SE amp: Please suggest a better operating point

[zobsky]

Well, I built this 12B4A amp which sounds very nice but every now and then, I get this nasty distortion on loud transients, .. sort of a sharp buzzzzzzz sound.

The schematic:


The input is clean (PC sound card outputting sine waves from winIsd, ... scoped it out).

I scoped the driver (past the coupling cap) and the output (at the secondary of the OPT) and it seems that the driver cuts off early (see pic). The driver is biased with a red LED @ 1.7 V .
The output tube eventually clips (or is that ringing?) and the cut off wave form of the driver is superimposed upon this.

I'd ideally like to increase the bias on the driver to avoid the cut off, but am not really sure what to do, as I don't have that much voltage to play with (about 210V). I have a CCS in place of the plate resistor. I tried increasing the current through the CCS
but that didn't help the issue. I'm currently at around 2mA through the 12AT7 driver.

Sorry for the poor cell phone picture quality.

Driver cutting off

[zobsky]

Output doing something wierd ????

[Predator864]

That`s clipping.

Adjust with ReplayGain your tracks at 89dB, and test the result.

[Ty_Bower]

Originally posted by zobsky
The driver is biased with a red LED @ 1.7 V .
The output tube eventually clips (or is that ringing?) and the cut off wave form of the driver is superimposed upon this.

I'd ideally like to increase the bias on the driver to avoid the cut off, but am not really sure what to do, as I don't have that much voltage to play with (about 210V). I have a CCS in place of the plate resistor. I tried increasing the current through the CCS
but that didn't help the issue. I'm currently at around 2mA through the 12AT7 driver.

At a constant 2mA through the 12AT7 and 1.7V on the cathode, you should expect about 110V at the plate.

If you get too close to the Ug=0V line on the curves, you're going to cutoff the tube. I'm sure this happens before you actually hit the 0V line, but I don't know how close you can get. I think you are on the right track with wanting to increase the bias on the 12AT7, but if you have only 210V to play with at the plate you're probably out of luck. Increasing the bias voltage even half a volt will put the idle voltage of the plate around 170V, and your power supply will have no volts left for the upswing. Get a higher voltage supply, or keep your input signal under 0.7 Vrms.

You might also want to think about what might happen if you ditch the CCS plate load and replace it with a resistor. If you use a relatively small resistor (10K?) I think you'll significantly reduce the gain of the driver stage. Maybe that's not what you want, but it'll probably keep it from clipping.

[zobsky]

Thanks
to summarize my earlier posts, I would ideally like the output to clip before the driver

reducing the gain on the driver might not be the best way to do this. I'll see if increasing c1 in the filter gets me any significant gains in voltage

I'll also see if I have any other 9 pin dual triodes that might work and deliver at least 3 ma

[kenpeter]

The drive can only rise 25V before the output's grid conducts.
With driver Mu=60, that equates to only .417V at the input.

Your flat spot may simply be the drive signal shorting to the
output's cathode through that grid's forward bias. That seems
the most likely.

If it was your input's grid doing the distorting, wouldn't your
drive signal's flat spot have been on the lower side of the trace?

[kevinkr]

Choke loading would allow you twice the voltage swing on your driver stage, and the 12AT7 has a low enough RP that this should work pretty well. The much higher effective Vp would also allow you much more flexibility in finding the best operating point for your driver stage.

[zobsky]

yes
choke loading may be something I'll have to consider in this case?

any affordable choices? this is a junkbox amp and a VERY good one at that
Thanks

[Wavebourn]

Take a small screwdriver, trim the CCS pot and see what happens.
You need to increase current until clipping starts to be more or less symmetrical.
Or, decrease resistor's value in cathode.

[zobsky]

2 ms was about the lowest I could adjust to

I think I'm going to pick up some Hammond 156c chokes