Discussion in 'Tube Audio' started by oldman55, Dec 19, 2017.
Glad to get that quad out of there. Voltage is 0.
If you are sure that that same measurement (with the 330 ohm resistor lifted) with the output tubes all in place causes the voltage as measured on that side of the 47 ohm resistor to rise (i.e. causes red plating), then it's either a wiring error somewhere in the output stage or wiring error in the bias circuit, or a bad output tube. It feels like something is dumping an inordinate amount of voltage onto the grid circuit, and you need to devise a few tests to determine where that voltage is coming from.
But before you try additional tests, I think I would take some 400 grit wet/dry sand paper and lightly sand each output tube pin, removing any corrosion or oxidation. Also if you can, make sure the output tube socket pins are tight and will hold the pins snugly when the tubes are inserted. When the amp is working properly, you should be measuring about a -25V difference or thereabouts between the grid of any one of the output tubes and it's cathode (positive meter probe on the grid and negative meter probe on the cathode). The more positive that voltage measurement, the more "on" the tube will be, and likely red plating will occur. If you are measuring at little as say -5V , 0V, or even a positive voltage difference then that is quite a sure sign that red plating is going to occur. Also if the screen voltage is vastly different than what the schematic says (350V) would also be indicative of a problem.
On the assumption then that the tube pins are clean, the sockets are tensioned appropriately, and the screen voltage is not out of whack, then I would think the problem is either bad wiring or a bad output tube. You can try a few tests like this:
1. With the 330 ohm resistor still lifted, put all four output tubes back in, but remove V1, V2, and V3 tubes. Power up. (output tubes should not conduct since there is no current path from cathodes to ground with the 330 ohm resistor lifted and with V1-V3 removed). Voltage between the high side of the 47 ohm resistor and ground should still read zero. A voltage measurement between the wiper of one of the channel's output stage DC balance pots and that same location on the 47 ohm resistor should read zero. If it reads anything other than zero, it points to a problem. If no problems detected, then try test 2, below.
2. Reconnect the 330 ohm resistor. Put V1, V2, and V3 back in. Randomly pick two of the four output tubes and put them in one channel's sockets, leaving the other channel's output tube sockets without tubes. Turn on the amp. If the tubes red plate, one of those output tubes is bad or the wiring for that channel is bad. If the tubes stay cold (as expected), put those same two tubes in the other channel's sockets and try test again. If still no red plating, then it might be one of the other tubes not currently in the amp that is the culprit. Swap out one of the two output tubes currently in the amp for a random one that is currently sitting on the side, and retry the tests. You are looking to identify if a swap in of a given tube tube causes the problem to suddenly reappear, or if the location of the tubes in a given socket suddenly causes the problem to reappear. If you had a spare set of output tubes, you could swap one tube in at a time from the spare set until you identified the bad tube.
These are just some ideas of things to try to attempt to pinpoint this problem. There might be other tests similar to this that you can try.
Thanks for taking the time to detail testing.
I put in an old set of 7355s that came with the unit. One of them gave out of range results and I assume it is bad though it passed emmisions, shorts and leakage test.
My results are:
The grid to cathode test was strange in that after warmup, they all started out about -20 to -25 but when leaving the clipped probe on, voltage would continually drift and after about 3 minutes would be about -12 to -14. Screen voltage was 367vdc. Pin 6 grid was about 1.5v but would drop down to about 1v after being measured for about 30 seconds.
Voltage between the 47 ohm resistor and the balance pot read about .5v on all of them.
I did not perform the tube swapping test as knowing that the grid to cathode was from -25 to -12 that they would not red plate.
After warm up, the bias voltage at the wiper settled down to -32v (started at -36v).
As a final test, I put in all of the old tubes from initial test and ran it for awhile. Music was clear (weaker in channel with not-so-good output tube). Bias voltage was -constant at 36.5vdc. The 330 ohm was still disconnected.
Put the 330 ohm resistor back on to wiper. Now playing at 35.5v.
So, with the original tubes in place and exact same parts and wiring, I cannot replicate the initial problem. Must have been a tube pin as the only thing different is that I cleaned them off better during the testing process.
The 330 ohm 6.5w resistor gets hot (about like a CL80) but not smoking and bias voltage not climbing. B+ rail is now also on the money (within 5v).
