Class B modulator current requirement?

[Edward Cain]

Hi all,
   My question concerns the plate supply current requirements for push/pull class B modulators.
   As a specific example, I see that a pair of 810's with 2250V on the plates has current requirements of 70 ma zero signal and 450 ma maximum signal.
   So, does the plate supply need to be rated at 450 ma or more, assuming a separate supply for the 810's?

Thanks in advance,
Ed, KJ4JST

[KE6DF]

Hi all,
   My question concerns the plate supply current requirements for push/pull class B modulators.
   As a specific example, I see that a pair of 810's with 2250V on the plates has current requirements of 70 ma zero signal and 450 ma maximum signal.
   So, does the plate supply need to be rated at 450 ma or more, assuming a separate supply for the 810's?

Thanks in advance,
Ed, KJ4JST

Some of the gurus may have more on this.

But, the 450ma figure is the peak current which only occurs on voice peaks.

So the modulator supply only needs to put out an average current which is much less that 450ma.

It does, however, need to be fairly well regulated so an occasional peak current demand doesn't cause too much voltage sag (and therefore distored output).

Something like a choke input filter with a large output capacitor would work.

[k4kyv]

My HF-300 rig runs about 2500 volts on the modulator stage.  I set the static plate current for each 810 to somewhere around 40 ma per tube (have a separate plate current meter for each tube). The power supply is capable of delivering well over 500 MA, but each of the meters kicks up only to about 100 ma per tube.  Most of the time when I talk, the plate current just barely peaks at about twice the static plate current.

OTOH, I can feed a sine wave tone into the modulator, and at 100% modulation, the tubes run as high as 200 ma per tube, and the plates turn cherry red.  With voice modulation, they just barely show a perceptible dark red glow.  According to RCA, I am still running them within the limits of CCS, although I may be pushing the plate voltage just a little. The tubes seem to be the limiting factor in my setup.

I would use a power supply that could furnish at least 500ma (including bleeder current) for at least a short period of time without damaging anything.  I wouldn't want a modulator that would go into melt-down mode with an inadvertent squeal from acoustical or rf feedback, or from whistling into the mic.

[steve_qix]

The supply needs to be able to furnish the maximum signal current.

I would strongly recommended against using separate supplies.  There are so many good reason to use a common supply.

1) The RF amplifier supplies a nice, constant load for the supply which help stabilize the voltage, particularly with a choke input filter.

2) The load is more constant so the voltage going to the modulator is more stable.

3) It makes going transmit/receive easier because there is only one supply to switch on and off.

4) A single supply is smaller than two smaller supplies

5) It is almost impossible to power the modulator without powering the RF amplifier because there is only one supply.

6) It is safer to use one supply - less switching, bleeding, etc to worry about.

7) It is probably less expensive to build a single supply.

But, to answer the original question - the supply should be able to furnish the maximum signal modulator current, because there are times when you will in fact see that much.

Hope this helps!

Regards,

Steve

[Edward Cain]

Thanks for the comments guys.

Steve, I'd love to use just one supply but I haven't found one yet. I want to use the 810's to modulate a pair of 4-250's. I have the tubes and I have a couple of T-368 plate transformers. One is rated 500 ma and the other 320 ma.

[WA1GFZ]

Put rectifiers on each transformer and parallel the positive and negative outputs of each rectifier into one filter. I did that with a pair of BC610 transformers and hung the primaries across a 240 line for a stiffer source.

[K3YA]

Thanks for the comments guys.

Steve, I'd love to use just one supply but I haven't found one yet. I want to use the 810's to modulate a pair of 4-250's. I have the tubes and I have a couple of T-368 plate transformers. One is rated 500 ma and the other 320 ma.

I would expect that a T368 transformer rated at 500mA continuous service under harsh conditions would work fine for a 50% duty cycle ham transmitter running 400mA on the final plus a class B modulator.  Just make sure the 240VAC primary wiring to the transmitter is up to the task. and be very, very careful not to exceed 1500 watts PEP.

[KM1H]

Since the voice only runs at about a 30% duty cycle all you need is enough output C to supply the peaks. A choke input and 30-50uF on the output should keep an anemic transformer happy. IOW the single PS needs to supply the Class C current plus the modulator resting current plus another 30%. A bit more if a lot of processing is used.

For a seperate PS for 810's a 225-250ma CCS transformer is sufficient unless you are a real windbag.

Dont cut it that close with the choke as excessive ISqR losses will negate the effect of the big C holding the regulation.

Carl

[The Slab Bacon]

I'd love to use just one supply but I haven't found one yet. I want to use the 810's to modulate a pair of 4-250's. I have the tubes and I have a couple of T-368 plate transformers. One is rated 500 ma and the other 320 ma.

IIRC, that is the tube compliment of a Collins KW-1. They used a common plate supply, so there is no reason that it shouldn't be doable. the 500ma transformer should be able to handle it, but....................

