Repair and restoration of my STAX SRA-12S

[bobkatz]

When I said that "Gilmore" may believe in floating the input RCAs it may actually be Justin, as that's how HeadAmp currently wires up its RCA input jacks on the KGSS. Anyway, I switched to grounding all input connectors directly to chassis, which includes XLRs, always ground pin 1 to chassis. This avoids the famous "pin 1 problem" outlined in a famous article from the JAES by Neil Muncy, reprinted in their grounding issue, which has sold out about 10 times. It's available as a reprint, highly recommended. Years ago I built an A/D converter with RCA inputs which I had floated from chassis. Without changing a single circuit, I removed the teflon washers, grounded the jacks to chassis, and the noise floor went down about 10 dB! Lots of RFI floating around in there!

[bobkatz]

By the way, Quad, if you still need an original owners manual and service manual I have both and can scan them if anyone wants. Let us here know....

[manaox2]

You have a 15 minute edit window to edit posts in the future also so you don't have to write a new post each time.

[luvdunhill]

Cannot load that schematic on my phone, but I'd guess that rail voltage difference is inconsequential. Assuming half of it drives a CSS which would have a change in the nA probably... I wouldn't assume more components is better than a good adjustable regualtor though, given how good those have become of late, when properly implemented.

[bobkatz]

You have a 15 minute edit window to edit posts in the future also so you don't have to write a new post each time.

I tried to edit some of my posts and it closed on me. I'd love to have the moderators increase that limit to a day.... By the way, will you sort my FETs for a bag of Raisinets? What's a CSS?

[luvdunhill]

I tried to edit some of my posts and it closed on me. I'd love to have the moderators increase that limit to a day.... By the way, will you sort my FETs for a bag of Raisinets? What's a CSS?

CSS is a typo. I'm not a huge fan of Raisinets unfortunately.

[spritzer]

He probably meant a CCS, constant current source. Edit: Indeed it was.

Ohh and I'd love to see the schematics.

[Quad]

Hello bobkatz,

Good to see that more SRA-12S amps are still in use.

You are absolutely right about the quality of the intermediate-stage amp and the phono board.

Normally indeed there is no need for the additional gain, and it causes audible noise as well.

Part of the cause of the noise is the high impedance of the volume control (250k if I remember well).

You did the right thing by choosing a much lower value.

Personally I feel that a balance control is indispensable. A lot of recordings are off balance, and my ears have some damage that needs correcting as well.

My volume control happened to have sufficiently good tracking, so I left it in place. More so, because I wanted to keep the balance control also.

I tried to keep my SRA-12S as original as possible, so I left all boards functioning and restored them all.

In practise I only use the direct input (via the vol/bal controls) to the output amps.

I think that the 2 uF is way over the top.

The headphone represents primarily a capacitor in the order of 100 pF. So you get a capacitive divider of the headphone and the coupling cap.

The (DC) charging time for the 2uF is also a bit long, since the charging resistor is 10 MegOhms.

I also feel that 400 VDC is on the low side. At most, 650 V could be across the capacitor.

During normal operation, the DC voltage across the coupling capacitor should not exceed much over 350V.

I do not know the voltage rating of the original output coupling capacitors, since they were already replaced by the previous owner (or most likely by his repair center for that).

The capacitors I used are rated at 2000V, but that is just because I had several of them already. And they, more or less, fitted on the pcb.

If you are in need of them, I can always mail you 4 pcs.

About the biasing the output stage:

At the time I did not have the equipment to measure distortion of a balanced system at a few hundred volts of audio voltage.

So I just adjusted for max output swing (as observed on an oscilloscope).

By the way, this setting corresponds closely with the prescribed value by Stax.

(At the time I only measured the distortion of one half section of the output, since you could treat that as non-floating. The distortion was quite large, but is virtually canceled out by using both output halves.)

Since then I have made an extention to my HP 339A distortion analyser, in order to be able to measure the output of this kind of circuitry.

