Bringing Up My First Amp

[donvan]

I'm in the process of bringing up my first amp, a 100 watt #183 circuit. It's going pretty well. I've had some issues but with some troubleshooting using my volt and ohm meter, I've been able to resolve them until now. I think to address my current issue I just need to get off my @ss and get an oscilloscope. But in the meantime, it's driving me crazy so I thought I'd throw it up on the forum and see if I get any ideas.

The issue is with the power amp stage.

Without the power tubes plugged in, the power supply voltages look good and I can come out of the preamp into another amplifier and I have a good, pleasant sound. Next, I set the bias voltage and installed the power tubes. At that time the bias current was low so I set it 70% power dissipation. One thing I noticed (and I think this is a clue) was that as I adjusted the bias pot and as my bias current approached 70% of power dissipation, the current would suddenly jump way up in a non linear fashion. I quickly turned off the amp, lowered the bias pot, turned it back on and tried again.

Once I had the bias set where I wanted it, I turned off the amp, switched from a dummy load to a speaker, plugged in a guitar, and turned the amp back on. I got a horrific, loud sound - the kind of sound that blows speakers. Also, my power supply voltages were now lower than I would expect them to be.

So based on the bias pot issue and this terrible sound I believe I have a really bad oscillation. So again, I think I just need to get an oscilloscope and troubleshoot this with a dummy load.

In the mean time, I'm curious what it sounds like when the NFB is hooked up backwards. It's hard to describe the sound that I heard but I would say it was a lower frequency. Similar to the hum you get when your amp is turned on with the volume up and you touch one end of the guitar chord with your finger when the other end is plugged into the amp except MUCH louder.

I really don't think I have the feedback of my OT reversed but I will unhook it as a test but I don't want to try it with a speaker hooked up and risk blowing the speaker. Therefore, as I keep saying, I really need to get that oscilloscope.

Until then, can we speculate?

[martin manning]

Disconnect the NFB and see if that stops the screaming. If it does, you probably do have the primaries or the leads from the PI to the power tube grids reversed.

[pompeiisneaks]

Disconnect the NFB and see if that stops the screaming. If it does, you probably do have the primaries or the leads from the PI to the power tube grids reversed.

+1

Edit: and to add the bias current shooting up could just be that you've gotten the tubes out of too cold range and they really start conducting BUT it could also be caused by the oscillation. Fix that first, and then see. What values of mA are you dialing in with the bias? Again, after fixing the oscillation, that's the first fix.

~Phil

[norburybrook]

you don't 'need' an oscilloscope to trouble shoot most things. I've built about 12 amps now and managed, with the help of these good people, to troubleshoot all the issues I've had with just a multi meter and some common sense


I'm nearly finishing my first #183 build too.


M

[donvan]

Thanks everyone!

Related to feedback wiring, in the files section on the layout, the feedback comes from the 4 ohm tap. On the schematic just below the layout it comes from the the 8 ohm tap. Later on in the thread there is an alternate schematic by mdroberts that shows the feedback on the 4 ohm tap. So the majority is going with the 4 ohm tap. Is that correct?

Phil,
I am getting my bias current numbers using a formula provided by Martin in an earlier thread:

"Typical bias setting for a Class AB amp is 70% of Pa max, and for EL34 Pa max is 25W. At 416V that would be 0.7*25W/416V = 0.042A. If you are measuring current at the cathode (with a 1R resistor, say) you will be conservative since screen current is included. Using 75% would compensate for that."

[donvan]

you don't 'need' an oscilloscope to trouble shoot most things. I've built about 12 amps now and managed, with the help of these good people, to troubleshoot all the issues I've had with just a multi meter and some common sense

Marcus, since you don't have a scope do you have a signal generator? How do you set your PI trim pots?

[martin manning]

Related to feedback wiring, in the files section on the layout, the feedback comes from the 4 ohm tap. On the schematic just below the layout it comes from the the 8 ohm tap. Later on in the thread there is an alternate schematic by mdroberts that shows the feedback on the 4 ohm tap. So the majority is going with the 4 ohm tap. Is that correct?

You can do either if you use the appropriate value for the feedback resistor. Looks like MDRoberts’ scheme is ok, the Jaysg scheme right below the layout is the one that seems have an error using the same 4k7 FB resistor from the 8R. I’d follow the layout. In any case I doubt that would cause the noise you are hearing.

