DIY 2204 negative DC voltage on grid of preamp tube

[jasonvilla]

‘Stable’ means that (almost) whatever conditions of input, load, control settings are applied, the amplifier exhibits no tendency to oscillate.

Got it, thank you. I can turn the preamp and master to 10 and it doesn't explode. It's a bit noisy all the way turned up, but it's a pretty small chassis for this circuit and the chassis is sitting on the ground.

[jasonvilla]

Do your other amps have 6V6 power tubes?

Yea pretty much haha. I think they all have JJ6v6s these days. One of the Marshall - ish amps I built has standard 50 watt iron in it, and I had el34s in there for a while. But it was just too damn loud. I don't love the sound of 6v6s being over driven, but you can get a bit of break up from them without my ears bleeding.

[jasonvilla]

This is the first time Ive done anything like this so be nice :)
Its hard to hear, and describe, but theres like some fizz when I play with the pre amp turned to ten. I just dont like how the distortion sounds, but I think it cleans up well with the guitar volume. This is with the 1.5k cathode resistor* Pre amp on 10, Master on 3. TMB on 8/5/8. P on 7. It was quite loud. Also, the cab doesnt have the chassis in it so the back is pretty open

[Willabe]

Do I see 2 or 4 sound absorbers, 2 on the wall, maybe 2 more on the floor, right side?

[jasonvilla]

Do I see 2 or 4 sound absorbers, 2 on the wall, maybe 2 more on the floor, right side?

Haha yea those are two DIY acoustic panels I made from mineral wool ( I have two on the other side of the room ) And the black thing on the bottom right is a DIY subwoofer. Actually, Ive built all my home theater speakers... they're a bit overkill 

[SILVERGUN]

You did a good job with the video and that is how I would expect it to sound.
Unfortunately, I don't have anything more hi tech than some headphones and I was unable to detect the issue.
It sounds like you have a good running amp and almost sounds as good as that YZ250.

Does the fizz go away if you play with the preamp gain down around 6-7? 

I would poke around more with the scope and see if you can see where the problem happens.

[jasonvilla]

You did a good job with the video and that is how I would expect it to sound.
Unfortunately, I don't have anything more hi tech than some headphones and I was unable to detect the issue.
It sounds like you have a good running amp and almost sounds as good as that YZ250.

Does the fizz go away if you play with the preamp gain down around 6-7? 

I would poke around more with the scope and see if you can see where the problem happens.

Dang, I was hoping that would help. I really need to find someone with an actual JCM800 to compare it to! Maybe this is what it's supposed to sound like 
The scratchiness, or fizz, is almost gone with the pre on 5.
I like the scope idea! How do I do that? :)

And damn, you caught me off guard with that YZ250 comment! I restored that a few years back, over the course of a year or two. Heres a before and after, just for fun :D

[SILVERGUN]

I like to keep things on the lighter side and I also like to show you that I'm paying attention.
In this case I'm guilty of leading you off topic so I apologize. That is a nice bike and project.
I would argue that it is relevant because it shows competence but since we're not in court it doesn't matter.

You can use the same sine wave input and just check every grid and plate throughout the amp to see what your distortion looks like. Make changes to your controls an see the results. You could then compare those results to what you see in your other amps. This would be a long process but might help you understand what you are hearing. You could take pictures of the scope traces at similar settings for reference.

Keep talking because it all helps. I think you'd agree that this stuff is cool to learn. We all started somewhere.

[jasonvilla]

Thanks again for your help :)

The scope is a lot of fun, but I dont know what Im looking for. I know what the sine wave looks like and what happens when it clips, but I dont know what that sounds like.  There is a math function on the scope that I can look at harmonics, which is cool to see even/odd order harmonics, but Im not sure how helpful that is in debugging the sound I hear. 

I took some pictures of the scope at different spots on the amp. Im not even sure if Im testing in appropriate spots. Do I need to test on the plate side? I assume my wave would just be elevated, right?  haha, again Im not sure!

I made a map of my test points :) Everything is to ground except across the PI. I kept the same exact settings I used in the video.

[jasonvilla]

2,3

[jasonvilla]

4,5

[jasonvilla]

6

[SILVERGUN]

I don't have much time today, but I can follow up later. You're doing fine. Great job with the map.
My suggestion was that you could take and log some measurements on this amp and then do the same thing with the same conditions on one of your other amps that doesn't have this issue.
By comparing the 2 waveforms you could possibly narrow down where your issue is coming from.

