MC40 power supply rebuild

Hi folks, new guy here - I've got some MC40s that I've been running for some years, replacing filter caps as they go bad. I'm starting to realize that it'd be less trouble in the long run, not to mention protecting these treasures, to go ahead and do a full rebuild on the power supplies. I've tried searching, and learned a lot, but a few questions remain ...

For the big multi-section can, there's a readily available (for $40) 80/40/30/20uF unit. So I'm tempted to up the capacitance on C16A/C16B/C16C, by tying some of the sections together. I'm curious how large of a capacitance people think is worthwhile, and if there's any downside to up'ing it too much. (On one amp, C16A and C16B are combined, the other is one of the older units where they're separate with a 10K resistor in between).

Similarly for the voltage doubler, I've got some dual cans (200 and 150uF at 350V). I'm wondering whether to use the 150 as per spec, the 200, or tie 'em together. (I'll also replace the diodes with 1N5408).

The older amp does not have the thermistor (and not in the schematic). No reason not to add one to it, when I'm replacing the one in the newer, no ?

On the bias supply, people say to add a 470ohm resistor between the transformer and the 1N4007. But there's mention of experimenting with resistor values to tweak the voltage (one comment that it may not be needed in an MC240). What determines this variability ? If it's your line voltage and/or your tubes, seems like you might want to retain the ability to adjust this, thus maybe a resistor with a trimmer pot in series might make sense ? Or is the variation due to more stable factors ?

Thanks !

P.S. I've edited the schematic for the two models (with the newer and older serial numbers) for clarity, if anyone wants a copy.

I can't help you with the work but I can give you a bump to the top! Good luck and welcome here.

Doubler caps are non critical, anything over 250V and 150 uf will work, for better bass I use 470 uf @ 400 V with no problem in all doubler supplies.

Here is an example of a restored MC240 with new capacitors, both filters and doubler.

dewickt said:

Here is an example of a restored MC240 with new capacitors, both filters and doubler.

Click to expand...

With Terry's endulgence, here's another, which is my first encounter with Mac tubes;


Here's the accommodation for the voltage divider on different capacitors with same-value sections;
Before;


After;

I have heard very good reports about using the capacitor kits that Jim McShane sells.He markets them for Citation amps.but they are reported to be a very good upgrade for Macs.

Rshackleford - I know you're new here, and so may not know his reputation, but Terry is the go-to guy for many AK'ers when it comes to amp rebuilding or restoration. When he gives advice on something like this you can pretty much take it as Gospel.

Also, I got my MC60 rebuild supplies including cans from Jim McShane & couldn't be happier.

Thanks, folks. Yes, I've gathered that Terry is "da man" on McIntosh stuff and I'm glad to hear from him.

I guess I'll tie the 200 and 150uF sections together for my voltage doubler (in the 2-section cans I already have). I'm surprised the photos above still show carbon resistors on the PS. Mouser has some 1% 2watt metal film ones:

http://www.mouser.com/Search/Produc...irtualkey61300000virtualkey71-CPF2100K00FKE14

They're "flameproof"; they are pretty small though, so unless theta-J is somehow magically higher, I guess they get pretty hot.

Thanks for the link to McShane: looks like the place to go when I do the next step of replacing coupling caps. Though I read a lot of good things about Dayton Audio's also.

Those little snap-in electrolytic caps are sweet. Compact and cheap. I'm tempted to jury-rig them in somehow like Terry does, but I've already got the cans I need. Still, it's half-tempting to make a DIY PCB board that uses all snap-in caps and has mounts for all the resistors and diodes and bias supply stuff too !

I'm still curious about the resistor for the bias supply diode: does it make sense to have a trimmer, in case it changes with different tubes and such ?

Im tuned in here too because I will be wanting soon to rebuild my 40's as well --- so any sage advice on where to source parts or other 'hints and tricks' will be greatly appreciated!

goatster said:

Im tuned in here too because I will be wanting soon to rebuild my 40's as well ...

