Homebrew 4-1000A plate modulated by a pair - Questions

[K1JJ]

A few questions - but first a summary:

I'm in the process of wiring up a 4-1000A modulated by a triode-connected pair of 4X1’s. I'm using a 1KW RCA BC modulation transformer with a 1.42:1 impedance ratio.  Have about 70H of Heising choke reactance. The plate power supply will be about 3500V, so the whole thing will be run conservatively.  It will have negative feedback from the modulator tubes to a low level audio circuit. Uses vacuum variable caps and turns counters for band agility

The modulator tubes and RF final will be on the same sub-chassis to permit cooling with the same blower.  I’ll post some pics when it’s working and finished in a week or so.


Questions:

Anyway, I'm wiring the RF and mod decks and had some questions on what values to use for bypass caps.


The triode–connected modulator tubes:  What value should I use for bypass caps to ground at the socket for the filament? (cathode)  There is audio there and I was thinking of about 0.1, but not sure. I remember someone once saying they didn't bypass the fils at all and just let the audio go through the fil transformer to ground. Any opinions?

The RF deck:  What value should the screen bypass cap be? I now have two .001 door knobs right at the socket pins to ground, then an RF choke going to the 20H screen self-modulating choke and to the 500v supply. Does the screen need further bypassing on the other side of the RF choke too?   Also, I remember someone once suggesting to bypass the screen circuit with an electrolytic cap that will pass audio. Where would that go in the circuit and what value, if desired? (This doesn't seem logical to me - maybe it's only for a linear amp screen)

What value caps are suggested for the RF tubes fil leads to ground?  .01 or .001?


Probably a lot of this won’t really matter, but it’s been a coupla years since I’ve built a plate modulated rig and have spent my time on linear amps. So the cap values are foggy as a result.

Thanks!

Tom, K1JJ

[KL7OF]

Tom...Very interesting...I had no bypass on the fils of triode connected 813's and then added .01 and could tell no difference .I had gobs of audio either way..That's my only experience...Not wanting to hijack your original question  but I am interested in the modulator negative feedback system you are going to use wid the 4x1s.....coupling technique and feedback voltage etc.I know you will keep us posted...Gud Luk   Steve

[N3DRB The Derb]

try this, OM.  

also sent U a e-type-mail with more.

[N2DTS]

As far as the screen bypass in the rf section goes, more bypass capacitance will lower the high frequency response of the rig, not bad up to a point.
.002 should be ok, unless you want one of those real wide signals....

The screen has to self modulate, and a lot of bypass cap will stop it at high frequencies, which would cut the high end some, not all, because the plate would still be modulating.

Strapping rig, a pair of 4-1000's in triode connection will take a lot of drive, what will you be using to do that?
How many watts will it take?

A real winter shack warmer!

Brett

 

[KM1H]

With the mod tubes on the RF chassis you may have to bypass the filaments depending on how hot things are. Just keep it in mind in case of cracklies.

I pretty much stripped a local 5KW BCB 4x1 rig in the 70's with a pair modulating a pair and built for 80-20M. Around 85-6 I sold it to a guy in CT but cant remember who. Does it ring a bell?

Carl
KM1H

[WD8BIL]

Vu, does that mod iron have a screen winding?

[WA1GFZ]

I agree with Carl on the mod bypass since they will share the same chassis and I think the transformer is external. On the RF side I would put a .001  or 470pf in parallel with the .01s to shunt any VHF resonant effects of the .01.

[K1JJ]

Excellent info!  Especially timely cuz I'm gonna finish the wiring on the sub-chassis today.

Derb:  Got the schematic. So looks like .002 at the screen pins AND a 4ufd on the other side of the choke, the power supply side.  I can see how that wud be fine, tnx.

Frank - Thanks for the bypass suggestions above - will do.

Steve/OF:  For audio negative feedback I'm using a solid state stage taken from the WA3KLR design for a cathode audio driver. See the schematic below. I input my 1 volt of audio along with the feedback from a resistive voltage divider taken off the 4X1 modulator plate cap.  They mix into the IC and then I feed the audio driver with it. I've used this same scheme on my last 4X1 X 833A rig and it worked perfectly.

Brett: The triode connected 4X1's will be fed by a backwards connected 8 ohm to 2K,  20-20K audio transformer driven by a 100W SS amp. I'm not sure how much drive it takes, but probably under 50w.  

