Pots at B+ Voltage

[defaced]

So I was looking at some variable voltage regulator circuits today for a preamp prototyping rig I'm building and noticed that all the circuits I looked at had a 1M pot in series with a 100k resistor, all between B+ and ground. My knee jerk reaction to seeing this was "%@#* that's dangerously stupid" because the pot was a front panel control with one end sitting at full plate voltage.

Am I being overly cautions or is this a really bad idea?

If this is a bad idea, are there any circuits out there that can allow adjustment of the B+ without putting a pot in this location? My two thoughts are replacing the pot with a resistor sting and jumpering them as required, or doing something similar with reverse biased zeners between the rectifier and ground. I like these two approaches, but I would prefer something that allows in-process control of voltage, but I'm not married to the idea.

Here's one of the circuits: http://www.rh-tech.org/index.php?topic=41.0

[R.G.]

I'm not sure I'm happy with the idea, but if either the pot shaft is chassis grounded, or the knob is solid bakelite/phenolic/epoxy/etc., it's probably not all too dangerous. It's no worse that some other things I've seen. I can think of some things to make the pot referenced to ground, not to B+, but it makes the circuit a bit more complicated. Things attached to ground are safer. The other thing is that all pots get scratchy sometime, and this on directly modulates B+, which is kind of a design flaw.

Looks like the 1M pot is intended to let you turn down the gate voltage on a MOSFETand have that lower the voltage to the output transformer and screens.

[olddawg]

Use a pot with a plastic (nylon, whatever) shaft?

[Steve Conner]

It's kind of ouch. Most pots are only rated at 250V. If the pot is mounted to a grounded panel, there is no shock hazard to the user. The danger is more that the pot will break down and go up in smoke.

I did it on my last build, but the HT on that was only about 250 to 300, and I used the biggest, meanest looking old pot I could find, with a case well over an inch in diameter. I imagine if you use a full size pot with a large element, and plenty of creepage distance between the element and the case, you'll get away with more voltage, but I'd start to get worried above 450.

Years ago I proposed a circuit that used 12V on the pot and boosted it with a couple of transistors, to give a variable B+ and tracking bias supply. I showed it to Kevin O'Connor and he said he had something similar in one of the TUT books.

In all of these circuits, it's a good idea to put a R-C filter between pot wiper and MOSFET gate. It filters out most of the pot scratching, and even removes hum as a bonus if you make it really big. Last but not least, the resistor in it saves the pot from burning out if the FET fails with all three terminals shorted together, as they often do.

[MWJB]

Probably better to use alternate Zeners or a string of them to adjust voltage. As well as being rated for 250v, most amp pots are only half a W dissipation. You can find some high wattage wire wound pots & cermet pots that might suffice for adjusting low current rails like preamp voltage. Series resistance in the main B+ rail will skew any results because it affects the sound.

Yikes! Something attached to the chassis panel, with full B+ on one end, my God! That sounds mighty dangerous...or at least as dangerous as the standby switch found on nearly every amp in the world, especially the ones that run high B+ ;-)

[defaced]

Thank you guys for the comments and sanity check. I've done some more digging and funny enough, ran across the "Power Scaling is Flawed" thread here which has some good ideas in it. I'm going to keep at it and see what else I can turn up and possibly prototype a circuit or two. I'm off to go learn about what "linear" and MOSFET have to do with each other.

Yikes! Something attached to the chassis panel, with full B+ on one end, my God! That sounds mighty dangerous...or at least as dangerous as the standby switch found on nearly every amp in the world, especially the ones that run high B+ ;-)

Yea, tell me about it. That's why my amp has a relay in place of the standard standby switch, which in the end, my just end up being a simple mute circuit. Once I knew which way was up, I stopped liking standard standby switches.

[redelephant]

That is a really bad idea. Didn't know those circuits were that stupid. It appears the circuit show in that link has some mistakes, like the 100kohm in series with the output.
I'd make a variable constant current source with the pot at the gnd side, loaded into a resistor that provides the required voltage drop, buffered by a MOSFET.

[R.G.]

That is a really bad idea. Didn't know those circuits were that stupid. It appears the circuit show in that link has some mistakes, like the 100kohm in series with the output.
I'd make a variable constant current source with the pot at the gnd side, loaded into a resistor that provides the required voltage drop, buffered by a MOSFET.

That's one I considered. I decided that a cascode arrangement with a second MOSFET to feed current to a resistor pulldown from B+ to feed the gate of the first MOSFET would be better. Actually, I guess that qualifies as a variable constant current source in some ways.

I also ran through all kinds of elaboration in my head, like a current limit on the main pass MOSFET, temp sense and shutdown, etc. as some of the things I tossed in under "complexity".

The thing is, most people would get away with it as shown. That ensures that the internet geniuses will keep propagating it because it's simpler than a good and more reliable design.

[Chuck H]

+++
Probably best to have a pot that controls a parameter that in turn changes the voltage as above and in Steve's post.

FWIW I use ordinary Alpha 100k linear pots as temporary 12ax7 plate loads to find values for split plate load circuits. Of course I replace the pot with a pair of fixed resistors. I've done this a lot with no pot failures. I did wonder if the pot would take it the first time I tried this though. I even tested for DC potential between the pot case and amp chassis just to make sure I wasn't going to shock myself.

[Steve Conner]

The "Internet geniuses" remark is true in a way. Remember this is supposed to be fun for hobbyists. 95% of people building their own tube amps are not EEs and have no patience to build and troubleshoot a complicated regulator circuit, let alone equipment to test if they actually got it working.

And even if you were an EE designing a commercial amp product, you have to think of the techs who'll be repairing it 10 years down the line. Sustainability, indeed the whole ethos of tube amps as timeless non-throwaway products, demands that we engineer things for easy repair, and 95% of the repair techs have no EE training or patience either.

In the face of this I think the proper response is to strive for Zen-like simplicity. MWJB, what do you say to a bunch of zeners selected by a rotary switch?

[guitarmike2107]

You guys are funny, ok maybe it is not the best circuit for doing this, but it ( or variations of it) have been used for years on 1000’s of amps with very few failures….and its dead simple and is hardly new anymore

Of course the pot you use has to be rated for the voltage it subjected too… so you have to shop about to get one that is

Would it pass CE safety testing .. I don’t know , but for most hobbyist that doesn’t even come into consideration.

Even if you used a metal shaft pot it should have a big ass plastic knob on it and the body should be connected directly to safety ground

[MWJB]

"In the face of this I think the proper response is to strive for Zen-like simplicity. MWJB, what do you say to a bunch of zeners selected by a rotary switch?"

Steve, I say you read my mind ;-)

[J M Fahey]

Really it's something to think about.
I'm also somewhat nervous about that high voltage, buy in no way can this be compared to death caps and the like, simply because a grounded metallic panel sits between said pot and us.
That, coupled to plastic knobs and plastic shafts, rounds the package.
I've been repairing my old faithful Kikusui scope last week, and I saw some pots, in the middle of the board, carrying way over 1KV voltages (focus, etc.) fully 4" from the front panel, coupled to the visible knob with a 3" brass bar, which was coupled to the actual pot shaft with a plastic bushing, which added around 1" separation between them.
Maybe that solution could be adapted here.

[Jazz P Bass]

If the chassis is earth bonded, does it really matter what voltage is "near" the panel?
Just a thought.

[Chuck H]

It wouldn't need to be a complicated regulator circuit. Why wouldn't any of the "power scaling" designs floating around work? Here's one I grabbed from an image search with the associated PS and bias supply removed: