New project: failure. Need advice about strange b+ behaviour

[naciketas]

Hi,
it's a new build: a 30W 6v6/el84 Pa with a top Boost/ef86 dual channel preamp.
To put it simply, whenever the screens are connected to b+2 node, the whole b+ drops down by a large amount: for example b+1 from about 440v down to 230 all the rest following accordingly.
Tried a different transformer, screen resistor (1k), new caps. Now I'm lost and no one around that could help me.
Any advice will be deeply appreciated.
Giovanni

[John_P_WI]

Assuming it is cathode biased? What resistor value do have on the cathodes of the power tubes?

[naciketas]

Thanks, yes it's cathode bias with one 220R (bypassed with a 2200uf cap) for each couple of 6v6/el84.

[Phil_S]

I think there is a short or you've installed a resistor of incorrect value somewhere that is causing a high current draw. You may have installed a 100K where you need a 1K or something like that. You may want to meter every resistor to see if you have a misfit. This may require lifting one end if you can't otherwise isolate it by pulling the tubes. I'd start with plate load resistors, screen stoppers, and B+ dropping resistors -- that's not in any particular order. You may also want to inspect using a magnifying glass for an unintended whisker making unwanted contact.

Can you post a photo and a schematic?

[naciketas]

Sorry, but I can't post any useful photo, The voltages go wrong only when the screens are in the circuit, so I assume the problem lies in that part, double tested the screen rez and it's 1k, and all the b+ dropping resistors.
By the way every signal triode (and pentode) is filtered and connected in parallel to b+3. Thanks

[John_P_WI]

Try disconnecting the power tube cathode caps.

If the problem persists, try using 1 pair of power tubes, first in the inner sockets, then again in the outer sockets. See if you can narrow it to a socket.

Do you have grid leaks on the power tube grids? What value? Master Volume? How are the coupling caps from the PI?

Are the tubes red plating?

[Firestorm]

If the screens aren't connected, pentodes and tetrodes will not be able to conduct very much, so I would look at the output tube wiring and connections. Are the grid loads of the correct value? Are they actually grounded well? Is there a short on one of the sockets?

[naciketas]

I'll disconnect cathode caps tomorrow, i'm about to go to bed.
I tried one pair of tubes at time, inner/outer sockets same results.
470k for the grid leaks to ground via tremolo's intensity pot ( I tried with alligator clips to bypass the pot ) and .022uf coupling caps. No master volume. PA tubes layout is 6v6 el84 ecc83 (PI) el84 6v6.
Thanks

[John_P_WI]

Probably not a shorted cathode cap IF the symptoms still showed when testing the individual pairs.

[M Fowler]

The layout is a little hard for me to follow but I tried to look the photos over but not able to figure out a solution for you.

Dumb question but you are running the 6V6 in pairs right same with EL84's in pairs.

[naciketas]

There's no dumb question when dealing with me.
The push pair is 6v6-el84 as the pull pair, as outlined in the layout attached.
Ciao

[Phil_S]

I don't see how that drawing can be right. It shows screen supply on both sides of the 67K B+ dropping resistor. Well, OK, maybe you are supplying lower B+ voltage to the EL84's. At least add to the drawing the bit that shows B+1 and how you connected it to B+2. Is there anything to separate the 6V6 screens and plates besides the 1K stoppers?

Can you tell us, with tubes in the sockets, how much voltage drops across each one of those 1K resistors? Simple math, applying Ohm's Law, will tell you how much current each screen is drawing. I think that is a good place to look for a hint.

Also, how much voltage drops across the 67K resistor? That strikes me as a rather large size dropping resistor, but you didn't show the whole circuit so it is hard to know if that means much.

Drops across means put one probe on each side of the resistor and read DCV on the meter. Forgive me, if I am telling what you know. I don't know your level of experience.

[John_P_WI]

Sorry, I don't have time to verify the pin outs on the sockets, but I think I would start by pairing the 6v6's cathodes and EL84 cathodes as pairs. That way you can bias the 6v6's and EL84 pairs separately. Also, after looking at your drawing I would disconnect the cathode caps to ensure that one of the caps is not shorted.

Unfortunately, your drawing does not list all of the caps, resistors, and connections from the PI.

[naciketas]

This is really a great forum! There's nothing to forgive. I appreciate your inputs. The 47k res is there to lower 6v6 screens voltage so that they have grid curves similar to the el84.
B+ is the same old recto to reservoir cap (B+1) connected to primary ot ct and 1k5 dropping res which is connected to a 33 uf cap (B+2), from here there are one wire to el84 and the the 47k res to the 6v6 screens. Now i'm at work, i'll take measures asap.

[M Fowler]

That 47k resistor doesn't seem right where it is and question if it is needed.

No grid resistors shown in drawing. The 6V6 are paired with it's own bias circuit. Same for the EL84's paired with it's own bias circuit.