Weird EL84 PP behavior

[alexg]

I breadboarded an EL84 PP amp today.

6n1p driver, EL84 triode strapped. B+ of 280VDC.

Am using a Lundahl 1660S as interstage/phase splitter, plate voltage on the 6n1p is 190VDC with two red LEDs on the cathode.

On the output stage, I first biased it using two paralleled 270R 2-watt resistor bypassed with 100uf cap (the cathode resistor is common to the two output tubes). I was getting 21VDC on the cathode, 155ma for both tubes, which is quite odd, I was expecting arund 10 to 11VDC. So I thought my loadline computation was wrong. Just to verify, I then used a 330R 9 watt resistor connected to both cathode of the PP pair and suprisingly, I was getting 58VDC on the cathode! The 9 watt resistor was getting hot! I thought I have a quad of bad EL84, replaced the tubes with another brand, same thing.

It is 2:40am here, am beat, I will check my work again tomorrow.

This is weird!

Any ideas? I just checked my cathode resistor and it reads 332 ohms on my Fluke digital meter.

[Svein_B]

Where are the grids of the EL84 connected? Are they at DC 0 level?

If the grids are at GND level, and you get that cathode voltage, you must have some serious oscillation.

I guess those EL84 were glowing quite red hot with the high current.

Svein

[richwalters]

EL84 is a high gm tube 11mA/v. Given even a half chance it will oscillate. Got an oscilloscope around ?
richy

[alexg]

The grids are connected to the secondaries of my Lundahl 1660S.

The amp sounds great, btw.

I also noticed this that the paralleled 270R resitors (2 watt) that I first use as common cathode resistor of the EL84 were burned.

Unfortunately I don't have an oscilloscope.

[jueic]

Hi, Alex

How about disconnecting EL84's grids from the transformers and ground them
with 1k resistors.. just to make sure the output tubes function ok by themself.

Jueic

[alexg]

Stupid me, I did not wire the Lundahl 1660S to ground!

Now bias voltage is spot on.

Thanks guys.

[kenpeter]

So, grid leak did not bias the secondary below the cathode voltage???

Where did a positive charge come from? I know that can happen if
heat from the plate cooks a grid until it emits. But how does that
runaway process begin here? I would have expected from my limited
experiments with floating grid, this node tends to bias itself a volt or
more negative. How in this circumstance does the plate ever get hot
enough to cook the grid into emitting more electrons than it captures?

[alexg]

Originally Posted by kenpeter

So, grid leak did not bias the secondary below the cathode voltage???

Where did a positive charge come from? I know that can happen if
heat from the plate cooks a grid until it emits. But how does that
runaway process begin here? I would have expected from my limited
experiments with floating grid, this node tends to bias itself a volt or
more negative. How in this circumstance does the plate ever get hot
enough to cook the grid into emitting more electrons than it captures?

Actually the amp before i grounded the interstage is quite stable, not red plates and it sounds great, no distortion. But the cathode voltage is too high for the bias that I set and the tubes are conducting too much current. I don't know where it was getting the bias.

The interstage is gapped, the secondary should not have any positive charge (or is there?). I think that the secondary should just have A/C on it.

It is quite confusing for me, as a newbie, that when I increase cathode resistance, the voltage across the resistor increases which does not follow the plate characteristic of a triode strapped el84.

[alexg]

Originally Posted by jueic

Hi, Alex

How about disconnecting EL84's grids from the transformers and ground them
with 1k resistors.. just to make sure the output tubes function ok by themself.

Jueic

I did this and the EL84 behaved as expected.

Somehow not grounding the CT of the secondary and the ground of the primary of the LL1660S interstage caused my problems.