White Amplification Little Rock

[Ted]

Anyone come across one of these?

4 x EL84
4 x ECC83
1 x ECC81

I believe its similar to a 70s/80s Matamp.

The valves are all very tired, barely showing anything on the emission tester. The output stage appears to be cathode-biased and ultra-linear. The cathode resistor is a blackened and unreadable 3W resistor which measures around 1K. This seems very high for a quad of EL84s. It has also cooked all the surrounding electrolytics - one of which appears to have a small zener soldered between its -ve leg and the pcb.

I haven't powered it up yet...

[1big1]

there's a store near glasgow that might be able to help you figure out what you've got. TwoTribesMusic. hope that helps.

[Ted]

Put in some old tested EH EL34s Powered it up - that cathode resistor glowed and then popped.
Tried a few other values 560R - gave me 42V on the cathodes which gives a combined dissipation (anode + screen) of about 9W per valve - which is safe enough for the time being.

Ran it with a sig gen into a dummy load - only seeing 12W into 8 ohm with 30V p-p drive to the EL34s - anodes began to glow on one side so I turned it off. Measured dc resistance of O/P transformer - CT to screen tap on each side 98R - CT to Anode 150R on one side and 108R on the other. Bad. Very bad.

[Robert M. Martinelli]

Hi Ted,
Sorry, but in your "thread starting" post you stated the output tubes/valves to be EL84s, then in the following post you referred to EL34s - I guess that's a typo....are they 34s or 84s?
Cheers
Bob

[Ted]

Sorry Robert. Finger trouble. They are EL84s. And they appear to have 520V on the anodes...

[Robert M. Martinelli]

Hi again Ted,
520 VDC at the anodes seems way too high IMHO; Don't have a schematic so I can't tell you if this is to be considered "normal" ( for that amp ); usually, EL84s work at lower voltages, in the 300 VDC range ( up to 400 in some amps )....It could be that the designers decided to run the voltages this high, but, if this is the case, I would have expected to find a "fixed bias" arrangement ( to get the most out of it ), not a "cathode bias" one, even though I guess it's fully bypassed ( BTW, is the bypass cap OK ? ).

As to the output transformer, a certain difference between the windings in terms of DC resistance is usually considered normal, due to the way OTs are wound; should you have any doubts concerning the OT's condition/symmetry, just remember that a quartet of EL84s need the primary impedance to be in the 4KOhm range. With an 8 Ohm speaker impedance, the OT ratio needs to be around 22 ( (4000/8)^0.5=22.36 ) and this can be easily tested with the amp off ( and don't forget to discharge the filter caps before working on the amp for safety's sake ), by disconnecting the speaker/dummy load, applying a 1 Vpp signal at the OT's 8 Ohm secondary and checking that the signal's amplitude at the primary is in the 22 Vpp range ( 11+11 Vpp with respect to the CT ) - UL taps are usually in the 43% range so you can test them the same way.

Hope this helps

Best regards

Bob

[Ted]

The first power supply capactitors are a pair of 100uF 450V caps in series, so it looks like it must have been running some pretty serious HT voltage.

It seems to be normal - I'm seeing 6.9V on the heaters so it might be a touch high from using the 220V transformer tap on our local 230V - I'll investigate.

[Ted]

Bob - thats a great way of checking O/P transformer ratio - thanks! i tried it and got a well balanced 18:1, with UL taps equally well balanced.

That gives an anode to anode load of around 3k with an 8 ohm speaker. On looking at the back plate I noticed that it says 16 Ohms by the speaker socket - and it means it!

With a 16 ohm load I'm seeing a pretty dirty 35W out. Now to work on that bias resistor....

[bluesfreak]

I would split the Quad of power tubes into 2 pairs and give each its own cathode resistor and capacitor analog to the Orange OD30

[Ted]

1K as a cathode resistor seems to work.

Cathode voltages settle to around 60V. Which is around 15mA per valve which is about 7W at idle.

The cathode resistor is dissipating around 3.6W at idle and its tucked between one of the main smoothing caps and its own bypass electrolytic - I'm thinking of replacing it with one of those aluminium clad bolt on resistors and moving it off the pcb and on to the chassis

What is the advantage splitting the Cathode resistor like that OD30 schematic?

[Robert M. Martinelli]

Bob - thats a great way of checking O/P transformer ratio - thanks! i tried it and got a well balanced 18:1, with UL taps equally well balanced.

That gives an anode to anode load of around 3k with an 8 ohm speaker. On looking at the back plate I noticed that it says 16 Ohms by the speaker socket - and it means it!

With a 16 ohm load I'm seeing a pretty dirty 35W out. Now to work on that bias resistor....

Hi again Ted,
glad I've been able to help, and, yes, that method is fast and easy enough...

35W seem in the ballpark for a quartet of EL84s, I don't think you'll be able to squeeze much more out of it, unless you switch to 7189s with a fixed bias. I've also been using Russian 6P14Ps with good results, they're very similar to 7189s ( mil-spec version of the EL84 ), their Wa(max) is 14W and, last but not least, they're pretty cheap .

Cheers

Bob

[Alex R]

All the Little Rocks I've seen (not a lot - two over the last few years) ran Gold Lion KT77s from new. I might expect a 1K cathode resistor on a quad of those. EL84 cathode resistors are usually lower values. Maybe ultralinear wiring changes things. Unusual amp, a schematic would be interesting - but I had no luck finding the KT77 Little Rock schematic.

[Robert M. Martinelli]

Ted,
I think you can safely lower the cathode resistor's value ( hotter bias ). 60 V across the 1 K cathode resistor means that 60 mAmps are flowing through it as you correctly stated, but the 15 mAmps contribution from each tube is not entirely due to the sole plate current, you have to take the screen grid current into account as well ( and subtract them from the cathode current reading ). I expect the screen grid currents to be in the 4-6 mAmps range on each tube, this means that only some 10 mAmps are now coming from each plate, so the actual quiescent power dissipated by each plate looks to be in the 4-5W range, ( which is very cold for EL84s ) rather than 7W ( still pretty cold ). I'd say some 9W would be good enough without the risk of prematurely killing your EL84s by biasing them too hot.

HTH

Best regards

Bob

[Alex R]

- further to that here's a brit amp that runs 2 x el84 ultralinear with a big cathode resistor.

[Ted]

Ted,
I think you can safely lower the cathode resistor's value ( hotter bias ). 60 V across the 1 K cathode resistor means that 60 mAmps are flowing through it as you correctly stated, but the 15 mAmps contribution from each tube is not entirely due to the sole plate current, you have to take the screen grid current into account as well ( and subtract them from the cathode current reading ). I expect the screen grid currents to be in the 4-6 mAmps range on each tube, this means that only some 10 mAmps are now coming from each plate, so the actual quiescent power dissipated by each plate looks to be in the 4-5W range, ( which is very cold for EL84s ) rather than 7W ( still pretty cold ). I'd say some 9W would be good enough without the risk of prematurely killing your EL84s by biasing them too hot.

Bob

I undrstand all that, but the fact that the scorched remnant of the original resistor measured 1K seemed like a good starting point. (I know that its not unknown for resistors to change value under extreme thermal stress)

I tried 560 ohms and although it was stable at idle a couple of the valves went into thermal runaway within 30 seconds when I ran it at full power.

Out of interest, I tried 150 ohms and the EL84s anodes were glowing within a minute at idle...