If proved good value during the development of 2 amplifiers of (partly) my own design.

If I come to it, I will do some measurements on the SRA-12S.

More current might not necessarily "bias them better", since each transistor is a single-ended circuit with just a resistor as collector loading.

For a momentary output voltage towards the (positive) supply rail, the source impedance (to the load) is the collector resistor.

If you set up for a higher collector voltage, the collector current actually drops. Because the collector current is (supply voltage minus collector voltage) divided by collector resistor. Or to put it simple: Just the voltage over the collector resistor divided by that resistor. So setting a higher collector voltage actually drops the voltage over the resistor, hence lower current.

Just try some different settings, and most likely you will find that it makes no audible difference.

The diaphragm biasing voltage for the phones was originally 200 VDC, and 230 VDC for later types.

In fact it was 200 VDC at the time of the SRA-12S.

But most normal-bias phones, if not all, could stand a higher voltage.

I added a "pro" biasing output as well, and I put that at the recommended value of 580 VDC.

Again, most pro-bias phones would be able to handle a substantially higher value. The SR-007 f.e. certainly would.

It also depends a bit on the air humidity, how high you can go. And the risk you want to take. Once a flash-over has happened, I suppose that there is conductive material (carbon) at that position, and the element will never be as it was before.

For my ears there was not much difference in sound (quality), for relative small variations of the diaphragm bias voltage.

What does change, however, is the efficiency (or sensitivity as you could call it equally correct) of the phones.

The bias voltage is just there to put, via a large series resistor, charge on the diaphragm. The more charge, the more sensitive.

All that to a certain upper limit of course.

The (very audible) higher output (with the same input voltage) gives the impression of higher quality, but in reality it makes very little (if any) difference, if you carefully adjust (reduce) the sound output to the original level.

It just gives mote "total" amplification, but that has no relation with quality per se.

Just do add a high value series resistor to the headphone connector, if it is not yet in your version. It is suppose to supply current, not voltage.

The do or don't of grounding stays an interesting subject. If it induces a hum component, it obviously is not the right way to do it.

If it doesn't, it is to your personal opinion whether you want to ground it or not.

The SRA-14S for example is sensitive to grounding problems of the internal wiring.

I usually keep the input connectors floating, but for the utmost in EMC immunity, grounding might be the better option.

But the circuitry and component (cable) quality should always be such that it does not make a difference in the real world.

The +/- 15V supply is fine at +15.2 and -14.7. There really is no need to replace the circuitry. I just took care of good electrolytic capacitors and added a polypropylene across each supply.

Integrated circuits sometimes are better than discrete, but not necessarily so.

I briefly tried a setup with IC-regulators (just to be on the safe side, audio-wise), but was not able to detect any difference.

Quite in contrast with popular belief, I feel that the SRA-12S is a very good sounding amplifier!!

And yes, I would very much like to have a high resolution scan (300, or better 400 or 600 dpi) of the user manual!

I will send you a PM with my e-mail address. Thank you in advance for taking the trouble.

I am looking forward to it.

Best wishes,

Jac

[spritzer]

I for one don't doubt it is a good sounding amp, in fact I've been looking for a cheap one. I always liked the old hybrid model, the SRA-3S.

[John Buchanan]

I for one don't doubt it is a good sounding amp, in fact I've been looking for a cheap one. I always liked the old hybrid model, the SRA-3S.

The SRA3s was a much better sounding unit than the SRA12s - I can't give you any details from nearly 35 years later, but there WAS a night and day difference.

[bobkatz]

Hello, Quad and fellows. How do I get to be a high roller ).

Anyway, first things first, my assistant carefully scanned the SRA-12S operating and service manuals and I have put them publicly on my dropbox as PDFs!

So, here they are the links:

http://dl.dropbox.com/u/9188783/Stax%20SRA-12S%20Operating%20Manual.pdf

http://dl.dropbox.com/u/9188783/Stax%20SRA-12S%20Service%20Manual.pdf

If you've never usd dropbox before, it's a great way to share files and also to keep the latest synchronized version on your computer(s).

