rx to tx that the system will wait before engaging the H.V. power supplies, and anything else you have controlled by the TX side of the sequencer.

There is no fine adjustment.  Just do it by ear.  I usually do a 1/4 to 1/3 second delay on the TX delay, and about 1/4 second on the RX hold time.  Adjustment is strictly by ear.


Just as a note, I built my first sequencer in 1975 and am still using THIS SAME UNIT (amazing !). 

The thing is, most transmitters don't actually shut down instantaneously - they take some finite number of milliseconds to really put out 0.

I like to give it at least 100ms to 200ms to "settle" to 0 output.  That's the receive delay in the sequencer.

A lot of times you will see arcing at the antenna relay, hear big pops in the receiver, etc.  when going from transmit to receive.  These are good indicators that you have a sequencing problem.

Coming up is similar.  Drivers, relays and other things do not settle instantly - they take some finite time to do so.  That's where the transmit delay in a sequence comes in.  This gives everything a chance to stabilized, and then the H.V. is keyed.

Yes, sequencing is important.

The dead giveaway on my 4-1000A rig was arcing at the spark gaps on the modulation transformer during un-key. It was nasty.  There was obviously a lot of audio power that needed to be dumped. Usually this power is dumped into the RF final and out to the antenna in a brief instant. This means kill the HV, kill the audio input, hold the antenna relay closed and keep the RF deck energized a little longer than the mod deck - this  usually works to quench the audio arc and dissipate the power.  (Along with the relay switched power resistor across the mod sec as I mentioned before)    

And as already mentioned, pops in the receiver during un-key are a tell tale sign of RF wanting somewhere to go. I have actually blown out the front end of an FRG-7 receiver as a result of this antenna relay arcing... poor timing.

Some rigs are easy to tame and some are a real bear.

This is why I made all six sequencer relays variable with independent key delays and un-key delays. I was able to actually adjust the pots on the 12 delays and see the arcing, popping and farting all go away to perfection once the timing was optimized.


The ultimate arc acid test, if you feel lucky, is to sock an atomic yallo into the mic while at the same time rapidly keying the rig in staccato. All of my rigs will do it, but it took some work, especially with plate modulated rigs with big audio power looking for a place to go.  I have done it and have seen several  others do it on the air on a dare. It WILL break parts if things are not dialed in right.


T

Here is the full-size schematic of the sequencer (this is .pdf, which is nice - can be resized, etc).

http://www.classeradio.com/2_stage_sequencer.pdf (http://www.classeradio.com/2_stage_sequencer.pdf)

Feel free to pass this around.  The circuit has been built many times.

In a QSO a few days back "tron" WA1HenryLellaaRrrr mentioned that Inca was a pre WWII mfr.

My 2 cents:

Try the existing interstage xfmr. You can test that fairly simply - but you'd want to know what input Z it is
"expecting" first. That can be determined by the put 10vac on the primary and measure the voltage on the secondary test. This tells the voltage ratio, aka "turns ratio". Which when plugged into the formula, and using
a calculator to do the sq rt or the squaring (depending on which way you use the formula) will tell you the impedance *ratio*. Knowing that you are driving grids of an 811a (for the original) once can then surmise
what the likely load side Z was, and so determine the primary side. The transformer either wanted to see a tube's
plate or maybe a low Z (~600ohm) source.

Knowing this info, you can put a test load on the secondary (the approximate resistance representing the presumed load Z the transformer was designed for) and then apply the requisite voltage on the primary - a fraction of the operating voltage is ok for getting in the 'ballpark" - and see what the frequency limits are. Especially the HF side, this will be somewhat of a valid test. On the LF side, you can expect a bit less, especially if the primary is supposed to sit on the plate of a tube and pass DC. If you wanted to get "fine" on this test, you'd put a cap to couple the swept audio signal to the xfmr, and put DC through the primary via a current limiting resistor to simulate the current through the driver tube.

Also, beware, driving a transformer from a low Z source (like an audio amp) usually produces a better response curve than when run off the plate of a tube (high Z source). But if it drops like a stone on either end of the spectrum, then you know that for sure.

