MC40 power supply rebuild

[Pio1980]

I used 50uFx2/500V and 100uFx2/500V e'lyticaps and mounting clamps from Angela Instruments for '240 and the pair of MC75 in progress. I ordered 4 more of the same clamps than caps and got the 330uF/250V Sprague snap-in e'lyticaps to fit them off the 'Bay for the doubler bridges, the clamps cost more than these caps.

 

[RWright]

Hello,
I have posted this elsewhere with an up-in-the-air reply. I have redone the caps and diodes in a pair of MC40s quite awhile ago, now. Only until recently I was concerned about the bias voltage being off due to changing to Si diodes. I have read that setting the 6L6 voltages with a variac, and then adjusting an inline bias trimmer to the right voltage is the way to determine what resistor to put in series with the bias line. However, after I adjust the pot for the correct bias voltages, the 6L6 voltage drops to 460v. I could readjust the variac, but to what point do I stop this? I don't want to over-biasing, or pushing unnecessary DC into the OPT. Anyrthing else could be wrong? Thanks.

 

[Beowulf62]

Hello,
I have posted this elsewhere with an up-in-the-air reply. I have redone the caps and diodes in a pair of MC40s quite awhile ago, now. Only until recently I was concerned about the bias voltage being off due to changing to Si diodes. I have read that setting the 6L6 voltages with a variac, and then adjusting an inline bias trimmer to the right voltage is the way to determine what resistor to put in series with the bias line. However, after I adjust the pot for the correct bias voltages, the 6L6 voltage drops to 460v. I could readjust the variac, but to what point do I stop this? I don't want to over-biasing, or pushing unnecessary DC into the OPT. Anyrthing else could be wrong? Thanks.

I assume you are setting the bias at -52V and this results in 460V at the 6L6GC plates? That would be a 2% difference from the original schematic specification (470V)...not much. Particularly when you consider that if the MC40 circuit uses carbon composition resistors, these will certainly have increased in value with age and at least some will be out of spec/tolerance. If you check the other B+ voltages in the circuit, I'm pretty sure you would find shifts from the schematic values as well.

You can check the resistors and replace those that are out of spec...but I doubt that a 2% drop in plate voltage is anything to worry about.

 

[RShackleford]

I have read good reports about "McShaned" electrolytic power supply re-builds on Mc240's.It was said that the extra capacity allows for more solid low bass reproduction.

I've rebuilt the power-supply on one (so far) of my MC40s, using a part set I got from McShane. It includes 820uF caps (snap-in type) in the voltage doubler and ultra-fast diodes. I also used his recommended caps for the bias supply, a 100uF electrolytic shunted with something smaller. The B+ voltage is now 485v, and the bias is -62v (because I put no resistor in series with the diode that replaced the selenium rectifier). He says this match of voltages is good. Bass is definitely improved. (One advantage of having two monoblocks - I can set my preamp on "mono" mode and then switch between the two amps, both getting the same input signal, using the balance knob on the preamp).

I also installed a mix of Russian K40 and 716P main-board caps that he recommended and provided (McShane is super-easy to work with), Vishay/Dale metal-film resistors (including in the PS) from Mouser, a CL80 thermistor, and a gold-plated input jack.

I glued (using silicon RTV) the doubler caps to the underside of the PS board (where the body of the cans was) and added a terminal strip there for the diodes and snubber caps; so that circuit is completely on that side, leaving tons of room (on the top side) for building the new bias supply. I retained the multi-section cap for the main B+ filter caps (though a single snap-in and an axial lead one would have sufficed, as I discuss on the first page of the "Hints and Kinks II" thread).

 

[RWright]

Hi,
Yes, I am setting the Plate voltage to 470v first, then adjusting for -52v bias. I'll check out the resistor values in the amps, too. Most of the resistors are still original carbon comps. The only resistors I ever changed were in the power supply dividers. Thanks for the info! Much appreciated.

 

[dshoaf]

No need to use RTV on the caps from Jim McShane. Here's a rebuild using the clamps instead. Same 820mf caps, too.

I miss those guys.....

