Repair and restoration of my STAX SRA-12S

[Quad]

This is my fist posting for this forum. I hope it’s to your liking !

Apart from the audio qualities, I am also interested in the technical qualities of equipment. So my posts sometimes will be a bit technical. Just skip the uninteresting bits. If that is almost all: so much the better. That way you have more time left for enjoying the music ;-)

Quite recently I finally managed to obtain a Stax SRA-12S pre-amplifier with built-in ES headphone amplifier.

According to the previous (first) owner the amp got very hot.

Well, since it is supposed to run quite hot, I was not alarmed and bought the SRA-12S unseen (well…. just a photo).

Unfortunately one channel of the ES-output was not functioning properly.

Spritser was so kind to mail me a photo of the schematics, and having a schematic and/or service manual certainly always makes repair a lot easier.

First I checked all electrolytics, and I only needed to replace 3 of them.

All the others were in excellent condition. Quite an unexpectedly good result, given the age of the SRA-12S.

As you all know, you can consider the power supply to be in series with the output (for the audio signal), so the supply quality has to be as good as the amps themselves.

Two of the defective electrolytics were at the final filtering of the + and – 15V regulator board, so I replaced these with MKP’s.

Not sure if it really gives an audible improvement, but it won’t hurt either. And film capacitors will outlive me (if I manage not to touch the 650V too often).

The defect in the HighVoltage amp turned out to be a bad 2SC1167 power transistor.

The fault was intermittant, drawing more current and blocking the audio signal. But in between blocking, the signal was not steady and gave a rustling, noisy, clicking sound. So a replacement transistor was to be found. To be absolutely sure I checked all (old) transistors I had, but alas…. no 2SC1167 amongst them.

It is a HV transistor in TO-3 package.

I considered using transistors in a different package and adding a heat sink, but decided againt it for aesthetic reasons. If I could find myself a pair of suitable TO-3 substitutes, that is.

I measured all HV transistors that I have, on my Tesla transistor tester, and compared them with the 2SC1167s in the amp.

Max current specifications of the transistor is not an issue, because all these large HV transistors are made for several Amps. And in this application the quiescent current is about 7.5mA. So the remaining important spec points are max C-E voltage and amplification factor at very low currents. For this I measured the DC collector current at a DC base current of 1mA.

A logical transistor would have been the BU208, but the Hfe measured way too low.

Despite the fact that I only have a limited number of transistors at my disposal, I was very happy to find 2 transistors (type Philips BUX88) that match and perform very well, and that can handle the high voltage. I am sure there are a lot of other types that will perform equaly well. These transistors have a large spread in the Hfe at such a low collector current, so it will probabply always be advanteous to select them with a tester.

Mounting and soldering them in the amp proved that they are working like a treat!

Problem solved, now there is room for some subtle improvements.

The original output capacitors of the HV amps had been replaced by the previous owner (or the importer) a long time ago.

The original value is 4.7nF. Since other STAX amps have higher value output capacitors (often 50nF), I decided that a somewhat higher value could not hurt. So I replaced all four of them with a high quality Philips MKP 10nF-2000V.

These are quite a bit larger than the originals, but can be fit in just about.

These are important capacitors, and only very good quality types will get you the very best performance and reliability.

The SRA-12S that I have is of the first version, all be it of a highish serial number. The later version looks (and probably performs) identical, but has input protection diodes (from input to +15V and from input to –15V) for all amplifier circuits. I added this to my SRA-12S as well, just to be on the safe side.

The later version also has higher capacity filter capacitors for the + and – 15V. So I added additional smoothing capacitors to the originals.

An MKP (polypropyene) capacitor across the +650V power line should possibly improve the quality of the HV, so I put a 220nF capacitor across the 650V.

The later version SRA-12S also has a 100nF-250V cap across the bias supply, so I added this capacitor also.

I am not decided yet if that is actually to be considered an improvement or not. Given the fact that the coating on the diafragm has a very high resistance, it should make no real difference except maybe filtering out that last bit of 100/120Hz ripple. The charge on the diafragm should be constant for low distortion, and due to the high surface resistance, this will be the case. If the surface resistance would be quite low, I would prefer not to include the 100nF cap. In that case the series resistor (4.7M) is in the charge circuit, acting with the capacitance of the speaker (approx 100pF) as a time constant, to prevent the charge on the membranes from changing to fast.

