McMartin MA-50 Troubleshooting

[whitfiel]

Hi again!
As I've written in other threads, I have been enjoying a McMartin MA-50 tube amp. But being a relative newcomer to tube amps, I would welcome some assistance in troubleshooting a problem.

Lately this amp has started to develop significant audio pops and scratches in the main output section, and ultimately blows the 0.5A fuse in the bias circuit on the cathodes. I've checked the bias and it's still at 100mA per the spec. The clicks and pops are pretty loud, and happen even when the preamp gain and master gain are set to minimum. This leads me to believe it's in the output section.

I've posted the schematic at: http://frontiernet.net/~whitfield/McMartin_MA50-w_cap.jpg

You will note a 47uF cap (marked in red) on the 470V output of the OD3regulator tube (and Grid2 of the 7027A output tubes). I installed this to remove some significant 60Hz hum.

All tubes test fine. I can't test the OD3 regulator on my tube tester, but I've swapped it with another with no change. Is my addition of a 47uF cap on regulator causing trouble? Is it's rated voltage (500V) too close to the operating voltage of 470V on the grids?

So where should I look to find the cause of the scratching/pops and the bias fuse on the cathode circuit? I welcome all ideas and help in fixing this. While I'm an EE, please keep this basic so I can learn more about this fascinating world of tubes.

Best regards,
Art

 

[anatech]

Hi Art,
Lose the cap. When the regulator fires it has to charge the added cap. Probably not good for the regulator.

There must be another reason for the 60 Hz hum, like C17A and B. Possibly a bad rectifier on top of that (SR-9). A problem in the HT supply would give you a 120 Hz hum.

-Chris

 

[PRR]

> 47uF cap (marked in red) on the 470V output of the OD3 regulator tube

No! Gas tubes plus caps make oscillators. Look-up the OD3, there is a maximum capacitor that can be directly connected across it. (Yes, your red cap is not across the OD3, but the 16uFd and 47uFd are effectively in series across it, and ~12uFd is surely too much.)

> to remove some significant 60Hz hum

60 or 120?

In either case, I would assume the 16uFd main filter cap has gone sour. It isn't big enough to begin with, and if it has gotten weak it might still make 10-20 watts without buzz in the plate circuit.

Now wait. Ripple in the G2 feed should not cause buzz. So what you really did by putting 47uFd on the bottom of the OD3 is add 47uFd to the +610V line, except the ripple is cooking the OD3. If that fixed your hum/buzz, C16 is sick.

Fix C16. If you cut it open you will find it is really two 350V caps in series in one can; get two 33uFd 400V caps to stay authentic, or a couple 100uFd 450V caps because the designer probably wanted to use 40-50uFd and that 16uFd cap was a cost compromise. (Myself, I've used 2*470uFd in similar amps with good results, but some call that over-kill.)

Still, your symptoms sound to me like a tube with an intermittent short. Tubes can test good and still have troubles.

Or the -43V bias supply is failing. THAT would pop a 0.5A fuse (nothing much else could). If C17 A/B have intermittent shorts (they never short while you are checking, of course), 7027 current goes to max which would be over 500mA (and also cherry-red plates).

I would in general assume ALL the electrolyic caps in a 40 year old amp are bad, but the main filter cap is your buzz and a short in those bias-caps is real bad news. At least for first-aid, get good-brand (like Panasonic) caps in 105 degree rating (Digikey.com stocks them).

 

[whitfiel]

Thanks to both of you who took the time to look at the schematic and post a reply. I'm starting to understand better how the OD3 may be compromized by my additional cap. Originally I didn't change C16 or C17A/B as testing by just adding a new cap in parallel didn't change the hum. I'm now supposing that when an old cap fails, just adding more (new) capacitance in parallel to the old won't necessarily fix the problem. Is that because the old cap is failing ohmic (shorted)?

Thanks especially for clarification on the main failure mode - the possibility of shorts in the 7027 power tubes causing pops and clicks. A replacement pair of tubes looks to be $45 or so (Parts Express, matched pair), which I can deal with.

Onward and upward with my education. I truly enjoy learning this stuff, and I appreciate your patience with my newbie questions and mis-guided repairs on the hum. I know the McMartin is not an audiophile amp, but it's easy to work on and fun to use for guitar. And that's the point - enjoy new things while learning.

