Asteroid unpleasing distortion

[arjepsen]

I'm not using the official layout, I made a PTP layout myself.
You can see it here:

PTP layout.pdf

Tried lowering the cap on the second grid to 47nF.
Didn't remove the buzz. :-/

[romberg]

What did your PI and speaker output look like with the high cut disabled? My guess is that they now look normal.

If they are normal and you still don't like what you are hearing, then it has to be one of a few things:

• Your guitar and or cable is producing an erratic signal. Do you have another guitar and cable to test with?
• There is something going on which has not yet been captured with the scope. Try adjusting the input signal in level and frequency a bit. Also try a range of control settings. Can you get a strange looking signal to pop out? I'd try input signals in the range 50-150mV and frequencies from 120-4.5k. The goal is to try and make something like an oscillation pop out.
• You just don't like the way this amp sounds. Which is also valid. And one could adjust the circuit to change the order each stage breaks up. If more stages start to overdrive at the same time then the distortion will cut in and out more abruptly than if the stages clip one after another. Attach your scope to say the spkr out, PI, 3rd and 2nd stage and see what the order is. Crank up the gain and or input signal and see when each stage starts to clip relative to the others. Once you know this, you can adjust interstage voltage dividers to change when this happens. Yea. It will no longer be the same amp. But it is better to have an amp you enjoy playing rather than a copy of some amp you don't.

[arjepsen]

Hey again.
Yeah, I've tried another guitar, and another cable.
Same thing: plucking the D A or E string produces a buzzy distortion that fades out rather suddenly, in a sort of raspy way.

I think I'm quite certain that it originates in the PI. I've been using a headphone amp to listen to various points in the circuit, and the preamp stays clean long after the pi start distorting
I'm trying to change various components around it, to see whether something seems to help.

[romberg]

What does the PI output look like with no negative feedback and no cut control circuit? My guess is the output will look "normalish". If this is so, and this were my amp,
these are the things I'd tinker with:

1. I'd try adjusting the amount of negative feedback and see if that can have an effect on the PI output "nipples" you have on page two. When the output signal flat tops then
   it no longer works against an increase in signal on the PI input. Which is what I think I see going on with your shots on page two. Increasing the 47k feedback resistor
   should lower the feedback signal and the reverse when it is decreased. One could try replacing it with a pot temporarily and see what various values do to the signal at
   the PI output.
2. Try lowering the signal entering the PI so that it beaks up later. I think I'd try this by playing around with that 220k-330k voltage divider before gain stage three. Again you
   could use a temporary pot (say 500k with a 47k in series). Once you find something you like better you can either keep the pot (if it is a trimmer) or measure the resistances
   at the setting you like and replace it with new fixed resistors.

[thejaf]

I'm not up to par with the actual theory at present (use it or lose it), but I ran into similar symptoms with a build once ~10 years ago. These two things helped my situation:

1. Lower the value of paired "grid load" resistors in the PI (typically 1Meg). I used 330K.
2. Change the type (not value) or the PI output caps. I had Mallory 150 type originally, which are widely assumed to be smoother. I changed to 715P orange drops, and there was noticeable improvement in breakup tone.

[arjepsen]

Still messing around, and getting increasingly frustrated.
I did try playing around with the 220K/330K divider, replacing it with a 500k pot. I could change when the PI would start distorting, but it still started distorting the same way - a buzzy distortion, like it's overdriven a lot, and then "fades" to clean rather abruptly. It's like the difference in signal level going from clean to full distortion is way to small.
Basicly the divider just controlled where the gain had to be for the PI to start it's distortion. The sound was the same - just at different settings of the gain knob.

I also played around with the nfb. I've tried both larger and smaller values - also using a pot, and even disconnecting it. It did change the overall tone some, but didn't seem to do anyting about the distortion.

I've tried upping the grid stopper on the PI to 100k. Only difference was less high end. Also tried 100k power tube grid stoppers.
I've tried upping the tail resistor to 56k from the normal 10k. Again a bit different tone, but still the same buzzy distortion.
I tried using a pot in series with the bias resistor, changing the bias resistance between 470R and about 1k5. Didn't remove it.
I've lowered the 100n cap on the second grid to 22n, and tried changing two output caps to 6PS type OD's - still no dice.
Tried different tubes.
Tried yelling at my wife - not recommendable...

