I changed again and now I have just under -43VDC applied to the grid. The result is that when measuring DC between pin8 and ground I get 0VDC untill the valves start to glow, then the speakers starts to humm and make popping sounds as the measured value starts to rise fast.
Just so that I got the theory right.
By applying a negative voltage to the grid we keep the electrons from rushing from the negative cathode to the positive anode. Then, by lowering the applied voltage and thus allowing electrons to pass we are raising the bias, so the fact that I get a rising voltage after a few seconds is because the grid stops fewer electrons and the tube is running hotter. Am I correct or did I get the last part backwards?
Just so that I got the theory right.
By applying a negative voltage to the grid we keep the electrons from rushing from the negative cathode to the positive anode. Then, by lowering the applied voltage and thus allowing electrons to pass we are raising the bias, so the fact that I get a rising voltage after a few seconds is because the grid stops fewer electrons and the tube is running hotter. Am I correct or did I get the last part backwards?
Is this a stereo amp?
If you test one channel with different tubes (rotate the tubes around/try different tube combinations) does it behave the same? Leave the tubes out of the second channel.
Are the tubes known to be OK?
Your description is correct; although with fixed bias the cathodes should be at (or very close to) ground potential. With a bias current of 60ma across the 10R cathode resistor, the cathode is at .6V. So when the amp is warmed up and idling (no tunes playing), you should see around .6V at the cathode if things are working properly.
So, for clarification, is the problem that the tube current is running away once the tubes warm up? Is this happening on both channels?
If you test one channel with different tubes (rotate the tubes around/try different tube combinations) does it behave the same? Leave the tubes out of the second channel.
Are the tubes known to be OK?
Your description is correct; although with fixed bias the cathodes should be at (or very close to) ground potential. With a bias current of 60ma across the 10R cathode resistor, the cathode is at .6V. So when the amp is warmed up and idling (no tunes playing), you should see around .6V at the cathode if things are working properly.
So, for clarification, is the problem that the tube current is running away once the tubes warm up? Is this happening on both channels?
Last edited:
Monoblocks, fully separated.
I did some tuberolling with the same result as earlier. A few seconds before bias starts to surge and a rythmic taktak-taktak-taktak from the speakers.
As sure as I can be, I got them measured and matched from a reliable person on a forum here in Sweden.
Well, if the current translates to voltage here then it is as you say. I measure voltage (2000mV setting on the meter) between the cathode and ground.
It is only happening on one channel I'm afraid. I tried the second channel just now and it seems like the secondary to the negative bias supply is broken. Since this is the first time I even used the transformer and the supply is identical to the other channel it seems like it has been broken from delivery (custom wound).
Ha, since I have been moving the top plate and tubes around so much the last coulple of days I had managed to break the solderjoint to the heater and that explains the non glowing tube. It is fixed now, the tubes both glow and with about -45VDC on the grid there in no bias and no strange sounds. Even the humming dissapeared.
But how come the feedback is the culprit (did I get the word right, thief, bad guy) and what do I do about it?
But how come the feedback is the culprit (did I get the word right, thief, bad guy) and what do I do about it?
Are you using 3k3 and 560pf for the feedback resistor and capacitor?
Are you testing the amp with a load on the outputs? Are you using a load resistor or a loudspeaker?
Do you have an oscilloscope?
Can you determine the open loop bandwidth?
You could try reducing the feedback by increasing the value of the feedback resistor until the oscillations are under control. Then you may need to adjust the feedback cap value to maximize HF response using a 10kz square wave input. You are looking for a sharp corner (as sharp as you can get it) on the leading edge of the square wave.
Also confirm that the grid stopper on the EF86 input tube is correct.
I have not built any amps with conventional feedback yet so I have zero experience here. I have been reading a lot though, and adding feedback increases bandwidth, sometimes to the point where a HF pole can cross 180 deg causing stability problems. Once you get the oscillations under control it's a good idea to determine stability margin, i.e. how close to the edge are you as far as oscillations go. Ideally, the amp should be stable with or without anything connected to the input and output. Testing the amp without an output load is not usually a good idea, though.
One of the experts may chime in here with a methodical way to optimize your feedback.
Are you testing the amp with a load on the outputs? Are you using a load resistor or a loudspeaker?
Do you have an oscilloscope?
Can you determine the open loop bandwidth?
You could try reducing the feedback by increasing the value of the feedback resistor until the oscillations are under control. Then you may need to adjust the feedback cap value to maximize HF response using a 10kz square wave input. You are looking for a sharp corner (as sharp as you can get it) on the leading edge of the square wave.
Also confirm that the grid stopper on the EF86 input tube is correct.
I have not built any amps with conventional feedback yet so I have zero experience here. I have been reading a lot though, and adding feedback increases bandwidth, sometimes to the point where a HF pole can cross 180 deg causing stability problems. Once you get the oscillations under control it's a good idea to determine stability margin, i.e. how close to the edge are you as far as oscillations go. Ideally, the amp should be stable with or without anything connected to the input and output. Testing the amp without an output load is not usually a good idea, though.
One of the experts may chime in here with a methodical way to optimize your feedback.
An old surround speaker, 6ohms.
No on both, I'm sorry to say.
Correct.
Yes, 4k7ohms.
I had to lower the resistor (Rvg1) again to be able to go higher than -35V but I didn't get an increased reading on the bias untill I got to -15V and then just 10mV between cathode and ground, not even close to 400mV. In this moment I had 10mV on one tube and 0mV on the other. I reasoned that I didn't have the DC-balance in the correct position and turned it a bit, with this a buzzing and popping sound returned, so I turnsed it back untill it stopped.
Zero experience or not, I am further along now than before and I am greatful for the help you have given me.
