Severe clipping in Ampeg V4 head later distortion model

[Redwood]

Before posting here I spent considerable time at the ampegv4.com forum but found that the vast majority of the threads there deal with earlier versions of this amp that have notable differences in architecture, up to and including completely different preamp tube selection.  Most notably, I have no 6K11 but do have the 6AN8 used in the PI section. 

Today I found this thread here:

http://www.el34world.com/Forum/index.php?topic=8920.0

And though it was not about the exact same amp I'm working on the schematic is much more similar.  My issues are not identical but certainly along the same vein and I am very impressed by the diagnosis that this forum was able to help the user through.  I'm hoping I can get anywhere near that level of help.

So what I have is a distortion model with the 300725-1 PT and 320822-1 OT.  The schematic that most resembles my circuit is 591761-1, which can be found here:

http://www.ampegv4.com/images/VT22_Schem2.jpg

The amp came to be completely dead, having just come out of storage for a considerable period, but not before the owner attempted to use it without bringing it up on a variac or any other preventive measures.  It almost immediately fried an output tube and when I opened it up I found a plate resistor on one 7027 had not only come unsoldered but got so hot it destroyed the socket pin to which it was attached.

I have since done a complete cap job and replaced all tubes in the amp.  I had a couple of bugs initially that I have mostly worked out but at this point I still have wonky voltages at a few preamp tube check points.  The amp does put out a signal and it's not bad at lower volumes with gentle pick attack but as soon as I hit the strings very hard or try to goose the volume very much past two I get really bad distortion that I'd characterize as clipping.

I have been beating my head against this thing for weeks now.  At this point I have replaced virtually every capacitor in the amp as well as a significant number of resistors.  I tried replacing the PEC in the tone stack with discrete components.  I have verified to the best of my ability the values of every discrete component on both circuit boards (though I only have a DMM, no analog or VTVM).  I do have a scope but I'm not very experienced with it.  I've tried applying a 1000 Hz signal and it looks OK at the grids of the first couple stages but it's totally whacked at the output.  I'm not really sure how to set the scope to look at it many points in between. 

I must admit I'm a little mystified by the use of 12DW7s in this amp.  I understand the first section in these tubes is similar to a 12AU7 and the second section like a 12AX7 but I don't really get what that first section is doing.  I know that at least some of the 12AU7 halves are cathode followers so the grid voltages aren't necessarily going to be zero but frankly I'm not getting perfect zeros anywhere, not even at V1 pin 7 closest to the input jack for channel 1.  They are not marked on the schematic so I have no idea what most of the grid voltages are supposed to be.

I suppose my next step is going to be lifting the grid side of every coupling cap in the preamp but I'd sure like to avoid having to do that if anybody has any other suggestions. 

For what it's worth I have tried running the preamp out to another amp and it sounds like crap.  I also tried running another preamp in and it wasn't any better.  It looks to me like neither of those approaches eliminates the 6AN8. 

All the rails voltages appear in the ballpark and I don't see anything funny happening with the output tubes, my bias voltage at point F is -32 compared to the -28.5 called for, and I even check the outputs with a Ruby BiasMaster, which showed the currents not dead on but all between 39 and 42mA, so I really think it's a problem somewhere in the preamp or PI sections.

V1      V2      V3      V4      V5      V201       12DW7      12AX7      6CG7      12DW7      12DW7      6AN8 Pin 1      140      180      95      195      300      236 Pin 2      4.9      0.001      0.002      43      86      66 Pin 3      10.1      1.1      3.6      65      175      87 Pin 4                                     Pin 5                                     Pin 6      160      220      196      220      112      150 Pin 7      0.001      0.001      0.012      0.008      .005      165 Pin 8      1.3      1.7      4.1      1.3      0.67      .018 Pin 9                                    3.6

I did not include heater voltages since they are all right around 6VAC.

Sorry this table looks jinky.  I attached it as a spreadsheet file wherein I colored particularly odd values in red, along with a second table pointing to what the values area supposed to be based on the schematic.

