Loose Socket Pin?

[David Root]

Phase inverter, conventional long tail design w/ 82K and 100K plate resistors. ~425V B+, 225V on the 100K plate, ~420V on the 82K input side. So essentially no voltage drop across the 82K. 9V on the input grid, 16V on the other side. 26V on the cathodes. Design is about 250V plates, 50V cathodes. This is also causing problems downstream as you might expect. Checked wiring, all seems OK.

This seems consistent with no connection at the socket No. 1 pin. It has opened up a bit, so I tried to move it around a bit, but even so I really think it should be making contact. What's the best way to make it work right? I should add this is a new build, not an old amp, and the preamp tube sockets are top class US mfg. I paid $7.00 each for, not cheapo or used stuff.

[dobbhill]

Double check the Cathode end of that triode. I once had a bad solder joint across the cathodes of a PI tube.
Good luck.
D

[David Root]

Thanx. That would do it all right. I'll check it out.

[paddy]

Hi David

I once had a similar problem which turned out to be a dry joint
on one of the heater pins.Only one of the heaters in the 12AX7
was operating,therefore only one triode conducting,causing the
high plate voltage.good luck

cheers
paddy

[David Root]

Thanx Paddy. I just spent five hours on it. Short story, I disconnected the input grid from the preceding circuit. Presto, PI now works more or less as designed.

So I have to find what's causing that in the preceding circuit. No leaky caps, checked that. In order back from the PI, there is a full FX loop w/one 12AT7 and a bypass switch to completely remove it when not needed, then an LC tonestack (Steve Bench's design), then an anode mixer preceded by a single clean stage and a triple stage distortion preamp in parallel between the input jack and the final anode mixer stage.

Then I find that in play mode I have 430VDC on the power tube side of the .047 coupling caps, also on the #3 input and wiper pin of each post pi mv pot (using the dual 250K audio taper pot arrangement called the Rich mod) on up to the power tube grids. Have to check that too but I think I wired the pot right.

Good thing I put all used tubes in it to test out!

[paddy]

hi david

sounds like you've got your hands full!
is it possible that you have wired B+ to the junction of the
power tube grid leak resistors which should go to ground?
this would account for the high voltage reading on the coupling
caps.

cheers
paddy

[David Root]

In the Rich 2 dual pot MV, the grid resistor value is set by the 250K pot in parallel (on the wiper and pot pin 1, where the bias voltage also comes in) with a, in my case 180K resistor, to make a ~100K grid resistor equivalent value. I measure 102K across wiper and pin 1. This arrangement replaces the actual grid resistors.

I can only assume this HT DC is coming from the plate or screen supply. The power tubes are old (original) Tungsol 6550s, labeled RCA, but obviously Tungsols from the glass shape, and test OK, although the two side getters are pretty much dead, the third getter is still OK. Is this a problem with these old tubes?

[paddy]

Hi David
sorry i can't answer the 6550 question but with regards to the
high voltage on the coupling caps i can only assume that you must
have a wiring mistake(ie supply voltage wired to the mv pot).
you could try disconnecting the coupling caps and pull both power
tubes and see if you still have high voltage at the i/p to the output tube grids.

cheers
paddy

[David Root]

It was a wiring problem. I had connected power tube cathodes to pin 1, and the other end of the grid resistor from pin 8 to pin 5, instead of from pin 1 to pin 5. That explains why I could not get any idle current out of the cathodes. The post pi mv also seems to work right too now, and no longer interferes with the bias or the grid resistor equivalent values.

"The REAL problem is usually right in front of the amplifier."

So now I think the amp is OK from the PI thru to speakers, as I get a thunk in the speakers when I touch the PI 82K plate to ground.

However, there is still the problem with the input to the PI. When connected, it seems to act as if there is about 145VDC entering and pushing the 82K plate from 279V to 424V, and depressing the other plate to 225V from 265V. There is no DC on the connecting resistor, which comes from the pre-pi mv, and I get no AC signal there either when I plug in a guitar.

Coincidentally the plate voltage on the FX loop recovery triode is 146V, and it is right ahead of the pre pi mv, but the connected .047 cap is working and not leaking.

Also in the buffer stage of the FX loop I am getting the right DC voltage on the junction of the 220K and 33K resistors between plate and ground., about 45V, but only 7.6V on the cathode which is supposed to be 45V too, sitting on top of a 10K to ground.

This is easily the most complex amp I've built, so I guess I should expect to have "prototype" situations!

[paddy]

Hi David

do you have a schematic for this build?

cheers
paddy

[jelle]

Hi David,

I like the Larry quote too! He knows his stuff very well!

Please post a schem with the voltages that you have measured. I'm sure one of us will fix it.

Jelle

[David Root]

The FX Loop is right out of TUT 1, page 7-11, Best All-tube Effects Loop. I can't upload anything as I don't have a scanner. It uses one 12AT7. Perhaps one of you who is less computer challeged than myself could put it up.

I have added a loop bypass DPDT switch which puts a straight wire from right in front of the 1M on the buffer triode grid to immediately following the .047 signal coupling cap from the recovery triode.

Another wrinkle is the B+ on the tube runs about 340-350V when the loop is switched in, and 310 or so when switched out. The feed is a parallel one, i.e. a separate line off the main sequential PS line at the PI B+ node. This is to get 340V or so on the 12AT7, as the rest of the line is over 400V.

I'll be taking more measurements tomorrow, I have to go to Phoenix today.

[David Root]

Nothing wrong with my wiring, I just figured the interfering DC was coming from the wrong direction. I didn't put a cap directly on the PI input grid because there is one in the FX loop recovery stage that I thought would do that. For some reason it didn't. I put a .022 on the PI input grid and that fixed it. The input grid DC was leaking back, not the FX loop recovery stage leaking forward. There was no DC on the grid feed from the pre pi mv.

Never designed with 6550s before, they seem to pull down the B+ some!. At -52V bias, 66 mA cathode current, I'm down to about 405V plate, 396V screen in idle. Using SS rectification with 320-0-320VAC PT secondary. If my math is right that is about 63% Pa, which is on the cold side.

[jelle]

Hi,

Glad you fixed it! I like those 6550 a lot! Have fun,

Jelle

[David Root]

Thanx Jelle! I ran it y'day. Where do you run your 6550s biaswise? Nice clean tone from the clean channel, needs a little extra attenuation to stay clean dimed. Distortion channel needs to be properly set, probably just a matter of finding the right gain pot settings. I have gain pots driving stages 2, 3 and 4, plus a pre and post-PI mv, so that may take a while. Sounds really nice with a Strat using a Fralin Vintage + set.

I borrowed a friend's vintage Marshall halfstack ('77 Greenbacks, original cones!) to help set it up, and he is going to "consult" on final tonecap values.

I think I will have to fiddle with the anode mixer tube circuit as I'm only getting 0.55mA current on each side and would like to have nearer 1.0 mA as this is the final stage of both channels feeding into the tonestack. I'll also take the Ik up to about 72mA, this would be about 70% Pa at 405 plate.

After I get it optimized, I'll post some pics.

Then I think I'll do something relatively simple next time, for a rest! Maybe a harp amp with switchable Gibson/Fender circuits!