The same thing can be done with the mod iron itself, although keep in mind that all transformers perform somewhat better at lower signal levels than near max or "peak" signal levels, and also with less DC rather than more.

A great way to drive 811 or similar tubes can be found by copying the Altec 1570B driver stage. It is perfect for this application and super easy to implement. In fact the secondary of the existing driver transformer can
be perfectly applied as the P-P cathode driver choke for this circuit. The limitation on that is the power/current rating of the iron. But it would likely support a pair of 3-6watt tubes, no problem. The great part of this is that using this method the frequency response of what is now a CT choke is irrelevant. In fact a transformer with "bad" HF response is just fine, and will have no lack of HF response as a cathode choke. Plus you could apply bias through the CT, need be.

The schematic can be found online.

I'd say that if you could boost the HV up to about 900-1000vdc you'd be in a good place to do ~150watts rms.

But if all you need is 100watts for you project then the 800vdc is more than enough, the tubes will loaf. Zero bias ought to be fine.

This method adds one stage to the speech amp, that being the actual driver, in this case a cathode follower. So, the voltage swing/gain must come before the follower. But for the 811s, you can look up the drive requirements in the tube manual, and not that much voltage needs to be swung, so nothing exotic is needed at all.

You might try the lower voltage primary taps, and see what the voltages look like.
The single rectifier in the schematic is rather anemic for anything of this power level. The 1570b, rated at about 125watts uses 4 dual section tube rectifiers (I forget what number, maybe 5U4) in parallel.

I'd go solid state and avoid the problems, just use a method for slow turn on of the HV on the 811s.
There are ways to do this and not need a sequencer for *this particular job*.

One of the worries about a restrictive frequency response due to limitations of iron in the signal path comes when
anyone tries to wrap feedback around the iron - the feedback tries to boost the extremes of response, which of course can't really be done without basically adding a boatload of EQ (what the feedback does has this effect). This extra feedback comes with issues of phase shift, and the resulting frequency response appears to be "flat", but in terms of transient response and phase response, it is often horrid and creates all sorts of *icky* things that can be seen using scopes and other test gear, and often sounds not so great.

But feedback is another issue, not actually mentioned WRT this modulator.

Generally speaking, I think anyone using transformers for audio is in safe waters as long as they do not try to extend the usable frequency extremes, and stay within those boundaries. In practical terms, trying to get a flat response to 20kHz (for example) and down to 20Hz from iron that does 80Hz to 6.5kHz flat is likely not going to go well, imho. So, the trick to making this sort of iron sound any good is to not "excite" it outside its passband. :D

Well, thanks for reading...

                 _-_-

Great comments.
Some EQ within what the iron passes can make it sound very good.
Or you can eliminate all the iron in the path and get really clean..

LOts of good things to consider. I'm starting to put everything back together and deciding what to keep and what not to keep. My amp I have has 4 output taps, 4, 8, and 16 ohms and a 70v tap. I'm not sure what the original speech amp used with this modulator was, there's just a 1/4 inch jack on the back that is the speech amp input, that goes straight to the driver transformer. The transformer is an Inca L-13, which I actually happened to find some information on. On page 22 of this PDF, it says this transformer was designed to go from the plate of a single 42, 45, 2A5, 46, 59, or 6F6 driver tube to class B grids.
http://dbase1.lapl.org/webpics/calindex/documents/15/522985.pdf

Maybe the 70 volt line will feed that transformer.
Single ended unbalanced step down most likely.
You could experiment.
Hammond also makes 8 ohm to 5000 ohm center tapped HIFI transformers (and others).
Back to back transformers is a silly way to do anything, output of a tube HIFI amp into another transformer seems the wrong way to go, but it can work.

The Hammond transformers were what I was thinking about using. Like I've said earlier, the goal of this isn't to sound like a broadcast station, more of just to build something that sounds decent and get it on the air and use it and have some fun with it.

