The Headcase Stax thread

[kevin gilmore]

Here is the biggest problem i have with that power supply. Anything bad happens, and the load

disappears and it goes nuclear. The heat sinks on the power supply are way to small. And absolutely

no protection whatsoever. No place to put a fuse to blow if something happens to the load.

The reason this is an issue is that people are already building and then quickly selling their exstata's

and the build quality is awful. A unknowing customer gets the thing, and something happens, and

flames shoot out of the power supply.

The heatsinks need to be bigger, and thermal cutouts have to be added to the heatsinks to shut

the thing down. And half the power going to the shunt devices needs to be taken up by power resistors

bolted to the chassis. This is diy done wrong. If you really want a shunt supply, i came up with the schematics

a couple of years ago. The current sources have shutdowns built into them, so if the load disappears the

current into the shunt supply also disappears. If you use the right devices, it is only one extra opamp per rail.

With any of the stax designs, standard cheap parts for the fets work fine as long as they are matched

to each other. Plenty of range to zero the amp in after that.

[spritzer]

How many PSU's have blown up? I haven't seen a single report, other than one from loose wires. Maybe NoNoNoNoNoNo screwed one up, but that is NoNoNoNoNoNo. And considering 50 J271's are still cheaper than it costs to buy a matched quad of 2SJ74, I really don't see a problem.

So you think any of this gets posted on the forums? I take it you just registered on HF yesterday? Once people realize the main thread is just one major wankfest with idiots like NoNoNoNoNoNo acting like experts they turn to people like me and KG.

Look, I realise that the Exstata isn't the be all and end all, but don't you think your engaging in just a little bit too much hyperbole?

Hyperbole or not this design falls way short of the intended target. It was never supposed to be the end all amp design but to come up with something that measures far worse then the Stax designs it is ripping off while being about as expensive in parts alone is simply retarded. Hell much more expensive considering what I just paid for that SRM-1 Mk2 P.P.

As for needing 50 J271 to get a good quad, that's crazy. Trust me, I'd wager I've matched more than anyone else.

Ahhh a good quad at 7-8ma wasn't a problem but at 10ma+, not so much...

[luvdunhill]

With any of the stax designs, standard cheap parts for the fets work fine as long as they are matched

to each other. Plenty of range to zero the amp in after that.

But how many Stax designs have as few gain stages? I really think the J271 is a great part here, just need Fairchild to offer binned versions...

Edit: maximing Idss is one way to maximize gm. I pointed out the proper way to choose devices in the HF thread.

[spritzer]

just need Fairchild to offer binned versions...

Exactly!! Knowing Fairchild though they've probably discontinued it already...

[kevin gilmore]

But how many Stax designs have as few gain stages?

I really think the J271 is a great part here, just need Fairchild to offer binned versions...

The srmxh, srm212,srm313,srm-mk2 and srm323 all have exactly the same number of gain

stages as the exstata. In fact my designs have the same number of gain stages too.

The first stage, with the voltage translator, and the main VAS gain stage. Output stage

is nothing but a current amplifier.

Alex says the input fets do all the work. Alex is wrong, all the input fets do is buffer the

input and subtract the feedback from the output. Voltage gain is no more than 5. Its

the second stage that does all the amplification (voltage gain of 200 or more) and this

is where non-linearities really end up going to the final sound.

There is absolutely no reason the amplifier cannot be designed such that it can take

anything in the range of 2 to 30 ma Idss. All you need is to replace a few resistors

with a constant current supply. Just like stax does. One extra transistor, 2 extra resistors

once you already have the constant current source for the Vas stage.

Edited July 24, 2010 by kevin gilmore

[Beefy]

So you think any of this gets posted on the forums? I take it you just registered on HF yesterday? Once people realize the main thread is just one major wankfest with idiots like NoNoNoNoNoNo acting like experts they turn to people like me and KG.

Well we're not on Head-Fi now. How many PSU's have blown up?

Hyperbole or not this design falls way short of the intended target. It was never supposed to be the end all amp design but to come up with something that measures far worse then the Stax designs it is ripping off while being about as expensive in parts alone is simply retarded. Hell much more expensive considering what I just paid for that SRM-1 Mk2 P.P.