Oh, and use the same rebiasing technique for all the Mac amps:

- Set plate voltage to what's called for in the schematic. Use a variac to do so.

- Adjust the bias resistor voltage divider for proper bias voltage readings called for.

- Plate voltage will change, accordingly.

- Readjust variac so to bring plate voltage back into spec.

- Readjust bias resistor value again to obtain bias voltage.

Repeat above 2 steps until both voltages agree. Best to use a resistor decade box to make finding the right bias voltage fast and simple.

Lastly, as I've posted here many times, run the amps for a number of weeks then retest that all operating voltages on each tube stage is in spec. I've seen all too many resistors drift out of spec once they're put back in normal use after many decades of dormancy.

Cheers,

David

I've rebuilt the power-supply on one (so far) of my MC40s, using a part set I got from McShane. It includes 820uF caps (snap-in type) in the voltage doubler and ultra-fast diodes. I also used his recommended caps for the bias supply, a 100uF electrolytic shunted with something smaller. The B+ voltage is now 485v, and the bias is -62v (because I put no resistor in series with the diode that replaced the selenium rectifier). He says this match of voltages is good. Bass is definitely improved. (One advantage of having two monoblocks - I can set my preamp on "mono" mode and then switch between the two amps, both getting the same input signal, using the balance knob on the preamp).

I also installed a mix of Russian K40 and 716P main-board caps that he recommended and provided (McShane is super-easy to work with), Vishay/Dale metal-film resistors (including in the PS) from Mouser, a CL80 thermistor, and a gold-plated input jack.

I glued (using silicon RTV) the doubler caps to the underside of the PS board (where the body of the cans was) and added a terminal strip there for the diodes and snubber caps; so that circuit is completely on that side, leaving tons of room (on the top side) for building the new bias supply. I retained the multi-section cap for the main B+ filter caps (though a single snap-in and an axial lead one would have sufficed, as I discuss on the first page of the "Hints and Kinks II" thread).

 

[RShackleford]

No need to use RTV on the caps from Jim McShane. Here's a rebuild using the clamps instead. Same 820mf caps, too.

I didn't consider using RTV to be any more trouble than using clamps (I may regret this when one of them needs replacing, but doubt it, I don't think it should be too hard). Mainly I did it that way because I like the way the layout ended up a LOT better. Basically, Jim's 820uF snap-in caps are WAY smaller that the stock cans, and if you just mount them in the same place with clamps, you're not taking advantage of all that freed-up space. I should have posted a pic, will later ... but basically my *entire* voltage-doubler circuit is on the side of the bracket where the cans used to hang. So there's all that space above it, which I needed to rebuild the bias supply with the dual-caps (big one plus shunt) that McShane sold me.

Which amp is in your photo ? I doubt I'd feel the same about the layout issue with it, but maybe.

I've seen all too many resistors drift out of spec once they're put back in normal use after many decades of dormancy.

I can't see why people don't just replace all the resistors when doing all the work the rebuild the PS, and usually replace the coupling caps. A few bucks in parts cost, and not that much more work ...

 

[RWright]

Hello,
I have since done the adjustments to get correct bias voltages. However, another issue has cropped up. I'm started to get some plate glow in V5 of one of the amps. I immediately shut them down. I ended using 910-ohm resistors in series, pre-diode, in the bias circuit. Any hints? Thanks.

 

[Steve O]

Hello,
I have since done the adjustments to get correct bias voltages. However, another issue has cropped up. I'm started to get some plate glow in V5 of one of the amps. I immediately shut them down. I ended using 910-ohm resistors in series, pre-diode, in the bias circuit. Any hints? Thanks.

Some thoughts:
1. What brand of 6L6s are you using? I've found some Russian 6L6s are prone to minor redplating when near rated dissipation.
2. Although Mac amps tend to be tolerant of unmatched output pairs, IMO they can benefit from matched pairs. You might have grossly unmatched pairs in the one amp.
3. Try moving the tubes around to other positions in the same amp and to the other amp. If the glow stays with a specific amp/socket, the issue probably lies with the amp. If the glow stays with the tube, the issue lies with the tube.

