5F6A/8417 80W

[David Root]

Just finished reworking a 4-6V6 30W design that was so-so into a sort of 5F6A tweed Bassman topology w/ reverb and 80 W or so from a pair of 8417s. It's not quite as quiet as I'd like but it is clean until you crank the preamp a bit and then it crunches nicely. The reverb feeds back into the preamp thru an anode mixer, V3 in the schematic.

Uses all octal tubes, 4-6SL7GT, 1-6SN7, 1-6K6GT, 2-8417 and 2-5AR4. Here's some pix and a rather hard to read as-built schematic.

The chassis is 27 x 10 x 2 1/2" and is 1/8" aluminum. It was overkill for a 30W amp and I thought this would be an interesting use of the over-the-top power supply.

Wiring's not as tidy as I would like, but I wanted to get it done.

[Structo]

Interesting.

That's a lot of toobs!

I have often wondered about running all octals like the old days.

Is this an original design David?

Hard to make out on the schematic but are you running your screens off of the choke?

What kind of bias setup are you using?

[David Root]

It's basically a 5F6A with an anode mixer mixing in the added reverb circuit between the tonestack and the PI, but using 6SL7s instead of 12AY7, 12AX7. The 6SL7s and the reverb driver are getting 6.3 VDC, you can't push gain on the 6SL7s unless you use DC heaters, they hum like crazy otherwise.
It goes in a large head cabinet w/reverb tank in it and sits on top of a slant halfstack with four Eminence Delta Pro 12-A PA speakers.

Screens are fed after the plate supply thru a 30H choke, and clamped to about 305 V by four 5W 75V zeners in series. This is belt and suspenders I suppose but I put the choke in before I decided to add the zeners.

8417s are tricky to bias properly especially GEs. I am using JAN Sylvanias which don't wander too much. Bias supply is completely separate, has its own little PT and a BR34 bridge rectifier wired backwards. Running at -15.5V, I have test points across 1 ohm Mills 5W resistors to read the cathode current and a separate adjustment pot for each tube..

There are lots of other oddities about this circuit, some thru necessity.

[tubeswell]

Cooool. Those 6SL7s must sound amazing.

I just finished a "5F6A+" (of my own adaptation) with 'verb and trem - only my verb circuit is like a standalone reverb in that it has a CF bypass stage and is between the vol controls and the DC pair. I'm using a cascaded 12AT7 (with a 1M dwell pot in front of the 1st stage) to drive the RT, and 1/2 AX7 to recover the tank and drive a tone control before going back into the mixer with the CF bypass stage. The range of 'verb is fairly good this way and amazingly the all the controls are fairly dynamic.

(I also built in a 6G9B bias-vary trem which is very cool with the verb).

[David Root]

Mine is similar but uses a 6K6GT reverb driver and a full double triode. It is straight out of Dave Funk's Tube Amp Workbook, page 117.

I have too much gain in the PI I think, 8417s don't need anywhere the same drive as 5881s. When I crank the post-pi master, the amp goes into oscillation. Either that or I need to dump more gain further back. Maybe both, I think I'll try a 6SN7 in the PI, that will cut the gain bigtime.

[C Moore]

Look at all the neat goodies in that second picture, and that bright blue Transformer!!!....

[C Moore]

Look at all the neat goodies in that second picture, and that bright blue Transformer!!!....

[FunkyE9th]

Nice! That's a lot of tubes! I'm curious...How do you like the Edcor OT? Is that one of their off-the-shelf HiFi transformer? I've been thinking trying their stuff out.

[David Root]

Yup, that is their CXPP 70, 3k5 primary. Rated 70W 20-20kHz <1dB down. It weighs 11 lbs. It is a hi-fi OT all right, but $92 plus shipping is a good deal! A replacement BF Twin OT costs that much. (I'm in Canada so by the time I paid $28 shipping, turned it into Cdn $ and added $18 in taxes, it ended up at about $145 Cdn.).

I haven't been able to crank it yet because the amp squeals when I crank up the post PI master. I think that is because there is too much gain in the PI so I am thinking of changing the PI out to a lower gain tube like a 6SN7. I changed out V3 this morning to a 6SN7 and that cut the gain of V3 from 22 to 12.

