Mystery Lafayette Amp 6973s REDPLATING!!!

[winterturtle]

Hello! I've recently obtained a pair of Lafayette amplifiers. Advertised by the seller as a pair of 1968s, they are a white-labelled Univox designs. They're both made in Japan & early examples of PCBs. The one I really wanted was a 1x15 piggyback that just screams when pushed & I love her -- I'm sure there are basic cleanup tasks but for now she runs just fine.

Then there's my current focus, a 2x10 with no model or serial number visible. This is the same unit based on their pics and description. https://reverb.com/item/184047-lafayette-by-univox-2x10-amplifier-w-6973-tubes-1960-s-black The board in mine says MUSICAL AMPLIFIER MODEL G 42.12.30 for which a circuit seems to be lost to time. I have been using this diagram as a guide which thusfar is close. https://www.tiffe.de/roehren/u255r.gif Most of the circuits match my design, some original components match, others don't, but it's at least a related design.

The previous owner did some... questionable work... which I assume was in an attempt to solve the issue. He swapped a few caps (overshoot on some values which seems to be normal) and a few resistors. Strangest and I assume to try and address the redplating he added a 10k 1-watt resistor to each power tube -- as the series connection between Grid 1 and the downstream node connected to the 220k resistors / .01 capacitors.

I've gotten the following values so far (warm unit, speaker load and instrument plugged in)
Grid 1 (pins 3&6) to ground : -7vdc/-5.5vdc cold; -1.5v/2vdc hot
Cathodes to ground : 1.34vdc
Plates to ground : 332vdc/338vdc
Grid 2 (pins 1&8) to ground : 315vdc
B1 to ground : 324vdc
Power transformer before diodes to ground : 270vac
Plate to Cathode, hot : 337vdc/342vdc

So from here I am not sure if I should be calling this cathode biased, but there's a resistor running the cathode to ground in series so yes?

On https://elecurls.tripod.com/biascalc.htm I entered 2 tubes; 1vdc measured across the cathode resistor; 10 ohm resistor value; 340vdc between plate and cathode. I got 16w plate dissipation per tube and 48mA plate current per tube.

I then jumped over to https://www.tubesandmore.com/tech-articles/tube-amplifier-bias-calculator and entered the 6973 tube type and a plate voltage of 335vdc which lets me know I should be shooting for 70% for a class AB (right, right?) amp which means then ~25mA is desired. Assuming my seat-of-the-pants process here is accurate I'm running twice the current I should be.

At this point I feel like I am getting too far out of my wheelhouse to identify how to proceed. I am kinda thinking the oddball resistors the PO added from the #1 Grid connections aren't doing much & I'd think removing those is probably wise. From there should I be getting the bias worked out by changing the cathode resistor value? Maybe a good opportunity to drop in a pot and have it adjustable? Not sure the current 6973 condition except that they're very old and have been abused some now -- may be swapping once I have the unit working properly overall so an easy rebias would be great. I realize there's probably some ultimately simple V=IR magic I should be doing to arrive at a gameplan but I'm mentally exhausted! Likewise if I should be doing more checks to verify the existing components are healthy please help me understand there too!

All in all I REALLY appreciate any help anyone can offer. I love old things & keeping them working - as well as obscure machinery and definitely tube based stuff. I really feel like I am on the edge of really "getting" it haha! =)

[astronomicum]

Looks like your bias voltage is low. For starters, a picture of the guts would help.

[astronomicum]

It is most likely fixed bias but work that out first before you start changing components.

[winterturtle]

Here we go. Took a while to unearth my imgur password =)

https://imgur.com/a/au6Ik0s

[PRR]

> Grid 1 (pins 3&6) to ground : -7vdc/-5.5vdc cold; -1.5v/2vdc hot

Should be -23. Hot or cold. With or without tubes. Find out why it isn't.

Some of the paper caps from that ghost factory are just bad.

[astronomicum]

Pics are great. The amp is fixed bias. As PRR suggests, trace your bias circuit and find out the cause of the low bias voltage.

[winterturtle]

Cool! So I added some values and the path the voltage follows is correct to picture (save that one resistor value)

Green is the voltage readings. Orange is notes of replaced items as I find them.

I really do not understand why the tube bias is controlled by a variable voltage signal from the term?

https://imgur.com/CkFI0p6

[winterturtle]

Still getting my head around the setup. The resistor that shows as 4.7k in the drawing looks original and is a 6k per the bands. Reads about 7.1k out of circuit.

