Delta Blues bias riddle

[firemedic]

This Peavey Delta Blues I'm working on was (badly) redplating 2 of the 4 EL84s. I finally opened it up & extracted the PCB. Both output gridleak resistors and all 4 gridstoppers have the correct resistance, 220k/47k(!), and I replaced both PI coupling caps to rule out DC leakage. Nothing looks burnt or melted.

Without any tubes in, I still get -16v on two grids (V4, V5), which is nominally correct, and -9v on the other two (V6, V7). With no current flowing in this amp, my tiny human brain can't figure out this discrepancy. I'm on the verge of buttoning it up & telling the owner to find a real amp tech. Unless someone else has a better idea.

 

[alerich]

I would just start at the grid pin of the socket of either of the redplating tubes (with tubes removed and power applied) and trace the path back through the bias circuit witha voltmeter until I found the -16VDC that is getting to the other two tubes. Once you find the spot that is -9VDC on one side and -16VDC on the other that should be the culprit.

[firemedic]

Yeah, thats the next step I guess. The way these Delta Blues amps are put together, the PCB has to be completely removed from the chassis to test components under power. It's a delicate operation.

[alerich]

I hear you. I don't do amp repairs for a living but I do help out my guitar player friends from time to time when I can. I am compiling a list of amps I just won't touch. There are several Peavey models on that list.

[guitardude57]

Did you check the tube socket tension on the pins?

I've seen bias run away, because of loose tube sockets.

Can make ya pull your hair out... I don't have enough to loose any more :)

[firemedic]

Nah, I'm getting a good connection with the socket w/ my DMM. It's weird, I get EXACTLY -9.81VDC on both sockets, none of the usual up & down business.

[sluckey]

The problem should 'logically' be on the board with the power tubes, since the bias comes to that board via a single jumper then splits to go to each set of power tubes. R49, R52, and R55 are the grid resistors for V6 and V7.

Turn the amp off and measure resistance from each grid (pin 2) to ground. Readings should all be about 300K.

Measure the voltage directly across R49. Compare to the voltage drop directly across R45. There should not be any voltage across either of those 220K resistors.

Report the results of those readings. BTW, I'm looking at this schematic. If this ain't the right one please provide the correct one.

    http://www.el34world.com/charts/Schematics/files/peavey/peavey_deltablues.pdf

[firemedic]

That's the one. This thing is such a PIA to work on.... I'll get back w/ results as soon as I'm done w/ the chores @ the house. 

[firemedic]

Ready? Begin.

I managed a hot check on the PCB without shocking myself. So far.

This PCB, as you have probably already guessed, has a lot of plug-in connections. The bias voltage looks good @-16v on all 4 sockets when the bias power (duh) and PT secondary are plugged in. OT primary plugged in, still good. OT secondary.... GAAAAH!

The same sockets are back down to -9.8vdc.

There's a wide ribbon connector to the aux. speaker/footswitch assembly that has abraded the V7 grid trace with its edge. I clean up the very light abrasion on the traces. I line the edges of the ribbon with electrical tape, then put the guts back in the chassis (which must be done to connect the jumpers). I'm extra careful to keep that ribbon off the board.

No change.
I'm not too proud to admit my relative inexperience. For a simple amp this circuit sure is complicated.

[Danskman]

Hi,
IMHO, you should:
- double check PCB traces around V6/V7
- redo the copper line(s) if any doubt, with small wire (22/24, depending the copper trace)
- do the voltage check WITHOUT any power tubes, note on piece of paper and publish them here
- have the 4 EL84 replaced and redo the voltage check with fresh new tubes, publish them here
Then, we should be able to help you further.
PS: did the two red-plating EL84 red-plated in V4/V5 sockets, too?
BR,
Danskman

[VMS]

Maybe the CR6 diode is bad.

Maybe someone here with more experience should confirm this first, but I think you can remove the diode and try the amp without it.




 

[firemedic]

I haven't put tubes in since there is no point: that bias voltage will redplate them in short order.

So I don't think the CR6 diode is the problem. Or the tubes.

Somehow, connecting the OT secondary creates a stable & resistive path to ground, from somewhere between the 220k gridleak and the 47k gridstoppers. Which drops the bias voltage. Just on V6 & V7.

One big problem is that I can only test so much with the PCB out of the chassis. When it's in the chassis, I have no access to the test points I need. I hate this amp. It had better sound good when I'm done.

[plexi50]

footswitch assembly that has abraded the V7 grid trace with its edge

Pin 7 isnt the grid. It is part of the high screen voltage.
Edit: Pin 7 is the plate voltage

Pin2 is the bias grid and pin3 is the cathode connection for all 4 tubes.

Is there continuity on all pin 3 traces to ground?

[HotBluePlates]

If the ribbon cable abraded one tube's grid trace, we might assume the paired tube is also impacted, at least indirectly. Which would explain why you have 2 tubes reading wrong bias.

