CT/Reservoir caps not grounded

[Smokebreak]

I'll post a report later, but I'm working on a '48 Epiphone Electar Zephyr(cathode biased 2 x 6L6, 6SQ7, 6SJ7, 6SN7) that I acquired a couple years ago, before I was into building. It was always in and out of the shop, and never was quite right, so in the corner it sat for years. I fired it up a few months ago, it sounded like the (field coil) speaker blew, mid-playing. That's another issue I'll address when I get to it..

So anyway the amp has "worked" over the years, and I finally opened it up today. I can't find a schematic anywhere for this amp(multiple versions were made) with this tube compliment, so I'm tracing it out. As I'm checking continuity, I find that the star ground that holds the PT CT, the reservoir caps, and the 6L6 cathode resistor are tied to the old ground lug(NOT connected to chassis) of the old cap can, that was replaced with individual, new filter caps a few years back.
There is no continuity to chassis ground.
The preamp filter caps ARE connected to chassis ground, which I believe rules out any floating ground scheme. Also, the 100K(??) grid bias resistors terminate to one another, and go nowhere...I see no diodes in this amp, so it couldn't have been converted to fixed bias

How is this even possible that the amp still produced sound?

I'll produce a schematic that should help.

[ampgeek]

I love headscratchers like this....I think!

No diodes observed but no mention of the rectifier type in the tube line-up.

What type of rectifier does it have? If FWB, the center tap of the PT "floating" is OK.

Cheers,
Dave O.

[Smokebreak]

Hey Dave, my bad, technically it is rectified by a 5U4GB (diode )

There are no SS diodes to produce any neg voltage for a fix bias conversion that may have occurred in the last 65 years. To make things interesting, the 6L6 pins 7,8,1 are tied together and physically connected to a ground lug on the tube socket!. So heater, cathode grounded. No centertap on the 6.3V winding, so I could see why they'd put the heaters on the cathode....if it was cathode biased, but with the cathodes tied to ground. I'm stumped. I'm just about done with the scheme and I'll take some pics.

[ampgeek]

Yep....headscratcher...for sure!

Tying one side of the filament heater string to ground isn't too unusual. I have seen that in some older Fender amps that functioned well.

However, no diodes for a fixed bias configuration is unusual. Could there be one of those old selenium rectifiers hiding somewhere? They don't look at all like our friendly 1N4007s or their close bretheren.

Quick question: Is the amp functioning well enough that you would be comfortable with firing it up and measuring some voltages?

That could go a long way in trying to understand what is going on here.

Dave O.

[Smokebreak]

Here is the traced schematic and pics.

To review the history, I got the amp a few years ago, before I had ever seen the inside of an amp. It worked for a day, sounded great, and the rectifier blew, I learned after taking it to the shop the next day. Got it back, it had a cap job done and tube job as well, and it never sounded the same. Loss of power, weird noises, etc. It made several trips to the shop until I just gave up on it. I fired it up a couple months ago and it played for a while(crappily) and then it sounded exactly like a blown speaker. Back in the corner until today, when I popped it open for the first time.
It has a FC speaker with the smallish OT mounted on the basket. I only intended to replace it with a PM speaker today, and a choke or large resistor to sub for the field coil. ifigured what the heck I'll trace out the circuit, as the only schemes I can find have all 6SJ7 tubes.
That's when I came across the 6L6 cathode issue, and possible grounding issue. The 6L6s pins 7,8,1 are all tied together and ground and the lug on the tube socket. This amp is supposed to be cathode biased, and there is no fixed bias circuit here.

Trying to determine what components were supposed to be grounded, and which were supposed to be tied to cathode, in a cathode biased operation, was difficult, as some things were grounded at the cathode, when they should be grounded somewhere else....because the cathode shouldn't be grounded. I hope that makes sense. Anyways, I disconnected the cathode from ground, to see what was what.

All the red circled components had continuity, but were not connected to chassis ground. The green circles had continuity, and were grounded to chassis, but only because the 6L6 cathodes were!

The OT/FC uses a plug/jack to make the runs to the chassis. I have no idea what that 22K from the 6SN7 is doing. I'm also at a loss with the 6SN7 cathode to 6L6 cathode network, but that 56K is definitely drawn correctly. Of course, again...it was grounded, as the 6L6 cathode was grounded. ugh.

I just realized I didn't draw the input, but the controls are at the top of the cab, with another plug run to the chassis, but i did trace it on another piece of paper, and it's straight up 68K>volume>tone>cap. Grid leak biased.

So...is this circuit making sense to anyone? I can't believe it ever made a sound. I figure I can ultimately just wire it like a "normal" cathode biased amp, but i wanted to see what you all thought. Then I can deal with the speaker...if that's even the problem

[Smokebreak]

Yep....headscratcher...for sure!

Tying one side of the filament heater string to ground isn't too unusual. I have seen that in some older Fender amps that functioned well.