Sincerely,
Art

 

[Burnedfingers]

Why play with a commercial amplifier? McMartin wasn't known for their superior circuit design but rather as a ho-hum 70volt line driver of limited bandwidth at best.

 

[whitfiel]

Why play with the McMartin? a) it was free and needed help, b) it has more than enough power for my son's guitar, c) the chassis is large enough to allow easy access to the wiring and components, and d) it was broken and I fixed it (initially). Since guitar is limited bandwidth, a PA amp like the McMartin certainly seems ok.

My primary audio amps are a pair of Ampex 2012 (6V6 output tubes) that I dearly love. A bit cramped inside to work on, but they sound great. So why risk messing up the Ampex as a learning tool when the McMartin was available. Please remember, I'm new to the world of tubes, so any tube amp is a new adventure. I don't have a basement full of tubes - just my Ampex (a gift from a friend), the McMartin, and a pair of 8W kits from Antique Electronics Supply (11MS8 output tubes). Sure I'd love a McIntosh or ST70 someday, and maybe sometime I'll build a pair of SE amps. But until I learn to maintain and repair the few I own, I probably can't justify something like a McIntosh.

Now if you want to discuss the complexities of various wind synthesizer controllers, I can get a deep as you'd like. But for tube amps, I'm just a novice.

Respectfully,
Art

 

[anatech]

Hi Art,
I don't see a thing wrong with what you are doing. I fix things just 'cause.

-Chris

 

[Burnedfingers]

But Chris, if he fixes a McMartin what will be next? A McGowan?, A Bogan, an Executone? My God where will it stop? A shelf full of commercial amps that only a mother could love? My God the horror of it all.

 

[PRR]

> just adding a new cap in parallel didn't change the hum.

Odd, illogical, but not unknown. And in any case, old electros can BURST. Not usually dangerous, but messy. If left unattended: rots the steel chassis. I know it is hard work, but get them out of circuit.

> to use for guitar

That trick with the gas-tube is strange, and IMHO not suited for guitar-amp duty.

Keep C16 in the 16uFd-32uFd range (not my 470uFd HiFi/PA overkill). A gitar amp should "sag": pass the strum of the transient clearly, drop during the sustain, rising as the note fades. Heavy tube-heads favor bottle rectifiers for this reason, but 700V 200mA is a big bottle and a major change. (Also less power, but measured power isn't so significant.) Not over-sizing the cap keeps the "sag" time constants about right for guitar work.

Use good 7027: made in the 1960s-1970s by a major-brand US or Euro company. I don't know if anybody has "re-issued" the 7027, but the majority of new-made tubes will not take the extreme high voltages that this amp runs. We used to run 6L6GC a lot harder than 450V, but I hear many new "6L6GC"s won't take that for long. That's a shame: you can get a replacement 6L6GC on a Friday night.

The old EL34(?) is a guitar favorite and rated 800V, but isn't as beefy as 7027: it might work well if you can put 16 ohms on the 8 ohm tap and accept about half the watts (which isn't half as loud).

At this lower power, self-bias (cathode resistor) makes at least as much sense as fixed-bias: less to go wrong, possibly more euphonic sound (guitarist's tastes vary so much, both set-ups have fans.) Even in fixed-bias, some cathode resistance (in series with F2) is wise: 10Ω gives a handy place to check bias; ~200Ω gives enough self-bias to reduce run-away while still mostly working like fixed-bias.

6550 was always good for 600V, and ElectroHarmonix has contracted for new production. EH tubes are usually good, and organists say the EH6550 works well in their Leslies (which is tough duty). But 6550 needs more heater power, will warm your power transformer.

All these tubes have similar but not identical pin-outs. You have to check closely: G2 moves around, and "NC" (NO connection) on one tube may be "IC" (INTERNAL connection) on another tube: watch for "NC" pins being used as handy tie-points and be sure they don't connect on the other tube-type.

Replace the wacky OD3 with 20K 2Watt. Use about 20uFd from G2 to ground. While the OD3 does have a lovely glow, its function here was dubious. It may have met a particular test-spec at a particular price. They may have had a crate of OD3 left-over from another product. But using a gas-tube as a screen-dropper makes the amp over-sensitive to supply voltage variation, risks failure of the OD2 on audio transients (you can't get a replacement OD3 on Friday night), and maybe cascading failure of the output stage. I may be fretting too much; but I know a resistor-divider WILL work in guitar-service and the slight drop in test-power isn't a big deal.

