KT88 PP amp.
I just biased the tubes class A at about 93mA (445 idle plate voltage). After playing for a couple of hours, I suddenly noticed that one of the 4 tube was crazy glowing. No change in sound, it is sounding great.
I turned it off for all of a few minutes at most and then back on and the tube was glowing normally again.
I rechecked the bias and it was about 84mA. It's PP partner was at 93mA still and for some reason both of the other channel's tubes had drifted to about 70mA.
So, I reset it all, trimming the bias down to about 85mA cuz I am nervous. So far no extra glow, everything sounds fantastic.
What the heck was that about? Am I just doing something wrong and being stupid or am I just giving myself too much credit.
I just biased the tubes class A at about 93mA (445 idle plate voltage). After playing for a couple of hours, I suddenly noticed that one of the 4 tube was crazy glowing. No change in sound, it is sounding great.
I turned it off for all of a few minutes at most and then back on and the tube was glowing normally again.
I rechecked the bias and it was about 84mA. It's PP partner was at 93mA still and for some reason both of the other channel's tubes had drifted to about 70mA.
So, I reset it all, trimming the bias down to about 85mA cuz I am nervous. So far no extra glow, everything sounds fantastic.
What the heck was that about? Am I just doing something wrong and being stupid or am I just giving myself too much credit.
In a simplistic way of looking at it, only one thing causes the plate to glow: too much current.
The tricky business is figuring out *why* there's too much current. Gassy tubes, setting bias voltage too low, too high screen voltage, heater voltages, short circuits...all can cause this problem. The list goes on. To me it sounds like your tubes cannot handle 93mA at your plate voltage for long. That is 41.385W of heat to dissapate, constantly in class A. Winged C/Svetlana lists 44W, Shuguang lists 50W abs. max. I think you may have a gassy tube, or the current was higher than you measured. Class A is much harder on big power tubes than a nice sounding class AB1. It also tends to expose a gassy tube quickly. The sonics are not radically different, depending on the bias point.
Some questions:
-Screen voltage?
-Screen current?
-Grid resistor value?
-KT88 variety?
-Old or new coupling caps?
The tricky business is figuring out *why* there's too much current. Gassy tubes, setting bias voltage too low, too high screen voltage, heater voltages, short circuits...all can cause this problem. The list goes on. To me it sounds like your tubes cannot handle 93mA at your plate voltage for long. That is 41.385W of heat to dissapate, constantly in class A. Winged C/Svetlana lists 44W, Shuguang lists 50W abs. max. I think you may have a gassy tube, or the current was higher than you measured. Class A is much harder on big power tubes than a nice sounding class AB1. It also tends to expose a gassy tube quickly. The sonics are not radically different, depending on the bias point.
Some questions:
-Screen voltage?
-Screen current?
-Grid resistor value?
-KT88 variety?
-Old or new coupling caps?
In amps like yours with negative-supply ('fixed') bias, every case of a runaway tube under no-signal conditions I have ever seen was due to loss of bias. I would first suspect that (since the other tubes are getting bias from the common supply) there is a loss of connection somewhere between the supply and the grid pin of the tube. Check the grid pin of the tube for bias (probably -40 volts DC or so), and if it's not there, walk backwards with your probe node to node until you get bias. That's the faulty node. Broken wire, bad pot, cold solder joint, something.
Aloha,
Poinz
AudioTropic
Aloha,
Poinz
AudioTropic
Thanks for the input. This is not a DIY amp, it is an older Sophia Electric. I purchased it broken and have fixed it, but have yet to delve into the circuitry much, just enough to figure out likely operating conditions.
I have been too busy putting together a schematic and trying to figure out another old Sophia amp, a PP EL34 (see another thread). I am still new at this and the EL34 is my learning piece. The KT88 was an opportunity not to be missed and is what I play now, since it sounds so fine. It will be the next project. I think there is a lot of room for improvement.