Considering people like me can build it, I think it nailed its target perfectly. Even though I am quite disappointed now by the revelations that it wasn't more thoroughly measured and tested before it was released, it was a great project for people at my level. I would have loved to build a KGSS, but I do try to be as independent as possible, and I just couldn't do it myself with the limited documentation.

So bring it on guys. Give people like me a better alternative, match Alex's level of documentation, use parts that everyone can get hold of easily, and we will build it!

[Nebby]

So bring it on guys. Give people like me a better alternative, match Alex's level of documentation, use parts that everyone can get hold of easily, and we will build it!

I agree and I don't even own a single electrostat headphone (yet?)!

[luvdunhill]

There is absolutely no reason the amplifier cannot be designed such that it can take

anything in the range of 2 to 30 ma Idss. All you need is to replace a few resistors

with a constant current supply. Just like stax does. One extra transistor, 2 extra resistors

once you already have the constant current source for the Vas stage.

good point in fact. As long as you avoid the extra power supply, but yeah definitely possible.

Beefy: let's not start that "easily available parts stuff"... I do agree about nice available documentation. Perhaps we can get TomB interested in distributing a 'stat amp

[mypasswordis]

Nebby, you're free to borrow whatever I have laying around. Though not for testing your amp builds. I should get around to fixing some of them up so there's more to go around but I never listen to any of my headphones anymore so it's not high on my list of things to do.

KGSSHV and KGST and like the 10 other designs KG posted up in the last few pages FTW!

[spritzer]

Have you ever even looked at the KGSS PCB pictures Justin posted? All values are printed on there and the majority are easy to find on Mouser simply by looking up the RN60D's and selecting the right value or doing a search. I mean 180K/5w isn't hard to find nor is 300K/3W etc.

I could also post the whole email dialogue between me and Robert (rbarth) which would be a complete noob guide to how one goes about building a KGSS, down to the smallest detail such as the K after a resistor number is really important...

Well we're not on Head-Fi now. How many PSU's have blown up?

There were three individuals who contacted me with blown PSU's. Always the same problem, it ran without a load and blew up. Then there is the whole issue which is probably down to flaky bias supplies which I will test once I get my PCB's... Now the KGSS supply can run for days without a load and even if you do something really stupid (like I did, mixed up the wires) only a single resistor blew up and the PSU worked just fine without it.

[HDen]

I could also post the whole email dialogue between me and Robert (rbarth) which would be a complete noob guide to how one goes about building a KGSS, down to the smallest detail such as the K after a resistor number is really important...

I actually think that'd be very helpful for a lot of people, it's always nice to have a starting point/walkthrough

[Nebby]

Nebby, you're free to borrow whatever I have laying around. Though not for testing your amp builds.

Thanks for the offer, I'll give them a try when/if I have a working setup. Have enough dynamic amps in my to-do list to last me a while first

[kevin gilmore]

So a default exstata (pair of boards) is 30ma on the +300 line, and 36ma on the -300 line. (actual measured)

If one of the two amp boards goes down for some reason, the regulators have to take

6 watts on the +300 line, and 7.1 watts on the -300 line. With the 4.7 C/W heatsink

and forced air cooling, the temp of the heatsink is 122F on the +300 line, and 132F

on the -300 line. Now if both boards are not connected, the heatsink reaches

172F, and on the -300 line about 192F. And of course with natural convection cooling,

well the temp is way higher than that... You would need at least 4 times that much

heatsink to prevent the thing from burning up...

[Dusty Chalk]

Stax bitches!!!!

Um...you need a comma in there, unless you're trying to say that you're a bitch for Stax.

I think you may have just earned yourself a custom title.

[mypasswordis]

Makes it kind of hard to isolate and fix possible problems when you need a load connected to the PSU, which otherwise just burns to a crisp. How is the PCB holding up?