Yes, by HiFi amp standards, this 20K||75K screen-divider is "too high resistance", and it won't make maximum sine test power. But with ~20uFd G2 cap, the G2 voltage will hold-up through guitar pluck transients, while letting G2 voltage dip under sustained overdrive to protect tubes and ears. Some playing-around will be called for.

A PA amp is not a guitar amp. It is too heavy, as I'm sure you know by now. You don't want to carry 50Hz iron for an 82Hz instrument through 120Hz speakers. OK, your son will carry it. But also, it has too much feedback for classic open-back guitar speakers. Raise R29 to 10K or 22K. Gain will increase; ignore that and listen to speaker bass resonance. High damping kills the "growl" of classic guitar speaker systems. Lack of damping gives blatty bass and screeching highs. The classic Fender formula was damping factor pertty close to unity, instead of the 5 or 6 this amp has (and >40 for HiFi). Calculations are amusing, but the true test is the tone with the specific speakers.

Guitar amps also usually have a rise: gain above 2KHz is 6dB higher than gain below 1KHz. You can't exactly duplicate this with those tone controls. Again many different tastes exist, and several very popular amps did use essentially this tone control scheme.

The concertina phase splitter is a clean reliable driver in HiFi. It is faultless up to and just past the edge of clipping. But when driven far into clipping (as guitarists play for effect) it goes asymmetrical, which is not the usual sound.

You might at some point consider stealing the plans for a classic Fender. Fender used a modified long-tail phase splitter, a strange dirty thing but its gross-overdrive waveforms are part of "the electric guitar sound". (OTOH, Ampeg sold a lot of boxes very like this one.) The Fender tone stack is also the gold standard of guitar work, and actually simpler than what you have.

Oh: your "MIC Inputs" (without any transformers) ARE guitar pickup inputs. (Except C20 has to be a typo. If it is a 50uFd 6V across R51, try with it, without it, and with 0.1uFd on the cathode resistor to lift the treble.)

You might try removing R19. Or reduce R13. This will skew the tone action, but also allow V2A to be driven into fuzz while keeping the Master control fairly low and soft.

"Program" input is for MUZAK (and burnedfingers complains about Executone?). It could take a CD player outout. Use two 10K resistors to mix the stereo.

> A Bogan, an Executone? My God where will it stop?

Right there. Bogens are excellent minimal amplifiers. Executones are %$#@!. I think this McMartin is a dull boatanchor, but good practice. It has usable iron (if you don't worship NFB), solid chassis, enough sockets: good mule for modifications. And nobody else in town plays through a rare vintage McMartin.

 

[anatech]

Hi Joe,
What's next is a McBogen. Stacks and stacks of McBogens.

PRR has made some excellent points. Get it working and then try his ideas. Never leave dead electrolytics in circuit either as they do like to leak. Sometimes they short as well.

-Chris

 

[whitfiel]

Great postings, especially PPR's detailed comments. Wow, there's more to guitar amps than I ever imagined. No wonder I stick with wind synth for my own musical instruments!!

Lots of info that will take me some time to digest. I truly appreciate all the advice. Especially the 50Hz iron for a 120Hz speaker.... Excellent way to think of it. I'm starting to understand the PA amp comments!

Prior to reading the advice to get good quality tubes to better handle the voltage, I ordered a Sovtek pair from Parts Express. More affordable than NOS and I didn't know about the voltage issues with new ones. I'll let you know if we have interesting fireworks. The original tubes are RCA, complete with heatskink sleeves.

I had no trouble finding a spare OD3 back when I first got the amp. It wasn't expensive, and readily available from Antique Electric Supply. I think it was a Friday night too. <g> The comment on using a resistor voltage divider instead of the OD3 was interesting. Someone in an older McMartin thread replaced the OD3 with a transistor in a undiscolosed circuit. I can see using a transistor for a current source, but this just looks like a voltage drop - despite the OD3 being called a regulator tube. So how does the OD3 work? I don't see how a two terminal tube will serve as a Vdrop and Vreg. Data sheets were not revealing to my newbie mind.

