Asteroid unpleasing distortion

[Reeltarded]

Because the transition is very short on large input. It's not easy for me to post examples but I am going to demonstrate the similar thing on my Liverpool.

Play delicately and partial gliding mute with my picking hand then open and very slightly harder goes from what sounds like deep clean (but it isn't honestly clean, it is actually umm underdrive..) to brash with lots of transient, like classic chordal and double stop bluesy stuff..

Give me a few hours 'til the neighbors would appreciate an unpracticed version of Ray Charles something or other.

Umm..

change the PI stopper back to 10k
put a 100p fizz cap across the PI
I use 5k6 grid stops on my big tubes similar gain amps

I'll be Bach as much as I'll be damned.

[stephenl]

Would raising the voltage at the PI node help?

[arjepsen]

change the PI stopper back to 10k
put a 100p fizz cap across the PI
I use 5k6 grid stops on my big tubes similar gain amps

ok, changed the pi grid back to 10k. I put the 100k's on the power tubes just to be extra completely sure
Put the fizz cap in (120pf) still same buzzy distortion.

Hmm... wonder if there's any idea in putting a grid stopper on the second grid of the pi?

Play delicately and partial gliding mute with my picking hand then open and very slightly harder goes from what sounds like deep clean (but it isn't honestly clean, it is actually umm underdrive..) to brash with lots of transient, like classic chordal and double stop bluesy stuff..

Give me a few hours 'til the neighbors would appreciate an unpracticed version of Ray Charles something or other.

Yeah - I do get the touch responsiveness. My Rocket sounds awesome. And again - it's not overall bad - I can play some stuff that sounds nice.
The buzzy distortion get's most apparent on single notes that are allowed to ring out.
Maybe just try plucking the A string alone - that's the best test I do right now to check whether the buzziness is still there.

@ stephen: the pi plate voltages are up just past 300 - I would guess that should be plenty?

[thejaf]

I know you're getting lots of suggestions. Did you perchance try a lower value for the 1M paired grid loads in the PI yet? Simple would be to just temporary tack in another pair (say 560K) to bring down the value and determine if that alters PI breakup tone.

[arjepsen]

please, by all means, keep giving me suggestions!

I actually havent tried that yet - will see if I can do that one.

[arjepsen]

Latest update - put a 1m across each resistor, lowering them both to 500k.
It's different - might sound a bit... "crunchier" in a good way.
But I still hear the buzz.

Anyways, for fun I just hooked the scope up again, and made a single trigger on the guitar - this is the output i got.
Yellow is the guitar input (plucking the a string), light blue and purple are outputs of the PI, dark blue is output jack.
Wonder if it tells anything interesting?

DS1Z_QuickPrint5.png

[Reeltarded]

Hey.. why does the amplitude of the attack pass but the next peak crushes on the output...

Is this simply a bad capacitor.. this is drving me as crazy as it is you, trust me.

[romberg]

If you look at the first pulse from the input, it is just a single positive peak. But the PI is putting out two pulses. Almost like a short delay. If this is going on (not exactly sure it is or why) then the rest of the waveforms make sense. The next two input pulses are a negative/positive pair. And the output have three peaks. The second peak which slams the power tubes would be the second input peak + the "ghost" peak. Then a third ghost peak (the echo from input peak two).

I wonder if this is some kind of short duration oscillation caused by parts of the circuit unintentionally coupling together via capacitive coupling (wires running in parallel too close together etc)? It would be interesting to use the scope and probe backwards in the amp a bit and see when this extra peak shows up. You can see it on the PI out. But it is not there on the input. So, when does it make a grand entry?

[romberg]

In your gutshot, near V2 I see what looks like two 0.1uf caps. Both seem to be connected to what I assume is the gradual/fast switch. The schematic only shows one cap in the circuit. What is going on there?

[arjepsen]

Yeah, it seems like the top peaks gets amplified without clipping, but the bottom ones clip.
From the image, I’m guessing the output of the PI isn’t clipped, even though one side is slightly larger than the other (a little weird to me, since I could hear the distortion in the PI when I used the headphone amp).
It seems like one side of the powertubes clips.
I actually tried the same thing with a different guitar, and it was pretty much looking the same.
I also tried to make a regular sine wave with the same amplitude as the guitar signal - however that simply stays nice and even, and the resulting waveform out of the PI was only about half the ampitude of what the guitar signal produced... no clipping.
Will try out a bit more when I come home after work.

Regarding the .1 caps:
They are connected to the fast/gradual switch. One is straight from the plate to the switch, the other goes from the 82k/22k divider junction on the plate. At the moment I only have the one from the junction to the switch in there. I took the other one out of the equation earlier, just to make sure that that didn’t do anything bad. In other words, right now the amp is constantly in gradual mode until I hopefully get it fixed.

[arjepsen]

Made another scoping:

Yellow is guitar input, light blue is grid on second stage, purple is phase inverter input and dark blue is across ot seconds.

DS1Z_QuickPrint8.png

It seems that there's a spike that gets clipped by the powertubes, every time the guitar signal goes straight from negative to positive - I guess that makes sense - this is where the signal "moves" the furthest, so that gives the spike. I'm not sure why the guitar signal looks like it does though, but it looks similar with the other guitar I tried. (the second guitar is a split humbucker).
The scope measures the input to around 220 mv pp.
I tried to make a similar sine wave with that amplitude and this gave me this result:

DS1Z_QuickPrint9.png

From what I can tell, the guitar signal gives the biggest swing from negative to positive, but every swing from positive to negative sort of pauses in the middle.... this sort of explains why there's clipping on only one side.
But does this mean the guitar is doing something wrong? I'm not sure about that, since it looks very similar with my other guitar:

DS1Z_QuickPrint10.png

[thejaf]

Have you double-checked bias on the power tubes?

[thejaf]

Also, I just noticed prior post where you mentioned 6L6 power tubes. Make sure you're not vastly exceeding the grid load (resistance to ground/bias) for fixed bias operation of ~100K per the 6L6GC tube data sheet.

[Reeltarded]

Swap power grid stoppers to something much smaller under 6k8 even....

[romberg]

I think you have two separate issues. One is that your power section is clipping in a non-symmetric fashion. The negative side of the waveform clips hard while the positive is no where near clipping. I'd re-check your bias (as has been suggested). I would also try swapping the power tubes and PI tube with other known good ones. If you don't have another set of power tubes, try swapping the tube sockets and see if the clipping on the output swaps as well.

The other issue is that extra peak that is clearly showing up on the signal after stage one. This peak is not there as far as I can see on the input signal. This peak is being amplified and becomes very apparent in the output. I bet it does not sound musical at all. This must be entering the amp on your first gain stage.

Looking at your gut shot, I can see no cathode resistor on stage one. I think it may be hiding under the cathode bypass cap. But I'm not sure. Also, I think I may not like your grounding scheme around stage one. I could be wrong on this, but how is the cathode grounded on stage one? To me it looks like the resistor and cap are connected to a terminal and then a wire is run in a big loop all the way around the filter cap, tube and input jack. Is this correct? The cathode ground carries *all* of the tubes current and should connect to the filter caps negative side in as short of a run as possible. If it is not, then the input may be picking up feedback from this ground current. Describe your grounding on stage one.

Mike