Last edited:
Sounds like you have something else going on.....If you change out your tubes with the other pair do you still get buzzing & popping when the DC balance pot is in the middle position?
Slowly double-check your wiring against the schematic very carefully. Then triple check it.
Do you have a variac or other means to reduce the mains voltage for testing?
With the output tubes removed, do you get the same voltage range for each tube from pin 5 to ground when adjusting the bias pot? You probably need to center the balance pot also to get the same voltages.
If you remove the small signal tubes can you adjust the DC balance pot to center with the output tubes in place and bias current flowing without problems? If so, you problem is upstream of the output tubes.
At this point, I would consider checking the voltages at anode to ground, grid to ground and cathode to ground at the pins of each output tube without the tubes in place. If you have clip leads for your meter, use them, especially on the anode as it is at B+ voltage. If you do not have clip leads or a means to test it without touching it, use the one hand method, as you are dealing with potential lethal voltages at the anode.
If you can reduce the mains voltage, it makes things easier and safer. You are looking for similar voltages between sides as well as reasonable voltages.
My voltages under load seems to be all wrong (tubes and speaker connected).
According to specs they should be as following:
A:450V
B:420V
C:180V
G:-60V
My values are:
A:166V
B:177V
C:84V
G:-84V
I get the correct AC values out from the secondary side of the power transformer and the resistors in the PSU are correct.
According to specs they should be as following:
A:450V
B:420V
C:180V
G:-60V
My values are:
A:166V
B:177V
C:84V
G:-84V
I get the correct AC values out from the secondary side of the power transformer and the resistors in the PSU are correct.
No kidding.........what's the current rating of your mains transformer?
Do the various voltages get close to the design values without the tubes installed?
Looks like you have a low resistance to ground somewhere. Double check all your resistors, i.e. make sure that you don't have a 47R where you should have a 47K R, etc.
If you adjust the bias pots for max negative voltage with the tubes installed (put the output tubes into cutoff), are the various B+ voltages still low?
Last edited:
kmj,
You are dealing with a couple of issues. The first being an instability in the amp, so before we start playing with feedback just leave it aside for the moment. I am afraid that the low voltages you are seeing is a direct result of the runaway oscillation. You should not let the amp run like this very long as you can damage the tubes.
I signal generator (even a soundcard one) and an oscilliscope are invaluable here. I had similar issues with my Opus when bringing it up.
I have briefed through the thread, but I am not clear on whether you can obtain the proper bias voltages without the output tubes in?
You are dealing with a couple of issues. The first being an instability in the amp, so before we start playing with feedback just leave it aside for the moment. I am afraid that the low voltages you are seeing is a direct result of the runaway oscillation. You should not let the amp run like this very long as you can damage the tubes.
I signal generator (even a soundcard one) and an oscilliscope are invaluable here. I had similar issues with my Opus when bringing it up.
I have briefed through the thread, but I am not clear on whether you can obtain the proper bias voltages without the output tubes in?
Okay I have read it in a little more detail.
Take your Output tubes out for the moment. We need to verify that you obtain the correct bias voltage on the grid.
Please also look at your caps on the bias supply. Make sure that they are correctly orientated. The polarity is backwards from normal since this is a negative supply. For this rail, the positive lead should be attached to ground and the negative to the rail. If you have them backwards you should replace them with new in the right polarity.
If the caps are correct, with the output tubes out, set your bias balance pot in the center, and the bias level pot in the most negative setting. Report back with your voltages like above but include voltage before and after Rvg. I also need to know the value of Rvg.
Like my previous post, I feel your problem is with runaway instability. But first we need to break down the circuit a little to isolate the source of it.
My OPUS did almost the exact same thing. The problem turned out to be a bad input driver tube. However, b4 we start swapping tubes, we need to eliminate variables. The output tube bias circuit should work fine with or without the tubes so lets start there and make sure it is working correctly. While you do this watch the behaivior of the driver and phase splitter tubes, make sure the plates do not start to glow or other parts start losing their magic smoke;-)
Boywonder, I suggested pulling the output tubes out since when I had this problem, it didn't matter. Once the output tubes started to conduct it took off and it wouldn't stop until I pulled the plug even if I set the bias back to max negative. Without the output tubes at least we can get some measurements without burning too much up. We should also be able to get some no load voltages which can point us to other issues if they don't come back reasonable.
Take your Output tubes out for the moment. We need to verify that you obtain the correct bias voltage on the grid.
Please also look at your caps on the bias supply. Make sure that they are correctly orientated. The polarity is backwards from normal since this is a negative supply. For this rail, the positive lead should be attached to ground and the negative to the rail. If you have them backwards you should replace them with new in the right polarity.
If the caps are correct, with the output tubes out, set your bias balance pot in the center, and the bias level pot in the most negative setting. Report back with your voltages like above but include voltage before and after Rvg. I also need to know the value of Rvg.
Like my previous post, I feel your problem is with runaway instability. But first we need to break down the circuit a little to isolate the source of it.
My OPUS did almost the exact same thing. The problem turned out to be a bad input driver tube. However, b4 we start swapping tubes, we need to eliminate variables. The output tube bias circuit should work fine with or without the tubes so lets start there and make sure it is working correctly. While you do this watch the behaivior of the driver and phase splitter tubes, make sure the plates do not start to glow or other parts start losing their magic smoke;-)
Boywonder, I suggested pulling the output tubes out since when I had this problem, it didn't matter. Once the output tubes started to conduct it took off and it wouldn't stop until I pulled the plug even if I set the bias back to max negative. Without the output tubes at least we can get some measurements without burning too much up. We should also be able to get some no load voltages which can point us to other issues if they don't come back reasonable.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.