I did some measurements on the SRA-12S, and this amp proves to be excellent.

The mains voltage in this country is 230V, so I set the voltage selector for 240VAC (the nearest higher value). It was still on the 220V setting, which corresponds with the local mains voltage when the unit was sold.

The maximum output voltage (without load) measured about 1080Vpp or 380Vrms.

This is dependant of the setting of the variable resistors on the output amplifier boards. I adjusted for max output voltage swing, and equal collector voltage of the 2SC1167s. (The collector voltage for max output turnes out to be 280VDC; with higer mains voltage, probable 290V will be best.)

The output with connected phones is a few percent lower.

Not bad results at all.

At this (lower than nominal) mains voltage, the high voltage supply is 630VDC. At 240VAC mains, the max output will be about 4% higher. Which, of course, is of no audible advantage whatsoever (<0.4dB).

I was not able to measure the distortion of the HV amp (because of the balanced output).

The I.S. (inter stage) amplifier gave the following results (measured @1kHz with HP339A with 30kHz and 80kHz filters engaged):

@ Vout = 0,5Vrms: d=0.009% (left and right channels). The distortion residue is all noise.

@ Vout = 1Vrms: d=0.0085% (left) / 0.007% (right). Noise and 2nd harmonic.

@ Vout = 2Vrms: d=0.0145% (left) / 0.011% (right). Predominantly second harmonic.

First class, certainly since the distortion is very low and almost only 2nd harmonic.

Bias voltage is 197V (@ Vmains = 230VAC), which is very close to the specified 200V.

The “normal bias” later changed to 230V.

Input sensitivity with balance control in mid position and volume max, is 95mV for 1V out. Both channels have identical sensitivity.

The sound is absolutely first class (on my SR-X Mk-III).

I woundn’t be surprised if the SRA-12S proves to be one of the better STAX units.

I will do some comparison with my SRD-7, which I have always liked.

I would be delighted to hear any experiences with the SRA-12S from the forum members!

Does anyone still have a SRA-12S and how does it sound to your ears?

Maybe there is someone out there having the user manual? A copy or high resolution scan certainly would not go amiss ;-) Thanks!

A test report from one of the magazines of the time would also be appreciated!

At the time (end of the ‘70s) it certainly must have been state of the art.

In the attached pictures you can see:

* the output voltage swing, just before clipping (200V/div)

* the inside of the SRA-12S

* the outside of this fantastic amplifier

[spritzer]

Great writeup Jac. I was sifting through my mountain of Stax printed materials and found a photocopy of a brochure for the SRA-12S which I'll scan or at least take a picture of when I can remeber.

[Tachikoma]

Interesting stuff

Since the SRM-1mk2's maximum output is also 380Vrms, maybe they have identical amplification sections?

[Quad]

Interesting stuff

Since the SRM-1mk2's maximum output is also 380Vrms, maybe they have identical amplification sections?

Yes, it would certainly be interesting to see the schematic of the SRM-1 and SRM-1mk2.

maybe Birgir has some info on these schematics in his vast collection?

Jac

[spritzer]

I don't have those schematics unfortunately.

[catscratch]

Fantastic writeup. But,

Spritser was so kind to mail me a photo...

Is that Spritzer's Egmont-building cousin?

*Jeremy Clarkson voice* Some say that his brain is biased to 580 volts, and "EL84" really is his middle name. All we know is, he's not Spritzer, but he is... Spritzer's Egmont-building cousin!!

or maybe:

There are those who say that "Stax" was his first ever uttered word, and that he can replace the pads on an O2 Mk1 in his sleep, with his toes. All we know is...

Bah, must make stig-style intros! Can't stop!!!

[spritzer]

....

[Quad]

Great writeup Jac. I was sifting through my mountain of Stax printed materials and found a photocopy of a brochure for the SRA-12S which I'll scan or at least take a picture of when I can remeber.

Hi Spritzer,

Yes, a high resolution scan would certainly be appreciated !

I was also wondering what the real differences between the SRA-10S and SRA-12S are.