For m6tt: everything is old, except the screen grid resistors and cathode resistors, which I replaced because one or two showed signs of prior burn. I can't remember the values. The tubes are the new issue Genelex Gold lion. Less than 100hrs so far at most, I imagine. I don't know the other values, since I haven't really pulled together the operating parameters, more than minimally. I have a heard time imagining that there was too much current, since it was the lowest bias of the tubes.
For John65b: I have no schematic. I will work it out, but give me six months or more. I am still playing with the EL34 and am about to be a dad for the first time. My plate is full, but I really want to do it someday and will post it when I have.
For Poindexter: I agree there ought to be something, but I was just happily playing music ( Matt Skellenger - The parentheticals, a local Denver Afro-jazz something band)
when I looked over and suddenly this one tube of four was glowing bright in the plate. I then turned it off for a few minutes, back on, readjusted the bias, increasing the bias for that particular tube from 64mA to 84mA and it hasn't happened again, I kept playing for another couple of hours. Go figure. Still sounds fine as before. Yeah I don't get it either. Am I wrong in assuming that since the bias had fallen to a lower value than the other tubes the current should have been less and the tube current should also have been less?
Thanks again for all the input. I will update if it happens again, but tomorrow is a work day, so no music for me...
I have been too busy putting together a schematic and trying to figure out another old Sophia amp, a PP EL34 (see another thread). I am still new at this and the EL34 is my learning piece. The KT88 was an opportunity not to be missed and is what I play now, since it sounds so fine. It will be the next project. I think there is a lot of room for improvement.
For m6tt: everything is old, except the screen grid resistors and cathode resistors, which I replaced because one or two showed signs of prior burn. I can't remember the values. The tubes are the new issue Genelex Gold lion. Less than 100hrs so far at most, I imagine. I don't know the other values, since I haven't really pulled together the operating parameters, more than minimally. I have a heard time imagining that there was too much current, since it was the lowest bias of the tubes.
For John65b: I have no schematic. I will work it out, but give me six months or more. I am still playing with the EL34 and am about to be a dad for the first time. My plate is full, but I really want to do it someday and will post it when I have.
For Poindexter: I agree there ought to be something, but I was just happily playing music ( Matt Skellenger - The parentheticals, a local Denver Afro-jazz something band)
when I looked over and suddenly this one tube of four was glowing bright in the plate. I then turned it off for a few minutes, back on, readjusted the bias, increasing the bias for that particular tube from 64mA to 84mA and it hasn't happened again, I kept playing for another couple of hours. Go figure. Still sounds fine as before. Yeah I don't get it either. Am I wrong in assuming that since the bias had fallen to a lower value than the other tubes the current should have been less and the tube current should also have been less?
Thanks again for all the input. I will update if it happens again, but tomorrow is a work day, so no music for me...
Good idea, I just need a tube tester...
Seems that the tube is working fine now, so if something were bad about the tube, enough to cause runaway glow, I would expect more consistent issues, not to mention a degradation of sound.
I wonder if there is a heat/instability problem with the bias pot...They seem hard to adjust and drift a lot even under stable conditions.
Anyone have a recommendation on what brand of pot to get for this?
Seems that the tube is working fine now, so if something were bad about the tube, enough to cause runaway glow, I would expect more consistent issues, not to mention a degradation of sound.
I wonder if there is a heat/instability problem with the bias pot...They seem hard to adjust and drift a lot even under stable conditions.
Anyone have a recommendation on what brand of pot to get for this?
Some ideas:
Coupling caps could start breaking down at temp, sending preamp DC+ to power tube grids. I'd replace the ones before the power tube. Make sure the grid resistors (i.e. G1) are datasheet max or below for fixed bias. A little high is OK, but it's bad for the tubes, and can make tubes run away by decreasing bias as the tube works harder. If they've drifted (and they may have) you could also end up with a variety of situations leading to red plate. Retension socket pins if the bias seems solid. Seriously check every carbon comp resistor to ensure they are 20% of value or better. They sometimes drift wild, up to a couple of megohms. Finally, make sure your speaker wires are connected well and have no loose ends that could be touching intermittently or when the amp warms them etc. An intermittent problem is the worst kind, since it "fixes" itself sometimes.
