Took some measurements. The output tubes are not in place. Are there any other places you want me to check?
The DC looks OK to me, but I'm not a tube guy so thers please chime in.
If you input a signal, can you follow it through the circuit to the outputs of the drivers?
Jan
If you input a signal, can you follow it through the circuit to the outputs of the drivers?
Jan
To me the sine wave is "perfect" all the way to pin 3 on the output tube sockets. KT88 not in place. When KT88 is in place the amp is stable as long as no signal is fed. When the signal is on the bias starts to flicker. Not so much in the beginning but after a minute or two I can hear it in the speakers. And the sound is of course distorted.
A couple of things:
1) When you remove the output tubes, the voltages on the input and phase splitter will be significantly higher, so that doesn't tell us much.
2) That's a Mullard circuit, and the plate voltage on the input stage is very high, even without the load of the output tubes. Typically it's more along the lines of 70-80 volts. Look at an Eico HF-60 schematic and you'll see what I mean.
3) It sounds like the amp is oscillating. There's a low-pass filter across the EF-86 plate resistor. Who calculated that? Is it enough HF cut-off for the output transformer? You also have no phase-lead cap across the feedback resistor, which would be unusual.
4) What output transformers are you using? Have you looked at some square waves (1K, 10K) to check for ringing?
Last edited:
Also, you're applying the feedback at an odd place. Look at the Eico schematic again, which is much closer to a standard Mullard design. The feedback is applied to a small unbypassed resistor after the bypassed cathode resistor of the EF-86. Have you checked to see how much feedback you're actually getting?
The transformer is Hammond 1650N. I`ll check with the output tubes in place and measure some square waves. The idea of the amp is taken from this one:
http://www.tube-amps.net/images/KT88_PP_01/KT88_Schematic.jpg
There is an unbypassed resistor in the cathode lead. It's on top, not bottom, but that's the same of course. 680R if my eyes don't betray me.
Same as the schematic he posted last.
Your other points are good ones.
Jan
Last edited:
The bias for the output tubes. I have set the bias to 40ma for 525Vdc to the KT88.
You mean the -65? That can only vary if your output tubes draw grid current, so that points to (violent) oscillations.
Or check this part of the circuit for miswiring.
Jan
Ah, okay. Well, the Hammond has a lower primary impedance than the Hashimoto, so you'll want to check the feedback level, you may need to tweak the value of the feedback resistor. But the biggest problem, I suspect, is that the Hammonds aren't playing well with the rest of the design and will need some tuning to get the amp stable. It would be helpful to see some square waves.
According to the schematic, as simon7000 pointed out, your Zero common tap is not grounded or is that a drawing error? I would make sure it's grounded. Check with and without feedback and you might need to tweak the feedback resistor value.
There's voltage discrepancy between the two circuits. In the original the B+ for the driver is 525VDC and yours is 400VDC.
? With the KT88 outputs not in place there can't be very much signal on their pin 3's. Perhaps this is an error of translation.
One interpretation of the bias flickering is that the amplifier is oscillating, or almost (conditionally) oscillating, at a very low, sub-sonic, frequency, sometimes called "motor-boating". By the way, does "amplifier" here mean both channels of a stereo amplifier, or one, or is this a single channel amplifier?
Yours is a very high gain design with a lot of loop feedback, so stability is not easily had. It has no high frequency response / phase shaping in the forward path or the feedback path, and very large coupling capacitors, which set the low frequency shaping, so you're not making things easy on yourself. If stability does turn out to be your issue, you might want to address those things for future reliability.
All good fortune,
Chris
I think there is a low-pass filter across the input stage plate resistor, but probably not enough (or tuned properly) for the Hammond outputs. You're right that the coupling caps are significantly larger than the original design. I had assumed HF oscillation but LF is is possible too.
I'm assuming that the series'd active current source dominates its combination with the bypassed 150K. That's only an assumption, but likely.
All good fortune,
Chris
All good fortune,
Chris