You might be hearing the effect of that spikey squared off wave at test point 4. See how it starts there and then stays that shape throughout?

If you've still got it on the bench add a test point at the grid of the cathode follower/plate of V2a.
Keep documenting like you're doing and it might become apparent to you.

Since you have a Ch2 input you could add a second test lead and overlay the grid signal vs. the plate signal. This can be useful for visualization.
Don't worry so much that you don't have a lot of experience with the scope. The only way to learn is to do it. It's the best way to see your sound.
 

[shooter]

since you're tinkering, like SG pointed out, the overshoot spikes can sound bad.  Measuring at the TS can be problematic.  A quick test, put another 470pF in parallel with the treble cap and see what the results are at point 4

[pdf64]

I suggest to also monitor the waveform at the DCCF, V2 pin8. As with the amp controls set up as they are, that’s the final clipping point. Whereas the waveform at ‘4’ will be sensitive to the exact settings of the treble and mid controls.

Also I’m wary of paying much attention to waveforms within a feedback loop (ie point 5); they can be confusing, look weird, as they’re a composite of the actual signal and the feedback generated error correction signal. And obviously the presence setting will also affect that.

[SILVERGUN]

I had a couple minutes and think I spotted an error.
It appears that the pic labeled #4 location does not match the mapped location for #4.

[mresistor]

SG I've been reading through this thread and wonder why you would ask about 6V6 equipped amps? 

[sluckey]

SG I've been reading through this thread and wonder why you would ask about 6V6 equipped amps?

Maybe you missed this...

     https://el34world.com/Forum/index.php?topic=26041.msg283529#msg283529

[SILVERGUN]

SG I've been reading through this thread and wonder why you would ask about 6V6 equipped amps? 

He mentioned having 6V6s in this amp but didn't say what the other 2204'ish amps had in there. IMO the 6V6 would be more likely to have a ratty sounding distortion than the EL34. Also, at some point I will recommend that he make a couple other changes. He already dropped the NFB resistor to 56K but it might have to go lower.
I'd also switch out those 220K grid leaks for 100K. sluckey handled all of this on the Plexi 6V6.
We're just not there yet and I didn't want to overwhelm him, but if he would have said the other amps have EL34s I would have pointed to that as a possible reason for what he is hearing.

[jasonvilla]

I don't have much time today, but I can follow up later. You're doing fine. Great job with the map.
My suggestion was that you could take and log some measurements on this amp and then do the same thing with the same conditions on one of your other amps that doesn't have this issue.
By comparing the 2 waveforms you could possibly narrow down where your issue is coming from.

You might be hearing the effect of that spikey squared off wave at test point 4. See how it starts there and then stays that shape throughout?

If you've still got it on the bench add a test point at the grid of the cathode follower/plate of V2a.
Keep documenting like you're doing and it might become apparent to you.

Since you have a Ch2 input you could add a second test lead and overlay the grid signal vs. the plate signal. This can be useful for visualization.
Don't worry so much that you don't have a lot of experience with the scope. The only way to learn is to do it. It's the best way to see your sound.
 

Thanks again for your help, for everyones help. I really appreciate it.
I understood your suggestion but I dont have an exact copy of this amp to compare it to. Anything close would be something that I built. I have a fender super sonic 22 that I could poke around at, but it doesn't have a CF and its set up a bit wonky compared to these "simpler" circuits. It's been very educational getting used to a scope and visualizing things. I havent taken the second lead out of the bag yet! haha Thanks again, this is the cathode follower/plate of V2a.

[jasonvilla]

since you're tinkering, like SG pointed out, the overshoot spikes can sound bad.  Measuring at the TS can be problematic.  A quick test, put another 470pF in parallel with the treble cap and see what the results are at point 4

I noticed that the tone controls really changes what the wave form looks like, which is pretty cool to see :)
This is TP4 alone and with a 500pf cap along the 500pf treble cap.

[jasonvilla]

I suggest to also monitor the waveform at the DCCF, V2 pin8. As with the amp controls set up as they are, that’s the final clipping point. Whereas the waveform at ‘4’ will be sensitive to the exact settings of the treble and mid controls.

