Fixed bias help

[goldstache]

I've got two identical builds in front of me, only difference is the PT and OT brand.  HT measures within 5VDC of each other.  Here's my rub.  I have bias test points for the Push Pull power section.  I've installed duets of the same brand and rated tubes.  Current draw on one amp is 32ma on both tubes.  The other amp is 32ma on one tube and 42 on the other.  Bias voltage is the same to both grids. 
What gives????  When I switch tube sets from amp A to amp B same 10ma off. 
Totally stumped.  I'm using 1 ohm 1% resistors on both amps tested and matched. 
Any explanations or pointers?
Thanks

[Mike_J]

Did you mean the same set of tubes have a 10 mA difference regardless of which amp they are in or that one of the amps has a 10 mA difference no matter which set of tubes are used?

[eleventeen]

You can and will find that kind of difference between same-brand and same type tubes....and more. If they were sold to you as a matched set, they clearly are not. If within one amp you can swap the tubes A > B and then B > A and the current mismatch follows the swap, there is nothing wrong with your amp.


I once bought a garage of electronic parts. In that pile of stuff were perhaps 30 6L6Gs, NOS, JAN Sylvania. Brand new in their boxes. I knew they would NOT be good choices for your typical Super or Twin Reverb because of their low max voltage ratings and in general, rough usage. But I had enough of them so that I figured I'd build a tube matcher that would show static current through two tubes in a way that I could take pix and thus sell the relatively undesirable 6L6Gs on ebay if I could find matched sets. Bot 2 identical new meters but otherwise made the tube matcher from junk parts.   





So, I started on my pile of 6L6Gs. I was rather amazed, they were all over the place, and I mean all over. Same conditions: 22 ma vs 43 ma, giant differences. THat's a solid double. Now I grant you, these were 60 year old tubes; but they were all Sylvania, JAN, identical date codes, whole thing. I'll also grant that tubes made in the 60's/70's are probably made to tighter tolerances and thus ANY old pair probably match better than the above 30-40% difference. But there you go. Arbitrarily, I decided that I wanted any pair to be within 2 mils of each other. Out of maybe 30 tubes, I think I was able to produce only 5-6 pairs. They were really all over the place. 

[goldstache]

The suspect amp always pulls the 10ma difference it seems. 
I walked away, had been wiring it up most of the afternoon.
I will check current and voltage measurements, from PI to power tubes, tomorrow.

I have to make me one of those current devices described above.  Great idea!

[sluckey]

Swap the 1Ω resistors. Does the high current stay with the same resistor?

Swap the OT plate leads. Does the current reading change? (If you have NFB loop disconnect first.)

[PRR]

> Bias voltage is the same to both grids. 

Same *AT* each grid? A leaky coupling cap will pull-up say -45V from bias supply to -40V (or worse!!) at grid.

[goldstache]

I lifted the leg to the coupling cap from PI to power tube grid and meeasured DC.  Nothing funny there.  Changed it out anyway, as well as my 1ohm 1% resistors on the test points measures closer now.
Thanks.