Also, FWIW, I don't generally worry about dull, stable redplating if the tubes are Russian. Emphasis on dull and stable. YMMV

 

[RWright]

MC40, again

Hi,
Thanks for the reply and info. I am indeed using some Russian 5881's that were supposedly matched at the time I purchased them about 10 years ago. I swapped the tubes, but didn't see any redplating in either, although I didn't have the amp on very long, nor was I playing any music, as my wife was sleeping. It was quite awhile ago that I replaced the electrolytics, but the amp seems to sound good yet, and is very quiet without hum and only slight hiss. The only electrolytics I might worry about is C10 and C11 if they are leaky. The main thing that worries me is not so much the tubes themselves as the output transformer. I wouldn't want excessive DC current flowing continuously and possibly damaging it. I went through the amp and checked out the resistors, and they all seem to be pretty good yet. When I replaced the electrolytics, I also replaced all the bumblebees with 761 Orange Drops. I also replaced the divider resistors in the power supplies. Up until recently, I had the -52v adjusted by placing a trimmer pot in parallel with the 56K and adjusted for -52v. It wasn't until recently that I was reading on here about putting a resistor in series with the diode to adjust the bias voltages.

 

[RShackleford]

I rebuilt one of my MC40s with McShane power-supply kit (820uF for the voltage doubler, fast/soft diodes, big bias cap with shunt) and like it. I replaced the multi-section can in the first one with the 80/40/30/20uF CE Mfg'ing one that's readily available, since the one in the amp (another of the CE parts) was blown.

Now I'm starting the other amp. I'm gonna use McShane's parts again (for the voltage doubler and bias), but wondering whether to replace the multi-section can, which looks fine (the one on the first amp was blown, so no decision to make). I'm a little reluctant to replace a good part, especially since the first amp had already gone thru one of those CE parts. But I know the one in this amp is quite old, so I guess I should just do it. Thoughts ?

Question I should know the answer to, but don't: how do I check the old one for leakage ? Just put a ohm-meter on it and wait for the value to settle and see how low it is ?

 

[dshoaf]

You'll need to put the old can under full voltage - most testers will not do this. In other words, don't trust it or the more passive testing available.

Swap it.

Yes, I've gone through a few of the CE can caps some years ago when they had a problem. Those, I understand, are gone now.

Why risk it?

Cheers,

David

 

[RShackleford]

Thanks, David. Maybe I should use a snap-in. Since I'm using the newer version of the circuit (which combines C16A and C16B, the separate 380v supply voltage goes away), I only need one. Plus, for C16C, an axial lead part should work (it's only 15uF originally).

 

[dshoaf]

Add the spare cap in parallel with the filter cap for the 125 volt plate voltage rail for V1. This adds a bit more decoupling for the voltage amp stage.

Cheers,

David

 

[RShackleford]

Duh, with the B+ now 485-490v, using a 500v part would be asking for it. Guess I'm stuck with paying $40 for the CE 80/40/30/20uF can.

 

[dshoaf]

Hey, its far better than we had it in the 80s....we've got so much more options than we used to.

Yeah, you'll have to bite the bullet, though, on the price, though.

Its like an old British car restorer explained, "Work your way back out of the project now or else you'll be tearing into it again later".

Cheers,

David

 

[RShackleford]

...we've got so much more options than we used to.

"Work your way back out of the project now or else you'll be tearing into it again later".

Which brings up the question, should a prudent Mac owner buy a couple of extra of the 80/40/30/20uF caps, or hope there's a better option when the currently-installed ones blow ?

Some years ago, I found what seemed like a good part for the voltage doubler cans and bought some extras. But I now realize the snap-ins (from Nichicon and the like), such as I got from McShane, are a much better option - smaller, higher value, probably lower ESR. Of course, that part is a lot less problematic, since it only has to handle 250v or so.

 

[dshoaf]

I've thought about that, too. I've decided that the current CE caps will be fine for the time the amps will be in my possession. I figure there's about 20 years in an electrolytic cap so that's good enough for me.

Cheers,

David