Need help diagnosing blocking distortion

[lgehrig4]

Hello!

I built a 2204 with the intention of applying the SIR #36 mod. The 2204 worked perfectly so I know I was starting from a good point. Not all of the schematics/layouts were exactly the same but in the end I went with this one. Only difference is that on V1b I'm using .68/2.7k and for the extra tube (V0) I'm using .22/2.7K, so same values but on different tubes. I also tried the cold clipper on V0 and initially thought that was the reason for the blocking distortion.

Levi-spec.png

Originally I had all 33K grid resistors going to the preamps but in diagnosing this I added another 33K in series for V1 pin 2 and V0 pin 7. I also decreased the grid bias resistors from 220K to 150K. This I got from Aiken's site. He has other potental solutions listed but I don't want to deviate far from the design. All the blocking distortion issues are happening on the preamp gain pot (one next to master).

https://www.aikenamps.com/index.php/wha ... distortion

Here are the preamp voltages

Pin...........V1...........V0...........V2...........PI
1.........278........N/A........188.......253
2............0.........N/A..........0.........27
3..........3.3.........N/A........1.1........44
6.........233........230.........332......254
7............0...........0.........183........29
8..........2.1.........1.9........185........44

IMG_3734.jpg

IMG_3735.jpg

[lgehrig4]

I just changed the grid stoppers on pins of of he power tubes from 5.6K to 56K and still no dice. I am at a loss.

The amp works better, less cutting off but not the volume or girth I'd expect, with the preamp gain set higher for some reason. I would expect that to be the opposite. I move this up to around 2-3:00 and as a chord was ringing out and the trailing chord sounds better but as soon as a struck a new chord.....splatt.

As you adjust the preamp gain around the lower settings there is more jumping in and out of the volume.

[Purgasound]

I replied in the thread over at gear page. Sounds like you're missing a grid leak somewhere in the preamp or a stage is being driven too hard.

56K grid stoppers on the output are massive!
Usually people go lower to 1.5K.

[lgehrig4]

I replied in the thread over at gear page. Sounds like you're missing a grid leak somewhere in the preamp or a stage is being driven too hard.

56K grid stoppers on the output are massive!
Usually people go lower to 1.5K.

Theey were 5.6K so having 56K is not the issue. From the Aiken site

"Depending upon the type of output tube, you can probably go up to around 50K - 100K max before the high frequency loss becomes too noticeable."

I can't draw a schematic. It would be easier for me to fix on my own

I'm just looking for suggestions to look into. I started with a perfectly working 2204 and the mods are aere exactly what you see in the layout I posted.

If you were experiencing the same how would you go about troubleshooting? That is all that I am asking for. I tried the suggestions that I found so far but I'm not sure how to go about troublehooting the circuit. What to measure and what the voltages should be? I have a DMM but not a scope.

Thanks for replying to both!

[Purgasound]

If I was experiencing this and I needed to begin troubleshooting I would be looking at a schematic of the circuit to see where the issue could possibly be coming from. I know that's not what you want to hear but I sketch out everything I build or modify as it's the only way to keep track of the signal path. Following layouts is fine when building but it's easy to miss ground connections or skip vital components when doing so. When you follow a schematic you have a reference to where each component is connected in the circuit, regardless of its physical placement.
It might sound excessive but it doesn't take but a few minutes to draw something out. I use KiCad or sketch it out in a notebook.
From my own experience I would get a lot of wiring wrong when following layouts when I started out. I currently will still draw layouts for my builds but the schematic always comes first and when I'm wiring something new, I'm following the schematic and not so much the layout. I just use the layouts to see where things are going to fit and how much space I need when making a board.
The grid leaks are crucial for each triode stage as the grid needs a DC reference to ground to stay charged close to zero. This works with the cathode resistor to bias the triode in a gain stage. I say DC reference as some may not consider their voltage divider as part of a grid leak but it will act as one provided it does not have a capacitor after the fact. A capacitor tied directly to the grid will block DC in which the grid will have no DC reference to ground. In any case, a missing grid leak will bias the stage too high and the tube won't pass signal due to excessive blocking distortion. You can temporarily tack in a 470K or 1M resistor from grid to ground of each stage and see if the signal comes back to life.
I see you've likely got the wires mounted to the underside of the turrets. While this may be a cleaner look it does make things more difficult when assessing what-goes-where. That's why I asked about the 470K to ground before one of the stages. I can't see if there is a physical connection from this resistor to ground.

[Roe]

I replied in the thread over at gear page. Sounds like you're missing a grid leak somewhere in the preamp or a stage is being driven too hard.

56K grid stoppers on the output are massive!
Usually people go lower to 1.5K.

Theey were 5.6K so having 56K is not the issue. From the Aiken site

"Depending upon the type of output tube, you can probably go up to around 50K - 100K max before the high frequency loss becomes too noticeable."

I can't draw a schematic. It would be easier for me to fix on my own

I'm just looking for suggestions to look into. I started with a perfectly working 2204 and the mods are aere exactly what you see in the layout I posted.