Click to expand...

Okey-doke. I'll summarize what I'm doing below. Hopefully it'll help you, and may Terry will critique. Let me know if you'd like my edited schematic (for a specific serial-number range):

For main PS caps, I got this 80/40/30/20 can:

http://www.tubesandmore.com/products/C-EC80-40-30-20

Used 20uF for C16C, 40uF for C16B, and tied the 80 and 30uF together for C16A (one of mine has the newer serial number where A and B are combined, so I tie 80/40/30 together). I'm using metal film resistors - not sure if there's a reason some pictures show the carbon type retained.

For voltage doubler, I have some 200/150uF two-sections cans, which I intend to tie together. I dipped one in that rubber coating stuff you can buy in a can (or maybe use Rustoelum's rubber spray ?) to insulate the can of the one not tied to GND. Of course you can use the cheap and small "snap in" type if you want to jury-rig the mounting. 1N5408 for the diodes (I think the faster version is discouraged). Ceramic disc cap for the "snubbers".

For bias supply, I'll replace the selenium rectifier with 1N4007 diode. Probably need a resistor between banded end and transformer lead, or voltage will be too high. A 470-ohm is supposed to work, but I'm tempted to use a 1K trimmer, since some people seem to find they need to tweak this value to get the voltages right.

I've read many opinions on which of the NTC thermistors to use. I believe I'll use CL-80 (3 amp and 47 ohms).

Different opinions on whether to retrofit a polarized 3-conductor (incl. ground) power cord too.

I'll gonna skip coupling caps for now, but when I do, Dayton Audio looks like a great value, as do the ones listed here:

http://www.mcshanedesign.net/citpage.htm#parts

... though he doesn't seem to have the 12uF for C10 and C11.

FWIW, I measured all the resistors as I went thru the amp with my project and found them all unstressed and in tolerance, probably the best carbon comp's on the market at the time which IIRC were Allen-Bradley. As I am on a time and cost restraint I'm replacing parts known for reliability issues only as well as the baddies.
For the added bias losser I'd probably use something more substantial than a trimmer resistor, I used a ca 3W metal film 470^.
I defer to Terry on any of the above.

With cost constraint, I'd sure as heck find some way to use those snap-in caps, or something besides that multi-section can I linked (about $40). But that's gonna add some time for sure.

The metal films I'm thinking of using are cents each; and shouldn't take any extra time, since all that stuff has gotta to be pulled loose anyhow, to replace the caps.

Pio1980 said:

For the added bias losser I'd probably use something more substantial than a trimmer resistor, I used a ca 3W metal film 470^.

Click to expand...

Looks to me like the bias current is in the neighborhood of a milliamp; even at 10ma that resistor wouldn't dissipate 1/4 watt. But maybe I'm missing something ? Certainly no harm in using a beefier one.

FWIW, on my circa 1958 MC60's only two resistors were out of spec.

Yeah this is what I am absolutely looking for from the gang here. Its super easy to say 'replace xxxxx because it can go bad' vs 'high probability to replace yy ". I for one am a total cheapskate and don;t really believe in wholesale changes of stuff -- unless there is a good reason. Carbon comp resistors from the 60's are generally pretty high-rel -- and there;s no 'physical' reason for time-dependent breakdown. Electrolytics --? well, every one on every piece of my gear is original. In fact, the only real fixes I have made were optocouplers in the marantz tuner in 1976, and one resistor in the scope vertical amplifier in 1977.
HOWEVER, I would have no problem finding (and at least keeping in stock) replacement can caps for the high voltage on the amps....
So cheapskates unite!

