The JCM800 is back in the shop. I had a hard time remembering what I did with it the first time? I actually searched on here to find it.
I am going to start doing a better job with invoices. Then I can search them based on the amp or owner. That way I have a record of what was done.
The customer said he wasn't happy with the sound. He connected up his bias probe, and measured the current to be 48mA. He said that was way too high. So he adjusted it and could barely get it down low enough. He got it to 38mA, but that was at the edge of the pot adjustment range. He tried another set of JJ EL34L tubes and those were at 60mA and he couldn't lower them.
I am in favor of people doing their own work. I just wish they would keep doing it. Instead of a little knowledge being a dangerous thing, and then they screw it up, and bring it to me. Because I never know what they did, because often they don't know what they did either. OK, enough complaining. Hey, it's Monday!
So I removed the power tubes and checked the voltages and the bias voltage range, to make sure it was working. It seemed OK, so I put it on the AC with the tubes in. Turned on the power. Waited 20-30 seconds turned on the Standby switch, and here was a brief sound, like a split second of hum, then silence. It blew the 3.15A main input fuse?
I traced it to one of the tubes he had installed. I happened to have a pair of JJ EL34L tubes. I put those in and cautiously powered it up, and no worries. I set the bias to around 35mA, and it seemed stable and all. I played it for a bit, and it seemed fine.
I looked into his claim that I had it set too high. The plate voltage was 440Vdc, and with a max power in a EL34L of 25W, that means the idle power is 70% of 25W.
Max idle power = 0.70 x 25W = 17.5W
So with a plate voltage of 440V and a bias current of 35ma the power is at
P= 440V x 0.035A = 15.4W or 15.4W / 25W = 62%.
The actual max current with a plate voltage of 440V is
Bias Current @ 70% = 17.5W / 440V = .0398A = 40mA
Now this would agree with his number of 38mA that he claimed was ideal. However, most bias probes measure the current at the Cathode as it exists the tube, just before it reaches the Chassis Gnd. At that point there are two sources of current into the tube. Current enters through the Plate connection but also through the Grid connection. that is the connection through the 1K resistors.
pa by
Dennis Kelley, on Flickr
By measuring the voltage drop across the 1K resistor, it's easy to know the current through the grid. Just divide the voltage across the 1K by 1,000 ohms and you have the current. In this case it was 4V, and therefore 4mA. Of course your meter is sitting at 400V when you do this, so extreme care is needed.
This is important because all this 70% stuff is talking about the Plate dissipation. That is the voltage on the Plate, multiplied by the current through it. The Plate current is the Cathode current minus the Grid current. So when I measured 35mA above, the actual Plate current was only 31mA. So the actual power dissipation at idle was P = 440V x 0.031A = 13.6W , not 15.4W as I calculated above.
Another thing happens when you increase the current through the power tubes at idle when setting the bias current. When you incease the current, the Plate voltage drops, due to the resistance in the output transformer, and the limits of the power transformer.
In the example above when I increased the bias current to close to 40mA, the Plate voltage dropped from 440V to 416V. So under those conditions the actual power dissipated by the Plate is
P = (40mA - 4mA) x 416V = 14.98W or 15W.
This is a percentage of the max of 25W of
Percentage = 100* 15W / 25W = 60%.
This is all well and good, if you want to wrestle with all this stuff. I actually enjoy the theory behind it all. Maybe I should have been an engineer? Oh yeah, that's right, I am. But the point is, there isn't just a number, of 40mA or 38mA, or 28mA. You have to take the voltage into account. Most tube amps don't have screen resistor. So you can easily measure the current. I usually just ignor it, and set it to around 60% to 70%, and rely on that current as a safety buffer. Quite often the tubes are out of balance a bit. One might be at 35mA and the other at 30mA. So I don't mind setting the hotter one to a higher current than ideal, in order to keep the lower one from being too low. I know there is some margin in the 60% number, and in the screen current that I haven't subtracted from my measurements.
So I am going to use the tube tester on his tubes, and see if one is indeed shorted. One of them did have a big black mark across it? Makes you wonder? Once I do that I will call the customer and have a discussion about tube bias, etc. Not to scold him or anything, just to try and give him the full story.
I've worked on several Marshall amps, and this amp is only the second that I really liked. The other was the newer DSL40CR. I've been told you have to crank the Marshall to get the right sound from them?