Also I’m wary of paying much attention to waveforms within a feedback loop (ie point 5); they can be confusing, look weird, as they’re a composite of the actual signal and the feedback generated error correction signal. And obviously the presence setting will also affect that.

Thank you, I wasn't sure where to test from, but that makes sense. I was just a little hesitant to start measuring high voltage DC. This is V2 pin 8

[jasonvilla]

I had a couple minutes and think I spotted an error.
It appears that the pic labeled #4 location does not match the mapped location for #4.

I dont think it's wrong? It's the outside lug of the MV, coming from the treble pot.

[jasonvilla]

SG I've been reading through this thread and wonder why you would ask about 6V6 equipped amps? 

He mentioned having 6V6s in this amp but didn't say what the other 2204'ish amps had in there. IMO the 6V6 would be more likely to have a ratty sounding distortion than the EL34. Also, at some point I will recommend that he make a couple other changes. He already dropped the NFB resistor to 56K but it might have to go lower.
I'd also switch out those 220K grid leaks for 100K. sluckey handled all of this on the Plexi 6V6.
We're just not there yet and I didn't want to overwhelm him, but if he would have said the other amps have EL34s I would have pointed to that as a possible reason for what he is hearing.

I actually started with 100k grid leaks. Ive tried 100k, 150k, and 220k (which is what is in there now). Actually, sluckey's 6v6 Plexi is the only example of a 6v6 Marshall that uses 100k grid leaks, that ive seen. Everything else Ive seen has 220ks in that spot. Even Rob Robinette's 6V6 JCM800 - which this amp basically is. I've tried different values in the NFB as well, I did all this stuff before starting to post on here - ran out of ideas :)  And the reason I havent been focusing on the power amp section is because I hear what Im hearing at low levels and I didnt think the power amp wouldn't come in to play. BUT I could definitely be wrong about that. I have thought about running EL34s at this lower B+ to see how that sounded, but I wanted to address the pre amp first and not get off topic.


...Now I want to try EL34s in there 

[SILVERGUN]

I had a couple minutes and think I spotted an error.
It appears that the pic labeled #4 location does not match the mapped location for #4.

I dont think it's wrong? It's the outside lug of the MV, coming from the treble pot.

Ok, I couldn't see the pot in the pic. It looked like it was attached post PI.

[SILVERGUN]

SG I've been reading through this thread and wonder why you would ask about 6V6 equipped amps? 

He mentioned having 6V6s in this amp but didn't say what the other 2204'ish amps had in there. IMO the 6V6 would be more likely to have a ratty sounding distortion than the EL34. Also, at some point I will recommend that he make a couple other changes. He already dropped the NFB resistor to 56K but it might have to go lower.
I'd also switch out those 220K grid leaks for 100K. sluckey handled all of this on the Plexi 6V6.
We're just not there yet and I didn't want to overwhelm him, but if he would have said the other amps have EL34s I would have pointed to that as a possible reason for what he is hearing.

I actually started with 100k grid leaks. Ive tried 100k, 150k, and 220k (which is what is in there now). Actually, sluckey's 6v6 Plexi is the only example of a 6v6 Marshall that uses 100k grid leaks, that ive seen. Everything else Ive seen has 220ks in that spot. Even Rob Robinette's 6V6 JCM800 - which this amp basically is. I've tried different values in the NFB as well, I did all this stuff before starting to post on here - ran out of ideas :)  And the reason I havent been focusing on the power amp section is because I hear what Im hearing at low levels and I didnt think the power amp wouldn't come in to play. BUT I could definitely be wrong about that. I have thought about running EL34s at this lower B+ to see how that sounded, but I wanted to address the pre amp first and not get off topic.


...Now I want to try EL34s in there 

Good stuff. That devil horn signal ain't pretty.

Really good communication across the board.

[SILVERGUN]

I'm on the run today and the pics are too small on my phone. I can't keep up.
I'll catch up reading later but maybe the others will chime in with some opinions.


If you have time, see if you can dial out some of the spike in the tonestack and then plug back in and see how it sounds. See if dialing out the spike makes the fizz go away.

[shooter]

If you're still tinkering;
monitor TP 5, adjust your input signal until the sine wave just starts flat-toping, then measure your input signal n post the #'s

[jasonvilla]

Good stuff. That devil horn signal ain't pretty.

