Also, are you sure that the output transformer is wired with correct polarity. Get the polarity wrong and you have positive feedback instead of negative.....
Greetings again. I received the 1000pF caps yesterday and soldered them in. The audible oscillation is gone, however when I measured the plate voltages, strange things were on the meters. With 0v signal input, the left channel is relatively steady at 304-6v, slowly moving across the range, while the right channel rapidly sweeps from 290-315v. May I assume this is oscillation?
Thanks
w
Thanks
w
Did you pull the PI tube as advised?
This interrupts the feedback loop and should stop any oscillation.
With a stable amp this makes no difference to power tube plate voltages.
This interrupts the feedback loop and should stop any oscillation.
With a stable amp this makes no difference to power tube plate voltages.
Hi, I saw your earlier comment, do you mean Preamp tube? Operating the amp with only power tubes installed to see if the problem persists? Will give it a shot.
Power amps sometimes start to oscillate when a meter is connected to a power tube plate.
Reason is a positive feedback effect caused by the capacitance between the hot meter lead and the amp circuitry.
It typically happens only at one side of the OT primary but not at the other.
The 12AX7 is your phase inverter (PI).
It's an integral/essential part of the power amp and not a preamp stage.
Reason is a positive feedback effect caused by the capacitance between the hot meter lead and the amp circuitry.
It typically happens only at one side of the OT primary but not at the other.
The 12AX7 is your phase inverter (PI).
It's an integral/essential part of the power amp and not a preamp stage.
Well, some progress. ;-) If both channels stabilize with the preamp tubes removed, then your next step might be to increase the feedback resistor bit by bit (thereby reducing the amount of feedback) and see if it settles down. You'll probably have to experiment with varying levels of feedback and different phase-lead caps to get a stable, working amp.
Maybe I wasn't clear enough.
There's a possibility that the amp only oscillates when the meter is connected, but actually is stable.
If both amps otherwise work fine I see no reason to (blindly) change the circuit.
The low hum could be caused be unbalanced power tubes.
I recommend to get out the scope before doing any changes.
There's a possibility that the amp only oscillates when the meter is connected, but actually is stable.
If both amps otherwise work fine I see no reason to (blindly) change the circuit.
The low hum could be caused be unbalanced power tubes.
I recommend to get out the scope before doing any changes.
right. Excuse my confusion, I've only built Single-Ended amps before this. Are both stages of the 12AX7 considered the PI, or is the first half the preamp? Either way, have to pull the tube to check. What's got me twisted is the tube grounding through the OT, so the Fb resistor is the K resistor, and the play between the feedback and the bias. And why one side of the amp is OK and the other is weird. Could mis-matched tubes be enough to make a gremlin?
Have to get ready to travel this wknd, maybe get to play with this again in the middle of next week. Thanks for your advice.
w
Last edited:
Yes, with a paraphase PI, the first triode section is a gain stage and the second one follows being wired for a gain of -1, meaning it inverts the signal of the first triode.
So the 2 plate outputs provide equal but out-of-phase signals necessary to drive the grids of the PP amp.
Yes, it's just a power amp meant to be connected to a preamp that can provide a signal of some volts.
The cathode bypass cap should be polarized. Positive goes to cathodes, negative to ground.
I often recommend using individual self bias resistors for each output tube.
A 250 Ohm resistor from one EL84 cathode to ground with a 100uF cap across it.
A second 250 Ohm resistor from the other EL84 cathode to ground with a second 100uF cap across it.
Why 250 Ohms? Easy to get, and 2X 130 Ohms = 260 Ohms, so 250 Ohms is close enough.
If the quiescent tube currents are equal, there will be equal voltage across each of the 250 Ohm resistors.
Try this with the PI tube pulled (the only chance for oscillation with PI pulled is a mis-wired output stage or output transformer.
If the output tubes do not have the same current, then the first rule of push pull output stages which have a single common self bias resistor needs to be followed: Use very well matched output tubes. All good kits used to follow this rule, with good reason.
And . . . your output transformers will work much better, no early saturation of the core laminations when the currents are equal.
Note: Un-equal currents will cause one output tube plate voltage to be lower than the other output tube plate voltage.
Once you have very well matched output tubes, if you wish, you can re-wire back to the single 130 Ohm and 100uF self bias circuit.
A 250 Ohm resistor from one EL84 cathode to ground with a 100uF cap across it.
A second 250 Ohm resistor from the other EL84 cathode to ground with a second 100uF cap across it.
Why 250 Ohms? Easy to get, and 2X 130 Ohms = 260 Ohms, so 250 Ohms is close enough.
If the quiescent tube currents are equal, there will be equal voltage across each of the 250 Ohm resistors.