[thejaf]

You may have already stated this, if so I apologize. Did you try replacing the PI tube? Try both brand and type (12AX7 vs 12AU7).

In my experience, with PPIMV amps the tube and tube type can obviously change overdriven tone also. I've also never had luck using a 12AT7 in PI slot on PPIMV circuits, always sounded raspy.

[arjepsen]

Yeah. I' tried a balanced JJ 12ax7, a balanced tung sol 12ax7, also a few "normal" jj's plus also an EH 12AY7.
With the 12AY7 it's like the sound is the same, biggest difference is where the gain knob has to be before breakup.

[Reeltarded]

What does the amp sound like when setup as normal with the PPIMV all the way up and the volume in full drive?

These amps have a very short range before breakup that develops in the next very short range then they go full sing the rest of the way up..

[arjepsen]

Well, it doesn't sound "bad" per se - it's a bit fizzy, but I think it's possible to dial that out with the cut and presence controls.
I've had the schematic and docs lying around for a while, and finally chose to build it after hearing those demos on youtube.
However - aiming for those sounds he uses at the edge of breakup, I don't hear those bad sounds in his demos.

Wonder if anyone here who has made one of these amps, would mind posting a short recording of plucking just the E,A and D strings - letting them ring out - with the settings I posted in the start of the thread? It would be fantastic to hear how it sounds from others' amps - whether mine is actually posessed, or it simply is the way the amp sounds. Allyn?

[Reeltarded]

My liverpool even has a short range at the end of clean where large inputs do almost exactly this.. the higher strings to a less extent because.. you know.. I think your amp might have the same problem as my pool did when I first got it..

look at the thread recently discussed in the TW files forum and at my comment about the values of the 150k to grnd after stage two.. then look at that layout in the TW files forum.. compare to your schematic and lets look at that stage for a minute and think about it..

Ken's first pools have the 150k there and that resistor got lowered to taste of the player later on. In my amp it's 56k now and that gives a lot more range to arrive at drive settings and makes that transition very wide compared to the 150k. The amp still sounds INSANE at full drive 6 to 10 on volume.

The fizzy thing only makes that problem worse.. it makes it just a little terrible... I think we are approaching resolution.

[Reeltarded]

This thread about liverpool mods...

http://ampgarage.com/forum/viewtopic.php?f=3&t=30226

R7 on schmatic marked 33-150k. 150k is terrible sounding at volume barely cracked.. barely at 1/2 of 1.. 56k sounds excellent and clean to dirty by bashing double stops on a strat.

[arjepsen]

Thanks. But that resistor isn't in the Komet circuit.
There is the 220K/330K divider there though.
However, I previously messed with it - and put a 500k pot in there instead - unfortunately it only made a difference in where the gain knob had to be, before the pi distorts.
Hmm.. wonder if it would make a difference if I just changed the 330k resistor to something lower...?

[Reeltarded]

Yep.. and on a hunch I am thinking about 30% so as close to 110k as you have around.

I looked at the schematic just how, that is the similar topology and that is where to focus. Not the pair, only the 330k.

My prediction is:

You are about to hear a beautiful clean sound at actual volume that turns into hellfire when the amp is cranked.



My amp went from 150 to 56k in its place. Its amazing. Similar thing on your's.

[arjepsen]

ok, tried it out: paralleled the 330k with a 180k for about 116k.
Left the 220k as is.
It's the same as described before - the PI distorts the same, just at a different setting on the gain knob. The buzzy distortion is still there.
(btw, I've taken the ppimv out for now - also the cut control. Also I've got 100k grid stoppers on the power tubes, 100k on the pi, 22k after the 220k/330k divider (now 220k/116k), and 39k on the second stage. Also I've lowered the coupling cap before the PI to 4,7 nF)
I still believe the buzzy distortion happens in the PI, based on my testings with the headphone amp, 'cause the preamp stays clean all the way up. The thing I'm not understanding is why the PI seems to have a very short distance from sounding clean to fully distorted.