I recently "rebuilt a pair of MC40s and this is what I did:

1. Replaced all PS electrolytics. For aesthetic reasons, I restuffed the stock cans using Panasonic radials and snap mounts of appropriate voltage. For C16 I used 2x120μF+1x47μF/500V 85°C TSUPs (500V NLA). For the doubler I used 820μF/250V 105°C TSHBs. 820μF is a value that a typical "McShaned" doubler uses. I was somewhat hesitant to use a value this large but reports from persons I trust indicated I’d have no issues and I’ve had none. Two notes: i. The practice of restuffing is controversial and goes in and out of fashion. I think it's out of fashion at the moment. My primary concern was self heating in a confined environment. I performed some tests using a voltage doubler circuit using 680μF caps set up to supply ~400V@300mA. I detected no heating over a 4 hr period even when the caps were stuffed inside an FP can and insulated with a rolled up paper towel. ii. Some believe that the mini electrolytics don’t sound as good as the full size FP style replacements from Authenticap, CE Mfg, JJ and so on. This isn’t my experience. Personally, I believe minis may actually be better than the FP replacements. If nothing else, detailed specs for the minis are readily avail and are quite good. I’ve never seen specs such as ESR and resonant F for the FP style replacements.

2. I replaced the doubler rectifiers with the 3A ultrafast/soft recovery Si units (don’t recall PN). If doing this today I might consider a pair of 1200V SiC schottkys although the TO220 pkg is PITA.

3. I redesigned the bias supply slightly to a “pi” type filter (CRC) using 2x10μF/250V film caps. I also replaced the Se rectifier with a UF4007. I used the R (don’t recall the exact value) to adjust the bias voltage per the procedure in the Mac “sticky”. If I were to do adj bias, I’d be looking at making the R of the CRC a pot/resistor combination.

4. I rewired the final sockets to take either 6L6GCs or 7027As. I did this because one of my 40s had 6L6GC/7027A silkscreened on the chassis. Curiously, the other one was 6L6GC only.

5. I replaced the input RCA jack with nice but not too fancy machined units.

6. Neither 40 had the typical NTC thermistor inrush limiter so I added them. I used the CL90 primarily because it’s got more thermal mass than the CL80 and provides a nice slow start-up. While the CL 80 is speced with a lower resistance, because the CL90 is operated closer to its max of 2A, the voltage drop is about the same as the 80 under normal operating conditions.

7. Electrically, my pair of MC40s was kind of rough with many out-of-spec parts so I decided to replace all passive components. Overkill but in the long run probably easiest. I notice that you’re not planning on replacing signal caps at this time. You should reconsider that decision if your 40s have any of the legendary striped “bumblebee” caps remaining. These things are notorious for being electrically leaky and sometimes even physically split open.

Good luck. When you’re finished you’ll have a pair of very nice amps that can compete with and possibly surpass anything on the market today in 40-50W power range.

Great input, folks. I hope I don't come off as argumentative in some of my responses. I just like a deeper understanding, and to be told why I'm wrong when I am. That being said ...

Goatster, I think you're making a mistake being too complacent about the electrolytics. I've run my pair of MC40 for maybe 10 years, and I've had at least four of the PS caps fail. People say it can be hazardous to your transformers and even your safety.

Terry mentions that it's not a good idea to use ultra-fast diodes in a PS, because it can increase peak loads (and thus possible reliability issues, arc'ing ?) on the power transformer:
http://www.audiokarma.org/forums/showthread.php?t=477882&page=3&highlight=diode+fast

Steve, interesting to read about your PS cap experiments. 820uF on the doubler, wow; I guess I won't have second thoughts about using 350uF on mine.

Can I assume you put your bias CRC downstream from the diode ?

On the thermistor, I definitely like the higher "cold" resistance of the CL-90. I just thought 2amps was cutting it close on operating current. But it sounds like your's is doing ok.

Thanks for the prod on doing signal caps. I'm sure I'll do it soon. Are "bumblebee" caps the ones that are striped like a resistor ?

RShackleford said:

Great input, folks. I hope I don't come off as argumentative in some of my responses. I just like a deeper understanding, and to be told why I'm wrong when I am. That being said ...