You wouldn't happen to have a schematic of the '10 as well?

There is talk about FET output transistors in the SRA-10S, but on photos I definitely see 2SC1167, the same bipolar transistors as in the SRA-12S.

Maybe the major difference between the two is the type-number

Edited April 27, 2009 by Quad
typing mistake

[justin]

^ This belongs in a Newbie's Guide to First Posts on Head-Case

[spritzer]

Hi Spritzer,

Yes, a high resolution scan would certainly be appreciated !

I was also wondering what the real differences between the SRA-10S and SRA-12S are.

You wouldn't happen to have a schematic of the '10 as well?

There is talk about FET output transistors in the SRA-10S, but on photos I definitely see 2SC1167, the same bipolar transistors as in the SRA-12S.

Maybe the major difference between the two is the type-number

They could have used the same circuit, wouldn't be the first time Stax did so. I don't think I have the schematic though to verify this.

^ This belongs in a Newbie's Guide to First Posts on Head-Case

Sure does.

[swt61]

^ This belongs in a Newbie's Guide to First Posts on Head-Case

X3

Excellent write up, and welcome aboard Sir!

Fantastic writeup. But,

*Jeremy Clarkson voice* Some say that his brain is biased to 580 volts, and "EL84" really is his middle name. All we know is, he's not Spritzer, but he is... Spritzer's Egmont-building cousin!!

Almost spit up my drink! That sounds just like him talking about "the Stig".

Edited April 28, 2009 by swt61

[spritzer]

Enjoy.

[Quad]

Thanks to Spritzer for the photos of the folder sheet!

Thank you also to the other members for the positive reactions to my first posting !!

I hope to find out some more about the SRA-12S, but the info definitely is thin on the ground.

On ebay I came across an old issue of "FM Guide" (1976), which appears to have a lab report on the SRA-12S.

But at over $ 27 (incl. postage costs) it is a bit over the top for me, for just 2 interesting pages....

I have quite a lot of old hifi magazines, but of course nothing on this fabulous amplifier

In the mean time I will just enjoy the sound and hope to get some more reading material later.

Thanks again to you all, and any experience with the SRA-12S (or SRA-10S) would be nice to know. Even hearsay...

[catscratch]

Almost spit up my drink! That sounds just like him talking about "the Stig".

Some say that he was born with a tube tester on each hand, and that all his account numbers end in "6sn7..."

[HeadphoneAddict]

Good job!

[faust3d]

<snip>

Enjoy.

I remember certain someone sometime ago promised to scan a lot of Stax brochures. Oh well, I am going to use this one to add to the "knowledge base" in the time being

[spritzer]

I still haven't got a scanner so I have some excuse...

[faust3d]

I still haven't got a scanner so I have some excuse...

Take your time I am sure the wait will be well worth it

[spritzer]

[oldstuff]

I've got my SRA12S sitting in front of me with the lid off and the output card supports removed. Noticed a lot of heat discoloration on the boards by the big resistors and some on the boards by the top side of the output devices (on the ground contact). I'm assuming this is fairly normal. Otherwise, I'm very impressed with the parts quality on this thing.

I'm getting a little bit of noise on the line level, and a noticable amount through the phono, which I guess isn't that unusual. Also, the #2 pre-outs are dead, which is kind of a bummer.

Spritzer, if you would be willing to email me the schematics, I would be greatly appreciative. Also, anyone with upgrade advice, I'm open as well. I've done some tube amp rebuilds and tube preamp upgrades, but this is my first run-in with solid state. The cards will simplify this whole thing.

Quad, that was an excellent write up. Too good, in fact. I read it and pulled mine off the shelf. I'm pretty happy with the sound once it warms up, but wish it was a bit more open on top. I imagine new caps would go a long ways to that. I had a pair of Dynaco MkIIIs that went from "blah" to completely amazing with a cap upgrade. Anyway guys, I'm open for suggestions and advice.

Mark

[Quad]

Hello Mark,

Thank you for your comment on my write-up.

I am glad to see that there are some more 12's still in existence .

