marlinster 6/1/2018 3:46 PM
Overheating negative feedback resistor
Hi,

The negative feedback resistor(47k) in my Selmer Vanguard is getting hot starting to smoke. What could be the cause of this?
Without tubes it is no issue. Bad output tube(s)?
 
The Dude 6/1/2018 4:40 PM
It's best to include a schematic when you start a thread. Here ya go.

[ATTACH=CONFIG]49043[/ATTACH]

I have to ask. You're sure it's the feedback resistor (lower 47K on schematic) and not the node resistor (upper 47K on the schematic)?
 
marlinster 6/2/2018 12:39 AM
 
marlinster 6/2/2018 1:15 AM
Thanks, I forgot about the schematic.

Yes it's the feedback resistor, I had another look
 
Mick Bailey 6/2/2018 2:05 AM
It's fine without the output tubes because there's no output without them. I would lift one end of the resistor and check its value - it may have dropped significantly. If the feedback resistor is overheating then check the connected 220R resistor as well (no need to lift that if the feedback resistor is out of circuit).
 
J M Fahey 6/2/2018 3:54 AM
Yes, it canīt be the feedback one.

And the screen drop one must drop over 200V just to dissipate 1 W, which would barely make it smoke, so itīs probably dropping even more (up to a possible maximum of 300V or so).

Pull the power tubes and reinsert them one at a time, and measure screen voltage to ground, I bet one screen is shorted.

In any case youīll have to replace *both* power tubes.
 
Tom Phillips 6/2/2018 8:22 AM
Another troubleshooting clue would be the answer to the question...
Does the "feedback" 47k resistor get hot when you are playing the amp or just when the amp is ON but sitting in idle producing no sound?
If the amp is ON but you are not playing then we have another clue that the 47k is the screen supply resistor.
It's worth a triple check.

Can you post gut shots?

Cheers,
Tom
 
marlinster 6/2/2018 10:50 AM
Hi, I just checked.
1. The feedback resistor measures 47K, so ok, lifted one end. The connected 220ohm resistor is also measuring ok
2. I assume pin8 is screen grid, it measures on both tubes 40v(?), tried also with one at a time then about 50v each. The resistor gets hot immediately when inserting the power tubes, regardless of order
3. The resistor gets hot when idle, without pre-amp tubes, directly when power tubes are in, does not matter which one is in

White cable on pcb side goes to output transformer and is connected to the speaker. So it should be the feedback resistor.

Many thanks guys

[ATTACH=CONFIG]49062[/ATTACH]
[ATTACH=CONFIG]49063[/ATTACH]
[ATTACH=CONFIG]49065[/ATTACH]
 
J M Fahey 6/2/2018 11:42 AM
Just measuring DCV to ground, at both resistor ends, will solve part of the mystery.

NFB one will have zero to at most a couple volts at each end ... screen resistor will have *hundreds* of volts, no way to mistake one for the other.

It takes all of 1 minute to check.
 
marlinster 6/2/2018 12:11 PM
Hi, just measured.

The 47k that I assumed is the screen resistor measures 39 volts on both sides.

The 47k NFB measures minus voltage on both sides, very high on one side.

[ATTACH=CONFIG]49067[/ATTACH]
 
The Dude 6/4/2018 5:47 PM
IMO, you should have more voltage at the screen node. Bad filter caps?
Also, you say, "The 47k NFB measures minus voltage on both sides, very high on one side.". How much voltage? Measured with no signal applied? This is a cathode biased amp with no negative supply, so the only thing I can think of is some type of strange oscillation that your meter is averaging to a negative DC reading. Do you have a scope? If so, temporarily unhook the feedback resistor and scope the output looking for oscillation and scope your supply nodes to check for ripple. If no scope, check/replace/ temporarily clip in known good caps across existing PS filters and see if things improve.
 
Mick Bailey 6/5/2018 2:27 AM
The screen voltage should be nominally 225v so something is amiss. These tubes are essentially a 6V6 crammed into a B9A envelope and run very hot, and this can make the paper-phenolic boards conductive around the sockets. This gives rise to all kinds of readings but mainly high voltages where you wouldn't expect.

Double-check your voltage readings and make sure your meter battery is good. A negative voltage is something I wouldn't expect to see anywhere unless you have oscillation as per The Dude's suggestion. You could disconnect the white NFB lead to see if this temporarily cures the fault.
 
marlinster 6/5/2018 4:27 AM
Quote Originally Posted by Mick Bailey View Post
The screen voltage should be nominally 225v so something is amiss. These tubes are essentially a 6V6 crammed into a B9A envelope and run very hot, and this can make the paper-phenolic boards conductive around the sockets. This gives rise to all kinds of readings but mainly high voltages where you wouldn't expect.

Double-check your voltage readings and make sure your meter battery is good. A negative voltage is something I wouldn't expect to see anywhere unless you have oscillation as per The Dude's suggestion. You could disconnect the white NFB lead to see if this temporarily cures the fault.
The Dude and Mick, many thanks for your hints. I am away on a trip and will check this weekend. Thanks!
 
marlinster 6/9/2018 9:38 AM
Hi, did this:
Changed all the PS filter caps, power tubes and the NFB resistor(tried without first). I looked like the 1st filter cap was to 100% to ground, but not sure. I re-flowed some solder joints that looked suspicious. I also discovered that one lead to the main switch was loose, but this was probably not the reason. Still not sure what the root cause was.

The screen voltage pin 8 for the power tubes are 315v, isn't this a bit high? Plate voltage pin 7 is 342v, pin 3 is 20v.

And yes, as you mentioned, the 6BW6 gets super hot.

(Input 2 dos not work though, have to check this also later)

Many thanks for your tips!

[ATTACH=CONFIG]49209[/ATTACH]
 
marlinster 6/10/2018 9:43 AM
Hi again,

I changed the selector on the PS transformer to 240v and now the voltages are lower.

They are still a bit on the high side though. Plate voltage is 315v (max 315), screen is 290v(max 285), kathode 17v, and from what I can see this is the maximum what the tube can handle.

http://www.r-type.org/exhib/aaa0343.htm

What do you guys think? Possible to lower the B+?

Maybe I should start a new thread...
 
Mick Bailey 6/11/2018 10:08 AM
Guitar amps often run higher plate/screen voltages on tubes when compared to the spec sheets. Those figures were determined to give the best service life and reliability for the tubes in a wide-range of equipment, but guitar amps sacrifice these to some degree for tone. You changed the selector to 240v, but what was it originally set to and what is your mains voltage? Just check also that your heater voltage OK.