Zener Diode in HT centre tap to reduce voltage on tubes: what's... happening?

Cirrus

Ok.

Zener Diode to ground in the centre tap of the high voltage secondary of the power transformer in a valve amp, to lower the voltage going to the valves. You pick a diode with a particular voltage, and it drops that voltage. Practical example: Just as an experiment, I put 3 5 watt 6.8v zener diodes in series, and my ac30 went from 345v to 325v on the plates. Pretty good.

But they are getting very hot, so I'm going to replace them with a 20 watt chassis mounted one.

I don't understand though. How do they do that?  Where is that 20v going? It's not going to ground, because ground is still 0v. How does it happen? How does it pull down the high voltage?

Cirrus

I'm probably posting on the wrong forum since by my reckoning there's one person who might answer if they have time. But I think I need to go and actually learn about current.

ICBM

The voltage is reduced because there's effectively a 'step' equivalent to the diode voltage in the circuit which must be exceeded in order for current to flow, so only the difference between the original HT voltage and the diode voltage is now available to the circuit. The rest is going into heat produced in the diodes, as you've found - the minimum power rating is the voltage drop multiplied by the maximum current draw of the amp's HT circuit, which should be safe with a 20W diode (from memory the maximum draw of an AC30 is around 300mA). In theory the 15W (total) you have now is enough, but they will certainly get hot in a valve amp chassis.

Cirrus

I don't think I appreciated when I worked out the power they'd be dissipating that even though they can take 15 watts between them, they're still 3 tiny parts just being cooled by air, dissipating what I think is 5 watts cumulatively - the wattage rating doesn't magically make them less hot.

The ac30 is happy though, and hopefully it'll help the el84s last a few more months (gets 15.5 watt dissipation down to 13.6). It's way cheaper than getting a new power transformer, too, which was my first thought (and indeed I'm not sure a more conservatively wound transformer with the extra 25v secondary for the IC chips etc exists), and presumably doesn't mess with the natural sag of the circuit in the way a fixed resistor would.

Cirrus

Post script: I got the wrong polarity 20 watt diode, so all it did was make the amp hum terribly. Ah well.

Cirrus

Further post script:

It turns out that I can't use a chassis mounting Zener, because it's dumping noise onto the chassis. The Zener needs to go between the Transformer and a circuit board, so I need to be creative to make something that works.

ICBM

Have you tried a suppressor cap across the Zener?

Keefy

I may be stating the obvious here but if you’re trying to reduce the plate dissipation in the EL84s, why not just increase the value of the cathode resistor?

Cirrus

Yes, though admittedly just by holding it across the cap.

I think it's a more serious issue;

On the CC series, the HT Centre tap goes via the standby switch (which I've removed) and HT fuse to the power section circuit board, connecting directly next to the ground side of the first (pre-choke) filter cap.

I read 0r resistance between that point and the amp chassis, so assumed I could just move that wire to the top of the zener connected to the chassis and electrically it'd be as good as the same.

Well I know better now. It might read 0v, but it's actually connected via this little doozy;



And I think that's keeping the rectifier hum off the chassis.

ICBM

Ah, yes I forgot that the modern AC30s lift the audio ground from the chassis.

The way around that is to still use the chassis-mount diode, but with an insulator, and a tag to make the ground connection back to the circuit - so the only functions of the chassis bolt are to hold it and heatsink it.

Cirrus

ICBM said:

Ah, yes I forgot that the modern AC30s lift the audio ground from the chassis.

The way around that is to still use the chassis-mount diode, but with an insulator, and a tag to make the ground connection back to the circuit - so the only functions of the chassis bolt are to hold it and heatsink it.

Yep, that's exactly what I'm going to do - ordered some nylon bolts and screws to isolate the bracket I'll mount it to from the chassis.

My other plan was just to make a bigger string of lower wattage diodes on vero board - 5 x 4.8v would do basically the same thing. But the bolt diode can be made more secure.

Cirrus

Well, things have got a bit out of hand. Bracket mounted 20w zener made the bracket hot enough to burn my hand after a couple of minutes, so I had a little cast around for things that could function as a heat sink.




I knew it was a good idea to keep this old Fender Mex Strat trem block/ bridge around after I replaced it years ago 

I've got a temp probe to assess the probably longevity of the 20 watter and a 50 watt zener coming today just in case. If that doesn't work I might need to just give up on the idea, and try to source a proper 245v mains transformer.