811A Modulator Build Starting

[W9ZSL]

  Ain't it wonnerful wunnerful when you can bring a schematic to the Jonque Box and start picking???  Then you find what you need?  Candy shop syndrome.

I've been tossing ideas off you for quite some time now and have been changing my mind constantly.  Don't tell me you never did that.  Finally, I found the key piece; a UTC S-22 in fine shape.  Here is my current "blue-print" for the basic audio/modulator chain.  It came from the West Coast 1949 handbook.  It allows for a 250 watt mud iron using an early UTC CVM-3 building block.

I got you.  Nurse a S-22.  I can't see more than 1250 V primary with 811A working at zero bias.  I'm basically duplicating the circuit but dividing it into two separate units.

OK.  Let's start with the S-22 and work back on the diagram.  Pair of 811As and a Kenyon 15 watt 500 ohm to class B grids input iron on the modulator chassis that can be SWITCHED to the 6V6 UTC S-8 driver transformer on a separate chassis. I won't need a cap between the Triad line input transformer and grid of the second amp if I can make the leads very short.  The S-8 has high and low Z sec taps so I'm going to switch the outputs high and low.  I can feed any transmitter with plate modulation up to 500 watts in.  Won't go that high.  I can see an 813 operating at 1800 VDC.

My first build will have only the 811A modulator section built around the S22 with 1250-1350 on the plates.  When you get that one key piece, to quote Mork, "Shazbot!"  

  

[N2DTS]

Its not going to be hi fi.
And you can run 1500 volts on 811a's without bias.
Books say you need it, and maybe you do in a broadcast setting, constant service with lots of audio all the time...

I put diode strings in for bias, but never use any of it.

[Opcom]

Do it! OK not hi fi but good clear voice is the thing.

I would use some bias if necessary to limit idle dissipation to 20W per tube, but that's just me, I am cheap and USA 811 are not.

You got lots of drive potential with a 6V6 there, but could use 2x more than what is needed to improve headroom or quality.

811's take 4W drive for 240W, so how about a pair of 6v6's or a 6L6?

BTW that S-8 transformer is only rated 1 watt!! How about trying a larger driver xfmr?

There are some out there for a pair of 6V6 or the like, and a 500 Ohm or universal output.

Look into a 25v or 70V line matching transformer, as a patch between a 'normal' push pull audio output transformer and the 500 Ohm line going to the Kenyon iron?

I believe you need more than a 1W transformer in that T1 position or your lower voice freq. will be very poor. Just my opinion on that. You might also cook that transformer? Maybe it was made for the triode connection of the 6V6 or something smaller like the 6K6..

Attached is a little suggestion on how to get a decent match, you may already have the parts. It requires another transformer but that should not be an issue and also gives more matching options, since you said you wanted 500 Ohms too.

Push Pull 6V6's or a single (possibly triode connected 6.5W) 6L6 in class A, feeds a cheap 10W OPT, feeds a 'speaker to 70V line' XFMR, and the output happens to be 500Z. The point of the scheme is that if you have 8W or more, the match is reasonably good. Just throwing an idea out there. I have used this, that is, a 70V line transformer fed by 4 Ohms, to drive a pair of 6146 grids. Those 25V and 70V line-to-speaker transformers have different turns ratios depending on the wattage -that is the difference between those and a Z-to-Z transformer.

[VE3AJM]

BTW that S-8 transformer is only rated 1 watt!! How about trying a larger driver xfmr?

The S8 transformer should work OK for your single ended driver circuit. Its rated for 5 watts. A push pull audio driver would work better of course with a S9 transformer, but if you already have the S8, use it. Good luck.

Al VE3AJM

[WD5JKO]

John Coleman WA5BXO has a good discussion on class B modulators, and has an innovative way around some of the deficiencies of the class B driver stage:

http://www.qsl.net/wa5bxo/driver1.html

Jim
Wd5JKO

[WBear2GCR]

As far as I am concerned a coupling xfmr is not required to drive the 811s. You can drive them off a P-P cathode follower. The preferred method is to use a single CT choke in the cathodes. The only consideration after that is how much swing you need from the follower to fully drive the 811As with 1500vdc on the plates. That's a tad high, but not outrageous, so those diodes to bias it back in quiescent current is not a bad idea.

