Bleed Resistor to Discharge High Voltage Caps in Tube Amp

[Boogie]

Hi,

I've got a Fender Red Knob Twin that Im going to be working on, I was wondering about some of the resistors in the PSU.

As far as I can tell, all of the really high voltage caps in the PSU have resistors across them (though the rightmost two on the picture have other resistors in series with the cap also). Most of them are 220K, but one is 30K 10W across 400V unless I have misread the schematic (its a little hard to read).

The schematic of the PSU is here

http://www.christianguitar.com.au/im...b-twin-PSU.gif

What are these resistors for other than to discharge the caps for safety when working on the amp? According to P = V squared / R the 30K one is using 5W of power all the time the amp is on? Which made me wonder what it is for....

About how long would you expect these caps take to discharge, with those resistors there?

The 75V winding doesnt have any resistors across the caps (bottom of the image), is that because 75V isnt dangerous?

Unless I've missed anything, is it correct that I don't really need to worry about making up a jumper with a resistor (like 10K 10W) to bleed the caps? Or maybe I should do that anyway for the 75V caps?

Of course I will check the voltages across the caps with a multimeter before doing anything, but I was interested in what some more experienced amp techs thought..... Also its a bit fiddly taking the metal chassis out of the wooden box, which I have to do before I can measure or bleed anything, it would be nice to know how long its got to be off for before its safe to dismantle it.

[Enzo]

One way to answer some of those questions would be to connect your voltmeter to the B+ and watch it discharge after power is removed. I would expect that high voltage to discharge in a few seconds to safe levels.

Your 220k resistors are equalizing resistors. You will note they are across caps that are in series. The caps themselves are rated 285v and 350v, both lower than the supply voltage they filter. The caps are stacked in series to combine them for a higher voltage result total. But the resistors insure they share voltage equally. If the two caps are across 400v, the resistors make sure the point in between them is at 200v. But they do also serve as bleeders.

That 30k is across the 400v node. It is not really a bleeder, it helps drop some voltage. Note it is a 10 watt part, necessitated by the 5w dissipation. They did not want to try to drop all that voltage with only the preamp tube currents.

Other than those equalizers, Fender doesn't generally use bleeders. So it isn't that 75v is not dangerous, it can injute you, but they just don;t bother with them. The fact the stacked cap equalizers act as bleeders is merely a bonus.

Having a dedicated discharge lead is always a good idea, you never know when you will need it. And other uses will occur to you at the oddest times. Just watch the circuit and see if you need more discharging.

The voltages won;t leap out of the chassis. You can unbolt that and slide it out, just don;t stick your fingers inside hte chassis until you know what is what.

[Boogie]

Thanks, thats really helpful.

Yeah I was thinking I could leave the voltmeter connected and turn off the amp and see how fast it went down.

Thats interesting about the resistors, wouldn't the 30K 10W also do some bleeding too though since its across the three 22uF caps?

I'll make up a proper discharge lead. I bought two 4.7K 5W resistors from Jaycar (they were the biggest they had) so in series that will be 10K 10W, I think that 400 V would overload/overheat these a bit but its only for a short time so it should be ok? Or I could buy some bigger ones online if thats not enough.

Ive got a crook back and the amp is a beast so Its always a bit of tug of war getting the chassis out, I was thinking that my fingers might touch something accidentally while I was trying to stop it falling onto the floor or something like that. (Actually before my original post I was reading a story of a guy who mishandled/dropped the chassis while getting it out, then grabbed it as a reflex, and zapped himself enough to get a blister).

I was looking up equations and I found discharge time as t = -R C ln (V/Vo), which for R=220k and C=220uF and V/Vo as 40/400 , this gives t = 111 seconds, or about 2 minutes to discharge down to 40 volts. But it looks like they would also be discharging through the 30K + 2.7K resistors?, and 33K is a lot less than 220K so it seems like it should go down much faster than that.

[Dave Curtis, dB AudioTech]

I know there's a requirement in the National Electric Code that says something to the effect that capacitors are required to have a means of discharging the voltage to a safe level (below 60 volts) within 60 seconds of power off. I'm sure that UL requires this also. This was not in effect when Leo started building amps...

[Enzo]

Of course the 30k will bleed the caps, but I don;t think that is its purpose. Fender traditionally does not use purely bleeder resistors.

[Boogie]

Thats interesting. I wonder when that requirement came in. My amp was made in about 1990.

[Randall]

I don't get why the 2.7K is a 10 watt. I figure less that 1/2 watt for that one? Is it because it is in series with B and C nodes as well?