The easy (very) way to do it is the Hammond or other transformer driven by a 20 watt solid state hifi amp.
With zero bias triodes, it does not get much easier.
Sounds great as well.

Biggest stumbling point I have found so far is trying to find tubes. I decided to go with 811As instead of the 809s I said about earlier, but finding a good pair of 811As that aren't chinese made for less than $100 or more is hard. Both tubes that came with the modulator are shot, the one's filament is completely missing, the other is broke on one side, half the filament lights up but the other half is dark.

Plenty around:

http://www.ebay.com/sch/i.html?_odkw=&_osacat=1500&_from=R40&_trksid=p2045573.m570.l1313.TR12.TRC2.A0.H0.X811.TRS0&_nkw=811&_sacat=1500

I have some Russian ones that still work after years.
The bases came unglued, but they still work.

572b's also work.

Have a look at the Hammond 117F32 70v line transformer. It would work great.

Al VE3AJM

What I want to do is use one of their push pull output transformers backwards by putting the primary side to the 811 grids and the secondary to my driver amplifer.

What does everyone usually use for wire on something like this. I know about using copper core ignition wire for the HV (and already have a source for it), but what about the AC wiring and everything else? Some of the wiring was cloth coated and still in good shape, so I reused it, but most of it is in need of replacement.

My favorite combo was my Bogen M60 60 watt amplifier and I used a UTC CG50AX line to pp grid.  That was ~40 years ago.  Today I would insert some sort of line level EQ.  It worked really great with my pp 811A modulator.  The mod iron was a Thordason T11M77 300 watt unit.  The final was pp 813s running really conservatively.

I used separate HV ps' - the modulator @ 1250 VDC @ 300 ma  the RF Deck ran off a 1500 ps rated at 350 ma.  I didn't have the smarts back then to optimize it but it was a good start.  Look for a UTC CG59AX and use the 70 volt line output on that amp

Al

Yeah thats what many have done with the Hammond 117F32. The secondary 8 ohm side is connected as the primary, to the 8 ohm output of the amp. The primary now secondary ends are connected to the modulator grids. The CT is the 8w power tap. Ever hear Borgi K9YQQ? He sounds great using one to drive class B grids using a Crown amplifier.

Available from Mouser, DigiKey etc. They stock it and it is a modern transformer. Its about 1/3 the price of the Hammond 1600 series audio output transformers of a similar power rating for the same performance. Could buy you some 811As.

Of course there are other transformers that will work.

Al VE3AJM

Ok, now that you explained how it works it makes sense. I even looked at the spec sheet but it didn't really make sense until now. I've heard Borgi's rig before, I believe its something like a pair of GI-7Bs modulated by a pair? I know it had a some kind of Russian tubes in it.

Yes that's the tube I believe. I found some information about his transmitter online since. Nice work!

http://www.aimnhi.net/kj9t/K9YQQ_AM_XMTR/pres0005.html

I was previously skeptical of the back to back transformer arrangement too. But after hearing it being more in use on the airwaves now, it can sound very good.

Al VE3AJM

Yes, that looks very nice.

Right now I'm waiting for some parts to get here so I can get the power supply section rebuilt, but I have decided to make a slight change in the plans for this modulator. I'm going with a pair of 814s for my RF deck instead of a single one, running at about 900 volts 240 ma off of a seperate power supply. I figure this modulator should still provide plenty of audio (since the 811A spec sheets give ratings for CCS service at 750 volts with something like 170 watts of audio), but the question is should I stick with running the RF deck straight through the mod iron secondary, or switch to modified heising to keep the current off, since it is higher than the 140 ma I originally planned on. I found a Millen 90800 exciter on ebay that I just couldn't pass up for the price (it was in great shape but is missing the tubes and coils) that I plan on using to drive the RF deck. I was thinking about using my icom 718 to drive it originally but the more I thought about it the more I wanted to make this into more of a standalone transmitter. The RF deck is just going to be a monoband 80 meter one, I also want to try adding 3rd harmonic resonators to the output network to try to get the efficiency up some, I figure at the voltage and current I'm running, if I can get it up to about 85% to 90% efficiency, I can get somewhere around 190 watts out of the pair of 814s. But that's a ways off, I'm still gathering parts for that.