The good news for swing is that there are a number of tubes that will work fine with 450-600v on their plates in the pre-cathode follower position and in the cathode follower position. Some small sweep tubes strapped for triode would fit the bill fine, although if you needed gain you could keep the gain stage (pre-cathode follower) pentode.

807s will be happy with up to 750vdc on the plates - and will match the look of the 811s too!

I haven't looked at the 811's curves in a while, but that will show how many volts the grid requires in B2 for full swing. It might be somewhat less than the 450v+ volts I was speculating on, in which case, "normal" plate voltage tubes will be easy to use.

A 6W6 or 6L6 (overkill for power) driving is quite sufficient at 1150vdc on the plates of the 811a, I use just that set up in an audio power amp.

I'd keep any coupling caps in the circuit before the UTC iron a little on the small side, to keep the iron from being saturated at LF - it doesn't like that much. You'd have to do a sweep to see where the -3dB point is, but somewhere below 100Hz is not a bad thing, how low is part of the balancing act.

_-_-

Depending on your chassis and layout, it might be a good idea to leave extra realestate around the area where the mod iron will go. That way later, when you stub your toe on some higher performance iron at a hamfest, you can drop it in, since there will be room. Just a thought...

[WBear2GCR]

Took a quick look at the WA5BXO page.

I'm not sure what exactly the idea behind this is. One of the issues is that the drivers are being used also as bias regulators, since they form part of a simple voltage divider between the +300v and the -300v shown. If these two supplies dip under load, I'm unclear how stable bias is maintained since the vbias adjust element has no feedback and is a simple divider off a zener reference. That point will stay solid, but the -300v and the + 300v are free to move about.

He speaks of "special feedback" but I'm not seeing any obvious feedback loop.
If there is some dynamic adjustment inherent in the design, I missed it so far, and he did not point it out.

I'd prefer putting the very same choke he specifies at the grid of the 811s and putting the bias on the CT. The idea of a low Z (at AC and DC) regulator there is fine. Then just drive the grids with cathode followers, tube or solid state. Much simpler, I think more effective.

The "distortion" that he speaks of is only when and if the tubes draw grid current. It ought to be minimal assuming that there is sufficient power/current capability in the driver (with the choke in the grid of the 811) sitting on top of the choke. IF there is any non-linearity, it wll be only on extreme voice peaks - in which case this is a beneficial feature, not a problem, since it acts like a bit of soft knee limiting, rather than hard clipping.

Also, don't forget the "3 diode limiter" is a good friend for almost all modulators.

[WBear2GCR]

yah, also he says that the all tube version doesn't work as well because the tubes are slower than the transistors? I find this somewhat odd since in reality tubes typically have much wider bandwidth than transistors (talking standard silicon power transistors here, not GaAn etc.)! I'm not sure why he found this to be so...

[W9ZSL]

Actually the speech amp/driver will be a separate unit.  I have a 15 watt Kenyon 500 ohm balanced to class B grids xfmr that will be mounted on the modulator chassis and fed from my HB limiter amp as the main driver.  I'll add a switch so I can drive the 811As with either unit.

[WD5JKO]

Took a quick look at the WA5BXO page.

I'm not sure what exactly the idea behind this is. One of the issues is that the drivers are being used also as bias regulators, since they form part of a simple voltage divider between the +300v and the -300v shown. If these two supplies dip under load, I'm unclear how stable bias is maintained since the vbias adjust element has no feedback and is a simple divider off a zener reference. That point will stay solid, but the -300v and the + 300v are free to move about.

He speaks of "special feedback" but I'm not seeing any obvious feedback loop.
If there is some dynamic adjustment inherent in the design, I missed it so far, and he did not point it out.

I'd prefer putting the very same choke he specifies at the grid of the 811s and putting the bias on the CT. The idea of a low Z (at AC and DC) regulator there is fine. Then just drive the grids with cathode followers, tube or solid state. Much simpler, I think more effective.

