NOS #102 Build

[dbharris]

Got the amp running late last night, but could only test at very low volume. I ended up having to replace the 4 shielded cable runs in the OD section. I removed each section of wire and replaced it one by one while tracing the signal with the scope. Each one passed no signal until replaced.

I tested each cable when removed from the amp and I don't really understand how they are bad. Each has continuity between signal ends and no continuity with the shield. Resistance between signal ends was never more than 0.2 ohms. I did not test for capacitance. I guess I'll chalk it up to user error on my first outing with this gepco cable.

Most likely going to keep the CTS pots in, at least for a couple of weeks.

Will take voltages, set the trigger and the PI balance today.

-Dan

[ijedouglas]

Got the amp running late last night, but could only test at very low volume. I ended up having to replace the 4 shielded cable runs in the OD section. I removed each section of wire and replaced it one by one while tracing the signal with the scope. Each one passed no signal until replaced.

I tested each cable when removed from the amp and I don't really understand how they are bad. Each has continuity between signal ends and no continuity with the shield. Resistance between signal ends was never more than 0.2 ohms. I did not test for capacitance. I guess I'll chalk it up to user error on my first outing with this gepco cable.

Most likely going to keep the CTS pots in, at least for a couple of weeks.

Will take voltages, set the trigger and the PI balance today.

-Dan

Does the Gepco cable have a foam insulation? I had the insulation melt on me when trying a few different types of coax. Both the Tandy RG-59 and Columbia Flexfoam 1389 that HAD used in 2nd/3rd gen amps have the foam insulation which melts if you look at it wrong

Congrats on finding the issue.

[dbharris]

Got the amp running late last night, but could only test at very low volume. I ended up having to replace the 4 shielded cable runs in the OD section. I removed each section of wire and replaced it one by one while tracing the signal with the scope. Each one passed no signal until replaced.

I tested each cable when removed from the amp and I don't really understand how they are bad. Each has continuity between signal ends and no continuity with the shield. Resistance between signal ends was never more than 0.2 ohms. I did not test for capacitance. I guess I'll chalk it up to user error on my first outing with this gepco cable.

Most likely going to keep the CTS pots in, at least for a couple of weeks.

Will take voltages, set the trigger and the PI balance today.

-Dan

Does the Gepco cable have a foam insulation? I had the insulation melt on me when trying a few different types of coax. Both the Tandy RG-59 and Columbia Flexfoam 1389 that HAD used in 2nd/3rd gen amps have the foam insulation which melts if you look at it wrong

Congrats on finding the issue.

Thanks Ian! It might be foam, there is a softer white inner insulator and then a harder thin black insulator wrapped around that protecting the signal wire. But this also has a double shield so you have to get it pretty hot when soldering that. Hoping these hold up because I don't want to have to run those wires again!

Burning in the amp now...to be continued

-Dan

[dbharris]

Played the amp a bit and it sounds good, maybe a little bright. Will keep burning in before I make any tweaks.

Switched all the glass. Quad of Sylvania 6L6GC with double side getter. Unfortunately one of the quad is really not matched to the others. 3 are within 10%, but the fourth compared to the one furthest apart from it is like 36% off.

PXL_20230903_205101117.jpg

RCA 12ax7a in v1

PXL_20230903_205920828.jpg

GE 12ax7a in v2

PXL_20230903_205925970.jpg

GE long plate in v3

PXL_20230903_205931704.jpg

Voltages are still a little low. I used UF4007 diodes in the rectifier, but I don't think that would that drop 20volts B+ compared to 1n4007.

AC input 122.4
Heaters 6.8
Relay supply 11.84
B+1 425
B+2 424
B+3 409
B+4 310.6
B+5 303.3

V1a 187/1.7
V1b 190/1.65
V2a 196.5/1.66
V2b 203.6/1.56
V3 input side 293 feedback side 295 cathode 49.2

Used the Ampeg method of balancing the phase inverter but I can't get very close to 0. At the end of the trimpot adjustment range I'm sitting at about 4mv. I am thinking this is because my power tubes are not closely matched? I definitely hear the ringing in the OT that some others here have mentioned. I'll try a few more tubes in the PI to see if I can do better.

