Crate Vintage Club 50 with EL84 getting very hot

[Stratfordade]

Owner reported had been playing fine but when powered up next day no power light and dead.

Getting access to chassis I found main T1A power fuse blown, but absolutely no sign of any visual damage to any component on the board. After replacing it powered up again with eye on PCB to see what might have caused fuse to blow. There was very noticeable 100Hz hum but with music plugged in it sounded loud and excellent but after couple of minutes wisp of smoke emerged from underneath where valves mounted! Powered off and then removed chassis completely. I found the EL84 in V6 (2nd from right in photo of chassis) position looked different from all the others with signs of heat damage — some blackening of glass and blistering of label at base.

I removed the burnt EL84 and powered up again. Hum was still there but amp still played OK albeit not quite as loud. With input removed I measured plate current of the remaining valve paired with the burnt one which was still out of the amp and it was 27.3mA. Grid bias at the still inserted valve and at the base of the removed 'burnt' valve was correct at -15.3V.

I then replaced the normal looking valve with the burnt one in the same position. After allowing it to heat up and settle down got very only slightly higher plate current of 28.4mA so no obvious fault with the valve that had been getting too hot.

What could cause that EL84 to get hot to point where label was smoking? Can I just replace it with a good one or do I need a matched quad for all of them? To the eye all the power supply electrolytics (Rubycons) look ok with no bulging. Any suggestions please of where to look to reduce the hum, or could it just be result of that burnt valve not operating normally and causing a mis-balance?

Thanks a lot.
Crate Vintage Club 50.pdf

[Jon Snell]

Check the mains voltage tapping. The UK is 240v!
Some people, when using 'Hot" valves remove Jumper 3 and fit a 33R 5W resistor. That cools them down as they, at best, are being over driven.
You could of course, select cooler valves or remove the lable

[Chuck H]

What Jon said. Crate does bias el84's HOT in my experience so cooling the bias is a geed idea in general. If 33R changes the tone of the amp outside the customers liking try a 22R. But remember that the point is to cool the bias so going smaller than that isn't going to help much. And...

Just to be sure about things I would test bias current for the known good tube on that side in both sockets of that side to see that the current is the same. That way you can eliminate the socket with more confidence. You asked if you need a matched quade to replace that cooked tube. I would say yes.

[Stratfordade]

Thanks Jon. I may try that 33R mod. Will certainly use your lateral thought of removing the label! Just wondered why it's only that valve showing signs of heat distress. The labels on the others (they are a matched quad of Marshall brand EL84s) are fine and no blacking on glass either. I am concerned that the overheated EL84 was responsible for the fuse blowing though as it's quite a coincidence that the label should burn up so soon after main fuse blew. For now I've popped in a Svetlana EL84 from Tube Amp Doctor in that position.

[Stratfordade]

What Jon said. Crate does bias el84's HOT in my experience so cooling the bias is a geed idea in general. If 33R changes the tone of the amp outside the customers liking try a 22R. But remember that the point is to cool the bias so going smaller than that isn't going to help much. And...

Just to be sure about things I would test bias current for the known good tube on that side in both sockets of that side to see that the current is the same. That way you can eliminate the socket with more confidence. You asked if you need a matched quade to replace that cooked tube. I would say yes.

Thank you Chuck. I'll see about measuring those bias currents. I did check plate current in both sockets on that side for the overheated EL84 and it was the same.

[Chuck H]

Thank you Chuck. I'll see about measuring those bias currents. I did check plate current in both sockets on that side for the overheated EL84 and it was the same.

Ah. Maybe I missed tha in the read. Since we know that tube is actually still "working", albiet drawing more current than the other tubes, I think your test confirms it's the tube and not socket.

[Stratfordade]

Check the mains voltage tapping. The UK is 240v!]

Interesting. The amp has the 94-270-30 mains transformer, wired as shown on schematic with only a 230V tap. I'll compare the voltage at centre tap with the test voltages in table but it can't help if the design runs things hot to begin with! According to schematic there were other export versions with a transformer that could be wired for 230V or 240V AC. Edit: Just checked TP19 which is OT centre tap and it's slightly high — 391V in test table and 397V measured so not quite 2% high. Think I'll install the cathode resistor Jon Snell suggested anyway.

[g1]

It's possible the tube in question had an intermittent connection or a temporary short that recovered.
For the connections, bad solder if the sockets are board mounted, or oxidized pins on socket or tube are all possibilities.
The hum is probably due to imbalance if only running 3 tubes or if one is now mis-matched.

[drewl]

I've had this problem a few times.
If you don't want to add a cathode resistor to cool it down, I've used a voltage doubler circuit for the bias supply to run the tubes cooler.

Or this mod for the bias circuit

[Stratfordade]

Thanks drewl that’s an elegant idea

[Mick Bailey]

Those look like the same labels TAD uses on octal bases. A sticky label of any kind isn't good applied directly to a glass envelope.