The ultimate DIY? A Stax SRM-T2!

[kevin gilmore]

one channel works, the other does not.

pretty sure that the tops of the batteries on the channel that does not work are about 10 or 20 volts

and should be 200. Will know more once i know the cathode and grid voltages of U2.

Could be that the servo is messed up.

Throw the circuit breaker if all turned on at once?

Actually all the sources are on at the same time. About 5kw total source power.

[Inu]

Q4 and Q5 are OFF?

Check Q1,2,3, D2, D3, R1,2,3 with multimeter at power OFF.

To confirm 1.2V across R1, R2 and R3 (5mA/2.5mA/5mA).

Swap 6DJ8’s...

[digger945]

Looks like the zeners are in correct. I can't quite make out the one on the right(if you're facing the front of the amp), and can't make out the color bands on the resistors, but it kinda looks like R1 and 3 are 740(transposed 470, which should be R2) ... it's hard to see.

On another note, it looks like you have both C12 and C13 populated. KG's notes claim use of one or the other, but not both.

Edited December 21, 2010 by digger945

[blubliss]

Thanks for all the comments, just got off the phone with KG.

Confession time, when we first fired up the amp section we had the +500v and -560v wires swapped on both channels so KG thinks this may have fried Q30 or 31. We are swapping one of them now.

Digger, we did remove C13 or C12, one of them.

[kevin gilmore]

blubliss goes to confession...

Gotta love current limited power supplies !!!

And one channel survived completely intact anyway.

Pretty strong circuit design.

[blubliss]

Ok, getting things a bit more in control but still problems on left channel. Replaced Q26 which got all the batteries up to voltage, cathode correct on the 6DJ8, and D24 lit, but output is high on left side. The 100v zeners seem to be conducting both ways and the other 2sk216s probably need to be replaced as Inu made reference to in his instructions to set things up.

Voltages at stax jacks:

L+ -432

L- -148

R+ -25

R- -25

B 397

Close but not there yet. It certainly could be worse with the level of fail we achieved with the wiring

[Kerry]

I'm guessing you are measuring the bias on the jack side of the 5M resistor. That makes sense since most meters are 10M ohms. If you calculate the potential divide it puts the bias at about 595V.

Once you have the zeners swapped out, I'd be curious what the gates of Q4,5 are as well as the test points at R99, R101.

[Craig Sawyers]

The 100v zeners seem to be conducting both ways and the other 2sk216s probably need to be replaced as Inu made reference to in his instructions to set things up.

I agree. Dead 100V zeners seems to mean dead K216's. Been there, got the badge (root cause for me, crap C3675's). Don't beat yourself up too much about the wiring error - the good news is that one channel survived the trauma. This circuit seems to be a fascinating mix as tough as old boots, and also kind of frail at the same time.

Incidentally, got 115 C3675's coming from three different sources now. If all are good, that is a result - plenty for other projects. But three sources hedges my bets - all three lots are unlikely to be simultaneously flakey.

[luvdunhill]

How do zeners die? Over current? Would higher wattage parts serve to protect the mosfets?

[Kerry]

I think the 100V zener's job is to protect the J79's, which have a voltage rating of 200V. I'm guessing they did that, but would love to know for sure (on my premise and result).

I'm not sure of all the root issues that could cause this to be triggered. Most everything is current limited so if we looked at most common failure points, then we could get to how much current & voltage the zeners would need to handle (without damaging the J79's) in a fail situation. At the end though, the zeners are pretty easy to replace so I wouldn't mind that too much. Seems like tube rolling with the wrong tube could be a likely fail situation.

I'd love to get other thoughts... Also, I haven't addressed Marc's question of how do they die.

[struts]

Hi guys and seasons greetings,

After a very long hiatus caused by work and waiting for my RK50 which has FINALLY arrived I am 99% ready with the build and almost ready to start testing. Few questions regarding final assembly:

1. Wiring the umbilicals. What are people doing here (in terms of assigning wires to pins)? I guess it's not too important that power supplies are interchangeable between units (at least not for me up here in Sweden!) but there must be some arrangement that works better/best with this mixture of high and low voltages, AC and DC.

2. Related to the above I was thinking of doubling or quadrupling up on the heater wires (looks like I have plenty of pins) as it will be carrying a couple of amps and I am worried about heat build-up around the pin if I only use one. Anyone done this?

3. Finally, max length? I was thinking of about 1m.

3. Is the 'collar' of the mains receptacle meant to be in the inside or the outside of the case? Seems to be enough room to mount inside and the switch is still reachable. Outside would work too but it would require a bit of forcing and I am worried about scraping off some of the anodizing and creating an additional earth path.

4. Similarly with the Amphenol chassis plugs/sockets. Seem to work fine with the chassis plate inside the chassis but is this what was intended? Any pros/cons I'm missing?

5. Finally, I have in mind that there was a late modification involving piggy-backing DIP sockets with a passive component in between but I cant for the life of me find it in the thread. Could someone post a short description, a link or the magical search terms?

Many thanks in advance.

[Craig Sawyers]

II'd love to get other thoughts... Also, I haven't addressed Marc's question of how do they die.

