Need some help with debugging an amp

[G._Hoffman]

This is the schematic I built it from (click for a bigger image):




With some of the typical tweaking after hearing it.  The cathode LEDs have been replaced by resistors (relatively high - 1k5 to 2k2 or so - I don't remember the exact values), and most of the bypass and coupling caps have gone up in value as the amp was too bright.  (It is still a bright amp, but it doesn't kill your ears anymore!)  If it helps, there are gut shots HERE.

The problem I've got is it distorts very easily when the volume is below about 9 o'clock.  Right now, it is being used as a test amp for our customers at the shop (our normal amp is on my bench at home, and I've not gotten around to fixing it yet), and it is frequently louder than we would like when it is playing clean (which is, really, what we need it to do).  It is slightly better with single coils than humbuckers, but if the volume control goes too low, the amp starts to break up, and not in a pleasant way.

Any idea what could be causing this?


Gabriel

[DummyLoad]

C27 of the PI needs to be grounded. IOW. one end of cap C27 connects to g1 and other end goes to ground.

--DL

[tubenit]

C27 of the PI needs to be grounded. IOW. one end of cap C27 connects to g1 and other end goes to ground.

Agreed.

Gabriel, you understand this much much better than I do, so I will simply offer what I've found from experimenting alot.

Are you committed to the tone controls as they are?  You can certainly change the tone with a different TS.

I quit using 22uf cathode caps on 12A_7 tubes. They sound muddy to me. I typically use 5uf or 2.2uf.  Maybe try a 5uf there?
AND then try a .001 across the plate of the V2a triode. My guess is that you will hear a clearer and smoother tone? 

I also would try using a much lower cap value on C23.  I'd try something closer to a .047 than a .630 (unless that was a typo on the value).  I realize the series values of the caps lowers the actual value prior to the LTPI.

Also on the cathode caps on the power tubes, I understand that the 100uf increases bass, but tonally ...... it seems to
increase the brightness or crispness also to my ears when I've tried it.  I am using lower values more on power tube cathode caps which leave a looser bottom end that sounds more compressed and less "crisp/bright" to my ears.  So maybe experiment with that. You could simply lift the ground on those caps and see if the amp sound smoother to you and less crisp/bright. If so, then try 10uf, 16 uf or 22uf?   Hope that makes sense.

With respect, Tubenit

[RicharD]

C27 of the PI needs to be grounded. IOW. one end of cap C27 connects to g1 and other end goes to ground.

--DL

That's that Stout tilted PI thang.

[G._Hoffman]

Actually, I'm very happy with the tone of the amp, as long as it is about 9:00.  Even that wouldn't be an issue if I didn't have the amp at the shop for people to try out their guitars - in my own use, it will be much louder than that, all the time.  To me, it sounds best in the 11:00 - 1:00 range, or at least that is where it does what I built this amp to do (nice clean sounds).  But a lot of our customers are pretty young, inexperienced guys, and either they don't want to play very loud because they are nervous, or we don't want them to play very loud because, well, they are less than melodious to listen to.  So I would very much like to figure out why it is distorting at lower volumes.  C27 IS grounded, just through R17.  Perfectly normal thing to do, and I very much doubt it is what is causing the distortion at low volumes. 

Oh, and it doesn't distort like that if you just turn down the guitar's volume, so I'm assuming it is something to do with the load put on the grid of the second half of V2.  Does that make sense?

Oh, and yes, I am committed to those tone controls, at least on this amp, because they sound good, and frankly I HATE the way most tone controls are interactive.  On this one, each control works the same way, regardless of what the other tone controls are doing.  It takes a few more parts, but it works the way I think an amp should work.  Honestly, the ONLY thing I'm not happy about with this amp, right now, is the distortion at low volumes.  (I love the distortion at high volumes - this thing is tight and kicking when it gets loud!)  I didn't want a real "traditional" amp, which is why I didn't build one.


Gabriel

[PRR]

> it distorts very easily when the volume is below about 9 o'clock

Then the problem is before the volume control.

What is before the volume control?

TWO stages of unspecified tube amp. Minimal loss between. First-stage is plate-mixer.

Plate-mixer is a pretty bad idea. It distorts awful easy.

Ass-ume 12AX7 though here.

Plate-mixer gain is maybe 25 and peak output may be as low as 10V. Then max clean input is 10V/25 or 0.4V peak.

Two-stage gain is maybe 25 * 0.5 * 50 or 600.