"Typical bias setting for a Class AB amp is 70% of Pa max, and for EL34 Pa max is 25W. At 416V that would be 0.7*25W/416V = 0.042A. If you are measuring current at the cathode (with a 1R resistor, say) you will be conservative since screen current is included. Using 75% would compensate for that."

D amps seem to be run a little cool, so maybe shoot for 70% using cathode current.

[norburybrook]

you don't 'need' an oscilloscope to trouble shoot most things. I've built about 12 amps now and managed, with the help of these good people, to troubleshoot all the issues I've had with just a multi meter and some common sense

Marcus, since you don't have a scope do you have a signal generator? How do you set your PI trim pots?

start with voltage difference then by ear


M

[pompeiisneaks]

Thanks everyone!

Related to feedback wiring, in the files section on the layout, the feedback comes from the 4 ohm tap. On the schematic just below the layout it comes from the the 8 ohm tap. Later on in the thread there is an alternate schematic by mdroberts that shows the feedback on the 4 ohm tap. So the majority is going with the 4 ohm tap. Is that correct?

Phil,
I am getting my bias current numbers using a formula provided by Martin in an earlier thread:

"Typical bias setting for a Class AB amp is 70% of Pa max, and for EL34 Pa max is 25W. At 416V that would be 0.7*25W/416V = 0.042A. If you are measuring current at the cathode (with a 1R resistor, say) you will be conservative since screen current is included. Using 75% would compensate for that."

I'm more shooting for you saying "I'm using 2x(or 4x) 6L6GC's or EL34's in this amp, and I'm reading XmA of current per tube" or something like that. I.e. WHAT are your actual readings, not 'I'm shooting for 70% Diss. If you mean by that quote that you're reading around 42mA per tube on EL34's (becvause you're quoting someone else, not saying that's your real measurement) then that does sound about right. I'm just trying to remove ambiguity and have actual data to go on.

As far as the NFB, we're not too concerned about the exact tap and exact rating, what we're trying to rule out is if you have the output transformer in phase with NFB turning it into PFB, and causing that howl. Just disconnect it temporarily turn on and if the sound is gone, you need to either swap OT primary leads on the power tubes OR swap the PI output lines to the opposite halves of them. Either will resolve the problem.

I think that martin answered that question perfectly anyway.

~Phil

[donvan]

Yep, it was the ol' swapped OT primary trick. I now have good sound coming out of the power amp!

My voltages seem a little low so I want to run them by everyone. When I say they *seem* low I am comparing to the mdroberts #183 schematic which has the power supply voltages listed. Does anyone see a problem with these voltages?

B1 457v
B2 455V
B3 431V
B4 308V
B5 299V

[martin manning]

Your 457 B+ vs. the 465 on the layout is less than 2%... Plenty close enough.

[donvan]

Your 457 B+ vs. the 465 on the layout is less than 2%... Plenty close enough.

That does sound plenty close when you consider the percentage. Thanks again Martin!

My next step is to get the FET board working. Before I get it biased I am wondering about my part selection. At the time I ordered parts I couldn't find any FETs that the BOMs in the file section were using so I ordered a 2N5486. It "looked" close but it doesn't have a metal case. Since then, I see that folks here are now using 2N4416A. I already have the 2N5486 soldered in the board.

Any opinions on the 2N5486 FET?

[martin manning]

It will work for sure. The only caution is it's Vds is only 25V, so be careful when you are getting it set-up.

[donvan]

It will work for sure. The only caution is it's Vds is only 25V, so be careful when you are getting it set-up.

I hope I didn't smoke it when I checked out my power supply voltages with no tubes present. We will see.

[donvan]

I'm looking at setting the PI trimmer. I have read through the sticky and there is a lot of good info there and many ways to do it. I have borrowed a low distortion signal generator and I now own a Rigol scope. I also have a voltmeter.

Gil has a nice method (similiar to the Ampeg method) where he uses a signal generator and a voltmeter. He measures the DC voltage difference between the each pair of cathodes (with 1 ohm resister to ground). My question for this approach is: Why is it necessary to tie together the cathodes of the power tubes that are in phase? Is that required on all of the methods?

Martin, you have an interesting method where you compare the *AC* difference with an oscilloscope. I was wondering why you do that at the 1 ohm cathode resister? That would be a small (millivolts) AC voltage at least for my lower resolution scope. Why not compare the AC voltages at the power tube grids and then do an A + B on the scope, then set the trim pot for a flat line?

Thanks in advance!