Power supply design for OB TX

[N8UH]

I've seen many posts on PS design here, but I still can't seem to find an answer close enough to get me going...

I'm getting quite close to beginning construction on my FB OM OB TX. Of course, I need to use the parts I have on hand, or ones I can reasonably expect to acquire in the near future. My plans so far:

Mod section:
- 810s driven by? (still working on that.)
- Modified Heising modulation using 5KW Dahl broadcast iron, 1.4:1 step down, 35H reactor.

PA:
- 810s or 8000s, balanced push-pull. B&W CX-series butterfly capacitor with HDVL jackbars and coils. Possibly using TVL or BVL coils in the grid circuit

HV power supply - here is where my question is. I would like to use a nice Dahl transformer I have on hand. It's a modern 1A 3KV affair meant for bridge use. I will be using vacuum rectifiers for the bridge. If I use 120V (I have to use 120V here. Dedicated 20A circuit though) on the 208 tap, I'll get approximately 1900V out of the secondary, and applying the normal PS calculations, I figure I will be able to get around 1650VDC out of the PS.

I would like to go with a common power supply for the Modulator and PA section, but a few questions have popped up in my head...

1. Given the fact that I will be using the modified Heising approach, the PS will be further filtered by the modulation reactor. I was originally planning on a 2-stage LC filter on the PS. 10H > 16uF > 10H > 16uF gives me pretty good results using the Duncan simulator, but that obviously does not factor in the additional filtering taking place from the modulation reactor. Would a single-stage LC filter, using series reactors and more capacitance be a better approach? Obviously starting surge current and voltage need to be factored in to this all, but the values above seem to be a good compromise. I'm not sure how to calculate the optimum filter values here. Thoughts?

2. I'm not quite sure how happy the class-B 810s will be at 1600V. The data sheet shows typical class-B CCS operation at 2000V, 420mA, yielding 590W of audio. I'm not sure how to calculate things from there, so I could use some education! Should I be looking for another solution here?

So far, I am not really deviating too much from 30's-40's design, which is good. I just have to use the modulation iron I have, and that transformer is the only thing I have that will give me enough current to supply both stages comfortably.

So what do y'all think?

Thanks!

[W7TFO]

Fancy meeting you here...

I'd set the PS up using 4 MV tubes in a bridge, and get yourself more grunt that way.

The only down side is it take three fil trans (or windings) to achieve it, but you already have those coming from me.

That Dahl plate trans you have has a primary winding on each core leg, and if they are in series, maybe do a little operating on them and reconnect in parallel.  You gettum' plenty strap that way.  *10's are 2kV+ tubes, much happier up there.

If you are still flummoxed, I will be coming thru Flagstaff next week and could trade you some iron with 120V primary and more outpoot than what you have now.

73DG

eGads, we are gossiping and carryin' on in public!

[KE6DF]

You can use the Duncan simulator to see the output on the final tube side of the mod reactor.

Just configure the mod reactor as another LC stage with an output capacitor of .01 uf.

Then you can see what the tube will see.

I bet you will find a single choke input stage will work just fine for an AM rig as the final presents a steady load on the supply well above critical.

Just put your two 16uF in parallel as the output from your single LC stage.

Dave

[Opcom]

I echo that LCLC vs LC sentiment, the overall result disregarding the way the supply charges seems alike. I tried both ways on the Tucker when testing the power supply and it didn't make much difference as long as the product of the total L and the total C were the same. I mean, going 4H-16uF-4H-16uF gave very close results to just going 8H-32uF.

To help figure the 810 performance at a lower voltage, compare with other tubes where several voltages are shown. If the tubes can draw enough current safely at a lower voltage to make the same power, then all you need do is provide a lower p-p load for them to work into. the kinds of questions may be:

1. how much more grid drive is needed to get the plate current up to make the same power? Will that endanger the grids?

2. How much will the plate to plate load impedance have to be reduced in order to get that power level (draw the necessary current)? The only way to really figure that is by the curves and the formulas for class B audio.


What about using a 'variac' o bost the line voltage for the Dahl? Or use a control transformer of some kind like a 48V 20A one?

120V = 1650V

120V
+ a 0-135V variac = 135VAC = 1856V

120V
+ a 48V 20A transformer = 168VAC = 2310V

How does that look?

[KX5JT]

I still don't know what "OB" means. 

[W7TFO]

Old Buzzard, you old buzzard!

73DG

[KM1H]

What a waste of good iron

OB plate transformers are almost free at hamfests and sell off that mod transformer for enough money to do it up with vintage iron with cash left over.

If you were closer Id do a deal for 1KW Gates xfmrs, I have the HV and mod ones plus the B+ and mod choke. Could use something a bit stronger

Carl

[N8UH]

What a waste of good iron

OB plate transformers are almost free at hamfests and sell off that mod transformer for enough money to do it up with vintage iron with cash left over.

If you were closer Id do a deal for 1KW Gates xfmrs, I have the HV and mod ones plus the B+ and mod choke. Could use something a bit stronger

Carl

A waste? Oh, I don't know about that... I have them and I'm putting them to use... They're just a little - oh how would you say - underutilized... 

Patrick: Great ideas on the variac/control transformers, but I don't have anything like that laying around, and it would add even more weight to an already ample chassis. 

I guess what I'm wondering is if I'll be able to get enough audio to fully modulate the 300-400W of carrier I'm planning on running (with headroom of course.)

If necessary, I suppose I could always go with split supplies...

[KE6DF]

2. I'm not quite sure how happy the class-B 810s will be at 1600V. The data sheet shows typical class-B CCS operation at 2000V, 420mA, yielding 590W of audio. I'm not sure how to calculate things from there, so I could use some education! Should I be looking for another solution here?

Thanks!

Here is a datasheet for GE 810's showing specs for class B mod use with 1500V on the plate. Shows 500W output. Should be plenty.

http://frank.pocnet.net/sheets/141/g/GL810.pdf

[N8UH]

Ahhh, thanks Dave!!! 

So it looks like there will be no issues at that voltage. That makes my life a bit easier.

You mentioned in an earlier post about using the mod reactor in the PS circuit followed by a .01uf Cap to make the simulation correct. Seems to work fine. Out of curiosity, why the .01 cap?

Thanks all for your help. I love this site!

[KE6DF]

Out of curiosity, why the .01 cap?

Well, the simulator program can handle a series of LC filters. I needed some value of C to make the simulator happy. I chose .01 figuring it wouldn't have an effect on 60hz so wouldn't impact the simulated ripple.

Plus, wouldn't you want some kind of RF bypass cap at the "bottom" end of the plate choke anyway?

What do people normally use for this? .01 or .001?

[KM1H]

Use .005 and compromise.

If you have a big plate choke in the 500uH and up size there is little RF to bypass. Its when 100-150uH is expected to work on 80/160 that you need to start clipping the highs as electrolytics do not like RF.

The Viking I and II used a pair of .01 in parallel seperated by a VHF choke. I changed the choke to a 15 Ohm WW and relocated one .01 as part of the process of getting rid of the space shuttle audio. The .01 was then added right across the pair of 100uF filter caps which replaced the 8uF oil filled.

In the amps I use .0047 6KV or 10KV discs in both locations and usually a 225uH plate choke from Ameritron since I dont limit my AM to the low bands.