Yes, there is some noise on the line inputs. When I use the SRA-12S with my SRX-mkIII, and use a cd-player at an aux input, the amplification is sufficient without the interstage (I.S.) amplifier. Just push the most right button. Then the noise is gone. The amplification in that case is the same as of the current Stax amplifiers: 60dB, or 1000x. (Some recent ones are 54dB, or 500x, to have the complete information.)

It certainly pays off to use the best quality capacitors you can find. No need to use exotic types, though.

A small investment in good quality polypropylenes, however, is justified.

And, what most designers seem to forget, the power supply in fact is more or less in series with the audio signal circuit. So not only the amps themselves should be good, but the supplies as well.

A parallel polypropylene capacitor across the 650V certainly does a lot of good. I used 220nF, just because I had one. But I guess that 100nF will be equally well.

And if you have an old type SRA-12S, it will still have 1uF electrolytics on the small power supply board. Replace them by 1uF (50V or higher) film capacitors.

Check all electrolytics, and when in doubt, replace them. Do use good quality types. in my '12 I only had to replace the 1uF and a 220uF-6.3V one. The others were in absolute remarkable condition, and measured (at capacity and loss factor at 1 kHz) better than many new samples.

I added 2 capacitors 1000uF-35V for the + and - 15V. The '12 I have "only" has 1000uF originally, and the later versions already have 2200uF.

And I put a 100nF MKP across the 24V.

The output capacitors had been replaced by the previous owner (I guess that this was done by the Dutch importer). The original value is 4.7nF. A larger value does not hurt, and I have some 10 nF 2000V Philips MKP capacitors, so I replaced each output capacitor by one of these. So, 4 in total.

In the picture you can see them on the output boards: the yellow blocks.

Philips has made some very good components. Note that not all of the yellow blocks are of the same dielectric (and thus audio quality...).

The large resistors and power transistors on the output boards get quite hot, indeed. The transistors dissipate about 2W each, and the resistors 2.4W each.

Take care in properly adjusting the mains setting and the output amplifiers (balance and voltage: the collectors of the transistros should have the same voltage, and the voltage between collector and ground should be between 280 and 300 VDC).

The #2 pre-outs are switched with the middle one of the three switches ("output"). Pre-out #1 is always on, pre-out #2 is off if you use the phones. So that appears to be correct in your case!

Finally check the bias voltage. If you do not have a meter with a very high input resistance, just measure before the 4.7M resistor.

The circuit of the amplifier looks relatively simple, but it has quite some interesting ingenuities embedded. Absolutely first class for the period and certainly, with some attention, still very very good indeed today.

You don't happen to have a user's manual, do you? If so, I would be delighted to have a (high quality) scan of it.

Succes with the restauration of this excellent piece of engineering!

If you have any remaining uncertainties, do not hesitate to ask.

Best wishes.

[oldstuff]

Sorry, Quad, no OM here. I bought this thing used from a dealer of mine who got it in as a trade. Years ago, I stumbled upon a pair of SR3As with the driver box at a church yard sale for the "princely" sum of $5. When I saw the 12S on the used shelf in his store, I just about freaked. What are the chances of finding one in a small town in Iowa? Needless to say, the price was also "right", since he had been unable to sell it and was happy that I wanted it. There was nothing but the unit.

I thought the SRA12S would make a good one piece headphone driver, never thinking of using it for a pre-amp when I bought it. However, not being able to leave things well enough alone, I had to try it. At first, it was kind of marginal, but after about 30 minutes, it really came to life, and showed a lot of promise. I've owned this 12S for about 3 years, but have really just started using it as a primary component. As you mentioned in your original post, it is an old piece, it's time for some upgrades.

Not sure where in the production run mine is, SN is low, in the low 1200s, which is probably not that low for a company like Stax.

Thanks for the information

Mark

[spritzer]

Here is the schematic:

[bobkatz]

This is my fist posting for this forum. I hope it

[bobkatz]

Oh I meant to ask about the +/- 15 volt supply. Mine measures about +15.2 on one side and -14.7 on the other. How important is that matching? Noise is about 1 or 2 mv AC RMS. I'm thinking about replacing the entire board with an LM-317/337 style regulation, but this regulator board has so many discrete components it has to be better than an integrated circuit, right?

I'm sorry to have revived this thread, but hope that the participants are still enjoying themselves!