You can always add the hy zing circuit later.
Much depends on the mod transfotmer.
I have some nice RCA ones that sounded better without the hysing.
I had a 50 henry Peter Dahl choke and tried various size caps and could get it to sound different but not better with various size caps.
I just ran without it and sold the choke.

Some iron sounds poor stock (cvm-5 was poor).
The Thordarsons sounded quite good to me.
Maybe try it the easy way first.

And I find most tubes work better at higher voltages and less current.
 

Dunno, the audio power amp to drive a reversed (small) output iron to drive grids of 811a tubes is a bit overkill, although it can be a fast-er way to get the job done. Using a larger solid state amp to run a backwards output iron to directly modulate the RF stage is a different deal, I expect.

The iron that works with the tubes you listed, including the 45 is pretty basic. Instead of a 45 you could use a triode strapped 6W4 or a 6V6, assuming you want to do it cheaply. Both tubes can be had for almost nothing by today's standards.

And, again, give some thought to that cathode choke loading method, and download the Altec 1570b schematic, you can basically just copy it. You get almost DC to light frequency response from the driver this way. Assuming I have it on my drive, I'll attach it to this post. Feel free to ask any questions about how to apply it. Btw, the Altec uses a triode strapped 6W4 (which is basically a 1/2 output power 6V6).

The 811s like a bit more than 750vdc, and I'd encourage you to get that voltage up to about 850-900vdc or so, if it is practical given what you have on hand for power iron.

                     _-_-bear

I should be able to get over 800 volts out of it using the 105 volt tap on the primary of the plate iron.

I'd not worry too much about going from 750vdc to 800vdc. Although I guess a little more is a little more.

Consider the basic circuit idea though.

             _-_-

Running too high a voltage on the primary is not a good idea.  You could saturate the core and over heat the xfmr.

Fred

I wasn't planning on anyway, the way I see it is what real difference is it going to make going from 750 to 850 volts? My line voltage here is high enough that the 125 volt tap is what I'll have to use on that transformer.

If you take a look at the suggested operating points vs. B+ voltage in the RCA manual (again easy enough to find online), you can see what the effect(s) of (in this case) going from 750vdc up to ~900vdc will have, especially on the max output. But also on the suggested load Z.

A bunch depends on how much power you will actually need based on the carrier your finals can muster.

                         _-_-

Well I finally finished it on Saturday. I got some big resistors to use as a test load today and was able to do a working test with it, using the resistors in series with a speaker to get the impedance up to something the mod iron could handle. The power supply puts out about 790 volts at idle and drops to 760 volts at full power (around 300 ma, it has the capability to go higher as I discovered during testing but thats all the higher my meter goes). The tubes idle at about 30 or 40 ma with no audio. Everything appears to be running cool. The power supply voltage doesn't sag under 760 volts, which I think is pretty good. At least to my ears, it sounds pretty good, the low end gets a little distorted but the midrange and upper end doesn't sound too bad, nothing that a little touch of eq couldn't smooth out. The real test will be once I get the RF deck built and try it out on the air. I have the power supply for that and the exciter started. I'll try to get some pictures of the finished product over the weekend.

which schematic/design??

perhaps you might put up the schematic with parts values...

A sign gen with a scope will give you an idea of the raw freq response (important) and using
the square wave you can see the general balance of HF to LF (by the tilt) and that will also show
the bandwidth via the shape of the square wave (1kHz. is a good place to start).

x10 probe on the scope is a good plan.

It might be prudent to adjust the coupling cap values to limit the LF below some reasonable
frequency so that you do not "crunch" the mod iron.

