3B28 Problems

[Barrie]

Yesterday, I drug out an old power supply that I built many years ago.  It's a typical design, using a CT transformer, swinging-choke input, smoothing choke and a couple of 8 uf oil-cap, with a pair of 3B28s as rectifiers.

I was surprised that it still worked (I used it to power a pair of 4-65As on 6M, contacted the Mayflower), but it did.  Into a 3500 ohm power supply load, it made 1250 volts at 275 mils.

That didn't impress me much, so tonight I drug out a couple of SS replacements for the 3B28s and plugged them in.

I expected a voltage increase, and I got it.  Now I have 1301 volts, under load.

What I didn't expect was the current increase.  I went from 275 ma to 350 ma into the same dummy load.

Could one, or both of the 3B28s be bad?

73, Barrie, W7ALW

[k4kyv]

Maybe one or more of the tubes is bad.  3B28's and 866A's are supposed to have a very low voltage drop that is nearly independent of the current being drawn, much the same as solid state diodes.  You should not see that much difference in voltage output.  The tubes are rated at 250 mills average current each, so a pair of them should be good for 500 MA minus the bleeder current, and they should handle output voltages up to about 3100 volts DC.

[KE6DF]

Something like what you saw might happen if one of the tubes was bad and you effectively ended up with a 1/2 wave rectifier.

The voltage drop across a 3b28 should be about 10 volts -- so not much more than across a silicon diode. But with half wave rectification you would see a significantly lower voltage out of the filter - hence my guess that's what might be happening.

http://tubes.mkdw.net/sheets/049/3/3B28.pdf

[w3jn]

Indeed.  One or both of the toobs is soft and you have a nice voltage drop across them under load.  Also, measuring PS output into a resistive load isn't quite the same as when it's feeding an actual xmitter.

[Barrie]

Indeed.  One or both of the toobs is soft and you have a nice voltage drop across them under load.  Also, measuring PS output into a resistive load isn't quite the same as when it's feeding an actual xmitter.

Thanks for the relies.  When I got the two tubes into good light, I noticed that one was very dark, almost black on the inside.  So it's possible that I was indeed running with half-wave rectification.

I've placed a quote above that I'd like to follow up on.  I've often jury rigged a power supply and used a dummy load to see (especially with an unknown transformer) what voltage and current I might expect.

I've wondered if whatever resistive load I'm using had the same effect on the power supply system as an actual tube-final load.

Can you tell me what's different, and why?

73, Barrie, W7ALW

[KE6DF]

Well I'm sure there are people on this forum that know more than I, and I'm a newbe here, but these are my thoughts.

If you are talking about a constant load from a final amp running CW with the key down, they the load would be the same as a resistor test load.

But, the real issue with power supplies, is how they respond to a dynamic varying load.

In AM, the finals draw current to generate the carrier. But the class B modulator current varies widely as you talk. When you are not modulating, like between words, the current is the low idle current of the tubes. But on voice peaks, the current can be quite high. Of course, on voice peaks a plate modulated final will draw more current as well. So the current demand on the supply varies considerably as you talk.

SSB is even worse, in that when you are not talking, the current is very low, and also rises sharply on voice peaks. The ratio of peak to idle current on a SSB power supply is much greater because there is no carrier being generated and providing a base load.

Now if you put a DVM on the output of the supply, you might find it doesn't vary much. But if you looked at the B+ with a scope, you would see large fluctuations at audio frequencies.

The result would be distortion of the transmitted signal.

With only 8mf of output capacitance, distortion would be significant with this power supply. It would probably be noticable on AM, and quite bad on SSB.

Now days people use much larger capacitance in power supplies. Paricularly those designed for powering SSB linears.

In fact, these days, most supplies do away with the choke and just use a large capacitor filter (20mf minimum, but more likely 40mf or more).

Sometimes you also can use a choke, but put a capacitor in parallel with the choke chosen to make the choke resonant at 120 hz (the dominant ripple frequency). this also provides good regulation and low ripple and is used in some high end linear amps.

If you want to use your supply for an AM transmitter with as few changes as possible, I'd suggest you replace (or augment) the final filter capacitor with one of much higher value. At least 20 mf plus.

But then again, back in the old days, supplies like your were used for AM transmitters just as is successfully.

[N3DRB The Derb]

git rid of that black tube. smash it to bits and sweep up teh glass.

866's are sure pretty, but they are a royal pain sometimes. I like 3B28's and thats what I used in the drb HB rig.

regarding cap vs. choke input supplies, choke iron is cheap, and you really just cant beat a good swinging munky choke input supply - I wouldn't use anything else for a modulator or a linear amp supply. For a class C RF final, cap input is ok fine, just make sure yer cap is rated for the max voltage they can soar to with no load, plus about 40% for overkill.

Build for double the trouble, and there will be no trouble.

[KE6DF]

Here is a good article on resonant choke input filters.

http://www.geocities.com/SiliconValley/Pines/5440/supply.html

I think a resonant choke input filter is a great way to go, but the high transient voltages across the choke present problems.

You need a high breakdown voltage choke.

and the resonating capacitor is hard to find.

It needs to be something like 0.25 mf with a voltage rating of 5K - 10K depending on the voltage of the supply. It has to handle AC voltages so you can't use electrolytics.

(The last time I posted a link on AMfone.com, it turned out the guy who created the circuit I posted a link to was a member. What a great forum this is).