I kinda assumed starting at the transformer was the way to trace - because there are caps separating the preamp from the grid connections in question thats not a source of negative voltage right?

[winterturtle]

Ok so maybe I'm getting it more.

Trem maybe is like a bias cut to make the effect? I know its not working presently so some of those junk caps you mentioned could be goofing maybe? On to study a bit.

Thanks guys!

[astronomicum]

Looks like you are on the right track. Univox used a 6K in place of the 4.7K in other designs. If you need to increase the bias voltage once you fix the other issues, you can lower that resister value. It would be helpful if you can get the trem to turn off. Pull the 12AX7 used for the ocillator (V1) until you can figure out the problem with the trem circuit. That will also kill your input but won't matter while you work on the bias circuit. Looks like there is a leaky coulping cap from the PI. Try replacing those.

[sluckey]

REMOVE THE POWER TUBES!!! Don't put them back in until you have approximately -23Vdc on pins 3,6 of each 6973 socket. I suspect the 6973s are already damaged.

Refer to the attached schematic. Check or replace the suspect caps, starting with the first suspects. The goal is to get approximately -23V on pins 3,6.

[winterturtle]

REMOVE THE POWER TUBES!!! Don't put them back in until you have approximately -23Vdc on pins 3,6 of each 6973 socket. I suspect the 6973s are already damaged.

Refer to the attached schematic. Check or replace the suspect caps, starting with the first suspects. The goal is to get approximately -23V on pins 3,6.

I set the power tubes aside, but from both your take as well as some of the noises, I suspect you're right =/ With my wiring I think I can get away with 6CZ5 but that's for another day when I have -23V! =)

SO I lifted one leg of the first suspects and neither is shorted, in fact I couldn't get a read at all on OHMS highest setting - just show as open on two good meters. I did the meter-style discharge for a few seconds then hook up and hopefully watch some resistance float gradually up to infinite. (Are these what one would call coupling caps? While I now know a decent amount about a phase shift oscillator controlling bias lol -- I only know of the 'phase inverter' that it often has a 12AT7 in Fenders haha. Guessing after the EQ / tone stack area then it's the phase inverter section? More reading later lol) *I am realizing now all the orange drop caps are new -- which a SBE716P600V I think is Sprague & the 103J is the 0.01 uF as appropriate to the drawing.

Same action on second suspect; small electrolytic 10uF / 160v definitely replaced. This case I saw a resistance number that then started gradually counting up. Seems legit, after a few seconds of it counting upwards I figured I'd let it be. The negative leg is directed toward ground presently.

Then the ceramic disc third suspect. Original appearance but .1uF rather than .5u as in the drawing. This one flashed a value then went to open pretty fast -- like half second kinda deal.

4th suspect much the same experience as the third -- and on closer look / the solder joints I guess this one was probably already replaced as well!

Inconclusive but my meter testing of course leaves some to be desired. Anything seem amiss with the above?

The trem system's .02uF and 2 .01uF caps all appear to be original maybe "Acushnet" ( https://imgur.com/YGYk1Yo ) but the solder points were all recently hot so perhaps not or NOS? These .02 and .01 trem caps -- the banded end is oriented toward the bias for the power tubes; with that being negative voltage does that make sense?    (The yellow one in the shot is the '4th suspect' =) )

In inspecting I noticed this! https://imgur.com/RhVkv8g I'm sure it is/was connected but by a small bit of solder ribbon. It was clearly a cold joint / pop off -- appears to be between on the incomming (banded) end of the second .01 trem cap. It this a plausible cause? I've left one side of the caps off for the moment and would other wise test haha.

I'm starting to wonder the most about the polarity, but the old noggin is hurtin' again, better take a break before my head makes the magic smoke!

[sluckey]

It would be more revealing if you check for -23v on pins 3,6 each time you disconnected those caps.

[acheld]

*I am realizing now all the orange drop caps are new

Looking at your photos, I think most (if not all) of the components on the board are new.   

As noted before, tracing your bias circuit and showing a schematic of what you have will help you in your quest.

When I see all new components, I'm always suspicious that the original circuit was altered.

[sluckey]

Let's approach this from a different angle. Looking at the schematic, disconnect the wire on the bottom lug of the INTENSITY pot. Now check the DC voltage on the 10µF/50v bias cap shown at the bottom of the schematic. If not approx. -23v then check the ACV on the cathode of the bias diode. Should be about 30VAC. Report your findings.