The mod iron is not set up to run speakers... using "resistors" in series with a speaker probably is not
a great idea. What you need is a load that is approximately the same as the Z that the modulator will see in operation, which is HIGH. So about 100 watts of resistors there, maybe 7-10kohm I'd guess. A scope on that, and then perhaps a BIG resistance value divider, something like >100kohm with something like 1kohm in series, taking the 1kohm at the GROUND side and using the junction with a coupling cap (like 1ufd or so) into an audio power amp... then you could listen to it. You'd want to set divider such that the output level at the divider junction is about 1vac, but a bit less and a bit more (like 4 volts ac) when the modulator is a full grunt is about right for driving the input of
a power amp. Or you can run this same thing into a mixer with headphones.

Put a 100kohm or so resistor to ground on the output side of the coupling cap... just in case any DC potential somehow build up.

This would give you something reasonable to hear...

                     _-_-

                      _-_-

the combined resistance of the resistors i used (big ceramic ones) plus the speaker all together equaled about 3700 ohms, which is about what i calculate the final will run at. The design i used was the the same thing that it was when i got it, that i posted before, just with a few modifications, mainly to the power supply. The only mods I did to the audio portion of the modulator were removing the cap across the primary that is in the original schematic, and changing the driver transformer out from the original Inca unit to a Hammond 70v matching transformer like what was suggested earlier on. As far as the power supply, I made it a solid state full wave rectifier, with a 10 henry filter choke and about 50 uf of filter. The original power supply had a two section filter with a 5R4 full wave rectifier.

If you want a true test, then after using the resistors and determining the best load and if you can make your RF stage act like that load or close to it, then use the RF stage tube and HV supply as a variable current source and load (by controlling its bias) and draw the carrier level DC current through the secondary and test the frequency response and power bandwidth again. Then you can see what it will act like in use which may be quite different as in worse. If the difference is too great then you may want to limit the audio power and frequency response or to use a modulation reactor to avoid having DC current in the secondary at all.

I'm trying to explain that the combined impedance of the resistor + speaker is moot.
That's just not a proper way to determine anything.
Whomever suggested this, is barking up the wrong tree.

The method I explained is far better.

What Patrick said about running current through the secondary is quite valid as this will effect the
saturation of the core. (if you use my connection method to monitor, then you MUST use the cap, and rate it
according to the voltage supplied to the secondary.

Regardless, monitoring the audio by directly connecting a low Z speaker is simply not a good way.

I'm surprised that 3700ohms is proper, perhaps PER tube?? Seems a bit low, since 811a run at 850+ volts typically
want to see more like 8k-10kohms. But if the tube manual says it's ok, then fine. Seems low to me. You'll be shy power and monkey swing if the match is poor.

                     _-_-

PS. the URL for your link is NG - what was it supposed to show??

If devlishly serious about a modulator, just get a 1570B and add the modulation transformer across the entire secondary of the amp's OPT. The amp is well tempered by its own OPT and feedback loops and the modulated load will be as well regulated. If the 175W amp has extra power not needed for the class C stage, swamp it if desired or to make match.

I do a similar thing with a 150W Stromberg Carlson AP-80 (PP-par beam tube amp) used to drive 3-500Z grids, which is a much nastier and less linear load than any class C stage. It works so well I have not bothered to put in the 1570B amp I had previously prepared for the job.

That link was put in accidentally by me bumping the keyboard. No one suggested that I series up the resistors and speakers, I don't know where you got that from, I did that on my own because I needed a quick way to test the modulator and see how well and if it was working. If I wanted a hi-fi modulator, I would've built one and not even screwed around with this but I got what I wanted to experiment with, and built it as such. The whole purpose from the beginning was to build a communications quality modulator to use for a homebrew transmitter.

And regarding the impedances I used, the 3700 ohms is the calculated impedance of the 814 finals I will be using, the impedance ratio of the iron matches this to the impedance of the 811s, which according to the datasheet is 5100 ohms at 750 volts.

Its a good idea to build a hifi modulator even if you do not intend to run it that way.
It will be clean and sound good no matter how you run it.

Sounds like you're making good progress with this 811a modulator project.

