Zener Diode Question

[David Root]

I've not used zeners before, so here's my situation:

I have a PT with two high voltage secondary windings, one is 770 VCT, the other is 520 VCT. I am using the 520 VCT winding as the screen supply and reverb driver and recovery tubes supply. The 770 VCT winding is conventionally switched and supplies the power tube plates, PI and preamp tubes, but the screen supply is not B+ switched i.e. it comes on when the amp is switched from off to standby.

I have two 39V 5W Zeners behind the two 1K sandbox screen resistors, connected to a 30H choke that feeds from a 100uF cap on the 520 VCT winding. With the amp in standby the zeners are working right and lowering the screen supply voltage about 35 volts to about 305V. BUT, when I switch in the 770 VCT winding, which I had assumed is independent of the 770 VCT winding, the screens elevate about 40V and I'm back where I started.

Since both of the windings use the same CT wire to ground, did I make a wrong assumption, or is this just back push EMF from the plate onto the screen, which could presumably fixed by moving the zeners ahead of the screen resistors?

[Wayne]

I'm trying to get a handle on what you've got and how you're using it - do you have a way to post even a crude drawing?

Does your PT have 2 separate windings, or is it one winding tapped 385-260-0-260-385?

W[/i]

[labb]

Try putting the zeners in the CT to ground and see what you get. May not drop the voltage enough for you but it should get both windings.

[David Root]

No, it is two separate windings, not a tapped off 520V. I don't really want to drop the 770 VCT by that much. Current Vp is about 495. It is a One Electron BFT-1B, www.one-electron.com. It is a 15 lb monster rated 340 mA.
EDIT--My mistake, it IS a 260-0-260 tap, I just looked at the PT diagram again. Duh! So I guess I'm stuck with the zeners on the CT.

My ideal voltages are 470 on the plates and 300 on the screen. That is a perfect primary impedance match to my OT, 3k5. I'm using a pair of 8417 JAN Sylvania.

Give or take 10VDC or so, maybe that might be close. I could try 30VDC zeners on the CT? I have those, but only 5W. What do you think? Two in parallel?

I have a schematic but my scanner makes 18MB scans and I can't upload that much, any hints?

[lewilson]

Another trick that works good is putting a chock in the centertap.
I have a box of them. I try different ones in a circuit till I find one that drops the voltage I need.. It doesnt get hot like a zener, and also a chock provides filtering .

[David Root]

I had forgotten about that one! I need to start reading old books again. So any voltage rated choke is OK because it's close to ground, it just has to be able to handle the max power DC mA, right?

But it doesn't clamp the voltage drop either like a zener. I need a steady screen voltage as 8417 screens can be a bit flaky, although the pair I have seem to be pretty stable.

Maybe change the 5W zeners on the screens now to 10V and also choke the CT?

[John_P_WI]

David,

I know this does not answer your question directly, but this would be a perfect place to use one of KOC's dc power scaling kits in which you could set the plate and screen voltages to your exact need. Yes, a little more money, but certainly less time spent. Set it and forget it....

John

[lewilson]

It sound like the zeners are in series with the screens since they are 39 volt zeners. so you end up droping 39 volts from the total voltage to the screens?
Im sort of fuzzy without a schematic, But maybe figure out screen current and usa a larger resistor to the screen?

[David Root]

The zeners drop only about 35 V, which is OK. They are in series with the screen resistors. The problem arises when I swich in the full winding to the plates, then the screen voltage goes back up from about 305V to about 345V.

I have a 25W 40V zener with a heat sink that I got from Ted Weber. That would likely be overkill on the plates but work OK on the screens.

I just checked out those power scaling kits. They are designed for circuits drawing plate and screen power from the same PS line so I don't think they would work on my arrangement which is two separate lines.

[John_P_WI]

David,

If you are happy with the plate voltage and don't care about changing that dynamically (power scaling) you should be able to use a single kit for the screens. Therefore, if you are not varing the plate voltage and and don't need to worry about bias tracking, a single simple regulator should work for the screens. You might want to send an email to KOC explaining your situation, he is a good guy.

If you have the original TUT, the Borris mod (or something similar) would work.

John

[Wayne]

I think I understand this correctly (not sure if I do), but I just re-read the thread including your original post.

You're switching the B+ to your plates, but leaving your screens energized. When the screens have voltage and the plates don't the screens, being the most positive electrodes, try to act like plates. Usually, more current than the plates can stand will flow under these circumstances. I'm surprised you still have tubes!

When you switch on the B+ to the plates, the regular (less) amount of screen current flows. Less current = less IR drop across the 1k sandboxes = higher screen voltage.

You either need to move your HV switch to the CT (I don't like this method), or use a multi-section switch so that your tubes never have one voltage applied without the other.

W

[David Root]

wayne, I think you could be right. I have only measured the cathode current with the B+ on the plates. They are biased -16.5V to 65mA each tube. I have variable fixed bias with separate adjustment and cathode monitoring jacks.

I'll measure the screen current with the B+ off. Fortunately 8417s are designed for high current but the screens draw about 2-3mA quiescent each with the B+ on. Data sheet says about 18mA each screen at max Po of 80W rms.

[David Root]

Yup, you're right. With B+ on, screen measures 2.5 mA across the screen resistor, which is normal. With B+ off, screen is 33.5mA across the screen resistor, which is definitely not OK.

I have to switch the screen supply too. I'm wondering if I can connect it to the same switch I'm already using to operate the amp. This is a so called progessive switch i.e tip down is off, tip horizontal is power on and tip up is B+ on. Somehow that doesn't sound right as that would be connecting the full winding with the tapped winding in parallel.

[Wayne]

Glad we caught that! Three cheers for tough tubes!

You are correct - what you suggested won't work - unless you've got an unused, completely isolated set of poles on your switch.

Do you have a part number for the switch you're using?

W

[tubeswell]

If your HT winding has a centre tap, put the zeners in series with the CT and the ground return path, with the banded end (cathode) pointing to ground. The zener(s) ideally should be rated about twice the expected dissipation that the zener will see. You can string a bunch of 5W zeners in series (e.g. 9V, 12V, 15V) to get a more accurate voltage drop (as well as increasing the effective dissipation capability).

The dissipation calculation should be based on 2 x (voltage drop of the zener(s) x the maximum current draw on the HT winding). For say a 12V or 15V drop, a single 5W zener is fine for something like a tweed deluxe PT.