Thanks to all for the help. If nothing else, I have gained more knowledge on how this cathode biased amp works.
Last questions; is the 35.5v bias determined by the preamp heaters? And if so, will swapping out V1, V2 and V7 possibly change the bias? And should I do so to get the bias down to 30v? Or can I replace a resistor to get it down?
Hi Oldman. Glad you got it working. Yeah, I suspected a bad tube in your first set of tests.
Swapping out any of the preamp tubes will not change the bias of the output stage, even though the filaments of V1-V3 are used in part to establish that bias. My experience is 12AX7s either new or old stock are very consistent in their filament parameters.
The 47 ohm and 330 ohm resistors form a voltage divider that feeds a few positive volts from the cathode back to the grids so that the difference between the grid and the cathode self biases the tube at approximately the correct value that the manufacturer intended. I expect your latest set of tubes run hotter than the original set, and that will create a self correcting condition that will increase the bias voltage appropriately (make it more negative) so that they run more or less at the correct quiescent current regardless of how old or new they are.
So I wouldn't change a thing (other than maybe make that 330 ohm resistor larger wattage to better manage the heat).
Enjoy the amp! Happy Holidays.
The small tubes won't change the bias enough to worry over. Its actually more determined by how healthy the 7355's are than anything else, assuming the 330 and 47 ohm resistors are on value. Definitely agree about leaving this alone for the most part as long as its working properly.
If you do need to change how it acts, making the 330 ohm resistor larger will drop the current through the tubes, which will make for less voltage at the cathodes.
Since I cant use a cathode resistor, how can I test the new 7355s for current?
Individually, you really can't unless you add a resistor per-tube between the existing wiring and pin 5 on the socket.
Disconnecting at that point leaves half the voltage doubler still in circuit and working.
It goes to the balancing pot. I thought that since this is cathode bias and/or the pin doesnt go to ground that it would be different on this amp.
it works, but you'd have to measure across each resistor, not to ground. Resistor would just go from pin 5 to the balance pot then. I was thinking how my Fisher wires, no balance pot on that, the cathodes all wire together and go to the phono tube heaters.
Which side of the resistor would get the test points or will it be about the same either way? I was going to come off of the pot.
You have to put a test point on each side of each resistor. It would be a total of 8.
personally I'd be tempted to just use mingrabber leads underneath, set the balance, then close it. if you change tubes later, pull the bottom off and set it again if needed.
I was going to add test points to balance each pair above chassis.
Was also looking for a way to compare tubes. But without fixed bias, wouldnt it be difficult to compare tubes anyway? Maybe just add one resistor and run all the tubes through that socket to test them relative to each other?
if you do that, it produces different resistance between the balance pot and the socket on that socket it would add an imbalance to it.
something like this would be an option if you just want to see what they do relative to each other. No permanant mods to the amp required if you do this, though you aren't going to be able to use it for balancing unless you build at least two.
7591 and 7355 run the same pinout. All that does is stick a resistor in the cathode and you measure across it. Easy enough to knock one together if you've got a spare socket and a dead octal tube you can rob a base from. You can make one for any pinout, the only thing that really changes is which pin gets a resistor instead of a straight wire.
One would come in handy. Unusual to me that the bias voltage climbs considerably when worn tubes are played. To the point where the resistors get hot enough to burn. So a tube failure (or worn out tube) could cause some damage?
if its a gassy tube, that might do it. Those tend to run away like that. Shorted tubes could do it too. Depends how its bad really, simply weak should not cause that sort of issue.
The tubes that are worn show strong emissions, no leakage and no shorts (and work in another 7355 amp) yet the bias voltage runs away. I shut it off after it gets above 70v or so as I know the 6.5w 330ohm resistor starts to smoke a little after that. As the bias rises, the HV+ gets lower (330v vs 370v spec).
Just put another 7355 in and the same thing happens. It takes from 2 to 5 minutes, depending on tube, to get to 70v.
So I guess I still have problems. It only stabilizes at 33v (30v spec) when I have in a set of nos Raytheons.
Cathode coating that sputters onto the control grid will cause bias cancelling runaway, something that doesn't show up in routine tube testing.
Separate names with a comma.