If you were experiencing the same how would you go about troubleshooting? That is all that I am asking for. I tried the suggestions that I found so far but I'm not sure how to go about troublehooting the circuit. What to measure and what the voltages should be? I have a DMM but not a scope.

Thanks for replying to both!

56k is fine really. try 68+k on both grids of the phase inverter

[Purgasound]

Sure it's fine but isn't just throwing up grid stoppers only going to mask the problem?
I wouldn't imagine the treble tamed by a larger grid stopper would prevent blocking distortion normally caused by excessive current or too large of an incoming signal.

[lgehrig4]

If I was experiencing this and I needed to begin troubleshooting I would be looking at a schematic of the circuit to see where the issue could possibly be coming from. I know that's not what you want to hear but I sketch out everything I build or modify as it's the only way to keep track of the signal path. Following layouts is fine when building but it's easy to miss ground connections or skip vital components when doing so. When you follow a schematic you have a reference to where each component is connected in the circuit, regardless of its physical placement.
It might sound excessive but it doesn't take but a few minutes to draw something out. I use KiCad or sketch it out in a notebook.
From my own experience I would get a lot of wiring wrong when following layouts when I started out. I currently will still draw layouts for my builds but the schematic always comes first and when I'm wiring something new, I'm following the schematic and not so much the layout. I just use the layouts to see where things are going to fit and how much space I need when making a board.
The grid leaks are crucial for each triode stage as the grid needs a DC reference to ground to stay charged close to zero. This works with the cathode resistor to bias the triode in a gain stage. I say DC reference as some may not consider their voltage divider as part of a grid leak but it will act as one provided it does not have a capacitor after the fact. A capacitor tied directly to the grid will block DC in which the grid will have no DC reference to ground. In any case, a missing grid leak will bias the stage too high and the tube won't pass signal due to excessive blocking distortion. You can temporarily tack in a 470K or 1M resistor from grid to ground of each stage and see if the signal comes back to life.
I see you've likely got the wires mounted to the underside of the turrets. While this may be a cleaner look it does make things more difficult when assessing what-goes-where. That's why I asked about the 470K to ground before one of the stages. I can't see if there is a physical connection from this resistor to ground.

I don'rt doubt the benefits of sketching out the schem. I didn't realized you were asking to see if I missed something, which is why I kept referring to the layout. I have compared the layout with my amp multiple times so to me the layout is my amp. At least the preamp up to pi.

Thank you for bringing up the wires underneath. Life would be much easier if they were on top. Never again! Not to mention I worry about overheating every time I solder/desolder a component and the wiring slipping out. Or at least losing a good connection.

Between yours and other's responses both here and on GTP I have a lot to digest.

p.s. Regarding "You can temporarily tack in a 470K or 1M resistor from grid to ground of each stage and see if the signal comes back to life." do you mean removing the existing grid resistors and conneection this or clipping with the existing grid resistors? For example, on the first preamp in the layout (V0), should I leave the current 33K connected to the pot but also connect this to ground via 470K or 1M resistor?

[Purgasound]

For the sake of the exercise you can leave everything in place. If it were permanent it would make sense to attach a grid leak before the 33K grid stopper. Otherwise it creates a voltage divider but it would only be a roughly 7% drop in signal to use a 33K/470K divider. It wouldn't be a noticeable drop in gain but it would provide a 503K grid leak path. This is just to rule out a missing grid leak. If nothing changes then it could be some type of parasitic oscillation.

Another scenario I didn't consider earlier would be two out of phase signal wires coming in close proximity with each other. Each time the signal leaves the plate of a stage it is 180 degrees out of phase with the signal that entered it through the grid. If a downstream wire is close to the input wire which is really sensitive it can couple through the air and cause phase cancellation. This wouldn't really be blocking distortion but it would result in a loss of signal. The opposite would be true for two in phase wires, this is where you could have a source of parasitic oscillation as well. You can use a chopstick to move some wires around to see if that has any affect.

[cdemike]

Cool build!

With this unpleasant sound persisting despite fairly large grid stoppers that large all around, I'm wondering if this is blocking distortion versus an inaudible oscillation. I'm particularly concerned because of the layout of the first three gain stages being crammed so close together. I already don't really love that Marshall placed the master volume on that side of the tone stack, but having 3 volume controls near the input seems like a recipe for an amp that's very susceptible to oscillation. That's especially true with the fairly lengthy leads required to get from that auxiliary board to the added tube between V1 and V2. Reversing the order of "V1" and "V0" to match the layout you linked may solve some of that, but I'd lead toward laying out components in sequence so that the lowest gain stage is furthest to the left and therefore is better protected from interference from higher magnitude signals from later in the signal chain.

Does changing any setting make the issue better or worse?
Also, I'd be interested to have a complete set of voltage measurements as well as whether you can detect any AC voltage at any of the grids or speaker jacks with the amp idling.

I also tried the cold clipper on V0 and initially thought that was the reason for the blocking distortion.