Try this with the PI tube pulled (the only chance for oscillation with PI pulled is a mis-wired output stage or output transformer.
If the output tubes do not have the same current, then the first rule of push pull output stages which have a single common self bias resistor needs to be followed: Use very well matched output tubes. All good kits used to follow this rule, with good reason.
And . . . your output transformers will work much better, no early saturation of the core laminations when the currents are equal.
Note: Un-equal currents will cause one output tube plate voltage to be lower than the other output tube plate voltage.
Once you have very well matched output tubes, if you wish, you can re-wire back to the single 130 Ohm and 100uF self bias circuit.
Last edited:
Good point.
A shared cathode resistor doesn't improve balance, rather it tends to make it worse.
sure....may have to wait a bit though...meanwhile, the individual bias resistors will be burning in a few years
Can't do individual bias resistors, I'm on a PCB. But symptoms sure do lean towards unbalanced tubes.
With the amplifier unplugged, and the B+ capacitors discharged . . .
Short the output transformer secondary windings (a jumper wire from Common to 8 will work). That will short out the inductances of all the output transformer's windings. (some times the inductance will cause a DMM set to Ohms to keep switching ranges, auto-range goes crazy.
Now you can set your DMM to Ohms, and measure from one output tube plate to the center tap (B+); then measure from the other output tube plate to the center tap (B+).
Those ohms should be close (perhaps something like 95 Ohms for one and 110 Ohms for the other, worry about near equality of Ohms, not the number of Ohms).
Then do the same Ohm measurements from one UL tap to the Center tap; and then from the other UL tap to the center tap.
Those should also be close to equal.
The UL taps to the center tap Must read less Ohms than the Plate taps to the center tap.
If the resistance is not close, then one of the plates might be connected to the Ultra Linear tap, and one of the screens wired to the plate tap;
or all the plate and UL taps might be swapped.
Equal plate primary resistances to the center taps will make for equal plate voltages.
Do not forget to remove the shorting wire across the Common and 8 posts, before you power up the amplifier again.
Short the output transformer secondary windings (a jumper wire from Common to 8 will work). That will short out the inductances of all the output transformer's windings. (some times the inductance will cause a DMM set to Ohms to keep switching ranges, auto-range goes crazy.
Now you can set your DMM to Ohms, and measure from one output tube plate to the center tap (B+); then measure from the other output tube plate to the center tap (B+).
Those ohms should be close (perhaps something like 95 Ohms for one and 110 Ohms for the other, worry about near equality of Ohms, not the number of Ohms).
Then do the same Ohm measurements from one UL tap to the Center tap; and then from the other UL tap to the center tap.
Those should also be close to equal.
The UL taps to the center tap Must read less Ohms than the Plate taps to the center tap.
If the resistance is not close, then one of the plates might be connected to the Ultra Linear tap, and one of the screens wired to the plate tap;
or all the plate and UL taps might be swapped.
Equal plate primary resistances to the center taps will make for equal plate voltages.
Do not forget to remove the shorting wire across the Common and 8 posts, before you power up the amplifier again.
A new set of Matched EL84 tubes arrived the other day and I finally had time to install em and put the amp back on the bench. Voltages are steady across the board now. I realized that the heater windings weren't referenced to ground and I added a pair of 100ohm resistors from the heater circuit to the top of the cathode resistor on each PCB, now heater voltage is elevated by 10v or so, and the amp is dead silent at full volume. Great.
After maybe 5 mins of listening, I realized I had a problem. I really had to crank the amp to get the bass I was used to, and soon after I noticed that my ears were ringing. So I guess the frequency response is off? The output is skewed to a higher range of frequencies. Perhaps this is related to the mismatch between circuit and output transformer. Can I improve bass response by replacing the cap in the feedback loop with a (larger?) value? I happen to have some 1800pF and 2200pF around here somewhere I could swap in, if it would help.
Or maybe there's another tree to bark up? Thanks for taking a look.
w
After maybe 5 mins of listening, I realized I had a problem. I really had to crank the amp to get the bass I was used to, and soon after I noticed that my ears were ringing. So I guess the frequency response is off? The output is skewed to a higher range of frequencies. Perhaps this is related to the mismatch between circuit and output transformer. Can I improve bass response by replacing the cap in the feedback loop with a (larger?) value? I happen to have some 1800pF and 2200pF around here somewhere I could swap in, if it would help.
Or maybe there's another tree to bark up? Thanks for taking a look.
w
No, the phase-lead cap doesn't work that way, in fact it can be dangerous to play around with it too much, the amp can oscillate with the wrong value. "The bass you are used to" could just not be there. It's difficult to say. At this point it really needs to be viewed on a scope to diagnose the remaining issues.