My apologies for also scanning 40 years worth of notes on these schematics. I'm sure you'll be able to navigate your way through them. The version which has the protection diodes and a few gimcracks I lost some of the pages, but there are VERY few if any changes in the updated versions and I'm sure you'll be able to figure them out. Plus there are board layouts of all the versions. And now for a complete round robin of Quad's post:

Hello bobkatz,

(SNIP)

Personally I feel that a balance control is indispensable.

By implementing two individual 10k ohm cermet input trim pots I was able to have my cake and eat it, too, and maintain increased channel separation, so I agree, a form of balance control is indispensable, just not Stax's version ). At this point in my SRA's life I don't care if it's original or just at its highest performance, I picked the latter. No one's going to be buying an old SRA, so its resale value is nill.

I think that the 2 uF is way over the top.

It may very well be, but the proper value of the time constant is R x C x S, where S is the Spritzer constant, which is at least 10,000 times what any normal human woud use ;-). I just followed Spritzer's comment that most later Stax amps used 2 uF. I'm sure your caps are fine, too.

I also feel that 400 VDC is on the low side. At most, 650 V could be across the capacitor.

Since I was only able to find a 400 v pair of MKP caps in this value I struggled through this choice. The original caps (as per the schematic you can download) are listed as 500 volt. Where does this 500 volt come from? Neither from the 650 volt B+ nor from the 200 volt bias, so I suppose Stax just found a number from thin air. I think that 650 volts can only happen if the transistor shorts and could blow up the phones. It is a possibility. Comments...

If you are in need of them, I can always mail you 4 pcs.

Thanks.... Let's see what Spritzer says about the 2 uF. Is it possible to measure frequency response accurately in situ? I guess with the load engaged. But you know, it was quite a pain to put these caps in, I'm not inclined to pull them out unless someone can confirm to me that we're in the danger zone with these 400 volters.

If you set up for a higher collector voltage, the collector current actually drops. Because the collector current is (supply voltage minus collector voltage) divided by collector resistor. Or to put it simple: Just the voltage over the collector resistor divided by that resistor. So setting a higher collector voltage actually drops the voltage over the resistor, hence lower current.

Right you are! I clearly was not thinking this through. I'll do your maximum voltage test with a scope soon and tweak the bias back down!

The diaphragm biasing voltage for the phones was originally 200 VDC, and 230 VDC for later types.

In fact it was 200 VDC at the time of the SRA-12S.

But most normal-bias phones, if not all, could stand a higher voltage.

Again, the church of Spritzer prevails and a 300 volt compromise seems very good to me for the normal and pro bias phones when used with this amp. Maybe I'll wire the second jack as Pro Bias if I get inclined. But soon I will be the proud owner of a KGSS so it will become academic.

The bias voltage is just there to put, via a large series resistor, charge on the diaphragm. The more charge, the more sensitive.

All that to a certain upper limit of course.

The (very audible) higher output (with the same input voltage) gives the impression of higher quality, but in reality it makes very little (if any) difference, if you carefully adjust (reduce) the sound output to the original level.

It just gives mote "total" amplification, but that has no relation with quality per se.

You are 100% correct that increased loudness causes the illusion of higher quality. But incorrect from my point of view that higher sensitivity means higher efficiency and more headroom in the amplifier for the same sound pressure level. And more headroom and running a lower drive voltage in these old SRA amplifiers translates to better sound to my ears. The farther from the clip point they run (within reason) the better they sound to my ears. Perhaps if we had a better supply regulation, etc. etc. etc. it would not be as much of an issue, but until then, I like to run higher sensistivity (higher output) phones, which means higher bias voltage.

Just do add a high value series resistor to the headphone connector, if it is not yet in your version. It is suppose to supply current, not voltage.

Agreed, but if you are speaking of the 4.7 M ohm buildout resistor, it's part of every Stax schematic.