At least vacuum tubes don't let the smoke (and flames) out as easily as the phase linear I'm learning SS on!
Coupling caps could start breaking down at temp, sending preamp DC+ to power tube grids. I'd replace the ones before the power tube. Make sure the grid resistors (i.e. G1) are datasheet max or below for fixed bias. A little high is OK, but it's bad for the tubes, and can make tubes run away by decreasing bias as the tube works harder. If they've drifted (and they may have) you could also end up with a variety of situations leading to red plate. Retension socket pins if the bias seems solid. Seriously check every carbon comp resistor to ensure they are 20% of value or better. They sometimes drift wild, up to a couple of megohms. Finally, make sure your speaker wires are connected well and have no loose ends that could be touching intermittently or when the amp warms them etc. An intermittent problem is the worst kind, since it "fixes" itself sometimes.
At least vacuum tubes don't let the smoke (and flames) out as easily as the phase linear I'm learning SS on!
Not sure of the age of this amp but maybe the coupling capacitors are on their way especially if they are waxes. As already said, a gassy tube or other kind of leak within the tube can make the bias go haywire. To answer your question about bias, this is a negative voltage and if it decreases it causes the tubes to draw more current. As there are a lot of positive voltages floating around in the tube, any leaky path to the grid will reduce the negative bias and particularly as the grid circuit is at a high impedance, you don't need much of a leak.
As only one tube seems to be affected mostly, I would suspect the tube. If another tube does the same thing in that position then suspect perhaps the coupling capacitor. In extreme cases, I have seen a faulty tube causing bias reduction to the others as well as itself and is a situation one tries to avoid
Les
Sorry M6tt, we seem to have been posting at same time
wrt the pots, I prefer good quality Cermets, especially the higher wattage multi-turn types
As only one tube seems to be affected mostly, I would suspect the tube. If another tube does the same thing in that position then suspect perhaps the coupling capacitor. In extreme cases, I have seen a faulty tube causing bias reduction to the others as well as itself and is a situation one tries to avoid
Les
Sorry M6tt, we seem to have been posting at same time
wrt the pots, I prefer good quality Cermets, especially the higher wattage multi-turn types
Last edited:
Thanks again for all the input.
Since this is a production amp that was working fine until yours truly played with the bias, I doubt there is a design flaw and rather, I pushed something too hard. Maybe a sensitive tube? Or an old capacitor? Or maybe I had set the bias wrong, when I re-set the tube currents after the glow incident, I had to stop and replace the battery in my meter. Maybe I was not setting things where I thought I was. Sounded good though.
Now it is reset, at a slightly more conservative setting and has been ok for a few hours play. I guess I will just have to see if if happens again. Need to play more music I guess. ;-) Can't complain about that.
I do want to get to upgrading and rewiring this amp. I am worried about some of the mentioned potential problems. It is a bit of a rats nest point to point design.
Since this is a production amp that was working fine until yours truly played with the bias, I doubt there is a design flaw and rather, I pushed something too hard. Maybe a sensitive tube? Or an old capacitor? Or maybe I had set the bias wrong, when I re-set the tube currents after the glow incident, I had to stop and replace the battery in my meter. Maybe I was not setting things where I thought I was. Sounded good though.
Now it is reset, at a slightly more conservative setting and has been ok for a few hours play. I guess I will just have to see if if happens again. Need to play more music I guess. ;-) Can't complain about that.
I do want to get to upgrading and rewiring this amp. I am worried about some of the mentioned potential problems. It is a bit of a rats nest point to point design.
Just go around and make sure the bias is good on all tubes before calling it good. You will find out once you set one tube, all the others will drift off. Its an iterative process that should be complete when all the tubes are biased.
You may notice that the bias drifts too much and you cannot keep the bias stable for any period of time. At this point I would wonder about a bad tube or circuit issue. But checking the tubes would be a good place to start.
You may notice that the bias drifts too much and you cannot keep the bias stable for any period of time. At this point I would wonder about a bad tube or circuit issue. But checking the tubes would be a good place to start.
