Re: EL34 power amp project
 

Now then...

That's different, as they say. It's not unknown for EL34's to do this, even Mullard ones. I guess it depends on the brightness of the hotspots.
I've certainly got one or two originals that are completely stable and which measure within "new" parameters (even as measured hot whilst working in the amplifier) that show a very dull but even red glow on one side of the anode in the TL25plus in a darkened room. i.e. You cannot see small rectangular blocks corresponding with where the electron beam hits the inside of the anode. (Corresponding to the blue patches seen on the inside surface of the anode).

I guess the Leak DC conditions just push the valve towards dissipation limits, and if there's a shortfall of current on one side of the cathode then the other side takes up the slack. The rating of the valve presumably accounted for this possible imbalance.

However, if you can see discrete (especially just one or two) individual bright blocks, I'd suspect it was due to poor valve construction.

Re: EL34 power amp project
 

The only time I have seen Mullard EL34's redplating is in origional BIG FAT FACE amps running 800 volt plates and 400 volt screens.

With regard to anti oscillation caps, ALWAYS use the best ceramics you can buy. NPO being recommended!! They are all of sixpence a dozen!! I kid you not.

The phenomena Dangerman describes is due to "shadows" cast by screen grid wires, and is 100% due to lazy/poor/bad construction. The "blocks" being caused by free electron flow where the screen doesnt "sheild" the plate.

Its kinda hard with self bias/auto bias / cathode bias, but a really excellent way to stop EL34/6CA7 from red plating, is to tie the suppressor grid direct to hard negative bias voltage. By this I mean,
when fixed bias is used, there is a negative supply available, usually "about" minus 50 or 60 volts, Without ANY resistance in circuit, the suppressor grid/s are connected directly to this negative supply!!!.
IF the valve is correctly biassed on the control grid, and the valve is overdriven, it can NEVER dissipate MORE than its maximum plate dissipation.
IF the bias voltage is lost, i.e. total failure of bias supply, the valve will be toast.
I have changed them out of Marshall amps, where the vacuum has "sucked" the glass envelope onto the plate!!!!!! when the glass melted !!!!!

Another useless piece of information for you :-D

Joe

Re: EL34 power amp project
 

Thats the phenonemon Dangerman, a diffuse red area rather than a focussed hot spot. Everything measures fine. I think when I get the chance I'll putthe extra measures in place, and see how it goes from there.
Could be a while though as out of the blue we had an offer for the house on Friday and so we could be busy packing stuff away in anticipation of a move in the next 4 months or so.

A.

Re: EL34 power amp project
 

a quick one:
I confess to a bit of a lack of understanding on this one.
I have an inkling that the EL34 gridleak resistors on this amp might be a bit too high in value.
I used the values as per the leak TL25+ circuit, these are specified as 680k.
I suspect this was to squeeze as much as possible out and reading various tomes on the leak products seems to suggest HJL ran the valves at or beyond their normal operating limits?
Mullard specify 470k on the 5-20 and i wondered if using this value or maybe even down to 330k would ease the load on the EL34's?
I am not worried about a reasonable reduction in output power as I measured 37W into 8 ohms when I had it on the scope when first built. In fact getting them down to 25-30W might be an advantage in terms of valve life?
Thanks guys.

Andy.

Re: EL34 power amp project
 

Reducing the grid leak resistors won't affect the output power but means the driver stages have to work harder to achieve the same output as before. Also, the loop gain falls a bit.
Net result is that even at normal listening level there is a little bit less feedback and hence slightly more distortion due to that reduced feedback. But only a little bit.

There's no harm in temporarily bridging another 680K in parallel to see if it makes any difference to the overheating. Or you could reduce the dissipation without affecting the gain by slightly increasing the cathode resistances, especially if reducing the grid leaks makes no difference.

Pete