Oh man, but thats soooo anti rock n roll!!!! :D

[jasonvilla]

I'm on the run today and the pics are too small on my phone. I can't keep up.
I'll catch up reading later but maybe the others will chime in with some opinions.


If you have time, see if you can dial out some of the spike in the tonestack and then plug back in and see how it sounds. See if dialing out the spike makes the fizz go away.

Alrighty, at TP4 I was able to make a square wave (no more spike) by adjusting the TMB like so:4/7/1. Pre amp on 10.


I made a video of what this sounds like (it sounds like poo) But you can hear the sizzle/fizz crap a bit better on this recording .

[jasonvilla]

If you're still tinkering;
monitor TP 5, adjust your input signal until the sine wave just starts flat-toping, then measure your input signal n post the #'s

Forever tinkering!!

Im not sure if I did this right. I kept the same TMB settings as the test before. Left the same 136mv sine wave going in to the amp, but adjusted the pre and master volume to get the largest signal i could get that was just starting to flatten. Should I do this differently? Thank you

[SILVERGUN]

Yup, I could hear the sizzle/fizz now. It seems to be external noise, not due to too much drive. I would try a new tube in V1 if you haven't already.

Pic 1 - Curious about this cap. I don't think it was on your schematic. If it is where I think it is clip it out and report back.

Pic 2 - I looked over your board and there is some missing solder around one of the black mica cap leads. Can you confirm a solid solder connection there?

[SILVERGUN]

While you're in there check this one too

[jasonvilla]

Yup, I could hear the sizzle/fizz now. It seems to be external noise, not due to too much drive. I would try a new tube in V1 if you haven't already.

Pic 1 - Curious about this cap. I don't think it was on your schematic. If it is where I think it is clip it out and report back.

Pic 2 - I looked over your board and there is some missing solder around one of the black mica cap leads. Can you confirm a solid solder connection there?

Ive tried 3 tubes in V1 but Ill try again since we've changed some things around.

Pic 1 - nah, thats a 100pF cap that wasn't in there when I made the schematic. At first I had no cap, then a 68pF, then no cap, and now theres a 100 in there. It doesn't seem to do much.

Pic 2 - theres a butt load of solder in there! It's hard to fill up those big turrets with only one small lead going inside. Ive done the tap test through out the board listening for cold solder joints and just not I put the output of that on the scope and watched as I moved that cap around and didnt see anything.

[jasonvilla]

While you're in there check this one too

Took me a second to find that one! That was cap I had put in there to check that location, it's not in there anymore. I have the Sozo back in that spot. So that pic is now old :)

[SILVERGUN]

Pic 1 - nah, thats a 100pF cap that wasn't in there when I made the schematic. At first I had no cap, then a 68pF, then no cap, and now theres a 100 in there. It doesn't seem to do much.

Where does it connect? It seems to come off of V1A plate, but to where?
Why?
Did you try it without it?.....I mean to see if it makes a difference with the fizz

[jasonvilla]

Pic 1 - nah, thats a 100pF cap that wasn't in there when I made the schematic. At first I had no cap, then a 68pF, then no cap, and now theres a 100 in there. It doesn't seem to do much.

Where does it connect? It seems to come off of V1A plate, but to where?
Why?
Did you try it without it?

Its just a snubber cap from pin 6 to 8 - one of the many things ive tried to get rid of the noise. I can take it back out.
Yea I built it without it, and have had it in and out a few times - trying various things.

[SILVERGUN]

Its just a snubber cap from pin 6 to 8 - one of the many things ive tried to get rid of the noise. I can take it back out.
Yea I built it without it, and have had it in and out a few times - trying various things.

Ok, I needed to make sure it wasn't a different mod that I have seen where they go from plate to grid. When placed there if the cap shorts you wind up with B+ on your strings.

Take it out or leave it. I was worried it might have been associated with the fizz.

[jasonvilla]

Not sure what else to try. I think ill replace the pre volume pot and maybe a few of those black micas.

[SILVERGUN]

Not sure what else to try. I think ill replace the pre volume pot and maybe a few of those black micas.

The caps, maybe...the pot, not.  You could try strapping a 1M resistor across the outside lugs of the pot and see if that helps. And you could always try adding a grid stopper to V1b.