This new project has been constructed from modular homebrew things from past rigs..  I used Quadzilla as the RF deck/mod deck. Also used the same vacuum variables and tank assy. Used already built screen and grid supplies as well as using the same negative feedback board and resistive divider.  Had to rewire a lot of stuff and make a new cooling system as well as wire the full modulator interconnects. The power supply is my old 5KVA pole pig that I’ve used for many years with a switchable 3500 or  6KV system.  The 4X1 would run on 6KV, but it wud likely crap out the mod xfmr – so I’m gonna stay wid 3500V.

Here’s a previous short thread on the RCA mod transformer:
http://amfone.net/Amforum/index.php?topic=6180.0
I don’t know what the p-p impedance is of the triode-connected 4X1’s but the according to the handbook, the 4X1 final should be about 5K ohms.  It usually takes a 2:1 step down for a pair of modulating one final of the same type.  This mod xfmr is about 1:1.42  step down, so reasonably close I’m hoping. I cud always adjust the final loading for a different impedance.

Bud – YES, the mod transformer does have a tap. I measured about 9 volts coming off it with 97 volts off the main sec. I was inputting 120V into the pri at the time. So maybe it cud be used instead of the screen choke? Any advantages to doing it this way? I’ve always used a screen choke in the past.

W5HRO:  What is the reason for the twisted leads on both the RF and mod fils?  I’m using Teflon wire and could twist the short leads under the chassis and also twist the incoming leads… but have never done this in the past. Right now they are not twisted. Please let me know what you think.


Carl:  I think I will bypass the fils of both the mods and RF tubes. Can’t hurt.

Steve, I  also saw no difference in the past when bypassing the modulators. When un-bypassed, you wud think it wud have an effect on the lower freqs, or cause some kind of phase shift with the higher audio freqs, but maybe not.


One more QUESTION:  The grid leak calculation calls for a 5K resistor. (200V grid leak bias / 40 ma = 5K)    It takes only about 10 watts to drive since I am neutralizing it and have found I usually don’t need grid swamping as a result.  What if I made the grid leak resistor a higher value like, 8K to send the final deeper into class C, thus more efficient? I have the extra drive available. Does this make sense and is there some disadvantage to doing this?  I wonder if there is a limit. It wouldn’t be right be knocking on the class E door using tubes… HA!

T

[K1JJ]

BTW, for whomever is interested:

In the past, I have found triode-connected 4X1's to be very stable and clean as modulators. Tron ran them for years and wants to duplicate his rig again. They require no biasing at usual plate voltages - just tie the grids and screens together and feed them directly with the hi-fi audio transformer. The CT of the audio transformer goes directly to ground.  3-4KV usually idles the tubes FB. If not, a small bias, either positive or negative could be put on the CT to place them right in the proper idling current.

In contrast, I've found the 833A's to be more sensitive to human errors. I've lost a few when an audio parasitic took off.  At 3500-4KV, they can be a handful. The 4X1's seem to handle abuse quite well, being bigger tubes with operating ratings as high as 5500V in mod service.  At 3500V, they will be running very conservatively, giving good IMD numbers and require low air blower settings.

So, that's my thinking for using the bigger modulator tubes. We'll have to see how the rig sweeps out.

T

[K1JJ]

W5HRO:  What is the reason for the twisted leads on both the RF and mod fils? I’m using Teflon wire and could twist the short leads under the chassis and also twist the incoming leads… but have never done this in the past. Right now they are not twisted. Please let me know what you think.

Twisting the filament leads should eliminate 60 cycle hum and also help further reduce any RF getting in as well.

Usually when you have a separate mod deck (sorry I missed that one before) some people will use bypass caps at the socket connections to eliminate 60 cycle hum not RF. However, using twisted pair is really all you need to do. If you look thru the old radio handbooks you normally won’t see any bypass caps at the modulator tube socket filament connections.

All filament leads in the RF deck should be twisted as well.

Makes sense. This also applies to directly heated cathodes, like the 4-1000A?   It will be easy to do on this rig. Thanks for the suggestion.

T

[N2DTS]

I always go past the recomended bias settings on my rigs.
I usualy use a big wire wound pot, and have the grid metered for voltage and current, and know you can go WAY beyond what the books say, with not much more drive needed.
Past a point, there is no improvement in the modulation, but it does not hurt anything, and might make things a bit more efficient?

Use the 8k....

Brett

[Steve - WB3HUZ]

Tom, consider doing some trapezoid tests to find the best bias point. The info could be informative for others. I've heard some claims of more linear moduation of finals when using mucho grid leak and heavy drive but I've never seen any measurements.