The "distortion" that he speaks of is only when and if the tubes draw grid current. It ought to be minimal assuming that there is sufficient power/current capability in the driver (with the choke in the grid of the 811) sitting on top of the choke. IF there is any non-linearity, it wll be only on extreme voice peaks - in which case this is a beneficial feature, not a problem, since it acts like a bit of soft knee limiting, rather than hard clipping.

Also, don't forget the "3 diode limiter" is a good friend for almost all modulators.

Those two transistors in the audio path are in an emitter follower configuration, i.e high input impedance, low output impedance, output at ~ 0.9 Vin, and in phase with the input. The current gain of the circuit is basically equal the HFE or gain of the transistor. The emitter follower has 100% current feedback, called "negative current feedback circuit".

http://www.mycircuits9.com/2012/07/npn-emitter-follower-tutorial.html

If the +/- 300 volts supply move around, this has little consequence on the emitter voltage so long as an overload condition does not occur.

Also, sometimes I find some old thread on this board relating to a topic such as this one from the technical heavyweights that seldom post here anymore. Here is one:

http://amfone.net/Amforum/index.php?topic=21901.msg156957#msg156957

There was a post in that old thread from "The Derb" that gave me goose bumps:

"Do the best you can with what you can run.

but remember, you're not talking with a static collection of lifeless metal. It's only a tool, like a hammer. The only thing that makes AM special is that other AM'ers are on the other end. I dont love doing this because the guys on the other end are running low distortion transmitters. It's the shared experiences and the lasting friendships that count.

Having said that, everyone likes nice tools in good shape. they make the work easier."

RIP Derb...

Jim
Wd5JKO

[W9ZSL]

Good suggestions.  For now I'm just going to start with the basic modulator deck; mod iron, filament iron, Kenyon input iron, controls and the 811As.  I'll have plenty of room to add on.  Power HV supply will give me 1250-1350 volts on the plates at 300 ma loaded.  I'm starting with zero bias.

As for the driver, I do have an NOS S-8 and I can see where you're coming from with the fidelity thing.  That's why the coupling caps are so small and with the transformer, replacing those little guys with .01 might be pushing it.  THAT DRIVER AIN'T WRITTEN IN STONE.  I do have all the parts including an NOS Stancor P-6012 for the PS.  As long as an independent speech amp/driver is in the future, who cares?  Make it up as you go along.  Now, if I could get my mitts on a Stancor A-4761 driver iron, we are talking serious!

Hi-Fi.  I was a broadcaster for 35 years.  Gotta have it.  A push-pull 6V6 driver would be the cats.  I would switch between it and the Kenyon iron.  I'll post a pix of the parts post-haste.

Got pix.  The Kenyon 500 ohm to Class B grids T261 iron is actually 7 watts.  It should work fine.  Remember this is a "bare bones rig" and I'm harvesting the Jonque Box. Fighting kooks, cooties and spideys all the way.  You may have Deja Vu.  

Mod deck with the S-22; will have a viewing window in front starring the 811As..  The Fil iron to the left will be under the chassis.  The Kenyon Type T261 line to grid is on the right.  All audio will be routed to the right side of the rack with AC and HV left.  
I'm still working on the layout for the PS. The SS rectifiers will be re-mounted.
Iron is Thordarson including a 400 ma swinging choke input filter.  Total 16 mfd at 2 KV filter caps.  Relay on the primary with slow-start.  It will be activated from the Modulator Deck along with the filaments. The deck will also contain a CW/AM heavy-duty switch and an audio input switch to select Line or Driver. I know, Those Johnson sockets should have 810s in them, but remember, these will look spiffy in the "window"!  

[KL7OF]

Look at website "playthings of the past"  The guy has a lot of audio transformers....I got a couple a-4761 from him....By the way, I piss beat a S-22 for many years at over voltage and over current and it still survives....Your mileage may vary...Good Luck .....Steve

[W9ZSL]

Yeah, Steve!  I was first licensed in '61 as a Novice and then the old Conditional with K9ZSL the call.  License lapsed.  I spent most of my High School years working for the local KW on 1240.  I went on the air for the first time on my 16 birthday.  I'm 68 now.  Point is, ever since 1961, I've wanted to build my masterpiece.  Disizzit!