Well in my case the C3675's in the battery went short circuit. That massively overloaded the bottom K216 which went short. That massively forward biassed the 100V zeners, which glowed briefly before blowing. The chain of failure also took out D6/D7 by massive overcurrent.

Which is why I recommend testing your C3675's for DC current gain (ie *not* small signal slope gain) at 5mA, should be 70-ish or more for good 'uns. 20V is quite enough to test at, and assume that the gain increases by around 10% by 740V (because of the slope in the collector characteristics).

[struts]

Sorry people. I seem to be suffering massive late-night brain-fade. I have figured out the answers to most of my questions.

Now bed.

[Craig Sawyers]

3. Is the 'collar' of the mains receptacle meant to be in the inside or the outside of the case? Seems to be enough room to mount inside and the switch is still reachable. Outside would work too but it would require a bit of forcing and I am worried about scraping off some of the anodizing and creating an additional earth path.

Struts - I know you've answered your own questions, but I thought this might be worth elaborating. The chassis components are all anodized, which is a very good electrical insulator. With such high and lethal voltages I was concerned that isolated chassis components could be a significant hazard.

The rear of the mains chassis plug is a case in point. I drew a pencil line around the ouline and then carefully used a dremel with grinding attachment to remove the anodizing behind where the plate mounts. I also dremelled the anodizing off areas when the chassis parts bolt together - the front and back plates where they mount to the heatsinks and top and bottom, the top and bottom where they mount to the heatsinks and front and back plates etc. So now all chassis components are decently electrically bonded. The ground wires in the PSU and amp were fixed with butch star washers to the spare holes in the heatsinks in accordance with safety standards.

I might blow the bugger up electrically (and have done) but once it is working and cased up it will be safe

[struts]

Good point Craig, I was thinking about the receptacle-to-chassis interface but not the interfaces between the chassis plates themselves. Of course the WHOLE chassis must be tied to safety ground. Doh!

[kevin gilmore]

I predict that blubliss should be listening to a wonderful staxmass very soon.

Both channels now adjusted to the right numbers.

(servo's are a wonderful thing when they work)

[Craig Sawyers]

I predict that blubliss should be listening to a wonderful staxmass very soon.

Both channels now adjusted to the right numbers.

(servo's are a wonderful thing when they work)

Fantastic news! Is that number 3? KG, Inu and now Blubliss?

I'd definitely be #4 if the XXXXing C3675's had arrived before a 4-day holiday. But they didn't, so I'll have to slum it with a BH for a while longer

[blubliss]

Closer now, but not listenable yet. We have a strange problem, static on the right channel and a hum on the left (left channel is where we replaced parts.). The voltages seem good.

The static is quite strange and I assumed it was tube related. It comes and goes, is not consistent. We moved all tubes side to side, same issue; changed output tubes, same problem. Two different headphones, same thing, so not dust in the headphones.

[Inu]

I have a same static noise problem. When tapped top of the 6DJ8 lightly, the noise stops.

I thought tube is bad, but no problem on the other amp... I am going to replace the tube sockets.

[blubliss]

Hmm, it really does sound like a tube problem. I replaced all the 6DJ8s and same issue. I will take a look at the sockets, got mine from DIY Hifi, you too?

Edited December 25, 2010 by blubliss

[Craig Sawyers]

Hmm, it really does sound like a tube problem. I replaced all the 6DJ8s and same issue. I will take a look at the sockets, got mine from DIY Hifi, you too?

Kevin - any insight from Stax on this? Is there any indication that they selected 6DJ8's, or that the screening can on the original T2 played a critical role?

Now 22:40 in the UK, got a belly full of excellent food and booze - so I'm retiring to a place of horizontal repose.

[kevin gilmore]

i've tried to get information on the tubes before. did not get anything useful.

On my unit which by now has at least 800 hours on it, i have 2 6dj8's that

went massively noisy. One happened at cj. But i am using old stock tubes

pulled from tek scopes. The two tubes that went bad, i put on the side

and recently put one of them back in, and sure enough it was definitely

toast. But it still tests good on the tube tester. I'm still on the same set

of output tubes, the chinese junk that i put in at CJ.

On some of the original production T2's, stax put shields connected to

ground around the tubes. Birgir's had the shields. I have a picture

of another unit without the shields. The center pin of the tube socket

is connected to ground, and the chassis is ground, so it should be easy

to make tube shields.

The T2 uses the input tubes in a way that i have never seen before, and

i'll never likely see it again.

Edited December 25, 2010 by kevin gilmore

[Inu]

I am using a trick that inserts tubes into the sockets completely then pulls tubes for 2 to 3mm upward.

Added 5PF Dipped Mica to the NFB resistors to reduce overshoot at 10KHz and...

[blubliss]

So, the static in mine does not seem to be affected by tapping the tubes. Could a failing semiconductor, transistor cause it? Or a cap?

[spritzer]

Could be quite a few things...

As for the the stock tubes, they were all Golden Aero which was an early 90's rebrand. I did some digging and they appear to have been matched Amperex tubes.