Goatster, I think you're making a mistake being too complacent about the electrolytics. I've run my pair of MC40 for maybe 10 years, and I've had at least four of the PS caps fail. People say it can be hazardous to your transformers and even your safety.

Terry mentions that it's not a good idea to use ultra-fast diodes in a PS, because it can increase peak loads (and thus possible reliability issues, arc'ing ?) on the power transformer:
http://www.audiokarma.org/forums/showthread.php?t=477882&page=3&highlight=diode+fast

Click to expand...

I don't understand the reasoning behind Terry's caution against using fast recovery rectifiers. I'm speaking of the reverse recovery issue in conventional Si rectifiers where the rectifier momentarily conducts in the opposite direction as the current reverses and then suddenly stops conducting thus possibly exciting resonant circuits composed of power transformer leakage inductance and stray capacitance. Fast and slow refer to the time period the rectifier allows a reverse current to flow. I assume that the fast rectifiers were developed for SMPS apps where the supply frequency is very much higher than 60/120Hz and slow recovery rectifiers won't work well if at all. IME, fast reverse recovery rectifiers are a major improvement compared to conventional rectifiers WRT noise generation, esp considering that were using them at such a low frequency. If anything would cause higher peak currents as Terry discusses, it would be the use of very large filter caps immediately after the rectifiers. This is one reason I was concerned about using the 820μF doubler caps. I'm assuming the transformer winding impedance keeps the absolute value of peak current to a workable value.

I suppose one way around Terry's concern regarding fast rectifiers would be the use of schottky rectifiers: they don't have a reverse recovery condition. OTOH, one might consider the no recovery time of a schottky as infinitely fast.............

Steve, interesting to read about your PS cap experiments. 820uF on the doubler, wow; I guess I won't have second thoughts about using 250uF on mine.

Can I assume you put your bias CRC downstream from the diode ?


On the thermistor, I definitely like the higher "cold" resistance of the CL-90. I just thought 2amps was cutting it close on operating current. But it sounds like your's is doing ok.

Thanks for the prod on doing signal caps. I'm sure I'll do it soon. Are "bumblebee" caps the ones that are striped like a resistor ?

Click to expand...

Yes, yes and yes.

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.

An explanation of my priorities;

I had the choice to upgrade or renovate to original standard and went largely with the renovation along with silicon and the comp losser for the bias as the best most conservative option under the circumstances, their next custodians can do what they will with them.
These are but two of 6 tired/defective/DOA vintage Mac pieces I'm doing totally on my own dime for a deserving elderly a'phile pushing 90 who'd like to have them back to enjoy while he still can (and does). He is not local (220 mi/4 hour road round-trip) and I'm doing the work in his upstairs bedroom, where the system is located, at a work station I've set up there as part of a weekly afternoon's visit. So far I have a C24 and MC240 done that he is enjoying immensely, there are now 2x MC75s belly-up and open waiting on parts orders to come in and my head cold to subside. After them come an MR-71 and MX110.

Terry was my first choice for him (he knows who Terry is, Mac-wise) but this is just not an option under these unusual special circumstances, I am his best and only plan 'B'. We also share multiple interests and experiences which make the visits particularly meaningful and therapeutic. Given his cultural contributions to his city of residence he is well worth it, Good Karma.

Being a little slow on the uptake, I've only recently become aware of the option of replacing twist-loc "can" capacitors ($30-40 and limited availability) by using special clamps:

http://www.tubesandmore.com/products/S-H122

.. with less expensive parts like the JJ ones:

http://www.tubesandmore.com/products/C-EC350-385

My question is this: instead of using caps like the JJ ones, why not use the "snap in" (but with the same clamps) parts like:

http://www.mouser.com/ProductDetail...=sGAEpiMZZMtZ1n0r9vR22QqzopUJ6WFX/j6xQ/8lWLk=

... which is both considerably cheaper and considerably more compact (still 1-3/8" diameter, but only 30mm tall) ?