As you mentioned, and with what you're working with there, you don't want/require hifi? response from this project. Some of the hifi proponents here don't sound so great on the bands when I receive them. Weak signals with mushy bassy/no sibilance type audio response, that don't cut through very well. Globe Kings and B&W 5100s with stock audio sound fine. Listen to Russ WB3FAU or Frank K3SQP. The Globe King with the terrible 6L6 single-ended audio driver and over coupled final, and the 5100 with the 6146 modulators.

Hear you on 40m over the holidays.

Al VE3AJM

Well, regardless, the method essentially does not work.

The speaker wants a low Z source, especially on bass.
That is the thrust of my suggestion. My suggestion was to permit you to know how to find out what your modulator is doing.

Dunno about the load Z that the 814s present, but from what you just said the load you use on the 811s should be 5100 ohms IF you were using the resistors in place of the transformer, which is what your text seemed to say.
If it was on the secondary, then fine. But still you want to use a divider to drive another amplifier or similar to "listen" to your modulators output.

I would expect the speaker to distort badly running it from such a high Z source.

                _-_-bear

I don't know what is so hard to understand about what I am doing with the resistor/speaker hookup. The transformer is in the circuit, hooked up exactly like I would use it with an RF deck, except instead of an RF deck across the secondary, there are a pair of resistors to absorb most of the audio power, and the speaker to reproduce sound. All I was concerned about was that the modulator did actually make power, pass audio, and didn't have any power supply hum. It met all those criteria on that quick and dirty test, the real test will be when I get the RF deck built and actually try it out under real world conditions.



I've noticed that too, a lot of bone stock rigs sound much better than the modified ones do. Sure the modified ones sound good too, when the bands are in great shape, but when conditions start to go south, the stock ones seem to have the ability to cut through the crud a lot better.

I have no high frequency hearing left, so most people sound stuffy and I enjoy d104 audio into a 1 or 2 meg resistor.
The thing is, if the modulator will pass 50 Hz it will not distort at 300 Hz.

Nothing says you have to run something to 20 Hz but if it passes 20 Hz, then 200 Hz is going to be clean and look like a sine wave.

We have all heard rigs that sound great with a D104 plugged into them, and some that sound bad, why not sound good?

I do not think you can tell what something is going to sound like till its in use.
DC through the mod iron, plate and screen bypass caps, even tank circuit Q and bandwidth might come into it, every rig I have sounds different..

It's easy enough to understand, but you complained about bass distortion...
...I said your test is not valid to see if you have a clean output or not, especially bass.

What I suggested does give you a pretty valid test. Of course weirdness in the plate of the RF output
tubes could cause unforeseen issues, but probably none will emerge.

Sorry that we're not quite on the same frequency.

Bear, a little side question, what is the value of that cathode choke on the Altec 1570B circuit. I was going to build that once upon a time, but could not find that information.

Now...back to the thread.  :)

The choke is rather non-critical since its role is to look like a high-Z at audio freqs, but it does serve to provide a DCR for "self bias" (if that is used) or a pass through for a bias voltage. You can see that as you go lower in frequency the inductive reactance present equates to an equivalent resistance. So, you can figure out the minimum required inductance based entirely on the equivalent resistance value needed to make the follower work "properly". The minimum value is going to be the DCR, but that is somewhere around 1Hz.

In practice the primary of a more or less random 5-10watt output transformer will have all the required parameters. Since it is a choke, generally speaking the quality of the transformer which would matter when used as a transformer (the HF performance especially would be of interest) makes little difference. But similarly to how the transformer fares at LF, where the inductance dictates the LF ability, it does as well here - but the impedance at the cathode is low vs. the impedance of the plate, which makes the inductance matter in similar but not quite the same way.

Also, since the "choke" is inexpensive, you can simply throw something in the hole and see how well it works at LF, and if it is ok then keep it, if not up the power handling (core size) and/or inductance.

It's neat that way.

The other aspect is that as one goes higher in freq the reactance increases, making the follower work closer to "ideal" - there are circuits that use resistors that once can compare with.

                    _-_-