-Dan

[Stephen1966]

A 1N4007 drops 1.1V and UF4007 drops 1.7V if I remember correctly. The 20V is probably going to include the 0.6V extra you need for each UF4007 but the rest is leaking somewhere else would be my understanding.

Do you have access to a scope?

[dbharris]

A 1N4007 drops 1.1V and UF4007 drops 1.7V if I remember correctly. The 20V is probably going to include the 0.6V extra you need for each UF4007 but the rest is leaking somewhere else would be my understanding.

Do you have access to a scope?

Hi Stephen, yes I have a scope. Please let me know where to start looking.

It also occurred to me I did not test the PT unloaded. I will do that today to measure the HT before rectification. Really appreciate everyone's help.

-Dan

[martin manning]

It also occurred to me I did not test the PT unloaded. I will do that today to measure the HT before rectification.

Just measure the reservoir voltage in "standby" and in "play."

[dbharris]

It also occurred to me I did not test the PT unloaded. I will do that today to measure the HT before rectification.

Just measure the reservoir voltage in "standby" and in "play."

AC is 122
In standby 455
Play 430

That is on a dummy load at 4 ohms. Voltages in my prior post were while connected to an 8ohm EVM 12L.

[martin manning]

~5% drop. Sounds pretty reasonable to me.

[dbharris]

Once the amp warms up fully the B+ drops to 424ish with preamp and OD volumes turned down all the way.

Maybe just a low wind on this PT?

-Dan

[martin manning]

I wouldn’t worry, your V1 and V2 plates look good.

[dbharris]

Thanks Martin!

[Stephen1966]

A 1N4007 drops 1.1V and UF4007 drops 1.7V if I remember correctly. The 20V is probably going to include the 0.6V extra you need for each UF4007 but the rest is leaking somewhere else would be my understanding.

Do you have access to a scope?

Hi Stephen, yes I have a scope. Please let me know where to start looking.

It also occurred to me I did not test the PT unloaded. I will do that today to measure the HT before rectification. Really appreciate everyone's help.

-Dan

Hi Dan,

With pleasure... I am thinking of this only with regard to the setting the PI and the story of that has filled many of the pages here so I will only add my personal experience and the rest can be deduced as you like. I think this is a great place to start and a great piece of advice that is worth revisiting: https://ampgarage.com/forum/viewtopic.p ... 89#p416389

After biasing the tubes to roughly where I want them to be (60% or thereabouts works quite nicely for me), I would begin by tying the cathodes on the power tubes together. Two on one side of the OT, two on the other. But, this will work even you attach the OSC probe to one cathode of the pair on one side of the OT and one cathode of the other pair. The scope should have both channels on display. Tying them into pairs gives a better averaged, result. Stop me, if I am not explaining this clearly enough...

I put the ground clip of the two probes as close to the chassis grounds for the two pairs of cathodes (either side of the OT) at the tail end of the 1R resistors and the probes themselves on the cathode pins of the sockets. I would be basically attaching the probe to both sides of the 1R resistors - probe to source and ground clip to ground as it were. A 1kHz signal fed into the front of the amp clean channel (no OD) and of course, a dummy load in the speaker jack. I turn up the volume and master up to "normal" playing level, quite high, as high as is comfortable without the sine clipping. If it starts to clip, dial it back a bit; it helps to have as clean a sine as possible. You will see at this point, the sine is not symmetrical but appears clipped on the negative side - this is the AC offset and you don't need to be concerned by this. There may be some disparity between the peak voltages on either side of the OT, such that even with the full rotation of the trimmer, you may not be able to get them to balance with equal voltages (peak). Instead, what I look for, is the voltage peak-to-peak and look for equivalence there. Your scope might be able to give you an accurate measure of the two channels side by side. I personally think this disparity between the voltages on either side of the OT could be down to the transformer itself or the ground scheme but either way, and because of imbalances in the OT that only someone like RG or Martin or Josip could explain, you're not likely to start or finish with a perfect balance between the two sides, rather, something "more or less".