The do or don't of grounding stays an interesting subject. If it induces a hum component, it obviously is not the right way to do it.

If it doesn't, it is to your personal opinion whether you want to ground it or not.

If your personal opinion is that you like EMI and RFI, who am I to argue -

The +/- 15V supply is fine at +15.2 and -14.7. There really is no need to replace the circuitry. I just took care of good electrolytic capacitors and added a polypropylene across each supply.

The latest opinions on paralleling supply capacitors from my friends at Prodigy are that they are not necessarily the be-all end-all. Especially considering ESR: Print Page - fun with paralleling film caps in a Mackie console

Theres more on this parallel cap controversy in other threads at Prodigy Pro. It made me take all the parallel caps that I had added years ago out of my SRA. And just use very low ESR Electrolytics in place of the old Electrolytic coupling caps. I certainly can agree that where a polyprop can be used to substitute there could be lower distortion in a coupling cap situation. Alone in a filtering cap situation I suppose it's better if it's a well-made cap. But parallel caps may not give you the results you think. At least do the research at Prodigy...those guys are smart.

Enjoy!

[Quad]

Hello bobkatz,

First of all: thank you for the manuals.

The notes on old manuals are always worthy of being read. Why invent the wheel twice?

So in fact you did implement a kind of balance control after all. Are they sonically of the same quality as the volume control or does that not make a difference?

I tend to feel that the sonic differences between volume controls are minimal. One of the main issues to me is very good tracking.

The output coupling caps: the listed voltage rating is probably just what Stax had at hand. Which indeed by no means implies a necessary value.

I expect that your output caps will sound similar to mine.

There is a school of thought that is against too high value capacitors. You could argue by using a largish value, you put a lot of (unnecessary?) dielectric material in the signal path. Which might have its own contribution to sound due to charge being kept in it.

Difficult to choose one way of thought of the other. I guess that the proof of the pudding is in the eating anyway.

About the 300V diaphragm bias: I am just too chicken to bias all my phones to a value which is much higher than recommended.

Since the room I use the phones in, is not humid, I suppose it could not hurt that much.

Since my amps can generate more than sufficient output voltages for the phones I use, that alone is not a reason for me to increase the value.

My loudspeakers are modified Quad ESL-63 electrostatics, so probably I am used to lower sound pressure levels than most people.

It would be nice to be able to learn from Spritzer if, by his knowledge, the higher bias voltage has ever caused harm!

It is relative simple to wire the second phone connector as pro bias. 300V is quit low for pro-bias phones, although they will of course work. But at a quite reduced output level.

I once planned to do some serious listening tests and decide for my self whether or not the bias voltage actually influences the sheer sound quality. Not talking about SPL.

But I have not come to it yet. I do, however equipped my home made amps with a variable bias voltage. I can set the value in steps of approx 2dB from 150V to 1500V.

(The steps of 1000, 1200 and 1500V are required for my Jecklin phones.)

Higher sensitivity indeed does mean more headroom in the amp, which sometimes is a major benefit. Depends on the amp of course.

Your action of making a feed-trough connection is certainly useful. I did the same on my own amps.

The bias series resistor is definitely not part of all Stax amps. F.e. see the service manual of the SRA-12S for s/n over1001.

C417 is directly across the bias output, and they did not put a series resistor in that line.

Concerning EMI and RFI: for high frequencies the current idea (in contrast to early thoughts) is indeed to groud as much as possible. Connect all connectors and screenings directly to ground.

For LF however, induction of flux from the mains or a nearby mains transformer quite easily induces a 50/60/100/120 Hz hum component in a ground circuit.

For European versions of audio equipment you see all kind of ceramic capacitors connected across input and output connectors. Just to be able to conform with the immunity regulations.

Usually ceramics do not a lot of good for the sound.

The induction of mains hum was a serious problem for Stax with the SRA-14S. The prototypes had a built-in mains transformer, but since the circuitry was too sensitive to induced hum, they decided for the final production series to use an outboard transformer.