If you paralleled the rectifier outpoots of both transformers through the rectifiers and a couple of steering diodes before hitting the choke inpoot filter, 800ma would have absolutely no problem kicking butt and taking names! !

[Edward Cain]

GFZ and Slab,
   Though both transformers are from T-368's, the 320 ma is 5760 vct and the 500 ma is 6336 vct. Seems like this would rule out paralleling.

Ed

[KC2ZFA]

If you paralleled the rectifier outpoots of both transformers through the rectifiers and a couple of steering diodes before hitting the choke inpoot filter, 800ma would have absolutely no problem kicking butt and taking names! !

Frank, what do you mean by this ? If I had two of, say, these http://www.antekinc.com/pdf/AN-8T800.pdf couldn't I just series each and parallel connect the resulting two 1600V 0.5A xfmrs to get 1600V@1A ?

Peter

[WD8BIL]

Though both transformers are from T-368's, the 320 ma is 5760 vct and the 500 ma is 6336 vct. Seems like this would rule out paralleling.

Variac the big one down or use voltage droppning resistors in the primary.

[W2PFY]

Those T-368 transformers were designed for continuous use so I don't think you have to worry about duty cycle. If your building a class B circuit you need a stiff bias supply since there will be grid current. You can get around the bias supply system by using a cathode follower driver or you could use zener diodes in the cathodes to obtain operating bias, if your going to use a transformer to drive the grids.

[The Slab Bacon]

If you paralleled the rectifier outpoots of both transformers through the rectifiers and a couple of steering diodes before hitting the choke inpoot filter, 800ma would have absolutely no problem kicking butt and taking names! !

Frank, what do you mean by this ? If I had two of, say, these http://www.antekinc.com/pdf/AN-8T800.pdf couldn't I just series each and parallel connect the resulting two 1600V 0.5A xfmrs to get 1600V@1A ?
Peter

Peter,
         You can series connect the AC outpoots, But I wouldn't directly parallel connect them. If you have any difference in the windings output voltages it will act like shorted turns in the windings and cause the transformers to heat up.

Connect the individual transformers to their respective rectifiers. Ground (or whatever you plan for B-) the negative side of both rectifiers, then take the positive outpoot from both rectifiers, put a diode in series with each one and tie the output side of each "steering diode" together and into the filter choke inpoot side.

This way if there are slight differences in the outpoot voltages of either transformer / rectifier, It will isolate each one from the other and will not be apparent to each other or the whole power supply.

[The Slab Bacon]

GFZ and Slab,
   Though both transformers are from T-368's, the 320 ma is 5760 vct and the 500 ma is 6336 vct. Seems like this would rule out paralleling.
Ed

Ed, If you used individual rectifiers for both transformers and steering diodes, It would probably still work. (at least to a degree) the 500mil transformer would most likely take the lions share of the load and the smaller one would probably start picking up more of the load as the secondary voltage from the bigger one starts to sag a little from the load. If you plan on running full wave center tap rectifiers, the voltage difference between the 2 transfomas really isn't that much, it's only a couple hundred volts. Probably between 2 and 5% of your outpoot voltage.

[W2XR]

GFZ and Slab,
   Though both transformers are from T-368's, the 320 ma is 5760 vct and the 500 ma is 6336 vct. Seems like this would rule out paralleling.

Ed

Ed,

I ran a T-368 6336 VAC at 500 Ma plate transformer (mine was made by Industrial Transformer Co.) for years in my hombrew single 4-400A modulated by a pair of class B 833As (before I upgraded to a pair of 4-400As in the final), and that transformer ran as cool as a cucumber. I never noticed any perceptible temperature rise in that thing. Yes, it was subjected to only ICAS operation, but that tranny is oil-filled and must weigh at least 100 lbs. It was most probably grossly overated for service in the T-368, and I think you will find it to be entirely satisfactory for any type of emission or power level of amateur operation. In the T-368, my guess is that the transformer is rated for CCS operation.

The only reason I replaced the T-368 transformer (when I went to a pair of 4-400As in the HPA), was because I acquired a NOS/NIB Basler 6200 VAC CT at 1.2 amp continuous-duty plate transformer for a Gates BC-1G kilowatt AM BC rig. I could not resist the temptation to put that beast in the rig, and I subsequently passed the T-368 plate tranny on to someone else.

I hope this provides some guidance to you.

73,

Bruce

[steve_qix]

That 500mA transformer should be plenty big - at over 100lbs, that should be plenty of capacity.  I've used similar transformers with bigger transmitters, and they do just fine and don't overheat even after those long evenings of old-buzzard transmissions.

You should be all set for a common power supply. 