I don't know cause the amp isn't in front of me (and the diagram doesn't show traces), but do you think that grid trace has a partial connection to some other point (that it shouldn't be connected to) as a result of the abrasion? A high-resistance short might explain the low bias voltage.

I'd wonder if you could cut the trace to eliminate the leakage path, and jumper the points before and after the cut with a wire. It would require you to scrape away a little of the solder mask on either side of the cut portion, tin the trace, tin the ends of the wire, lay the ends flat along the trace, and heat with the iron to make the connection.

Worth a try. Otherwise the alternative seem to be ditch part of the board and handwire that section.

[firemedic]

At great personal risk, I did a more complete hot check. I replaced R49, no change. -7v is being dropped across the resistor.

Having disconnected the C35 coupling cap at the junction of R52, R55, & R49, the bias voltage is back where it's supposed to be. So it's not the PCB trace.

I have already replaced C35 with a Mallory, and in any case it comes off the cathodyne cathode which shouldn't have DC anyway with no tubes. That end of the cap reads 0.6v with the OT secondary connected, 0.3v with secondary disconnected.

I get negligible DC voltage on the OT secondary when it is isolated and B+ on the primaries.

Somehow, a small positive voltage is being generated on the OT secondary post connection not feeding the NFB. When the OT secondary is plugged in to the PCB, this small voltage follows the OT winding to the post which does feed the NFB loop, thence somehow to the PI junction I have disconnected.

I'm tired and quitting for now. Hot testing this board is extremely tedious and kinda jangles my nerves.

[sluckey]

Having disconnected the C35 coupling cap at the junction of R52, R55, & R49, the bias voltage is back where it's supposed to be. So it's not the PCB trace.

I have already replaced C35 with a Mallory

If the bias voltage is OK with C35 disconnected, it will be OK if you replace C35 with a good cap.

There doesn't have to be any voltage on the other side of C35 to knock the bias voltage down. If the cap is leaky (resistive) you will have a voltage divider from the bias line thru the cap and that 68K cathode resistor.

Just replace the cap again.

[firemedic]

The schematic may not be giving the complete picture; there are a few connections at that junction, I need to trace out, that aren't on it.

[sluckey]

Having disconnected the C35 coupling cap at the junction of R52, R55, & R49, the bias voltage is back where it's supposed to be.

If this is a true statement and C35 is the ONLY thing you disconnected to put the bias voltage back where it's supposed to be, then replacing C35 will fix it.

The resistance checks I asked you to make will tell you if there are other issues at the junction of R52, R55, & R49.

[firemedic]

All right everyone, here was the problem:

A solder bridge from C35 to C54 in the V3a cathode. Looks like when the board was manufactured, C35 had a lead bent over just far enough to form an intermittent bridge with C54.

When I replaced C35, the solder bridge looked like it was supposed to be there so I faithfully reproduced it, turning an intermittent connection into a solid one.

I got rid of the solder bridge and now everything is gonna be OK. All nominal voltages are... nominal. Nice sounding amp. I hope I never see it again.

Thanks for helping me work through this problem, everybody! Next up, a Carvin XV-112E. See upcoming thread.....

[HotBluePlates]

All right everyone, here was the problem:

A solder bridge from C35 to C54 ...

Good job finding it!

I once briefly working in a p.c. board manufacturing plant. Sometimes one board of a product made there didn't work, and we'd troubleshoot it. Solder bridges and other manufacturing related problems were generally the cause.

What really sucked was when you got a board that didn't work, and there were no bridges. Then you start suspecting faulty, shorted parts. We had a couple cases where every single part was removed, but there was still a short from the power rail to ground. The short was internal to the board, between trace layers when the raw board was silk-screened. The most complex circuits had 7 or 8 layers of traces, not visible from the outside.

Be thankful even the most horrible guitar amp p.c. boards aren't that difficult.

[firemedic]

This is a hobby for me. If at all possible I'd like to avoid PCBs, but people keep bringing me their broken stuff. It gives me a headache but I can't say no....

[plexi50]

Great job finding that problem. That is one of those issues that unless you are familiar with that board you could spend a lifetime searching for the problem. I see a lot of SS amp boards that have extra long leads all over and that any one of them could easily touch and make contact with other parts. I snip them off when i see them. The next one you do will be easier now. I also keep some relays in stock just in case

[vibratoking]

...The short was internal to the board, between trace layers when the raw board was silk-screened...

That's why electrical test is usually a mandatory item for most companies that purchase PCBs.  It catches power/ground shorts among many other things.  Also, why you should always check continuity on power and ground of any new board or amp build before applying power.  Power/ground shorts are very common occurences.

[plexi50]

For the past few years now i have been doing continuity checks on all circuit paths right off the bat that i suspect ruling out a broken connection. But! That didnt work on the 5E3 a month or so ago when it turned out to be a very dirty pin inside the 12AX7 socket preventing the cathode pin 8 from biasing and firing up that tube