Common Cathode Resistor Bias Question

[CANT]

Ok, I have a Sherwood S5000 (7591) that has been modified to allow for individual tube bias/balance adjustment but still keeps a common cathode resistor per pair. The amp came with a matched quad of JJ 7591S and I was told that the bias was set with a spectrum analyzer for the best distortion figures. When I inquired as to how I would need to bias the amp for a new tube set I was told...

"We critically set the bias when we restore the amplifier to best suit the output tubes to the specific circuit in the amp. If you need to replace the output tubes, any new closely matched pairs of output tubes will work just fine in the amp with the bias settings as-is. It will be closer to having the best harmonic signature than setting the bias point to the factory setting, which will have greater odd-order harmonics and a higher bias setting."

I have a matched quad of TS 7591A that I thought would be fun to try so prior to installing them I checked the current bias voltage settings and got, left to right, -18.3 /-20.6/-19.4/-18.5

I measured 0.7V across both 8.5ohm (schem shows 6.8ohm) cathode resistors which does look correct (41mA) for what the JJ's are marked (43mA) but was curious as to why the neg voltage values were so all over the place? Does this even look right? The schem seems to say -19V is the correct value for all 4 tubes, though that would be pre mod.

I did try throwing the TS quad in after taking the initial measurements but the amp cooks them at around 50mA. So the "any new closely matched pairs of output tubes will work" is obviously not exactly true... The research I've done seems to say that this could be a common issue with the TS specifically so I am debating on just giving up on them and leaving everything as-is? But what would be my best strategy for how to bias for the TS if I wanted to use them and don't have the ability to perform the task the way the original shop did?

Hope this all makes sense... Thank you for any assistance you may be able to provide

 

[gadget73]

If it has individual pots, the easy thing to do would be to install a 10 ohm resistor per-tube and simply adjust the bias on each tube to get it correct and equal across each resistor. Having 4 pots and only 2 resistors can work but its a much less convenient way of getting it done. It can be done without a spectrum analyzer and a lot of fancy words too. The old way was to simply adjust for minimum hum at the speaker. When the current is equal, the power supply noise cancels and its quieter. Those usually only had a balance pot, which is what it looks like Sherwood provided originally. No adjustment for current, you bought Sherwood-matched tubes from your neighborhood Sherwood dealer and the current would be correct for the bias circuit in the amp. In the modern world you can't do this, hence the individual bias controls and test points. The Russian tubes in particular need more negative bias voltage than old production do. I don't really know much about JJ tubes in particular but I'm under the impression they bias similar to old production.

If the grid voltages are that different, that tells me the tubes aren't especially well matched for idle current.

If thats 41ma per-pair that seems low to me for a 7591. Somewhere in the 30-35ma per-tube range is more typical. 0.7v across a 6.8 ohm would put it more like 50ma per-tube which is too hot for the voltages in this.

 

[Tom Bavis]

R95 (4.7K) could be changed to a 7-10K pot to allow bias adjustment. 10 Ohm 1% resistors in each cathode would allow measuring actual idle current. The TEST switch introduces 3V at the common cathode point; adjust BALANCE for minimum hum on each channel. This may be better than matching idle current exactly, as it matches gain, should minimize distortion. But matching currents may give the best bass response. Proper bias voltage is whatever it takes to get proper idle current.