My own amps have there input jacks isolated, but my main ground connection is somewhere in between the connectors.

Sometimes it is required to make an additional ground connection at a different location. Thinking things through combined with some trial and error work, will give a solution.

The are different schools of thought about paralleling capacitors, as you indicate. Audio Research for one, used to have all kind of paralleling in the deep past, but now often choose to have just the supply electrolytic.

Since the power supply actually is in series with the output, a very good quality supply will not go amiss. Which does not at all necessarily means electronically regulated...

I tried several power supply concepts for my DB Systems DB-1 pre-amp and could detect not a major audible difference. So I restored it to the original concept. My measurements do not confirm this outcome.

But I still feel that a good MKP does not do any harm

Thanks again for making the manuals available, and have a lot of fun listening to your Stax equipment!

Best wishes,

Jac

Edited August 31, 2010 by Quad
typo

[bobkatz]

Hello bobkatz,

First of all: thank you for the manuals.

The notes on old manuals are always worthy of being read. Why invent the wheel twice?

You're welcome. It's fun looking at history, especially your own scribbled history in the margins of schematics.

So in fact you did implement a kind of balance control after all. Are they sonically of the same quality as the volume control or does that not make a difference?

Unfortunately they are not as good quality as the stepped volume control I built. Anything that has a wiper running against a resistive element will introduce more potential distortion than a wiper on a gold-contact switch. Maybe someday I'll build some switched rebalance attenuators for the SRA, but for now the SRA project is finished. Also, predicting attenuation to an exact portion of a decibel requires knowing the impedances and since the controls are in a passive part of the circuit and the balance control would be in series with the volume control, I cannot work out how I would be able to add an accurate stepped balance control in series with the stepped volume control without adding a buffer amp, which I do not want to do.

However, for the KGSS I already constructed two individual 1 dB/step volume controls which I hope solves both the volume and balance issue. But if 1 dB proves not to be enough I'll construct another Goldpoint-switch resistive balance control with 1/4 dB steps up to + or - 3 dB, IF I can work out the impedance issue and not add a buffer amp.

I tend to feel that the sonic differences between volume controls are minimal. One of the main issues to me is very good tracking.

Sonically, it's all very subliminal and cumulative. Once I started listening to and building the best line-amp circuits for my various audio systems I became more of a purist than ever. Would I pass a blind test between a cermet trimmer and a gold-plated Vishay-loaded switch? Doubtful. But do I feel better having nice clean connections between input and output? Absolutely.

The output coupling caps: the listed voltage rating is probably just what Stax had at hand. Which indeed by no means implies a necessary value.

I expect that your output caps will sound similar to mine.

There is a school of thought that is against too high value capacitors. You could argue by using a largish value, you put a lot of (unnecessary?) dielectric material in the signal path. Which might have its own contribution to sound due to charge being kept in it.

Difficult to choose one way of thought of the other. I guess that the proof of the pudding is in the eating anyway.

I also think that a good direct-connect DC servo amp sounds better than any capacitor, so I'm looking forward to my new KGSS when Justin gets it done. In a line driver amp, I have done the shootout between a superb DC servo amp and a superb audiophile grade coupling cap and the DC servo ate the cap for breakfast. So I'm a fan of direct coupling wherever possible.

About the 300V diaphragm bias: I am just too chicken to bias all my phones to a value which is much higher than recommended.

I can't blame you. Take a look at Spritzer's enlightening post in this thread which covers both the cap issues and the diaphragm bias issues. It led me to believe that 300 volts is a reasonable value for normal bias phones as long as I don't play too loud (which I do not):

http://www.head-case.org/forums/headphones/785-headcase-stax-thread-342.html#post379470

The bias series resistor is definitely not part of all Stax amps. F.e. see the service manual of the SRA-12S for s/n over1001.

