Power amp essentials?

[briane]

That is basically what I do. Bob I was over and never was a believer in PI trimmers until I adjusted his with him watching and listening. Sometimes it takes the right PI tube as a baseline!

Ha, ha, your the one that originally described it to me on this forum!

The tube is critical, but while subtle the proper adjust of the PI 'OPENS' up the sound with bloom. Thats what to listen for. Theres always one place that sounds best, check the voltage, 5-6 off one leg, voila.

[Mark]

Okay, the madness continued. I installed individual bias pots, and set the bias at 34mA for 475VDC rail. I used my Fluke 77 to measure the ac voltage on the PI and adjusted the trimmer to get a matched output from the PI stage.

The first thing I noticed was how quiet the amp was, not a peep, no hum, no hiss!

The clean tones sounded much the same (no basis for comparison), but the Overdrive channel was awful. I will have to check if I haven't bumped something whilst soldering, it works okay, but the tone is compressed and buzzy. I was wondering why the clean channel still sounds good, I suppose it could be the overdrive channel generates a lot of harmonics and the balancing has wiped a lot of them out. I do intend to see if the transconductance is okay or not.

I'll try and find more time over the weekend to tinker.

[Structo]

Is the OD tube high or low plate?

If high, double check the cathode resistors.

Set the OD trimmer to 26K wiper to ground when amp is on and OD is switched on.

[dogears]

You can't use a voltmeter to balance the output.

Might as well put any old random tube in. Might be as close or better.

Okay, the madness continued. I installed individual bias pots, and set the bias at 34mA for 475VDC rail. I used my Fluke 77 to measure the ac voltage on the PI and adjusted the trimmer to get a matched output from the PI stage.

The first thing I noticed was how quiet the amp was, not a peep, no hum, no hiss!

The clean tones sounded much the same (no basis for comparison), but the Overdrive channel was awful. I will have to check if I haven't bumped something whilst soldering, it works okay, but the tone is compressed and buzzy. I was wondering why the clean channel still sounds good, I suppose it could be the overdrive channel generates a lot of harmonics and the balancing has wiped a lot of them out. I do intend to see if the transconductance is okay or not.

I'll try and find more time over the weekend to tinker.

[ampdork]

Ear is the easiest but Scott why couldn't one use a meter to balance AC?
Doesn't Andy use a meter method?

[Mark]

Thanks Tom, I try re-adjusting the OD trimmer.

I used my Fluke 77 to measure the ac voltage on the PI and adjusted the trimmer to get a matched output from the PI stage.

I seemed to have glossed over things here. I put a 1Khz signal into the PI, I then measured the ac voltage on either side of the PI. I did this as it is relatively quick compared to using the CRO. I thought the meter measuring the ac voltage would be doing it the same way for both sides of the PI.

I should also point out, this isn't the end of the process, it is the beginning, the way the output tubes were any adjustment to the PI was futile.

I needed to reach a point where everything is balanced again, I wondered whether the transconductance of the tube was as big a deal as we'd made it out to be?

I'm not sure, but I seem to recall output tubes don't have a lot of gain, but the function of the output tube is to pull current through the output
transformer thus the significance of transconductance.

As I mentioned earlier I will have to see if the same bias current means a similar transconductance, and when does it all go pear shaped?

The bias does set the operating point which is in effect the plate characteristics, does the different forward conduction points really matter, I'm thinking that you shouldn't really be driving the output stage all that hard in a Dumble?

Who knows I might learn something.

I believe adjusting the symmetry of the PI makes no difference to the PI itself, it is about how it drives the output stage.

I also wondered why HAD didn't use a 12AT7 which could provide better drive to the output stage if biased correctly.

So progress for me it to check the OD stage again, measuring the transconductance of the tubes (in a primitive way) and then try adjusting the PI by ear to hear the bloom.

I hope I've covered it all a bit better?

[pula58]

thought I'd re-open this useful thread!

I am trying to figure-out something about this issue.