I almost had it when I scored a special edition Hammond that had a hard-wired 15 K Ohm CT and three sec taps rated at 175 watts.  Somehow it didn't feel right.  Finally ended up with this.  Of course, I also have an NOS S-20 good for 55 watts and the VM-3 rated at 125.  I don't want to buy any more!  It's gotta end some time...Damn Compulsions!  A POX on you!

[Opcom]

BTW that S-8 transformer is only rated 1 watt!! How about trying a larger driver xfmr?

The S8 transformer should work OK for your single ended driver circuit. Its rated for 5 watts. A push pull audio driver would work better of course with a S9 transformer, but if you already have the S8, use it. Good luck.

Al VE3AJM

Which one is that S8 transformer, and from which catalog? It looks like there is a discrepancy in the catalogs here.
'49 gives no power rating.
'55 says 1W
'56 says 1W
'60 says 5W
'70 says 5W
'74 says 5W
They are all in the same G-3 case/size, what gives? If it is indeed the same part, then maybe 1W is for a wider power bandwidth than the 5W rating. Good thing this was brought to light.

[Opcom]

. error

[VE3AJM]

I have the 1974 TRW/UTC catalogue here.

I used the S-8 as the driver transformer, shunt fed, in a 75 watt modulator using triode connected 6AV5s. The audio driver tube is a triode connected 6AQ5. The mod transformer is a Hammond 2100.

Al VE3AJM

[W9ZSL]

According to my "S" series data sheet the most common tube used was a 45. They only run 2 watts output so.....  A S-9 definitely would be better.  No hurry on the driver.

[VE3AJM]

Just found a thread on significant changes to some UTC transformers over the times of manufacture discussed in a post from 2011 here.  http://amfone.net/Amforum/index.php?topic=29210.0

Apparently this also happened with the S-8 driver transformer amongst some others that UTC made. There were case style and color changes as well. This might help one determine or identify the vintage of the transformer in question. Interesting.

Al VE3AJM

[W9ZSL]

  Know what I think, AL?  I'm betting that UTC had a crapple load of parts left over from WW2.  I've seen S-8 thru 10 with a darker "crackled" finish.  I've seen most with the smooth.  I had an S-9.  Never sell anything!

Mine has the grey but the paint job is for crap.  It's baked on but there are hairs and hints of tarnish unevenness in the finish.  It wasn't repainted.  That leads me to believe UTC used surplus parts to retool for civilian use.  To come up with product fast, they didn't worry as much about paint finish quality as quantity.  If it was in storage, drag it out and sell it.

Correct me if I'm wrong, but the S-8 "wattage" question has a simple solution.  The 811 modulator shows an ART-13 with provision for a CVM-3 125 watt up to a 250 watt iron.  There is no difference in T-1, and the S-8 is listed.  I have the Stancor P-6012, so I can do this driver.  I'm eventually going to use the UTC S-22. This is from the 1949 West Coast Handbook.

Let's see.  We were talking about the S-8.  Logic time.  Both CVM and "S" series dates back to at least 1949.  The 6V6 will do 5.5 watts using the power supply shown.  The S-8 HAS to be at least 5 watts, probably under-rated.  The 811As need 4.4 watts.  Makes sense to me.  Maybe the 1 watt rating was a misprint after all so Opcom might be right.  The same handbook shows another modulator identical including the S-8 to drive P/P 807's.  See my post about building a practice mudulator.  This is it.

My 5933 replacements for the 807s require just .2 watts drive!  Wozers.  Higher Fi.  I'll probably replace the .003 with a .01 and the.005 with .05 because I'm going to stick the Triad Line to Grid transformer in front of the 500K pot so I can feed it from the studio.  S-20 iron mint.  Should be able to modulate my DX-60B just fine.