Thanks again for the excellent comments and education. I truly appreciate it.

Best regards,
Art

Wind synth stuff at: http://members.aol.com/whitfiel/artwind.htm

 

[PRR]

> So how does the OD3 work?

It is a hollow-state Zener diode.

If you have a raw or too-high B+, you run a resistor and Zener in series to ground, and take a steady voltage across the Zener. Same with the gas-tube, almost. The gas has to be taken a bit above the rated voltage to start conducting, like a Zener it has a maximum current (in Zeners, usually given as Power) but it also has a MINimum current below which it does not ionize and regulate well (Zeners degrade gently at very-low current: soft-knee instead of intermittent ionization).

OK, now reverse the resistor and Zener. As supply voltage varies, what happens to the voltage on the resistor? It varies the same number of volts, but that is a larger percentage of the voltage. If B+ varies from 700V to 500V, and the OD3 is 150V, then the screens can run 550V to 350V, quite a large change.

One interesting angle is that a pentode's current is proportional to both G1 and G2 voltage. G2 is traditionally "fixed" and G1 varies from bias-point up to about zero grid-cathode voltage (few audio drivers can push G-K voltage positive). So the maximum possible current is set by G2 voltage (unlike transistors where the maximum current is essentially infinite). By dropping G2 this way, and using a B+ with a lot of sag (pretty common in this type amp), they could give a measure of protection from installer mistakes (like wiring 2Ω across the 8Ω tap, which is too-easy to do in large 25V/70V systems; or the popular dead-short). Load sucks big current, B+ sags, G2 voltage sags worse, peak current is limited, amplifier survives the warranty period.

> original tubes are RCA, complete with heatskink sleeves.

Heatsink sleeves?????? Can you snap a picture? I have never seen that on vintage gear. Yeah, someone now sells a damper-cooler, and it isn't a dumb idea, but In The Day we just relied on convection over the glass. (Larger tubes were fan-blown and even water-cooled, but not 7027-class stuff.) (Oh, wait: there is a ham-radio trick of putting metal 6L6s in a gallon of oil. One air-cool 6L6 is good for 40-50 Watts Class-C RF; they say four oiled 6L6 would put-out nearly the legal-limit, 600+ Watts, if you did not keep the key down too long.)

 

[whitfiel]

> Heatsink sleeves?????? Can you snap a picture?
> I have never seen that on vintage gear.

Here's a photo:
http://frontiernet.net/~whitfield/tube_heatsink.jpg


BTW: got the matched Sovtek replacement tubes today. The clicks and pops are now gone and the amp is working fine. I still have to replace the original power supply filter caps and ultimately remove the 47uF cap I put on the OD3 tube. But for now the amp is working and it's confirmed that one of the tubes was indeed arcing or similar.

How do I tell if the new tubes don't tolerate the high voltages as suggested by PRR? Is this just an early life issue, or will I see something like a faint glow looking inside the perforations on the square structure? I can't tell if that faint glow is a 'cherry red plate' or just glow from the fillaments (which are more orange looking on the top/bottom per usual).

Thanks again for all the interesting thoughts and teachings on this thread. I truly appreciate the education.

Best regards,
Art

 

[PRR]

I have never seen such coolers on octals. Quite odd.

> How do I tell if the new tubes don't tolerate the high voltages

7027 is rated 600V, you have 700V. On old tubes, I'd shrug it off. There would be a slightly increased infant mortality: tubes that arc or fail fast. And if the factory was buying standard-spec 7027, they could not return tubes that failed at 700V (they may have been buying special-spec tubes; big customers could negotiate selection or even special material). If they seemed OK after an hour, I'd ship them.

Since the tube racket shrank from an Industry to a hobby (and small change for poor Russians, Czechs, and Chinese), special nickel alloys and selected mica is no longer readily available. New tubes may not stand over-volting so casually. You didn't arc at fire-up, good. I'd watch the idle bias: it will drift a bit in the first hours as gas getters-out, but any long-term or large drift is a sign that the tube is going sour.