That's very surprising. I've lost that part of the page of my service manual and it's not in my PDF which I just posted. But it shocks me (literally) to think of building a headphone amp without a large series resistor, at least as shock protection (current limiting) for the user. If I were Stax, I'd hate to be the focus of a wrongful death suit. You don't want lethal current-capable voltages coming out of a headphone socket. Anyway, it seems that all standardized electrostatic headphone biasing circuits I've seen include that series resistor as a standard.

C417 is directly across the bias output, and they did not put a series resistor in that line.

Just shocks me! Can you point me to a copy of that schematic, please?

Concerning EMI and RFI: for high frequencies the current idea (in contrast to early thoughts) is indeed to groud as much as possible. Connect all connectors and screenings directly to ground.

For LF however, induction of flux from the mains or a nearby mains transformer quite easily induces a 50/60/100/120 Hz hum component in a ground circuit.

And as you described, the potential solution where you can solve both problems (for balanced circuits) include using a very small bypass capacitor between pin 1 and chassis. I believe there is a Neutrik 3 pin connector with an integral capacitor for this purposes. The RF rejection though, becomes nill at extreme frequencies unless you integrate this capacitor expertly.

I hate debugging hum problems!

[bobkatz]

Hello, Quad. I see that I did have enough of a portion of page 13 and did scan it to the service manual. In page 13, the "modified" Stax schematic for serial numbers that have the protection diodes, I do see that R403 4.7 Meg is installed. Are we talking about different things?

[bobkatz]

Hello, Quad. EDIT:

Pity I can no longer edit my post with the links to the manuals. Because the manuals are complete. Pages 13 and 14 are the pages which cover the modified schematics which include the protection diodes. I just didn't tape them together to make a large picture like I did with pages 11 and 12 (the original schematic version for older serial numbers).

Anyway, I see that in the newer schematic, I do see that R403 4.7 Meg is installed. Are we talking about different things? Perhaps someone rubbed this R out of your copy of the schematic (

[Quad]

Hello bobkatz,

Yes, we are talking about the same part of the schematic.

But... after the resistor R403 (4.7M) you find the capacitor C417 (0.1uF). After that there is is direct line to the bias-pins on the headphone connectors.

So in effect: no series resistor. Just an additional RC-filter stage in the bias supply.

The schematic prior to s/n 1000 does not have this capacitor. And thus does indeed have a series resistor in the bias line.

Things can easily be corrected for s/n over 1000 by just adding a 4.7M (or an other large value that you happen to have lying around) between the bias wire and the connectors.

Best regards,

Jac

[spritzer]

Stax did experiment a lot with the bias supplies and they still vary a lot from amp to amp. I can't say I'm a fan of having a cap after the ballast resistor though...

As for output caps, the Woo Audio GES (which is the all triode amp Dr. Gilmore designed and can be found on Headcase) has 2.2uf/630V output caps but Stax have always used smaller caps in their amps. Their latest capacitor coupled amp (the recently discontinued SRM-300/310) used 100nf. Now I know Woo has been playing with the 2,2uf value as at least one amp had 3.3uf which is also the value used in the larger WES.

I do like the Spritzer constant. It says a bit about that I just got a large shipment of Stax stuff from Japan and I had no recollection of buying some of the equipment. Still it's not bad to find a mint SR-Lambda I never knew I had...

[Bidoux]

As for output caps, the Woo Audio GES (which is the all triode amp Dr. Gilmore designed and can be found on Headcase

Headwize ?

[spritzer]

Yup...

[bobkatz]

Hello bobkatz,

Yes, we are talking about the same part of the schematic.

But... after the resistor R403 (4.7M) you find the capacitor C417 (0.1uF). After that there is is direct line to the bias-pins on the headphone connectors.

So in effect: no series resistor. Just an additional RC-filter stage in the bias supply.

The schematic prior to s/n 1000 does not have this capacitor. And thus does indeed have a series resistor in the bias line.

Things can easily be corrected for s/n over 1000 by just adding a 4.7M (or an other large value that you happen to have lying around) between the bias wire and the connectors.