[WA1GFZ]

Tom a triode connected 4-1000A with zero bias will be the the same resting current as a GG. At 3500 volts I figure 100 to 125 ma per tube. That is well into AB and should be easy to drive. You could use a Zener at the CT of the heater transformer to knock in down a bit unless you need a nice room heater.

[KC2IFR]

Tom,
Working with Jamie on my 4x1 x 833's we didnt find any problems with the 833'3 at 3400 volts. I think 833's in the rf section is a different story. Jamie designed feedback around the 833"s and the whole transmitter is a joy to operate......very forgiving no surprises.
Bill

We do have separate bias adjustment for each 833 and separate cathode meters for each toob. Cant have to many meters!

[K1JJ]

Steve,

Yes, I think I'll experiment with different grid leak bias and even modulator idle current. Using a trapezoid is a good idea to see what's happening. I could even vary the audio neg feedback to find the overall sweet spot.

Frank, 125ma at 3500V is around 1/3 plate diss, so that wud be a good idle, esp during the cold winter months... :-)

Bill, yes, the 833A's are time tested in BC transmitters for years, so they work well when set up correctly. I'm always changing things around and experimenting, so things tend to get stressed. I was using new Chinese 833A's and made a few operating "human errors" with them.  I've never had a pair of 4X1's blow out though.

Well, got the subchassis all wired up now with the bypassing we all talked about above - and now concentrating on the plate tank and interconnections. Got on AM briefly tonight with the Tron and company. He said the RCA mod transformer will be OK up to the limit of 3500V, so that's the limit or bottleneck to higher power...  It will be nice to be back on with a plate modulated rig after a coupla years of using linears - they're so frickin boring. As Tron said, linears just lie there like sick whores...

Maybe in a week I'll have it working.

T

[KC2IFR]

Tom,
The research and careful construction u do with anything u build is first rate. U are not afraid to ask questions about something u are not sure of.......that is a quality not often found!
JMHO....

Bill

[K1JJ]

Tnx, OM...

I do axe some stupid questions at times...     It tends to get the thread going and some people involved who would normally not respond.   In the end I always learn a ton of new stuff and proceed with confidence on the project - as wid this 4X1 rig tonight.

BTW, I heard your 4X1 X 833A's on the air about a week ago. It sounded flawless.  Nice BC audio and full quieting over here. You guys did a good job getting it worked out. There's so few of them on that I'd decided it's a good choice. It's always been my favorite lashup, the 4X1 final modulated by 833A's or 4X1's. Over the decades, I keep coming back to it after going with simpler linears for awhile.

At this point for amplifiers, I'm down to the HB solid state eight FET linear, my big 8877 linear for ssb and now this 4X1 X 4X1's for AM.  Everything else has been either torn down for parts or put off line in moth balls. My life was getting too complicated maintaining all this other stuff and I started making operating mistakes after being away from a particular rig for a while. Just a big rig for ssb and one for AM that are band agile is my new goal.  I'll bet you know the feeling.

T

[KL7OF]

So Quadzilla is an organ donor for this new rig? ....... Mod Quad?....I follow with great interest...Steve

[KM1H]

Tom, consider doing some trapezoid tests to find the best bias point. The info could be informative for others. I've heard some claims of more linear moduation of finals when using mucho grid leak and heavy drive but I've never seen any measurements.

I believe Ive mentioned that on here in the past and Tom asked me about it at the May Nearfest. I can drive the 250TH's on CW at around 50W and for AM I cranked the bias to deep Class C (thats the term Ive been using) so that it takes 75W from the Viking I and the scope looks much better than at 50W. Im also burning a little power with resistive swamping and series grid resistors because the Viking doesnt look that good when backed down. The other amps are run similar.

I should be back on the lower bands in a few weeks after a few more tower climbs before the snow arrives.

Carl
KM1H

[WA1GFZ]

Tom,
That is exactly why I'm planning the dual 8 erbs for the future. KISS
After the big rig is back up.

[N2DTS]

Tom,
Heard you on with Tim asking about the mod iron.
I use 2 of those same transformers (at lower voltages) and they sure seem great, much better than the cvm5 I used to use.

I suspect its a bit light for your rig, but should work well if you limit the power and voltage a bit.
Who knows what it can actualy take, but it has no problem at 2500 volts, 400 ma, 600 or 700 watts of audio.

Brett

[K1JJ]

Steve/OF:  Yes, Quadzilla was a major organ donor.  I actually removed one 4X1 and left three - one for the final and the two modulators.  The air flow is better without the end tube close to the blower.   I never did like having two 4X1's in the final in the past. Too much fil, screen and grid current - a waste. Heck, one 4X1 will do 2600w carrier out with 5500v on it, so why go wid more than one?      I just finished up making a new band switch and tank coil to cover the major bands. Also installed a pair of those good metal turns counters. Those black plasdick ones seem to fail after a while. Their turns counter gears jam up.