[PRR]

> I'm not getting perfect zeros anywhere, not even at V1 pin 7 closest to the input jack for channel 1.

In most circuits, Grid is "near zero", which can really mean "much less than cathode voltage".

You have 1.3V on cathode, which is not too different from 1.6V specced. You want to see grid 10 or 100 times lower, 0.1V or 0.01V. However if grid resistor is over 1Meg, a little offset is normal. And Ampeg used 5Meg here. However your meter shunts this, and if there is no plug then the grid is near-shorted (1K).

0.001V is zero for all practical purpose. Half-volt is not a big deal, though you might try another tube to see if the first one has unusual grid leakage.

Several tubes here *are* biased-up to non-zero voltages. V4 123 should have cathode near 45V. The grid id surely near +44V when not looking, BUT applying meter will pull-down the grid severely. If cathode voltage is good, don't fret the grid.

> table looks jinky

Agree.

Here is a nifty tool for noting tube pin voltages:  http://www.el34world.com/charts/valve/ValveData.php

However if your concern is the 6AN8, leave out the "probably close enuff" tubes and just show that. (I been behind a plow and am too tired to parse those un-columns.) I'm unclear about V4 123, list those clearly also.

> preamp out to another amp and it sounds like crap.  I also tried running another preamp in and it wasn't any better.  It looks to me like neither of those approaches eliminates the 6AN8.

Pre-out does not go through the 6AN8.

[Redwood]

Thanks very much for your response, PRR. I filled out the ValveData table you linked to and it can be seen here:  http://www.el34world.com/charts/valve/ValveData.php?e=view&f=14528

There were not options for most of my actual tube types so I made notes to that affect.  The 6AN8 in particular has a value for pin 9 which isn't even available in the tool but I listed all those values in a separate note.

You hit the nail on the head regarding V4 123.  While pin 1 is OK, the grid and cathode on that particular circuit probably represent the biggest anomaly I've got.  V4 pin 2 should be 95V but I'm only seeing 43V, whereas V4 pin 3 should only be 45V but I'm getting 65V.  This is probably the only one that is very obviously a cathode follower but as I said my understanding of this type of circuit is too weak for me to divine what this imbalance means.

I should perhaps also mention that if you look at that part of the circuit, where the plate is connected to V5B grid through C14 and C15, there is a line labeled H that drops down to the reverb pot and to the other ground rail supposedly through C22 - 100 pF ... but for the life of me I cannot find that particular cap anywhere.  I see where it should be near the junction of R35, C14, and C15, but it's not there nor does the board appear to have a place for it.  I don't know if that's significant or not but figured I should mention it.

[PRR]

Sorry to overlook this.

> V4 pin 2 should be 95V but I'm only seeing 43V, whereas V4 pin 3 should only be 45V but I'm getting 65V.

The drawing is wrong. You can not measure V4 pin 2 and get a right number, the meter loads this point. A correct number at pin 3 (Cathode) would be good enough. Yes, it says "45V"(?) but I calculate 58V at cathode pin 3. 65V is not obviously wrong, especially if todays wall voltages (and B+) are higher than back-when.

> line labeled H that drops ... supposedly through C22 - 100 pF

With an asterisk. The note says "some models". I can't make out which, but it isn't too important. 100pFd at point H (pretty-much a 150K impedance node) rolls-off highs above 11KHz. This is twice the top of the guitar band. "No-effect" on sound. Perhaps to kill some supersonic oscillation in some variants of this too-complicated beast. And when an amp is oscillating supersonics, it can't do full justice to the intended audio. If you don't have it, presumably Ampeg thought this build didn't need it.

Though, if you have that point in sight, you might try 100p-500p here. Just in case your build has whatever problem that caused Ampeg to slag that point in some amps. 500p is probably "too much", will shave the top-top harmonics which are a reason we might pick Ampeg over Gibson. But if otherwise it makes a big difference, you might try various values 100p to 300pFd.