[winterturtle]

THANK YOU all so much for the input & apologies for running slow. I've had customers in town the last few days so my evenings have been tied up. Thursday evening I'll have time to get my caps resoldered down (I am planning to hook-jumper them into place and power up - switching which one is out of circuit to get the voltage readings that sluckey suggested above before resoldering as well. Not perfect but may offer some insight?)

To get reliable readings & insight I'm guessing I need the preamp tubes in place yeah? I'm trying to conceptually get my head around the removal of the power tubes not throwing off my readings. Probably obvious but I know a reasonable amount about the sub components but am slowly seeing the beast as a machine rather than a bag of parts!

[winterturtle]

Let's approach this from a different angle. Looking at the schematic, disconnect the wire on the bottom lug of the INTENSITY pot. Now check the DC voltage on the 10µF/50v bias cap shown at the bottom of the schematic. If not approx. -23v then check the ACV on the cathode of the bias diode. Should be about 30VAC. Report your findings.

I get a bit confused with Anode / Cathode in the polarity sense. Would the upstream side or the downstream side be considered the cathode? On the downstream side of the diode supplying the negative voltage I'm getting -23.1V back when I measured with everything still soldered to the board / as I received the unit.

[astronomicum]

Upstream side is the Cathode, and downstream is the Anode. Refer to attached. On the Cathode, you should have about 30VAC, on the Anode, you should have about -34VDC. After the 6K resister, you should have about -23VDC (when all is working). All voltages will be a bit higher than the schematic due to the higher line voltage (120VAC vs. 115VAC). As suggested, leave all your tubes out until you figure out the bias circuit. The tubes draw very little bias current so having them out will not affect your troubleshooting. Once you fix the problem, then you will make final adjustments in the bias voltage if needed with tubes in (including a new pair of 6973s).

[winterturtle]

It would be more revealing if you check for -23v on pins 3,6 each time you disconnected those caps.

Ok! Mulligan time.

Here's a table of readings as I un-jumped each of the components!
https://imgur.com/fhTgGMx

And the original name diagram
https://el34world.com/Forum/index.php?action=dlattach;topic=29048.0;attach=99073;image

[winterturtle]

As noted before, tracing your bias circuit and showing a schematic of what you have will help you in your quest.

Agreed and I'm keeping a hand written notes schematic to try and track both what is obviously changed vs what looks original and doesn't match values. The schematic I have also seems to subtly not match my circuit on a few original components with clearly aged solder as well -- the drawing number / revision on my board and the drawing are the same era but don't match =/ Pretty close though =)

The only change I have noted other than somewhat upped capacitor values is the addition or two new resistors between the power tube grids (pins 3&6) and the normal circuit, one on each power tube -- 10k 1 watt on each. They're in series / the first component the grid touches after it's connection with the tube socket pins. I am assuming the previous owner trying to tame the bias?

[winterturtle]

Let's approach this from a different angle. Looking at the schematic, disconnect the wire on the bottom lug of the INTENSITY pot. Now check the DC voltage on the 10µF/50v bias cap shown at the bottom of the schematic. If not approx. -23v then check the ACV on the cathode of the bias diode. Should be about 30VAC. Report your findings.

-16.94- -17.47vdc on the circuit side of that cap with the wire to the intensity pot / trem circuit removed.

-23.0vdc on the anode; 30.6vac on the cathode.

Now some time to think haha!

[winterturtle]

I pulled the center / wiper / yellow wire off the pot above but maybe that's not the intent.

I resoldered the center & removed the wire now that leads to the resistor then cap at the head of the trem circuit. I guess I would call this the ground of the pot, terminal c / 3 in different diagrams.
-17.8vdc on the cap; -23.6vdc on the bias diode anode; 30.9vac on the cathode.

Still in this state of disconnected pot wire; I am measuring VDC on both sides of the resistor that shows as 4.7k in the drawing and is a stock-looking 6k in my machine. -23.2vdc on the leading side (directly connected to the bias diode / anode) then -16.84vdc on the end that connects to that capacitor, 47k resistor, and the rest of the circuit.

Seems to indicate that resistor is the bottleneck. I know the water and an orifice metaphor for resistors but in application here... Maybe that means the resistor is too high a value or maybe the circuit is drawing too much & suppressing voltage for current? I've desoldered one leg and got about 7k on a meter I believe - this drawing is 4.7k and the banding [blue {maybe even violet for 7k?}, black, red, silver] so indicates 6k...

