Abnormal Current in SE EL84 on first power up

[JoeTele]

I powered up my Mullard 3-3 build for the first time tonight on the variac. After frying/replacing a couple of power resistors (much higher wattage), I realized that my single EL84 power tube was pulling no less than 120 ma of current with 276 volts on the plate (I was aiming for the 310 stated in the build plan, whose schematic values led me to calculate around 50ma through the tube). I did rework the power supply for the build, using a GZ34 in place of the stock EZ80, but I modeled it in the Duncan program and have to believe that the 120ma (67v on the cathode of a 560 ohm resistor, bypassed with a 25uf cap) is more than slightly off resistor values in the power supply. Could the tube be bad, or am I maybe dealing with oscillation or something else? Most of the voltages seemed reasonable given this excessive draw, the fact that I didn't have real-world currents to work with when I modeled the power supply.

Thanks!

Joe

[strelok]

A scope on the output would tell you if its an oscillation. However I'd first suspect the tube. Kinda surprised its still alive! At the numbers you quoted its dissipating about 33 watts!

Something seems very much off. I would double check the wiring as well to make sure its correct and something is not connected to the wrong pin, before you change out the tube. You should be seeing about 24v on the cathode, based on your plate voltage and Rk.

[Stevem]

With speaker hook up and a voltmeter set for AC and hooked up across the speaker power up the amp and look for voltage above .250 at idle, if so then yes, the amp is oscillating !

Also to save your OT from failing install a fuse in the OTs red lead!

[labb]

Possible bad tube. Sometimes the EL 84 will fail with a short of screen to grid internally. When it does you get a run a way tube. Check tube pin 9 to pin 2 with your meter and see if they are connected.

[JoeTele]

The tube did not seem to be shorted and I did not get a reading across the speaker terminals. Aware of the stress on the components, I variac'd the amp up. The behavior is a bit unpredictable, and may change slightly when I pull the tube and then put it back in. Today, I was able to push the mains to almost 120 with about 60v on the cathode rather than 120(though it was creeping up). I could barely get .1 mv across the speaker. I did try one other oscillation test I found on another forum whereby I grounded the EL84 grid via a 2.2k resistor and this stabilized/lowered to cathode voltage to about 15v at full mains power. At this point is it still worth trying (i.e. buying) a new tube? If oscillation is the culprit, would a powerful scope be helpful if not crucial to track down the source? This is my most ambitious project in terms of not following a schematic/component layout exactly, I'm in this for the long haul and would not mind an excuse to upgrade from my 7mhz single trace Leader. I've attached a pic of the chassis. Are there any glaring lead dress no-no's that anyone can spot? The schematic is here.
http://www.r-type.org/articles/art-003h.htm

BTW the main power resistors will be changed and properly mounted when I'm past the troubleshooting phase.

Thanks yet again!

[Phil_S]

T...The behavior is a bit unpredictable, and may change slightly when I pull the tube and then put it back in...

This makes me think there is a cold or missed solder somewhere, or a bad tube socket, or something like that. I dunno...hope you get this sorted out.

Also, it looks like your board is wood? Wood can be conductive. It is not a good choice of material.

[sluckey]

I grounded the EL84 grid via a 2.2k resistor and this stabilized/lowered to cathode voltage to about 15v at full mains power.

You should have a resistor from grid to ground. Usually about 470K. What resistance do you measure from grid to ground?

Also, the grid voltage should be zero. What do you measure?

[martin manning]

Hold on there Steve, the power stage is DC coupled to the stage driving it. See here: http://www.r-type.org/articles/art-003h.htm

[JoeTele]

Yep, grid voltage is an issue. It's about 110v with only 40 on the variac. Thing is, this is a directly coupled circuit, which the claim they get away with due to the "starvation" conditions that the EF86 is running at. I'm going to re-measure/solder my power resistors but I wonder if I need to add rc coupling? Would that gut power significantly?

Joe

[sluckey]

Yep, grid voltage is an issue. It's about 110v with only 40 on the variac. Thing is, this is a directly coupled circuit, which the claim they get away with due to the "starvation" conditions that the EF86 is running at. I'm going to re-measure/solder my power resistors but I wonder if I need to add rc coupling? Would that gut power significantly?

Joe

Guess I should have looked. Sorry. The EF86 is 'supposed' to be running about 20V on the plate. But you have 110V on the EL84 grid even with the variac set to 40. That explains the high current in the EL84. Gotta get the EF86 plate down somehow.

[JoeTele]

Makes sense! Will re-work my power supply resistors. I assume I would NOT want to change the plate load resistor, but perhaps the one in front of it?

Joe

[martin manning]

I'd look for a wiring error. The EF86's screen voltage comes from the EL84's cathode voltage, and that feedback loop should be regulating the EL84's current. I don't like the wood circuit board either.

[JoeTele]

Will ditch the wood. I should have researched that! In any case, I had only modeled up to the plate resistor, not across it. Having played around with adding it to my model, and looking at some of the measurements I took across it at full power/overload, I think I do need to increase resistance in front of the plate resistor, so I'll add that, if only to protect the tube etc in the next phase of testing, and see if things line up better. I always check my wiring several times before power up and all my solder joints for continuity, but of course I'll review again when I get her rebuilt with the new board.

Joe

[JoeTele]

Martin thanks for highlighting that cathode to screen feedback loop! I looked again and realized that I ran that joining resistor from the *ground* side of the cathode resistor. Will redo that first, remeasure, and go from there.

Joe

[martin manning]

That should make a huge improvement. Let us know what happens!