New Amp: Purple Flashes, No Sound

[The New Steve H]

Right, that's what I meant.

More info, from power tubes. Voltage at pin 6: 471. Voltage at pin 4 (after 470 ohm resistor) 359.

[The New Steve H]

I have not checked the voltage with the tubes removed.

The only resistor I see in the circuit near the filter caps is a 4.7K, and it's between caps, not from cap to ground.

[The New Steve H]

I am getting suspicious of the general area of the bias circuit, since everything else on the way to the power tubes seems to work. I have been known to put diodes in backward in the past.

[Firestorm]

Voltage at pin 6: 471. Voltage at pin 4 (after 470 ohm resistor) 359.

You sure? That's 240mA!!! Even a 20W resistor should be on fire.

[The New Steve H]

I don't understand how I'm getting voltages this high; I assume my meter isn't wrong. I'll check again and make sure I read it right.

[The New Steve H]

Okay, here is my new theory.

I got a nasty pop out of the speaker while I was measuring voltages, and it seems like I can generate respectable noises while poking around near the preamp tubes. I am thinking I may have the input jacks wired up wrong. I was starting to think something was amiss between the preamp and power tubes, but I can't find any problems.

These input jacks are the weird ones with grounding switches, so it is completely possible that I misunderstood how to wire them up.

Will check after lunch.

[Gibsonman63]

I would double check the bias circuit. Have you measured the bias voltage and/or current to the power tubes?

I have been tripped up on this one a couple of times becuase you are producing a negative bias voltage, so the diode and capacitor are backwards from the rest of the amp.

Google Half Wave Rectifier. Your bias rectifier circuit is a very basic example.

As far as the flashing rectifier tube, it could it be that the power tubes are turning on and off due to biasing issues. The bias voltage is what turns the tube on. If your tube is only on for part of a cycle, then the power will suffer as well.

There is nothing magical about the shorting input jacks. They just short in input to ground until you plug in a cable. Otherwise it is like having a guitar cable plugged in with no guitar.

[Structo]

Yes, make certain the bias capacitors have the + end to ground.
This creates the necessary negative bias voltage to the grids.

[Firestorm]

I'd suggest pulling the output tubes and measuring the negative bias voltage with them out. It isn't much affected by presence/absence of tubes. If there's a bias problem, you could be cooking expensive stuff.

[The New Steve H]

The bias voltage was -54V. I am pretty excited about my input jack theory. It can't be the output jacks, because the speaker made a loud noise. I am getting more distortion than I realized, so I think the amp is really trying, and it's just getting a seriously crappy input signal.

I put a 12ax7 in and took the 12ay7 preamp tube out, and the distortion increased.

Seems like every voltage is maybe 20% higher than expected. I was pretty sure I was supposed to use the outermost wires on the transformer, but maybe I was wrong. The AC here is right on 125V.

[The New Steve H]

If the 470K resistors on my output tubes weren't big enough to take a lot of current, could it cause my output signal to crap out? It turns out I used small ones instead of the big ones I bought.

Edit: I am looking over the pinouts and layout. Correct me if I'm wrong, but if one of those resistors fried and ended up with lower resistance, could it screw up the signal by making the screen grid too positive?

Embarrassing second edit: I think I found the problem. I measured the resistance across one of the resistors, and it's 233K. This would explain a couple of things. First, the high voltage didn't fry the resistors, because they were too small to pass much current. Second, the amp doesn't work. Arggh.

I must have put 470K resistors in there instead of 470 ohm resistors. I do not know why it now reads half that amount; it's as if the resistors are paralleled.

I know I have to put the right resistors in, but with the high voltages I'm getting, won't they melt? Is it safe to start the amp with the current limiter and see what happens?

[Gibsonman63]

You're pretty much always safe with the current limiter. Put the 470s in there and give it a shot.

[martin manning]

I must have put 470K resistors in there instead of 470 ohm resistors. I do not know why it now reads half that amount; it's as if the resistors are paralleled.

I know I have to put the right resistors in, but with the high voltages I'm getting, won't they melt? Is it safe to start the amp with the current limiter and see what happens?

You could put a 470k right across the HT and it would only have to dissipate about a half-watt. As it is they are lowering the screen voltage and throttling down the current through the power tubes to the point where they are barely conducting at all, which is why your voltages are high. Put the right ones in and try again.

[The New Steve H]

I changed the resistors. In the process, I accidentally roasted a 4700 ohm resistor, but it still works, and the resistance is not changed much, so I have not put a new one in yet.

Anyway, I now have 470 ohm resistors on the power tubes, and the amp acts about the same. The voltages have come down somewhat. Still a little high, however.

The voltage on both sides of each 470 ohm screen grid resistor is 454V. I was not aware that it was possible for a resistor to have the same voltage on both ends. I assume that's what happens when the current is negligible.

I found a better 5f6a layout picture, much sharper than the crappy ones that are all over the web. Uploading in case other people want one in high resolution.

[The New Steve H]

Wow. I finally got it running.

Apparently, the letter "k" in a resistor designation is important. For example, 470k looks almost exactly like 470, yet it's actually different by a factor of ONE THOUSAND. And unbelievably, amp designers aren't considerate enough to make designs that will work with or without the k.

Not only did I have 470k resistors in the power tube sockets; I somehow stuck one on the phase inverter cathodes. I was supposed to have a 470 ohm resistor there.

I had no 470 ohm resistors left, so I crammed a 442...I started to type "442k"...in there. The amp now works. There is a certain amount of noise, but I can't make any judgments about that until I have the amp completely finished, with tube shields and whatever.

Thanks for all the help. This is extremely cool.