Ya know: at 700V and 100mA you are AT the plate dissipation rating and 17% over the plate voltage rating. You are sometimes told not to run at more than one maximum rating, so this is really pushing the poor tubes. At least for smoke-testing, I'd bias down to say 70mA total (35mA per tube) to get Pd down from 35W to 25W. The lowest distortion would be around 160mA idle and we already had to miss that for power dissipation reasons. The minimum is broad: going to 70mA won't make the sound awful, and is kinder on the tubes. After you have a few hours use, you might try pushing 30 or 35 watts and see if the sound improves; if not, idle it cooler.

BTW: orginal 7027 is rated 450V 25W. 7027A is rated 600V 35W. IIRC the early model was not out for long, and all production soon shifted to 7027A specs even if sold under the 7027 part number. I have no idea what specs Sovtek is working to, but there are enough hard-worked 7027 amps that if they only gave you 450V and 25W a lot of tubes would die within warranty, so I assume they make 7027A types.

http://www.triodeel.com/7027.htm has an interesting essay on the 7027, 6L6/6L6GC, 6550, EL34, and other big bottles. It appears that while each was designed for a specific niche, once in production they evolved toward each other and may be the same guts with different markings and a couple variant pins. "it was a case of juggling numbers to claim more 'horsepower' for sales purposes." (Note that a true 6L6 or 6L6G is a very different tube than 6L6GC and its kin. You can put the GC in a G socket and it will loaf; but a socket made to push 6L6GC to its ratings will melt 6L6/6L6G. Also the 6L6 has a kink that the later tubes lack, a convergence from KT66 detailing.)

The 6L6GC/7027 clan differs from the 6550/KT88/KT90 clan in heater demand: 0.9A or 1.6A. The bigger heaters tend to claim somewhat higher dissipation and voltage ratings. If your power transformer runs not-hot after several hours of operation, it can probably stand another 9 watts of heater power and use the "big" tubes.

> see something like a faint glow looking inside the perforations on the square structure

A tube has a bottle, a big wrap of metal, a small (~1mm x 2mm) central core, and some fine wire grids between.

The central core must glow dim-red. The exact red/orange color temperature is not critical: if the heater voltage is +/-15% of nominal and it glows, it will be fine. (Oxide filaments are a little critical, non-oxide filaments are very fussy; you won't find these in commercial non-transmitter gear because heater/cathode tubes are much more convenient.)

The outside wrap must not glow red (a few seconds is not fatal; but prolonged plate-glow ruins the tube).

The grids usually can't be seen well enough to judge glow, anyway the central cathode glow would overwhelm any grid glow.

Then there are the blue-purple glows.

It is very normal to have a ghostly purple glow on the inside surface of the glass, especially around holes in the plate. A few electrons miss the plate and hit the glass. The glass has trace impurities and some of them glow when hit by fast electrons.

It is very bad to have purple glow inside the plate. That is gas inside the tube. The electrons must move freely in perfect vacuum. If there is enough gas to glow from electron-hits, it is messing up tube bias and gain. After long storage, a little gas may be loose and it may clean-up after a few hot hours. But if it doesn't, the tube is probably leaking and will die fairly soon. Possibly catastrophically: tube current rises toward infinity despite bias, and with bad luck it lingers at some current that melts transformers but does not quite melt fuses.

The "chrome" inside the glass (top of your 7027, side in many tubes) is a thirsty metal that sucks gas. As it absorbs gas it turns brown (just like steel rusts). Brown all around is not a problem, any more than a little dirt in a vacuum-cleaner bag is reason to throw-out the bag. If a tube was made clean and sealed well, the getter will have at least a little silver even 80 years after it was made. If a tube leaks, it usually goes quickly from brown to all-white, which is the final stage.

 

[whitfiel]

PRR: loved the detailed description on tube basics, including the info on the silver getter. I am familiar with getters in flash lamps, so this was very interesting info on the silver getter patch in tubes.

But I truly enjoyed the article from
" http://www.triodeel.com/7027.htm on the 7027, 6L6/6L6GC, 6550, EL34, and other big bottles" as you put it. Fascinating trivia on tube numbers, market specsmanship, and compatability. Thanks for sharing that.

The Re-tubed McMartin continues to play well. I've ordered some new caps and will replace them sometime this summer. But for now the clicks, pops, and blown bias fuses are gone. My sincere thanks to everyone who helped in my ongoing education. I hope I can return the favor in the future, or pay it forward.

Best regards,
Art