A digital scope has all the bells and whistles and can give you a more stable and accurate reading of the voltages pk-pk but if you use an older analogue scope you may also sum the A and B channels together - this gives you (visually) something like a flat line, but because of slight differences between one side of the OT and the other, and the ground scheme, and the AC offset, and the fact that the two sides are most likely not exactly 180 degrees out of phase, and perhaps even a hint of crossover distortion, that summation of the channels, that line, is going to look less than straight. But it is the analogue alternative and gives you a visual indication (flatter is better) of the balance, in the absence of legible figures.

The problem you are going to have is if you have one tube that is seriously out of bias with the other members of the quartet, it's going to sing flat and throw the whole performance. To deal with this you might set the PI from a single good tube of each side and then look at it again when you have found a replacement tube that is better matched. If it's only one bad tube though, you can set the PI using only one good tube, from each side, and when you put the replacement in, it will be very very close to maintaining the balance. Others, might counter this with the idea of injecting or exploiting a little imbalance between the signals in order to encourage more second order harmonics, note bloom etc. But, I go for as close a balance as possible. And then, after setting up like this I switch everything off and get set up with a guitar to play. This is the listening part of the test.

At this stage your PI trimmer is more or less where it needs to be so bear that in mind when you go playing and making very small incremental adjustments around that point, you may be making very small turns a couple of degrees for audible results either way. It's time consuming but I am listening for that note bloom, that increased higher harmonic content to become audible and for sustain to develop. Go LOUD, it's much easier to hear than if you try to please the neighbours.

The ringing in the transformer you mentioned was an observation I believe I was the first to make but I have to say, when I installed a Tube Town OT in my Delta (#183) build, I didn't get it at all, so it's confirmed you won't hear it with all transformers or set ups. That's a very gross first way of finding the balance. The second is the scope, looking for the quantifiable measure of balance. The third, is using the most sophisticated instrument known to man, our ears, to determine the qualitative measure of balance. You can omit the first, the second (a luxury) is easier than trying to deduce the readings dancing around on a DMM, but the third is indispensible. In any case, whatever method you use, I would always look for the balance towards the end of the signal chain (at the cathodes of the power tubes) rather than looking for it at the PI tube plates (after the coupling caps if you don't want to overtax your scope, of course). Again, it might get you close, so it has some merits, but we are more interested with what the signal looks and sounds like at this later stage when the power tubes and the OT have come into play.

Have a look at that post about finding the PI balance though - if my method doesn't work for you, there are plenty of ideas there for you to play with.

[dbharris]

Thanks Stephen this helpful.

That is the method I used previously just with a DMM only. I will try with the scope (mine is digital by the way) and I like your suggestion to try just a pair of tubes if I still have no luck. I also ordered some 20K trimmers from Mouser in case I need more range there.

I may also swap the JJs back in just to compare.

-Dan

[Stephen1966]

Thanks Stephen this helpful.

That is the method I used previously just with a DMM only. I will try with the scope (mine is digital by the way) and I like your suggestion to try just a pair of tubes if I still have no luck. I also ordered some 20K trimmers from Mouser in case I need more range there.

I may also swap the JJs back in just to compare.

-Dan

You're welcome. If your JJs are a known good set it can't hurt and will almost certainly help you get your eye in. Only you might need to repeat the process when you put your matched NOS tubes back in. The trimmer is more effective with regard to the PI tube though so you might have an imbalance there. It's worth rolling the PI as well, but otherwise, a 10k trimmer works for me.

An afterthought but something that always gives good results for me, is to inject the signal at the PWR AMP IN of the passive loop. Several other builders bias this way as well. Bypassing the preamp stages like this avoids any preamp distortion "fogging up" the PI and power amp readings.