Ass-ume V2a can make 60V peak output. Then max clean input is 60V/600 or 0.1V peak.

Most gitar amps will take 0.5V peak. 2V peak is not unknown.

Input headroom is slim to negative for most guitar levels.

Looking past the volume pot: V2b has grid at ground and huge cathode resistor. Cathode won't be any higher than 3V, current is 3V/100K or 0.030mA. That's way low. You could maybe run the tube that thin if going only to the longtail, but there's that Middle network with 100K to drive.

Max output from this V2b stage may be well under 2V. I'll take a wild guess (EL84?) and conjecture that V2b is just-barely smacking the longtail well.

If the longtail drives with less than 1V, then with gain of 600 in front of that the full-up input sensitivity is 1V/600= 1.6mV. That's WAY into mondo-shredder territory. That's for testing the neighborhood's tolerance, not for testing guitars. No wonder you have to run the Volume "below about 9 o'clock". Sensible sensitivity for a non-thrasher amp would be 10mV-30mV, not under 2mV.

The Bass knob only cuts, and not much at that.

Treble knob, when VOL is midway, can boost or cut quite seriously.

The Middle knob tries to be a Bridged-Tee, but without checking, my guess is that it is more a bass-cut.

What does SW3 do?

[G._Hoffman]

> it distorts very easily when the volume is below about 9 o'clock

Then the problem is before the volume control.

What is before the volume control?

TWO stages of unspecified tube amp. Minimal loss between. First-stage is plate-mixer.

Plate-mixer is a pretty bad idea. It distorts awful easy.

Ass-ume 12AX7 though here.

Plate-mixer gain is maybe 25 and peak output may be as low as 10V. Then max clean input is 10V/25 or 0.4V peak.

Two-stage gain is maybe 25 * 0.5 * 50 or 600.

Ass-ume V2a can make 60V peak output. Then max clean input is 60V/600 or 0.1V peak.

Most gitar amps will take 0.5V peak. 2V peak is not unknown.

Input headroom is slim to negative for most guitar levels.

Looking past the volume pot: V2b has grid at ground and huge cathode resistor. Cathode won't be any higher than 3V, current is 3V/100K or 0.030mA. That's way low. You could maybe run the tube that thin if going only to the longtail, but there's that Middle network with 100K to drive.

Max output from this V2b stage may be well under 2V. I'll take a wild guess (EL84?) and conjecture that V2b is just-barely smacking the longtail well.

If the longtail drives with less than 1V, then with gain of 600 in front of that the full-up input sensitivity is 1V/600= 1.6mV. That's WAY into mondo-shredder territory. That's for testing the neighborhood's tolerance, not for testing guitars. No wonder you have to run the Volume "below about 9 o'clock". Sensible sensitivity for a non-thrasher amp would be 10mV-30mV, not under 2mV.

The Bass knob only cuts, and not much at that.

Treble knob, when VOL is midway, can boost or cut quite seriously.

The Middle knob tries to be a Bridged-Tee, but without checking, my guess is that it is more a bass-cut.

What does SW3 do?

Yes, 12AX7 and EL84s.  SW3 is Local Negative feedback - makes a big difference, and I'm not sure I like it, but that's why I made it switchable.  The tone controls work very well, though you are quite right about the treble control being perhaps too effective. 

I get the too much voltage bit, but then why does it clean up above 9:00?  From there to about 12:00-2:00 it is a very clean amp, and as I said, if I didn't need it for the shop right now I would be happy there, since it is plenty loud at those levels.  The values for the voltage divider after the plate mixer on the schematic was a starting value, and R38 is now 470k.  (Sorry, I forgot I had made that change.  It's hard to keep track of all these things, you know!)  Should I maybe bring that down to a 220k?  It just seems so odd to me that it would be distorting at lower volumes, but then not at middle volumes.  I expect it to get pretty crunch at higher volumes, so that's of course cool.


Gabriel

[PRR]

> why does it clean up above 9:00?

Below 9' the input distorts first; above 9' the output distorts first.

You really want less gain before the volume control, more after.

[G._Hoffman]

> why does it clean up above 9:00?

Below 9' the input distorts first; above 9' the output distorts first.

You really want less gain before the volume control, more after.

So, I'm getting grid clipping, then?  Huh.  It sounds kind of gross.  I'll tweak down that first voltage divider, then, and see how that works.


Gabriel

[LooseChange]

If you are getting grid clipping, enlarge the grid stop. Be sure its on the tube socket.