[bob p]

That twin has to have some pretty high plate voltages. They're not listed on your schem, but I think amps from that era ran the anodes up around 475VDC. that's higher than you need for the preamp tubes, so high voltage amps typically used a voltage divider to bring the voltage down to more reasonable levels for the preamp stages. Take a look at R128 / R129.

[Boogie]

I think because everything powered by A,B,and C all run through it (which is basically the whole preamp section including reverb and the phase inverter). Also, I think that some of the big caps in the PSU , the 200 uF ones, will bleed to ground through it and the 30K 10W once the power is turned off, which would be a few watts of power. If you look at the two big caps and the 30K and 2.7K resistors alone, they ground across them. Also the 220K resistors are across them but 32.7K is much less resistance so most of the bleeding would happen through them? Unless the centre "tap" in the middle of the two 220uF caps changes things a lot?

[Boogie]

The two wires going up off the picture on either side of the choke power the 6L6s (the power amp section). The wire to the left of them is the power amp bias. The A node powers the phase inverter - so I guess it needs more power than the rest of the preamp,since its like halfway between the preamp and power amp? (Technically i think it counts as part of the power amp?)

[Boogie]

I can put up the rest of the schematic when ive got a few more minutes . The two high voltage lines for the 6L6s are switchable voltage depending on whether the amp is hi or lo power (there is a switch on the front panel). +469V at hi and 236V for lo. The bias line is also switched and -65V on hi and -29V on lo.

[Boogie]

This is the whole schematic

http://www.christianguitar.com.au/im...-knob-twin.gif

It's funny because last night right before I fell asleep I was thinking about how the bias measuring points might work (since I was wondering how hard it would be to mod another brand of amp to have them), and thinking that (from what Ive read) the voltage measured at the points (eg 40mV) corresponds exactly to the bias current (eg 40mA), then it seemed like there must be an exactly 1 ohm resistor in series with the bias current, with the measurement points on either side of the 1 ohm resistor . (So that V=IR across the resistor becomes V=I since R=1 ohm).

I forgot all about that until I saw the replies to this thread, and looking at the schematic, and guess what - there are 1.0 ohm resistors across the bias measurement points (surprise, lol)

I'm gradually learning about how amps work, its great

[Enzo]

Well that is an application of Ohm's Law. Amazing how many places we use that.

Please use R numbers instead of "the 10k resistor".

All the node A, B, and C current flows through R128. Schematic says voltage drops from 469 to 397 across R128 - 72v. I get I=E/R=72/2700=27ma. Disipation would then be (72x72)/2700 = 1.92 watts. It doesn;t matter where the current goes, you have the resistance and the voltage drop, and that lets you calculate current and power. Those are schematic voltages, your measurements of actual voltages would be more accurate than those estimates. At close to 2 watts, I;d use at least a 5 watter, and 10 watts, why not?

bob - the schematic shows those voltages outside of his clipped area, but you were close. And considering the climbing mains voltages we all have, your 475 is probably closer than the 469 on the drawing.

ANY path to ground will bleed the caps. You better hope the caps themselves don;t do a lot of the bleeding, that would mean they were leaky. Oh wait, you are thinking of the main caps gpoing over to find ground through R129. The bleed or discharge current lasts a very short time and is no more than it is constantly when the amp is on. Think about it. The amp has the full voltages on it al the time, when you turn it of, they start to fall. SO no additional dissipation will occur from bleeding.

You can easily calculate current through each node by the voltage drops on their resistors. Remember that C current runs through B and A, and B current runs through A. So the R130 drop includes current for both B and C nodes.

R129 is 30k, and has 397 volts across it. 13ma. Dissipation is over 5 watts. SO 10w is a good choice.

The current through R129 and R130 ought to add up to the current through R128.

The A node powers the phase inverter - so I guess it needs more power than the rest of the preamp,since its like halfway between the preamp and power amp? (Technically i think it counts as part of the power amp?)

It doesn't need more power, but it gets more voltage. Power and voltage are not the same thing. They want the higher voltage on it for headroom, so the signal won;t clip. The PI is part of the power amp, yes, but the power supply doesn;t care where the tubes are, it just supplies current.

[Randall]

Please use R numbers instead of "the 10k resistor".

Please explain this, or point to explanation.

[52 Bill]

Please explain this, or point to explanation.

All he means is to reference the R part number on the schematic so that we all know which one he is talking about, as there can be many 10K resistors, etc.