But, like pdf64 mentioned before, these are mods to what should be a working circuit.

There was another point of interest to me from the pics. The solder joint on the negative side of the preamp filter cap looks possibly cold. I would reflow that joint to be sure. Don't forget to use a heatsink on the lead.

[shooter]

Forever tinkering!!

the scope shows ~40Vac rms for clean drive to the 6V6.
that setup should give you nice chime-y clean sound.

It seems to be external noise

If there is something sneaking in, your scope "should" see it, should be quick, sharp spikes, intermittent somewhere "on" the sine wave.  Using a chopstick tapping n wiggling you should be able to get it to happen.  quick twist knobs also

[pdf64]

Randall Aiken mentioned recently that after numerous early failures, he’s never going to use silver mica caps ever again. And he will have been buying kosher brands from bonded parts wholesalers; it’s not like he would be sourcing stuff off ebay.
When you’re making high end amps for touring pros, a $1 cap can cause an urgent warranty claim costing $hundreds to get fixed.

[SILVERGUN]

While the iron is hot, here is a couple items to address.
The ground connection on this (I'm assuming presence) pot doesn't look all that great. I would try to get a more secure path to ground.
The other thing is that it is bad form to just float resistors off of tube socket pins. That higher wattage resistor is heavier than needed and would eventually snap off at the 90 degree bend.
It looks like there are a couple other instances of this throughout the amp.
A common trick is to use the socket hold down screw and attach a little terminal strip to give you a mounting point.

[shooter]

silver mica caps

+1
gave up on them a few years back, fail rate was way up compared to other caps

[jasonvilla]

I will try swapping out the 4 silver micas on the board for ceramic caps (all I have) to see if it makes a difference.

I will reflow the areas SGUN mentioned as well. Im not sure where I would put them now, but in future builds ill make sure not to hang resistors off of tube sockets. Thank you for the advice.

I just spent a good ten minutes looking at the scope, bending the leads on the silver micas, tapping every solder joint, and I didnt see anything move :(

[SILVERGUN]

Looking and thinking more at the pic above.
You definitely want to create some distance between the orange and yellow grid wires and the green and black wires that are going to the speaker out jack.


This is the pic that made me question the ground on the back of that presence pot.

[shooter]

I didnt see anything move :(

then you have no problem 

EDIT:
try this;
set your sig-source to 880hz, set all your amp knobs to 5, play through speaker

If it's clean, play with knobs,

[jasonvilla]

Well, I think it might have the silver micas ...I think the attic fan above my living room is causing noise as well.  I have been playing with more NFB and I think it sounds better with more. Still working on that.

I switched out the 4 micas for cheap ceramics AND I took out the 1.5k cathode resistor and put back the 820R. Interestingly  TP4 shows -4.5v instead of the -5.5 it saw with the 1.5k.  The 470k grid stopper is still in there because its hard to get to :)

I know it looks like the wires are on top of each other/touch but they're not. It just looks like it in the photos. They're as far apart as I can make them!

I have been reading through old threads and It seems like I could try out the EL34s and keep the 6.6k OT I have right now. And it should be ok with this lower than normal B+... Is that correct?  I have pins 1 and 8 tied to ground*

Thanks again everyone!

[pdf64]

The negative Vdc appearing at tp4 with heavy preamp overdrive is a total non issue in my view, so its magnitude is somewhat irrelevant.
If without a big grid stopper, the next stage’s bias is shifting excessively, such that sound is blocking out, then by all means reintroduce sufficient grid stoppering to avoid that.
Maybe the Vdc shift at tp4 is a good indicator of that, but the Vdc at the actual grid would seem a better point to measure/monitor  it?

[jasonvilla]

The negative Vdc appearing at tp4 with heavy preamp overdrive is a total non issue in my view, so its magnitude is somewhat irrelevant.
If without a big grid stopper, the next stage’s bias is shifting excessively, such that sound is blocking out, then by all means reintroduce sufficient grid stoppering to avoid that.
Maybe the Vdc shift at tp4 is a good indicator of that, but the Vdc at the actual grid would seem a better point to measure/monitor  it?

I’ll measure it at the tube socket tonight :)
I’m now wondering if I should reduce the power tube grid leaks to 100k and/or reduce the nfb resistor and/or change the PI cathode resistor from 470 to 820...