Monkeymatic KT88 SE Bass amp

[xtian]

A gentleman hired me to make a very interesting bass head. It's a 25-watt singled ended design, based on AX84's SEL. For bass guitar?!? Yup. Client's idea is to use the head as a low-powered practice amp, and then, by running line level signal out of my amp and into a 300-watt, solid state amp, to power his large speaker cabs, making good use of preamp and power tube distortion to color the signal.

Finished wiring and powered up today. Looking at 450v on plate of KT88. Getting bad distortion, though. Look at the scope: I have a clean sine wave exiting V1 after passing both triodes (top of photo). But look at the wave if I just move the probe across the 220K resistor (R16 on the schematic)! Why would it clip like that?

[surfsup]

What's pin 3 read voltage wise? is that a 820R for sure?

It looks like you have a resistor tied to the cathode with a green wire attached going off picture. What is that? It is not in the schemo.

[xtian]

What's pin 3 read voltage wise? is that a 820R for sure?

That 820R resistor (value confirmed), just to the left of my signature, is the V3a cathode resistor (R15 in the schematic), and it reads 2.8v to ground. 98v on V3a plate (with 12au7 installed...12ax7 had ~200 on plate).

I replaced both 12ax7s with 12at7s, and it made no difference in this behavior.

It looks like you have a resistor tied to the cathode with a green wire attached going off picture. What is that? It is not in the schemo.

That's the NFB resistor (hand drawn in my schematic). I disconnected it before making latest measurements. Same behavior, with the radically and asymmetrically clipped sine wave.

I'm going to lift R17 to ground and try again.

[Leo_Gnardo]

If you were using that point to drive a power tube grid and it went into the "positive" zone aka A2, that would load the driving circuit into clip like we're seeing. Perhaps driving the V3B grid positive wrt the cathode is the problem. Try a larger R15, say 3K3 to 10K and see whether that clip goes away, or at least lets you get a larger signal amplitude before it does.

On another matter I'm not so sure the presence circuit will work the way you have it penciled in. High frequencies will be shunted to ground via the 1 uF C10 bypass cap so shunting them via the presence control won't matter. Take away C10 and voila', the pres control makes sense.

[xtian]

Thanks, LG, but I did not install the presence control (see my "delete" mark). I just used the NFB loop. But now I have disconnected it for further testing.

I lifted R17 to ground, and that allowed my sine wave to look normal going into the grid of V2a. Now, at the output (plate) of V2a, the wave is radically clipped.

I'm going to take a careful look at my wiring around V2, and will take your suggestion on the V2a cathode resistor (R15) if I can't find anything. Oh, but here's a thing: If I make the sine wave entering V2a very small by turning the gain pot way down, the sine wave exiting V2a STILL looks very clipped, just very small in amplitude. See photo.

[Leo_Gnardo]

The nearest schemo I had ready to hand that's similar to your pre was a Marshall 2203.

You mean V3 on your schemo? I don't see any V2.

I'm suspecting that circuit between the V3 triodes is what's foxing us, R13 and D3. Can't say I've seen the like. What happens if you lift the end of one of those components?

[xtian]

Sorry, Leo, I was naming the tubes V1 and V2 in my first post. Sorry! From here on out I'll use the schematic designations. So the cathode follower is V3a.

R13 and D3 are a safety measure, explained by Merlin here: http://www.freewebs.com/valvewizard1/dccf.htm

At any rate, I thought a diode is a likely source of clipping, so I removed it. Same exact behavior. No change.

[xtian]

Hey! Ain't larnin' fun?

I noticed that, even with the gain pot at zero, I was still seeing a perfect sine wave out of the second triode. I thunk and thunk and figured it out. Anyone noticed I have a shared cat resistor on the first preamp tube...and no bypass cap? Well that signal was shaking the cathode good, and being amplified by the neighboring triode, and out came the sine wave at the other end, sneaking right around the poor gain control. Sneaky bastard.

I just put a 25uF bypass cap across the 820R, and things are MUCH better. That asymmetric clipping must have been phase cancellation. A little half-wave, 100% neg feedback. Hey, I invented something!

I've got a few more issues to sort, but she's starting to sound pretty cool.

[xtian]

Here is a recording of my Tele, from the line out, with internal dummy load engaged (so no speaker!). Direct into MOTU audio interface, no FX, no EQ.

[Leo_Gnardo]

Anyone noticed I have a shared cat resistor on the first preamp tube...and no bypass cap? Well that signal was shaking the cathode good, and being amplified by the neighboring triode, and out came the sine wave at the other end, sneaking right around the poor gain control. Sneaky bastard.

Hmm, yes, see "Push-pull without a phase inverter" thread, appearing now. Yup, that would do it.

[xtian]

I was getting blocking distortion (motorboating) when turning up the gain past 8, even with no signal plugged in. Think I ruled out parasitic oscillation. I adjusted the voltage divider (R20,21) until I was able to dime the gain control without motorboating. Ended up with 1M/220K, giving me about 18% signal. Overall, amp sounds great like this; plenty of gain, plenty of output volume. Just thinking the overall distortion character sounds not quite as good as it did before, but don't know if I'm imagining this.

Given that the coupling cap is 0.022uF, is there a way to calculate the highpass value? Other reason I might want to use lower value resistors in my voltage divider?

[JoeCon]

Xtian

I got this from Leo and Martin in a recent post. See if this works.

F = 160,000 div. by R x C, where C is in uF, R in ohms & F in Hz.
I simplified it a bit for yez so you don't have to calculate the 2 x pi in the basement nor deal with the x .0000001 for C in Farads. You can also calculate for C, given an F.

C = 160,000 div. by R x F .

Whatever C you happen to have, close enough for rock and roll. Parallel 'em if you have to fine tune.

Leo, I use 159155, k-ohms, and nF. It's a little more accurate, but mostly I like it because 1) that number is all on the diagonal of the keypad, and 2) it's easy to get to nF from either pF or uF (only three decimal places either way).
There's another one of those so-called "treble peakers." These things are a high-pass shelf filter, so there are two break points. The lower frequency +3 dB break is easy because it is always determined by the parallel RC (the upper one depends upon the surrounding circuitry).
For the Marshall values 470k and 0.47 nF it's 159155/470/0.47 = 720 Hz
For the Fender dry bypass it's 159155/3300/0.01 = 4822 Hz

[rock_mumbles]

Was I lead astray when I was told to never share a cathode unless you have parallel triodes that are in phase ???

Inquiring minds want to know

[xtian]

Was I lead astray when I was told to never share a cathode unless you have parallel triodes that are in phase ???

No, I think you are right! And for good reason, as I discovered.