GA20 homebrew isn't quite right

[Phil_S]

KT66: Many thanks for these very incisive observations. I think you are pointing me in the right direction. I took a fresh set of readings this morning and did a bit of tinkering. See in between....voltages are today's readings.

Just for background, I think this is my 11th amp build, but my first PTP. I have learned a few things, but am far from expert. I can obviously use help here. This is not my first recycling job. I have built several amps from junque.

A couple of things stand out to me that don't add up. One of them is that you state that before you plugged in the preamp tubes that the voltage measurements at the 2nd and 3rd filter caps are 246 and 148 ; However, with no current draw there should be no voltage drop across the 47K power resistor.

I misstated. With no tubes, HT secondary is 625VAC. With the 5Y3 in socket, it drops to 618VAC. Output is 402 at cap#1, 399 at cap#2, and 390 at cap#3. I think this makes more sense and is in line with expectations.

Also, the PI has a common cathode and identical plate resistors so the plate voltages should be the same, but you recorded 151 and 41.7. I would suggest that there is some kind of short that is causing both the voltage drop across the PS resistor with no tubes in and the diminished plate voltage on pin 5 of the 6SL7.

Sorry, again, I botched the report, but just a bit and it doesn't change your comment. PI Readings:

Input side of 6SL7
Grid 45V (yes forty five, not point forty five)
Plate 78V
Cathode (common) 1.4V

Power tube side of 6SL7
Grid 0.8
Plate 150

Without the tube in the socket, plates voltage is the same for both sections. Same result with two different tubes, one used RCA, one NOS Sylvania.

I'll return to this below.

I also agree with Bob that dropping 100V when the 6V6 tubes are put in seems excessive. Did you measure the plate voltage on both tubes and are they about the same ? If you look at the schematic for the GA-20T, which has the same power section, there is 90V difference between the screen and plate, your amp has half that which might point toward the excessive 100V drop.

Va on both power tubes are one volt apart, 295, 296.
Screens (common) are 246v
Cathodes are 12.1
Grids are 0.44 (fed by the side with both the 220K grid leak + 4.7K) and 1.6 (fed by the side with only the 220K grid leak.

A word of caution also, the 5U4 draws 3 amps of current and if your tranny was designed for 5Y3 use (which uses 2) you might be in danger of toasting the 5V winding.

Yes, thanks for the reminder. I thought it wouldn't hurt to give it a try. Aside from dropping 20V less than the 5Y3, there is no difference. It is possible I've got a bad rectifier tube, but I don't think so. Otherwise, I'd have seen even higher B+ with the 5U4.

Sanity check on the rectifier? VAC=618 or 309-309. 5Y3GT should produce about 1.1x that number or about 340VDC? I'm seeing 300 with the circuit load. We agree this drop (see discussion on 5U4) is too much, but I don't think it is the tube. The output transformer show 56 ohms on each side and I think it is not shorted. Besides, it does pass a signal to the speaker...I realize this only means so much. I didn't test the OT for turns ratio. Perhaps I should?

I traced the circuit to the schematic early today. I did not find a wiring error. I will have to go back with a shop light and magnifier to look for a whisker, but am not expecting to find one. Assuming I do not find a whisker, I'm going to have to uncover the wiring error.

On the 6SL7, when I lift the .05 cap on the low voltage plate, the voltage rises to be equal to the other section. The cap meters OK as to value. I am using a Tripplett DMM for this and it does not measure leakage. I reworked the 220K+4.7K which were sitting on pin 6 of the 6V6. I used pin 6 as a tie point and it should be OK to do this. Moving it changed nothing. I suppose I should change the cap? Do you think a leaky cap would account for this?

The PI behaves the same way, with our without the 6SJ7 in it's socket.

Back to screen and plate difference. You say it should be 90V drop and I'm sure you have a firm basis for say this. I am seeing half that. To get 90V drop, we need screens to pull 4.5mA each across the 10K ohm dropping resistor under zero-signal. The tube data sheet says zero-signal screen current is 4mA for two tubes in AB at Va=315. The drop is see is 296-246 = 46, or 4.6mA screen current. I am wondering about whether I understand this correctly. I am not seeing where we will get a 90V drop. Is there something to pursue here?

Many thanks.

Phil

[Phil_S]

Well <staring down at shoes> it helps a great deal to use a 470K plate load resistor for the 6SJ7 and not a 470 ohm. Now, I've got real sound. Not good sound, but the volume seems appropriate. It is very buzzy, fuzzy. Sound cut out for a bit and I'll be checking for bad solders (Hi, Bob-I!). Plate voltage on the 6SJ7 is not right. I'm only measuring 5V, dropped from ~150 at the B+ node across 470K. That couldn't be right and I'll be looking at that.

I've got a couple of used 6SJ7's on order. I'd like to see if it's a dud.

I've got a nice filament tranny on order, too.

The plate voltages on the PI are still not symmetrical. One side is still 150V, but the other side is now 110V.

I am guessing there are two problems. One related to the 6SJ7 and one for the PI. It's progress, though.

[Structo]

Heheheh, those darn color codes strike again!

Sounds like you are making progress.

[Phil_S]

Heheheh, those darn color codes strike again!
Sounds like you are making progress.

Yup, got another that looked red-red-yellow, but it must be brown. I reworked the 220K plate resistors for the PI and replaced the tone pot. I also flipped the filament supplies and now have 5.1 on the 5Y3 and 5.6 on the 6.3v tubes. I ordered a filament tranny. The 6SJ7 has Va=4, yes, four volts, so I think it is shot and am awaiting another. The only thing that isn't looking right is the PI plates at 78 and 92. 92 is close to expected. Don't know about the 78. I'll see what happens when I get a working 6SJ7. I'll swap the 6SL7 later, too. It makes plenty of volume now. I'm in the home stretch.

[KT66]

Glad to hear you are getting it all figured out. My thinking on the screen / plate voltage discrepancy was going in the wrong direction, I believe your analysis to be correct.

I've done the wrong resistor thing myself, so I always give a quick measure before I solder one in. Kudos to badtweed for catching the heater
reversal.

I like paraphase amps, hope you can post some clips once you get it dialed in.

[Phil_S]

Glad to hear you are getting it all figured out. My thinking on the screen / plate voltage discrepancy was going in the wrong direction, I believe your analysis to be correct.

I took a good hard look at the GA20T voltage chart and that pointer proved to be quite helpful.

For the power tube plates and screens it the GA20T says Ep=340 and Escr = 270, a difference of 70V and I've got a difference of 50V. Plate voltage on my amp is also nearly 45V lower. I often doubt myself when I do the math, so I feel good that I was able to articulate the "argument" for this one. Nevertheless, what you said was helpful because it forced me to take a close look.

I've done the wrong resistor thing myself, so I always give a quick measure before I solder one in.

Yeah, but one has to read the meter properly. I meter every resistor. I wasn't watching the indicator for the K-ohms, even though I thought I was. I am really beyond help when I do this sort of thing.

Kudos to badtweed for catching the heater
reversal.

Yes, kudos to badtweed. Funny, in looking over the PT more closely, I made a pencil marking that the seller's indications were reversed. Pencil is the operative word, as it was quite faded and I only noticed it when I had the magnifying glass in my hand. Once again, I am my own enemy. Anyway, the winding is not 6.3v @ 2.5A like the seller said and I'm going to scab in a 6.3v@3A to cure the problem properly.

I like paraphase amps, hope you can post some clips once you get it dialed in.

I'll try to work up the nerve. I'm not exactly a competent guitar player and after 40+ years of noodling around, I don't actually have aspirations.

Again, thanks for looking in on me.

Phil