Best regards,

Jac

I see. I think that we are not exactly arguing about the concept of there being no ballast resistor, but rather there being an extra filter cap after the 4.7M ballast resistor. So let's say, for purposes of discussion, all the Stax amps contain this resistor, but not all of them have the extra cap. And so the discussion begins. What would be the SONIC EFFECTS of this capacitor being there or not being there? It does "regulate" the bias voltage, but it also makes for a VERY slow charge up of the bias to the phones.

[Quad]

The charge up is not that noticeably slow, since the RC-time is only 0.47 s (4.7 MOhms x 0.1 microfarad).

Before you start the CD, the membranes will be already fully charged if the charge build-up time would be due only to the series resistor in the bias supply.

But there is a low impedance (w.r.t. the impedance of the headphone elements, which is about 1000 times larger than that of the final capacitor) in parallel of the output of the bias supply.

So it certainly is not just a current source, but much more a voltage source, due to the reservoir capacity of the capacitor.

I suppose that the sonic effects are minimal, since the defining factor is the speed of the charge migration on the diaphragm itself.

Reason: The time-constant of the diaphragm and the series resistor is only about half a ms, so well in the audio range of the phones. So in order to have low distortion, the series resistance of the diaphragm itself, must be very much larger than 4.7 MOhms.

If the charge stays "put" during the slowest mechanical cycles of the diaphragm, the distortion is minimal. If the membrane would be (good) conductive, which is the case on some electrostatic loudspeakers), then an improvement can be made by using a very high value series resistor in the bias supply.

I think that the function of the series resistor is primarily a matter of safety. If you touch a connector (or pin) which is at a high potential, a very noticeable and potentially dangerous current can flow through your body.

Most likely you touch the (grounded) cabinet of the amp when plugging the connector in or out, which makes up for a good return path for the current.

Furthermore, due to the small shock, you might make an uncontrolled movement and hurt yourself or damage something.

If, however, there is a high value series resistor, the current is limited to a safe value.

[spritzer]

The resistor is indeed there just to protect the user.

As for the cap, most of the older amps have it. I've been working on a SRM-Xs today (introduced in 1990) and it has it, same as the SRM-1 Mk2 PP (circa 1990 as well). I do think it is there just to provide some extra filtering as for instance in the SRM-Xs the bias supply is taken directly off the B+ rail. With the minimal filtering found in the amp it should help.

[eureka125]

Hi, Jac,

I have one of these and it has one channel that is very weak.

I had it repaired 10 years ago By a gentleman in Toronto who has since passed away. They rebuilt one board completely. I'm thinking that it's the other board that is now failing.

Would you feel capable of repairing this unit or know someone who could?

Rick Yurick

PS the reason I want to continue to use it is because I have a decent pair of headphones which I purchased at the same time as I had the unit rebuilt. My original headphones from 35 years ago died

[spritzer]

I'm bumping this since I have a SRA-10S here in for complete restoration. The unit is in rough shape with a note from the Japanese seller that it kept blowing fuses. No wonder when resistor are clearly failing all over the place, diodes look suspect and even the RCA's on the back are utterly fubar. Transformer is 100V only (even ripped it apart to look for hidden windings) so I need to have a custom unit made but the amp was cheap enough so why not have a little project.

Now my issue is what exactly differentiates a SRA-10S from a SRA-12S? These look all but identical too me and the schematic I posted earlier in the thread is a perfect match for the SRA-10S. Sure some parts are missing and/or the wrong value but that is par for the course with Stax schematics. They draw the same power from the wall and all the SRA-12S amps seem to have the model number on a sticker on the back, not silk screened directly on the metal. Could be the same amp but different names for different markets?

[spritzer]

The work continues to rebuild this beast from the ground up. The input sockets were way beyond salvage and so was the power cord so time to break out the drillpress:

Now the PSU was utterly fubar with scorch marks next to the HV diodes so I'm doing a 100% rebuild. Same deal with the amp boards, they will be completely rebuilt except for the transistors which I hope are ok.