Since the voltage before the resistor is close to the desired value, would it make sense to jumper it? Not sure if the intent of its inclusion here is to limit current or voltage... if the intent is to depress voltage then maybe it's not necessary? Even a 4.7k like the drawing in theory would suppress voltage if there's sufficient current I think.

[winterturtle]

Upstream side is the Cathode, and downstream is the Anode. Refer to attached. On the Cathode, you should have about 30VAC, on the Anode, you should have about -34VDC. After the 6K resister, you should have about -23VDC (when all is working). All voltages will be a bit higher than the schematic due to the higher line voltage (120VAC vs. 115VAC). As suggested, leave all your tubes out until you figure out the bias circuit. The tubes draw very little bias current so having them out will not affect your troubleshooting. Once you fix the problem, then you will make final adjustments in the bias voltage if needed with tubes in (including a new pair of 6973s).

OOOOOOOOOOOOOOOOOOOOOOO

Sorry, catching up on things comment by comment here. Based on this feedback and my consistently getting -23 to -20 vdc on the bias diode anode in my testing could mean the diode is failing and limiting the voltage coming through it.

This seems to make the most sense of my stream-of-consciousness processing so far eh? =)

[winterturtle]

Hey everyone and sorry for going quiet! I'm actually travelling at the moment so while I don't have meter readings I thought I'd give a general update haha.

I replaced the single diode that makes the negative voltage for the bias circuit -- not perceptible change in my voltages (thought I don't have specifics documented atm.) I used a 1A 600v rectifier diode which I believe is an equivalent of the SD-1.

Feeling frustrated and ignorant (hahaha) I decided to review a Princeton circuit's bias for a sanity check. On it, the bias capacitor polarity was labelled -- and it was the reverse of mine! The 10uF/50V cap on the bottom of the diagram had the negative leg going to ground. Given with this cap totally disconnected I did not (seem?) to have the correct voltages I'm trying not to get my hopes up BUT went ahead and ordered a new electrolytic from antique electronics supply rather than flipping the existing. Maybe one of you kind gentlemen has run into this before!

Many of the caps aren't labelled on the circuit; should all the electrolytics in the B+ be negative to ground and in bias be positive to ground as a rule of thumb? I am also guessing the negative side of the caps in the trem oscillator should point toward the source of negative voltage too? I have not checked their orientation yet but did order replacements there as well =)

[tubeswell]

Hey everyone and sorry for going quiet! I'm actually travelling at the moment so while I don't have meter readings I thought I'd give a general update haha.

I replaced the single diode that makes the negative voltage for the bias circuit -- not perceptible change in my voltages (thought I don't have specifics documented atm.) I used a 1A 600v rectifier diode which I believe is an equivalent of the SD-1.

Feeling frustrated and ignorant (hahaha) I decided to review a Princeton circuit's bias for a sanity check. On it, the bias capacitor polarity was labelled -- and it was the reverse of mine! The 10uF/50V cap on the bottom of the diagram had the negative leg going to ground. Given with this cap totally disconnected I did not (seem?) to have the correct voltages I'm trying not to get my hopes up BUT went ahead and ordered a new electrolytic from antique electronics supply rather than flipping the existing. Maybe one of you kind gentlemen has run into this before!

Many of the caps aren't labelled on the circuit; should all the electrolytics in the B+ be negative to ground and in bias be positive to ground as a rule of thumb? I am also guessing the negative side of the caps in the trem oscillator should point toward the source of negative voltage too? I have not checked their orientation yet but did order replacements there as well =)

The polarity of the main power supply filter caps are +ve to the high voltage side -ve to ground. Bias supply filter cap polarity is the other way around because we're working with negative voltage (if you flipped the bias caps, they'll blow and you won't have -ve bias voltage

[winterturtle]

The polarity of the main power supply filter caps are +ve to the high voltage side -ve to ground.

Looks like this was it! I reversed (and replaced) the 10uF/50V cap. I still have extra resistors before the grids -- but at any rate my redplating is solved. I've forgotten the exact but close to -23 up there. I need to better learn the math for calculating dissipation; 6973 / 6CZ5 isn't listed on any of the convenient calculators I've found yet.

My trem still doesn't work at all & that's my next mystery - though polarity seems like the place to start. Unfortunately the princeton schematic where I got the polarity idea previously doesn't call it out. Can anyone confirm the polarity of the three capacitors in line to the trem tube's grid? The bands on all 3 are oriented toward the plate / away from the grid. The 4th, on the trem tube's cathode, is oriented with the negative leg to ground.