So, I have a 12AT7 PI, 6L6GC outputs into a Hammond 40W OT.

I measure the OT turns ratio from each half of the primary to the output and they are very very close. I put in a 300Hz sine wave, and from the secondary I get 239mv driving one half of the primary, and 238 mV driving the other half.

I trim the PI so that the sum of the output tube grid signals (as seen in the O-Scope) is AC zero. Then, I look at output of the amp using the FFT feature of the scope. Lots of second harmonic, plus I can visually see mismatch between the pos and neg arcs of the sine wave test signal. So, I re-trim the PI to get the 2nd harmonic down as low as I can. The position of the trimmer is, in this 2nd case, completely different compared to when I had done the trim matching the PI outputs. I thought, maybe the gm's of the two output tubes don't match, but when I swap the tubes the trim pot stays in the same position to get the minimal amount of 2nd harmonic as seen at the output.

Being that the OT seems balanced, why should the trim pot setting be so different to get matched AC signal on the PI outputs vs. seeing minimal 2nd harmonic at the speaker (dummy load in this case)? Recall that swapping the output tubes with each other made no difference, and I measured the turns ratio of each half of the OT primary to seconday and each side matched very closely. I have also checked the coupling caps, and all resistors in the PI.

So, if the OT is balanced, and the tubes are balanced, and the PI is trimmed to good balance as seen at the output tubes grids (or PI plate loads), where is the 2nd harmonic creeping-in? That is, why is the AC trim to make the PI outputs be balanced NOT minimizing the 2nd harmonic at the output of the OT??

[martin manning]

How about putting a signal into the OT secondary and then see if the two halves of the primary will cancel out on your scope? I don't know how you measured the voltages to determine the turns ratio, but the response may not really be the same on both sides.

[pula58]

How about putting a signal into the OT secondary and then see if the two halves of the primary will cancel out on your scope? I don't know how you measured the voltages to determine the turns ratio, but the response may not really be the same on both sides.

I drove one half of the primary (between center tap and one (brown) lead) with a signal generator and measured the rms output at the 2nd-ary. Then, I drove the same signal into the other half (between primary, CT and other (blue) primary lead). The output voltages matched to within 1% or so.

Thanks, I like the idea of driving the secondary and looking at the primary voltages. I'll try it...

[pula58]

I tried driving the secondary and measuring the voltages between the two halves of the primary. They match to within 1%.

I checked the screen R's, they are the same for each of the 2 output tube tube sockets. I also scoped the voltage of each screen with the amp at max power. The screen waveforms are nearly identical. The DC screen voltages are also identical. The average screen voltage when the amp is at full power are also within a 1/2 V or each other.

The only thing that doesn't match is the DC resistance of each half of the OT primary. One half is 82 ohms, the other half is 110 ohms

[martin manning]

When you compare the PI outputs and null the voltage you are balancing the entire waveform from each side against the other. Then, when you look at the output you are looking at a combination of the top halves only. I suspect that is the difference.

As I recall, some say that the right spot for that trimmer is where the second harmonic in the speaker output is high.

[pula58]

I just tried the amp with another OT (different manufacturer). With the new OT the trim point for the least 2nd harmonic is more toward the middle of the trim pot range. With the previous OT the trim pot had to be turned all the way in one direction.

So, this is interesting, my measurements of the OT revealed no assymetries in terms of apparent turns ratio. Yet, the OT is clearly not acting symmetric. The only thing non-symmetric I could measure was that 1/2 of the primary winding was 110 ohms series DC resistance, and the other was 82 ohms. Which leads one to suspect that the turns ratio from each half of the primary to the output are not matched. Yet, as mentioned in previous couple of posts in this thread, when I drove the OT primaries and measured at its output, or drove the output and measured at the primaries the OT seemed balanced.

[martin manning]

The DCR is different because the length of the wire needed to make the turns is different. The no-load voltage ratio doesn't reflect that because there is almost no current flowing. I think if you try the turns ratio test again (driving each half of the primary) with an appropriate load on the secondary it will show up.