Totem pole filter cap Q

[trobbins]

I've often put a zener somewhat down the RC power supply chain - wherever it makes good use of the upstream resistors, and maximises the number of downstream caps that can be a readily available voltage value (eg. 350V or 315V). It just takes time to solder together 11x 33V zeners :-)

[J M Fahey]

You needn't !!
Just one 33V Zener, 2 resistors and 1 cheap HV transistor (MPSA42)

[Wombaticus]

You needn't !!
Just one 33V Zener, 2 resistors and 1 cheap HV transistor (MPSA42)

It depends where it is in the chain, of course, but you wouldn't need to dissipate very much power before you toasted an MPSA42. Probably something in a TO-220 package that you can at least slap a miniature heat sink on would be the more conservative choice. Personally I use MOSFETS for such amplified Zener applications.

[J M Fahey]

True.
I just pointed to the minimum: a 300V 500mW "Zener", for the first preamp tube, the most sensitive one.
From there, up as much as you wish.
A TO220 can easily dissipate 5W and a TO218/247 over 20 W.

[trobbins]

That's somewhat the advantage of eg. 11x 33V (apart from that I've got a large stash of them gathering dust) - the power rating is nicely higher than a few 100mW, which allows the position to be 'more upstream', and allows cooler operation normally, but also copes with unloaded B+ with enough margin. And it folds up and sits on a tag strip easily without having to contemplate a heatsink or a chassis mounting placement.

But I'll happily use a TO3P FET (as I also have copious IRFP450) heatsunk to chassis for more onerous 'zeners' - although one application I had needed a TL431 to achieve the desired HV regulation.

[J M Fahey]

Well, if you have IRFP450 (which I use too) , you can forget shunt regulation and go for more efficient series pass one.
Or regulare the main +B voltage and make your own Power Attenuator.

[trobbins]

Ahh, but there is something about a KISS zener solution, especially when there is no need for exotic performance (I just restore old amps) and the issue is only about the voltage stress on supply caps when B+ is unloaded (I get this whoozy nice designer feeling knowing that caps won't get stressed if someone pulls out valves) :-)

[Tage]

In regard to the Original post about totem pole caps, is there a downside to taking a screen and preamp B+ from the center point of the ballancing resistors? if say the Main B+ was 600V, then @ the resistors meeting point it would be 300V. there wouldn't be too much current draw? with say 100k balancing resistors? This is with a non-ceter tapped bridge rectifier.

[Gingertube]

Tage,
Drawing a bit extra current from the lower capacitor will drop its voltage marginally and make the voltage share resistors "work a bit harder" - depends on how much current you are drawing. If you are just feeding screens of the output tubes and the preamp then the voltage shouldn't be disturbed too much. You can actually trim this by adding a small cap across the bottom cap of the main filter but it is probably not required. If say you have 2 off 100uF in the main filter then you could try adding another 4u7 or 10uF across the bottom cap to trim the voltage up a bit untill you have equal voltages across the voltage share resistors again.
EDIT\
Woops forget that - bad advice, the voltage will share according to the capacitance (inverse relationship) so if you add capacitance across the bottom cap as I suggested then you will actually REDUCE the voltage across the bottom cap as the same charge current will produce less volts across more capacitance.
You could try trimming the top voltage share resistor down in value a little - else just leave it.

Here is a schemo that I drew for a Hiwatt and now seems to appear on a lot of sites. Note the unequal capacitance of C6a and C6b compared to C7. They did'nt seem to think it worth adjusting the voltage share resistors even though the 100uF of C6a and C6b in parallel will have 2 x the voltage across them compared to the 220uF C7.
http://www.webphix.com/schematic%20h...watt_sa412.pdf
Cheers,
Ian

[Chuck H]

FWIW I'm was hoping to feed the trem (2 triodes, one operating as a voltage gain amp and the other a cathode follower), the reverb driver (1/2 12au7 driving the tank, so more significant current draw. But not more than power tube screens). More significant to the issue is that I want to use three caps (w/balancing resistors) in series to derive a 2/3 voltage from the node. Problem???

[trobbins]

The hassle may be more if there is a bad part in your preamp. It would pull the voltage down on the parts and lower cap, but the top cap may suffer from over-voltage. Similarly, if you disconnect the preamp parts, then the cap voltages could go the other way, and again suffer. By the time you end up putting enough margins in to cover both sides of the coin, it may work out a bit clunky.

[J M Fahey]

Bad idea.
Your current load will be equivalent to a low value resistor in parallel with the lower divider leg, so goodbye equal voltage division..
Or, to see it from another point of view: you will have the upper 100K in series with your screens !!!
Plus the rest of the preamp.
Do you think it will work?

[Chuck H]

I didn't think it would be a problem because I've seen screen taps done this way as power reduction circuits. But it occurs to me now that the load for a reverb driver is much greater. So I see the problem. It complicates the rail, but I'll just branch and divide instead. Thanks

[Tage]

yeah I kinda thought that the upper ballance resistor would have all the screen and preamp current going through it, which for big tubes is a significant amount of current, especially through a 220k resistor. And then there is the issue of imbalanced voltages on the caps...