[sluckey]

Can anyone confirm the polarity of the three capacitors in line to the trem tube's grid? The bands on all 3 are oriented toward the plate / away from the grid.

Doesn't matter. Those caps are not polarized.

The 4th, on the trem tube's cathode, is oriented with the negative leg to ground.

That's good.

[winterturtle]

Can anyone confirm the polarity of the three capacitors in line to the trem tube's grid? The bands on all 3 are oriented toward the plate / away from the grid.

Doesn't matter. Those caps are not polarized.

Ah / uh-oh. Mine appear to be polarized. The three AccuCaps are the oscillator main 3. The yellow is the cap the connects to the bias supply and trem tube plate. https://imgur.com/79NktLI

I'm thinking here again about the Princeton. Those (vintage) units have ceramic disc and people seem to recommend using polypropylene orange caps - both of course non polarized.

Looks like time for another antique electronics order eh? Get myself three polypropylene caps and see how they perform.

[sluckey]

Ah / uh-oh. Mine appear to be polarized.

THOSE CAPS ARE NOT POLARIZED. The band simply indicates the outer foil. No need to mess with them if the oscillator is working.

[winterturtle]

Ah / uh-oh. Mine appear to be polarized.

THOSE CAPS ARE NOT POLARIZED. The band simply indicates the outer foil. No need to mess with them if the oscillator is working.

I don't entirely understand what on the foil would be marked if they're not polarized? I had thought the band was intended to mimic the + crimp side of the electrolytics I'm used to. More to learn for me and capacitors!

I did some basic voltage / ground testing and will post second -- but how does one validate that a LFO is working? I've been googling / reading but haven't found a direct process. With the exception of cathode voltage kinda 'sagging' with a meter on it, I'm not seeing any values 'wiggle' yet. I'm using a digital meter so far but the values are stable & I'm anticipating a shifting value under and oscillation.

[winterturtle]

Was doing a little pickin' on the piggyback Lafayette Friday night and she started to give me some voltage through the instrument -- SO I expedited my three prong update.

While I was at that I went ahead and did the same to the patient in this thread, originally redplating and still no trem working. While in the shop & already drilling and controlling metal flakes in the chassis, I added a speaker jack since the original connection was hardwired and had been butchered over the years. Opportunistic, I added a second jack/line off the OT and though tight, I have nice 4 & 8 ohm jacks & it's grounded without a death cap.

https://imgur.com/5meth3Q
https://imgur.com/nI2UbZh

I checked what I know to check inside the trem circuit. It had the vintage looking GE 12AX7 to today so I went ahead and put in a known good Chinese 12AX7B that's recently removed from an AC15C1. I figured making sure it wasn't the tube made sense but no change overall. The trem pedal jack is also a shunting jack (what I have been calling it anyway) so I have a pedal jumper cable that I've cut and is a known open circuit... Kinda the opposite of what people do with old RCA cables for Fender trem without a pedal...

Anywho here's the values I got. (Intensity at full high and speed control at noon)
https://imgur.com/sdRkXDW

The plate seems most interesting. The schematic has 190v listed -- but I have ~154vdc with the pedal grounded and 304vdc (basically B2) with the pedal open/on. I have a stable 0VAC when the trem pedal is open/on -- my thought is perhaps I could see oscillation here as ac but no dice.

Cathode may be a clue also -- with trem grounded / off it reads 1.078vdc to a schematic value of 0.9v which seems ok... but then with the trem pedal open/on, the voltage goes to 6.6vdc then gradually falls. I waited maybe 45-60 seconds and it had steadily reduced on the meter to maybe 5.5vdc and I disconnected given I don't understand the phenomenon hah.

I'm going to keep with the youtube videos and googling but if anyone can point me down the right road (for the twelfth time lol) it's very much appreciated.

Happy Canada Day and-or 4th of July everyone =)

[winterturtle]

Against my better judgement I took the 'throw parts at it' approach since electronic components are relatively cheap. Looks as if one or more of those odd looking capacitors must have been bad. I slapped in some new polypropylene capacitors and the trem is off to the races -- and how nice she sounds! I've put some different value resistors and a bias pot on order to make her adjustable and able to hit a solid -23V at idle as well; she's still running a bit hot. Its been an ordeal and I wanted to put a bow on this package as resolved =)