Changing the PI to a 6SN7 means redesigning the entire PI too. I wish I knew what the gain of an 8417 is in pentode mode. The Sylvania data sheet shows 16.5 in triode mode but does not list it in pentode mode.

Maybe I have the OT plate leads out of phase with the NFB, but in the past that has caused oscillation immediately, not when I get above 10 W or so.

Any helpful observations appreciated!

[FunkyE9th]

Could be lead dress? Try moving some wires around by the PI and see if the frequency of the squeal changes, or if it makes it squeal at different PPIMV level.

I had problems with an EF86 squealing when the volume and tone gets to a certain level. I scratched my head for a few days trying to figure out how to fix it. I fixed it by shielding the grid and plate wires.

Regarding the gain of the 8417, given that you have the amp built, why not just measure the gain of the 8417? Put a 1KHz tone measure it at grid and then at the plate.

[David Root]

I tried that with a 400Hz tone and there doesn't seem to be any.
Problem is I can't put more than a few hundredths of a volt into the 8417 grids before they squeal. Maybe I have the OT plate connections out of phase?

I have shielded a lot of the early signal lines V1, V2, V3.

[FunkyE9th]

Why not temporarily disconnect NFB and see if you still get a squeal?

You can use a dummy load so that you don't have to listen to the squeal. Then measure the gain quickly coz that dummy load will get hot quick with the amp squealing.

[David Root]

Thanx Funky,

I checked, it doesn't oscillate w/out NFB, so I switched the plate OT leads and reconnected the NFB, and no oscillation.

I then measured the power tube gain, it is 22! Usually we are used to 6-12 in the common power tubes we use. So I don't need another 16 gain in the PI, which is what I have.

25.3VAC on the OT secondary 8 ohm tap is 80W RMS with a pair of 8417s, so that means (3500/88) exp 0.5 = 20.9. 20.9 x 25.3= 529VAC at the plate, 529/22=24.1 VAC from the PI output. The PI gain is 16, so 24.1/16=1.5 VAC into the PI for 80W out of the OT. With about 8dB NFB that becomes maybe 2V or so at the PI.

If I use a 6SN7 PI instead of the 6SL7 and redesign, I can drop that PI gain to about 8 using 39k/47K plate resistors, so I think I will try that.

Once again all comments appreciated.

[Merlinb]

The right side of your common anode mixer doesn't appear to have any grid leak?

Also, you seem to be mixing wet and dry reverb in the anode mixer, but you are also feeding dry signal forward through the 3M3 resistor? Is that deliberate? I'm sure there's some positive feedback in there...

[David Root]

You're right Merlin. I finally figured out what was wrong yesterday, after splitting plate resistors on V1, V3 and V5 to no avail. So I disconnected the reverb and the problem disappeared. I do need to put that grid leak on V3B before I reconnect it though. It's funny how you can look at a schematic hundreds of times and not see something so obvious is missing! Must be Oldtimerz creeping up on me. The dry signal feeds into V3B out of phase with the wet signal, I will probably just eliminate it.

Part of the problem is this was a rebuild of a previous design and I kind of took the reverb for granted because of that. I will review it and simplify it.

I have been rereading and referring to your book thru all these tribulations, so I now have 470k gridstoppers on the main preamp tubes including the anode mixer, and I slugged the power tubes with 150 pF from grid to cathode.

The amp sounds nice and clean and I have to put the gain back in V1 to get the 80W back. It should be around 50. I cut back V2 to about 23 gain and it now has 68k plate and cathode resistors, still has just over 205VDC on that CF cathode so I put in the 100 ohm resistor and the arc protection.

I cut V5 down to 2.2 gain by splitting the plate resistors in the ratio of 1:4. I think I will leave that because the 8417s don't really need any gain from the PI, I measured them at 22. They are quite stable and the bias doesn't wander about as I had been told it might.

Your book has been very helpful to me. Thanks for all the work you put into it. One suggestion for the 2nd edition--you need a better proofreader.