I know! I have been following the glowing tube picture thread and thought of this too late! I t hasn't gone haywire a second time, so no picture. I was too freaked the first time and turned it off right away...
It was pretty spectacular. I am not sure I have the cajones to force it to do it again. Maybe I will try with a lesser quality/older set of tubes. Boom....
I'm a bit late on this but here is my take:
The vacuum of many valves/tubes on sale today is not as "hard" as some of the originals. Sometimes it's from new, sometimes it's due to gas emission from the electrodes after some use. Whatever the reason, when run hard, the device can run-away. This is especially true in fixed bias arrangements. Cathode bias or automatic bias arrangements suffer less. You will notice that on spec sheets there are 2 different values given for grid-leak resistors (Rg1). The fixed bias one is lower, indicating that the manufacturer takes the run-away issue more seriously in this mode. Often the fixed bias value is less than half the auto-bias value.
So check the value of the grid-leak resistors (with an ohm meter) and compare them with the max value on the manufacturers data sheet (the actual manufacturer of your valves). Choose a value that is at least one value lower than the max. Caution: If too large a downward change is made this can affect the distortion of the driver stage as well as the frequency response. This can be a problem on older amps designed around original valve types.
This is just one element of ensuring that you are running within the design values for the device. If you can ensure that all are within ratings (remembering that the mains might by a percentage fluctuate too) then assuming no component failures you will get the best out of the valves/tubes and can expect a long and trouble free life.
Other reasons I have found for this behaviour are leaky coupling caps and poor contact between valve g1 pin to socket.
I repair quite a number of valve amps for people and always check the above points before fitting new output valves.
Hope this helps.
The vacuum of many valves/tubes on sale today is not as "hard" as some of the originals. Sometimes it's from new, sometimes it's due to gas emission from the electrodes after some use. Whatever the reason, when run hard, the device can run-away. This is especially true in fixed bias arrangements. Cathode bias or automatic bias arrangements suffer less. You will notice that on spec sheets there are 2 different values given for grid-leak resistors (Rg1). The fixed bias one is lower, indicating that the manufacturer takes the run-away issue more seriously in this mode. Often the fixed bias value is less than half the auto-bias value.
So check the value of the grid-leak resistors (with an ohm meter) and compare them with the max value on the manufacturers data sheet (the actual manufacturer of your valves). Choose a value that is at least one value lower than the max. Caution: If too large a downward change is made this can affect the distortion of the driver stage as well as the frequency response. This can be a problem on older amps designed around original valve types.
This is just one element of ensuring that you are running within the design values for the device. If you can ensure that all are within ratings (remembering that the mains might by a percentage fluctuate too) then assuming no component failures you will get the best out of the valves/tubes and can expect a long and trouble free life.
Other reasons I have found for this behaviour are leaky coupling caps and poor contact between valve g1 pin to socket.
I repair quite a number of valve amps for people and always check the above points before fitting new output valves.
Hope this helps.
Might be a good idea to change the suspect tube over to a position in the other channel. That way should you again see a problem, it would help isolate whether it follows the tube or stays fixed to a particular socket and associated supporting parts.
Do reset bias if you move it. It would be interesting if bias changes much between the positions.
Do reset bias if you move it. It would be interesting if bias changes much between the positions.
Thanks! tons of great advice. Of course, just like a car, it hasn't done it again since i started this thread. Go figure. I must have just screwed something up.
Just to check, KT88 with quiescent plate voltage 445V, 10R cathode resistor, trying to bias class A, 42W dissipation. I should be measuring somewhere around 95 - 100mA?
So i measure 0.95 - 1.0 A on my meter (10R). Right now I am at 90mA, because it seems I shouldn't push that hard.
Just to check, KT88 with quiescent plate voltage 445V, 10R cathode resistor, trying to bias class A, 42W dissipation. I should be measuring somewhere around 95 - 100mA?
So i measure 0.95 - 1.0 A on my meter (10R). Right now I am at 90mA, because it seems I shouldn't push that hard.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.