New amp build problem

[dearmond]

Hi!
Got a problem with a new amp I just built.
I want to switch between to different pairs of output tubes. Specifically EL84's and 6L6/EL34's. Both sets are cathode bias with their respective grounds going to an on/off/on switch. I did it this way through research and posting it on other forums which nobody questioned it. However, it doesn't work-lots of noise! But when I pull out one of the pairs, it sounds great. I obviously missed something. Can anybody help me? I've attached my schematic.

Thanks!

[Badside]

So if I get this right, the switch lifts the cathodes on one pair at a time?

As long as it's properly wired, that should work pretty well I guess... (Mesa uses the same scheme for their multiple power options)
I'm assuming you double and triple checked the way it's wired?

Also, I find the use of a shared 10K grid stopper odd. I always put them right on the tube sockets, one for every tube. And I start by 1K5 and leave it at that if everything works fine (haven't had to increase yet)
Sharing a 10K for 2 tubes actually makes it appear as 20K if I'm not mistaken, although in theory only one of the 2 tubes work at a time.

Is the lifted tube wrecking havoc in the other one because there is nothing impeding the interference from one to another? Maybe... but I can't guarantee it.
(remember that the lifted tube still has B+ on it and active filaments)

[Phil_S]

1) Put grid stoppers right on the sockets, otherwise don't bother because they won't function as intended.
2) Pull the unused pair of tubes...does that cure the noise problem? If so, then I suggest you need a more complex switching arrangement that shuts off the whole tube, not just the cathode.

What are you doing about the impedance mismatch on the primary side of the OT? EL34, 6L6, and EL84, under "perfect" conditions, each wants to see a different primary resistance.

[Structo]

Also, usually with 6L6 tubes the screen resistors are 470R.

[triodecr]

I have done something very similar only with a quad of EL84s and a pair of EL34s and two switches. I wired it as your schematic shows using the same type of switch that I use for power and standby. Are you sure your switch can handle the current?

I do recall that I used 1k screen resistors on the '84s.

triodecr

[Firestorm]

Since you're leaving the screens on the unused tubes energized and you don't have the resistance of grid stoppers between the used and unused grids, I *think* the unused tubes may be acting like diodes and trying to rectify the signal. Put grid stoppers on each tube (where they should go anyway as Phil says) and see if it helps. If not, change the switch so it both lifts the cathodes AND disconnects the unused grids.

It looks like you have parallel 270Rs and 250uFs on the EL34/6L6 cathodes. That would make a 135R bypassed by a 500uF. Doesn't seem right.

[dearmond]

Thanks everybody for your input.
As far as using 10k grid resistors, that's just a starting point. I'm still learning so I based some of my decisions on what other amps out there have used in somewhat similar setups and start there. As far as the placement of the grid resistors, I'm gonna take your advice and mount them at the tubes. Easy fix. I don't know why I didn't do that in the first place. However, when I pull out one or the other power tube pairs, the amp sounds great! Some tweaks are in store of course, but I'm quite pleased with with the initial results. I'm just frustrated that I can't use the switch with all tubes installed. This my first cathode bias build so as for the values used it's the same reason as above. I've had to upgrade the resistor for the 84's and built a AC line reduction box to lower the wall socket voltage, which helped a lot particularly for the 84's. Transformer selection was a real challenge. I went with the closest impedance match I could find. So far, it sounds ok. I also thought about shutting the grids off too. Just didn't know if that was a good idea. If the grid resistor change, I'll try lifting the grids also. If that don't work, maybe a new switching method to take all power off the unused pair.
Thanks again! And keep the inputs coming!

[triodecr]

I missed the shared grid stoppers on the schematic. I used individual grid stopper resisitors in my amp as Phil and Firestorm are talking about and the switching works fine.

[dearmond]

Yeah. An oversight in my design. I'll get it right one of these days!

[dearmond]

Putting the grid resistors at the tubes worked! Works great! There is some tone tweaking, but it sounds pretty good! I feel like dumb-ass, but I'm happy!
Thanks guys!

[Phil_S]

Putting the grid resistors at the tubes worked! Works great! There is some tone tweaking, but it sounds pretty good! I feel like dumb-ass, but I'm happy!
Thanks guys!

No need to feel that way. This is why we have on-line forums. Someone told me, then I tell someone else. It's how we learn the do's and don't's. This was a solid learning experience. I'm glad it worked out.

BTW, IMO, the dumb-ass is the one who doesn't ask the question.

[dearmond]

Phil, you absolutely right. No dumb questions. I can't thank you guys enough for your input. I'm so happy I've been dancing around the living room!
The overdrive is great, the headroom is great and the volume is great!
I got a slight hum and too much low end, other than that, it sounds good!
I SOO HAPPY!!

Thanks again everybody!

[Structo]

Another success story!