Did the issue improve after this change? Blocking distortion occurs when gain stages start to draw grid current, which happens when the positive signal peaks exceed the value of that stage's bias voltage. Because warmly biased gain stages have less potential between the grid and cathode, they are more prone to blocking distortion. So in other words, a cold clipper in that position should be nearly immune to blocking distortion.

As an aside about the schematics, it might be easier if you're uncomfortable drawing a schematic from scratch to print out a 2204 schematic and made notes on what you changed. Even if you posted a picture of the annotated schematic like that, it'd be a lot easier to diagnose.

[lgehrig4]

Cool build!

With this unpleasant sound persisting despite fairly large grid stoppers that large all around, I'm wondering if this is blocking distortion versus an inaudible oscillation. I'm particularly concerned because of the layout of the first three gain stages being crammed so close together. I already don't really love that Marshall placed the master volume on that side of the tone stack, but having 3 volume controls near the input seems like a recipe for an amp that's very susceptible to oscillation. That's especially true with the fairly lengthy leads required to get from that auxiliary board to the added tube between V1 and V2. Reversing the order of "V1" and "V0" to match the layout you linked may solve some of that, but I'd lead toward laying out components in sequence so that the lowest gain stage is furthest to the left and therefore is better protected from interference from higher magnitude signals from later in the signal chain.

Does changing any setting make the issue better or worse?
Also, I'd be interested to have a complete set of voltage measurements as well as whether you can detect any AC voltage at any of the grids or speaker jacks with the amp idling.

I also tried the cold clipper on V0 and initially thought that was the reason for the blocking distortion.

Did the issue improve after this change? Blocking distortion occurs when gain stages start to draw grid current, which happens when the positive signal peaks exceed the value of that stage's bias voltage. Because warmly biased gain stages have less potential between the grid and cathode, they are more prone to blocking distortion. So in other words, a cold clipper in that position should be nearly immune to blocking distortion.

As an aside about the schematics, it might be easier if you're uncomfortable drawing a schematic from scratch to print out a 2204 schematic and made notes on what you changed. Even if you posted a picture of the annotated schematic like that, it'd be a lot easier to diagnose.

I also have a thread going at TGP and most seem to think it is oscillation too. In the end I may end up switching the sockets and placement of the input/gain2. Would leaving the heaters the way they are be an issue?

So far nothing I have done made any difference, including switching from the cold clipper. Last night I poked around with a chopstick, moving wires with the guitar plugged in, and no changes. One thing I have noticed is with the preamp gain almost full up the issue does get better. No more cutting in/out while playing or adjusting the pot in that range. I had it on the bench connected to a 12" speaker facing down so I still couldnt tell if all the volume and punch was there, but at least is wasn't dying out when I pluck a string.

At the suggestions of many I purchased a scope(Hantek DSO2D10), that just arrived today so I'll be focused on learning how to use this this before I get back to troubleshooting this mess.

Aside from AC at the jacks and and grids, I can tell you there is none at the input jack. That is my first test before I touch that amp after first firing up. There is a very low bit of DC on the input before I take the amp off standby (fraction of a volt) and that goes away off standby. No AC.

I'll check this with the grid but besides the voltage measurements in my first post, what else should I check?

[cdemike]

Would leaving the heaters the way they are be an issue?

Nope! Heaters can stay as they are unless you're getting 60hz hum. It's possible it'd be quieter if you redressed the heaters to make use of twisted pairs on tube I'm suggesting be reassigned to become V1 (assuming you keep the "V0" assignment; it'd be V2 if we rename the tube with earliest gain stage "V1" to avoid confusion), but heaters wouldn't be meaningfully contributing to this suspected oscillation.

One thing I have noticed is with the preamp gain almost full up the issue does get better. No more cutting in/out while playing or adjusting the pot in that range. I had it on the bench connected to a 12" speaker facing down so I still couldnt tell if all the volume and punch was there, but at least is wasn't dying out when I pluck a string.

This sounds more and more like an oscillation to me. Increasing the volume should make blocking distortion worse, not better, by increasing the signals' magnitudes at the grids, which would increase tubes' tendency to draw grid current at the positive peaks.

Aside from AC at the jacks and and grids, I can tell you there is none at the input jack. That is my first test before I touch that amp after first firing up. There is a very low bit of DC on the input before I take the amp off standby (fraction of a volt) and that goes away off standby. No AC.

I'll check this with the grid but besides the voltage measurements in my first post, what else should I check?

AC would be the culprit for an oscillation. With the input grounded, is there AC at any of the grids at different settings? The scope may be the best way to better diagnose this oscillation, but sometimes you get lucky with things like AC showing up somewhere in the signal path despite the input being grounded, which can tell you which gain stages are interacting to produce the oscillation. If you find AC without any input signal, you can send different points in the signal path to AC ground using a large film cap to see if you can disable the oscillation (100n or 220n usually work well for me; make sure you discharge the cap between moving it to different test locations it if you use the cap to send a point to AC ground that's not at DC ground such as the power tube grids or the phase inverter grids).