Carl: That deep class C is intriguing. I located the grid leak resistor  (a big 100W windwound) in an easy to get to clip-on board so I can change it. I might axe Tron what he thinks of the idea too.  I've found that running more than recommended grid current always made the tube come to life in plate modulated service. Better audio peaks. Maybe deep class C will work out too.

Brett: Yes, I like the look of that transformer too. You have two of them? How are you using them in service?  I’ll bet it weighs at least 130 pounds.   What amount of Heising reactance are you using and how low do the lows look good until it falls apart? I found with 70H, sweeping the old 4X1 rig was good to about 40hz or so. My old big RCA transformer was better. What have you found?


I used to have an RCA 10KW 1:1 piece of iron here. What a beauty. It had a matching Heising reactor. I got it from the Apeman W2VW.  Sold it off maybe 10 years ago. Too big and heavy, like 1300 pounds for the pair.



Another question, this time real stupid….  I realize that to modulate say, a 1KW carrier 100%, it takes only about 500w of audio power, though, the RF power increases four times or to 4KW pep.  Now, where is all that extra peak power (2500W) being drawn from? … is  the power supply supplying it thru the Heising reactor if it’s not coming from the mod xfmr?   I can’t get it outa my head that if the modulation iron is causing the voltage to double in series with the plate circuit, why it is not supplying more than 500w of that 4KW?  I’ve read the reason a few times in the past but can’t remember it now that I need to apply it to this new rig… sigh.  Don’t beat me up too bad now….

T

[Steve - WB3HUZ]

realize that to modulate say, a 1KW carrier 100%, it takes only about 500w of audio power, though, the RF power increases four times or to 4KW pep.  Now, where is all that extra peak power (2500W) being drawn from?

Yes, it takes 500 Watts average power for 100% modulation. But it takes much more than 500 watts peak power. Apples must be compared to apples. If the plate voltage is doubled (the definition of 100 percent modulation), the power must increase by 4x.

[N2DTS]

Maybe I have a different mod trans, but it sounded the same.
Mine is RCA, its big, maybe 75 pounds, and had the impedances you have I think, but its designed to have DC through it, I think they made two versions of it.
I tried the reactor, and got no improvement, so leave the dc on it.

One is in the 2x813 rig, modulated by a pair of 4cx250b tubes in ab1, at 2000 volts, or 4x100th tubes.
The other is in the 3x4D32 rig, modulated by 811a's all running 1200 volts, 300 watts carrier out.

I had a third one, in a push pull 812a modulated by 811a's rig, but I sold that rig on ebay.

I friend bought a warehouse of RCA transformers, lots of those mod transformers, and various matching power transformers, some quite large.
I guess the guy that had them worked at RCA, since it was close (in Camden NJ).

I never swept it, last time I tried that I blew out the mod iron in the 30K1 when it got down to 30 Hz!
I should sweep it (at low power) to see how good it is, but it SOUNDS good...

My rule of thumb is to have as much audio power as RF output power, things are real clean that way, with no limit of the positive peaks. More is better.

The power comes from the fact that you double the plate voltage (or more on peaks) to the RF final.
So say my 2x813 rig runs 2000 volts and 400 ma, under modulation, the plate voltage swings up past 4000 volts on peaks. If I put out a carrier at 2000 volts and 400 ma, then turned the variac up to 4500 or 5000 volts, that would be the same as modulating it....

A far cry from these crappy linyar amplifiers, I have four 813's in grounded grid and can get about 150 watts carrier out at rated plate disapation!
Four 813's in plate modulated service would do 1200 watts carrier and 4800 watts pep all day long without getting hot!

Now, for sideband service, the duty cycle is low, so I could crank the power up to 3000 volts and get 1500 or more pep watts out with the tubes being cool....

I would never consider running an amp on AM, but the flex demands it....
Lots of power in for little result, but it will make a good space heater for the shck in the winter.


Brett
 

[WA1GFZ]

looking at 100% modulation on a spectrum analyzer you have 1 KW carrier with 2 sidebands 6 dB down. Each sideband is 250 watts since it is 6 dB down from 1 KW. It takes 2x voltage peak modulation to create 250 Watts RMS sidebands. The final is a high level mixer in this case. F1+F2, F1 and F1-F2
F2 (audio) won't come through the plate coupling cap