Suggestion: Use solid-state rectifiers.  Less heat, MUCH more reliable than the venerable mercury vapor tubes (and safer, too).  No filament transformers is also a big bonus.  The book says for a full wave center tapped rectifier, the PIV rating must be 2.8 times the RMS voltage.  I use at LEAST 4x the RMS voltage.  The 1n5408 is a 1000V, 3A diode and they cost .25 each at Mouser.  You could to put 16 of them in series in each leg (they are small, so no problem) - giving you 16kV per leg, and the whole thing would cost you about $8.00.  Spend an extra $2.00 and put 20 in each leg for even more of a safety factor 

(sorry if I offended the tube purists  )

[k4kyv]

Everything said so far about the merits of a common power supply is true.  But there is one important merit to using separate supplies; this gives you some control over the positive peak capability of the transmitter with a given modulation transformer turns ratio.  For example, my HF-300 rig runs separate supplies, and the mod xfmr turns ratio is 1.55:1. Running the same plate voltage on both modulator and final, I can make just a little over 100% modulation and that's it, regardless of how much or how little plate current I load the PA to.  But running the modulator plate voltage at 2500-2600 volts while the PA runs at about 2000 plate volts, gives me up to 130% positive peak capability.

With separate supplies the modulator power supply must be designed for the best possible dynamic voltage regulation - exactly the same requirements as exist for a slopbucket leen-yar.  I use a 8-30 Hy swinging choke, a 25K bleeder resistor, and about 25 mfd of filter capacitance. This seems to be a good compromise. Another way to go: a "contemporary" HV supply with solid state rectifiers and no filter choke would work just as well with a class-B modulator as it does with a SSB linear.

I use an alternative to separate supplies with the Gates BC1-T.  I run the main power supply at the full stock 2600 volts, with full voltage applied to the plates of the modulator tubes.  But I replaced the big rheostat that fine-tunes the plate voltage to the final, with a bank of fixed wirewound resistors, which drop the plate voltage to the PA to around 2000 volts.  That cuts back the carrier power of the transmitter, but provides enough additional headroom to allow up to 130% or so positive modulation capability.

[KC9LKE]

 

[KC9LKE]

That 500mA transformer should be plenty big - at over 100lbs, that should be plenty of capacity.  I've used similar transformers with bigger transmitters, and they do just fine and don't overheat even after those long evenings of old-buzzard transmissions.

You should be all set for a common power supply. 

FWIW:

IIRC the Gates BC500-T Plate xformer I have setting on the shelf is marked
4750 VCT @ 460 ma. Thats 500w carrier 24-7-365 CCS.


Regards

Ted / KC9LKE

[The Slab Bacon]

this is for Peter and Ed who did not completely understand the concept of isolating the 2 transformers with "steering diodes". Please see the attached PDF sketch. It is a very simple concept when you see it drawn out. All of the diodes shown in the drawing should have the same voltage / current ratings.

This eliminates any issues caused by slight differences in the transformer voltages and / or primary phasing issues. It is actually quite simple and can be used with either FWB or FWCT rectifiers. I drew it with FWCT, but it doesn't really matter which you use as long as your transfomas are up to it.

[KC2ZFA]

me like !

tnx

[KE6DF]

I drew it with FWCT, but it doesn't really matter which you use as long as your transfomas are up to it.

I guess I'm not seeing it.

If the transformers were almost exactly the same voltage, or off by only a few volts, then yes I can see how it works.

But if the transformers differed by several hundred volts as was stated earlier in this thread, then one of the stearing diodes would always be reverse biased and the higher voltage transformer would carry all the load -- no matter what the phasing.

Dave

[k4kyv]

The way I see it, is that if one transformer sags below the output voltage of the other, then the other transformer takes over.  But then that takes the load off the first transformer and allows the voltage to recover, so they reach an equilibrium point where each transformer is loaded down just enough that they sag to equal voltage.  If there is a big difference in the unloaded output voltages of the two transformers, they might not share the load equally, so it would still be necessary to make sure the current rating of neither transformer is being exceeded.  Two nominally identical transformers, for example a pair of BC-610 plate transformers, should for all practical purposes share the load equally.

Sometimes variacs use two or more separate autoformers ganged together and wired in parallel to deliver more current.  They don't just wire them directly in parallel, but have small inductors connected between them so that small voltage differences don't cause excessive current to be drawn, heating up the transformers and brushes, and reducing efficiency.

[The Slab Bacon]

Dave,
        You are right if the voltage difference is excessive. But, the difference between the 2 transformers that Ed has is only around 2.5% + after the rectifiers. As Don said, after you start loading it down the difference will be even less.

Granted, one transformer will be working a little harder until you start loading down  the supply and sagging the outpoot, then the other one will start to work harder as needed. Depending on how heavily it is loaded down, it will eventually reach a point of equalibrium.

If you measure the HV as you increase the current load on the power supply, you will be quite surprised how much it moves around under load.