However, no diodes for a fixed bias configuration is unusual. Could there be one of those old selenium rectifiers hiding somewhere? They don't look at all like our friendly 1N4007s or their close bretheren.

Quick question: Is the amp functioning well enough that you would be comfortable with firing it up and measuring some voltages?

That could go a long way in trying to understand what is going on here.

Dave O.

Ok reconnected the cathodes to ground and fired it up. Took me 10 minutes to balance the chassis upside down with that super short OT plug!

There is indeed -21V at the 100K grid bias junction, and -22V at the grids. So fixed bias it is. I unscrewed the board as much as I could, but couldn't find any rectification.

Here are some other pertinent voltages:
6L6 : plates 315, -22 grids
6SN7 : plates 224(pin 2),36(pin5) yikes
cathodes 36(2), 93(6)
6SJ7 CF : 232 plate, 35 cathode
6SK7 plate: 23V ouch

B+ off rectifier is 435VDC,
"A" node(screens, OT CT, after FC) I've got 320V. Is a field coil supposed to drop that much??
"B"node is 154V
"C" node is 232V(470K Rp gives that wicked 23V there)

[Smokebreak]

Aha. I believe back-biasing is used to generate the negative voltage .
http://www.aikenamps.com/index.php/what-is-back-biasing
The original 150r cathode resistor (the only thing i replaced today)was used as the back-bias resistor, and the grey cap(reverse oriented) is the smoothing cap, and the 39K was "set" as opposed to a pot, to set proper neg voltage.. I think.

[Smokebreak]

I'm replacing the field coil speaker with a PM speaker. I've got everything traced out from the chassis, through the 6pin plug, to the output transformer(mounted on speaker), as well as the field coil connections.
I'm going to try a huge power resistor there.
The only thing I'm not understanding, looking at the schematic and having double checked my tracings, is the large(1W) 22K CC Resistor that comes off the cathode of the 6SN7. There is about 2V there. The other end of it winds up at the + side of the speaker(8ohm tap of OT).
Any idea as to what this accomplishes?

[ampgeek]

Nice! I have been pouring over this in the back of my mind and would have never come up with that back biasing concept.

What triggered you to investigate that approach? My conceptual troubleshooting "approach" needs some more fuel!!

I believe that the 22K between the 6SN7 cathode and the 8 ohm tap is a rudimentary form of NFB. I have seen something like that in a few old Fender schematics.

Rock on,
Dave O.

[Smokebreak]

Yes! NFB. Good call!

I was staring at the schematic, meter in hand, and realized that the neg voltage trail stopped at the "?" at the 39K going to the bias leaks, then found continuity to the CT. At first I thought "maybe it's a non-CT tranny with a bias tap?" Well that would be rare, so I googled "Non-centertapped fixed bias", and somehow that pulled up the Aiken article, which I then remembered reading years ago, but no had real use for at the time. All the components whose destination didn't make sense were in his schematic, more or less. The concepts were there, at least..
I believe the circuit that someone(?) put in here years ago is even a cruder form. Here's a quick drawing of the exact circuit I have.

[Smokebreak]

Ok I got it running! I subbed a 300R 10W for the field coil and installed an 8Ohm speaker, though I haven't tested the OT for the correct impedance. Now I'll see what sort of power it's actually dissipating and how much voltage it's dropping compared to the FC, and what the 6L6s are dissipating, then try it with cathode bias. First I've got to figure out where the OT is gonna live! Sounds pretty good for starters, real warm and honky https://soundcloud.com/jeremyslemenda/e ... tar-zephyr

[Smokebreak]

I didn't intend for this to be a rebuild thread. Just wanted to replace the damn field coil! The more I get into it though, it's becoming a great learning experience, so I'll post my findings. Maybe it will help someone else out in the future. Plus it's a great opportunity for me to learn about and tweak pentode stages.

It took me hours to get a sense resistor on the cathodes of the output, as the cathodes, one side of the heaters, and pin1 were all tied together, and all used for tie points for grounding. After sorting out the rats nest directly under the board, I ended up installing a new heater run and moving connections around, as it turned out one side of the heaters actually came in on pin 8 of the first 6L6.

So, 1ohm resistor goes in, and both tubes are pulling 200mA at 340Vplates! Assuming closely matched tubes, that's about 31W dissipation each. The amp is nowhere near as loud as you'd expect with 2x6L6 in fixed bias, and I imagine it's because of the lowish B+. The 5U4 is putting out 410V, and the 660ohm field coil dropped it to ~300V for the plates. With a 300r 10W in place of the field coil, plates came up to 340, and now I have a 100r, and plates are up to 380. I'm just trying to get as much clean volume as I can out of this thing.

I'm gonna get the fixed bias right, and convert the input 6SK7 to cathode bias, then covert outputs to cathode bias, and observe changes.

As I'm moving into circuit analysis a bit, can someone explain why the designer put the CF in the middle of the circuit? Im having a hard time understanding it's usefulness there. Thanks!