[KrypticMind]

First off, take the top off the amp. You will notice two PCB's standing upright in the amp, one for each channel. These have pots marked TVR1 and TVR2 which you can adjust via holes in the side of the chassis. Now the TVR1 pot is the balance and the TVR2 pot is the offset. What you have to do is take a voltmeter, set it to about 30VDC range (or if you have an auto-ranging unit, just set it to DC) and measure at the Stax sockets on the front.

Use the diaphragm above and put the red and black probes into + and - for one channel and adjust the balance pot (TVR1) for 0VDC. Now remove the probe from the - output and hook the lead to any exposed part of the chassis for a ground connection. Now adjust the offset pot (TVR2) for 0VDC. You will have to go back and forth between the two adjustments a few times and then repeat for the other channel. Normally it is best to do this after the amp has warmed up for a bit (say and hour or two) but you can set it instantly after power on and then readjust later on.

Now there is a relay bank on the output of the amp (the soft "tick" sound you hear when powering up the amp) and if they don't engage then you can't measure at the Stax sockets. This makes things just a little bit trickier though...

Thank you for the detailed instructions. I'm not currently with the amp, but I'll report back with what happens, which should be sometime this week. Hopefully, I don't blow up the amp or kill myself. Or both.

[spritzer]

I think it is time for an electrostatic amp build thread...

Um...you need a comma in there, unless you're trying to say that you're a bitch for Stax.

I think you may have just earned yourself a custom title.

Yup missing the comma but not being at CJ you don't get the context either i.e. me drunk as a skunk yelling at Ed while holding the camcorder...

[Beefy]

Beefy: let's not start that "easily available parts stuff"... I do agree about nice available documentation. Perhaps we can get TomB interested in distributing a 'stat amp

If it isn't available at Mouser or Newark I get antsy, and shipping costs spiral out of control. This was one of the really good things about the Exstata.

I could also post the whole email dialogue between me and Robert (rbarth) which would be a complete noob guide to how one goes about building a KGSS, down to the smallest detail such as the K after a resistor number is really important...

Yeah, I don't need quite that much info, but there is more than just the resistors. A simple BOM with at least one recommended part value, specifying preferred cap tech and lead spacings etc. would be very desirable. A few people describing problems and pitfalls they had is always incredibly useful too, and saved me huge amounts of heartache on my first M3 and Millett Max builds. Your other thread is definitely a great start.

If one of the two amp boards goes down for some reason, the regulators have to take

6 watts on the +300 line, and 7.1 watts on the -300 line. With the 4.7 C/W heatsink

and forced air cooling, the temp of the heatsink is 122F on the +300 line, and 132F

on the -300 line. Now if both boards are not connected, the heatsink reaches

172F, and on the -300 line about 192F. And of course with natural convection cooling,

well the temp is way higher than that... You would need at least 4 times that much

heatsink to prevent the thing from burning up...

I've got an AMB E24 controlling mains power to my Exstata, and this has provisions for an open-on-rise thermal breaker to cut the power. From the start I was considering using this but never bothered to follow through, because I didn't need to place a Newark order. Do you think that this sort of scheme would be quick enough to save the shunt transistors from frying?

I was looking at 75C or 80C TO220 models, hopefully I can find a way to secure them to the heatsinks. Unless anybody has a better suggestions?

[kevin gilmore]

Well for example the fqpf3p50 seems to be out of stock at newark,digikey,mouser...

So just because every single part is listed as a mouser part, it may do no good.

ksc5042 has been discontinued, mouser is getting one more batch in, then that

is done too. Exstata will have to change for the future anyway.

thermal breakers are a real good idea on exstata's. Not a bad idea on a lot

of other stuff too.

Edited July 25, 2010 by kevin gilmore

[Beefy]

So just because every single part is listed as a mouser part, it may do no good.

I guess I got lucky. I got everything except for the big cap on the high bias supply.

thermal breakers are a real good idea on exstata's. Not a bad idea on a lot

of other stuff too.

Cool beans. They will go on my next order

[luvdunhill]

If it isn't available at Mouser or Newark I get antsy, and shipping costs spiral out of control.

I'd wager you can get everything from SumR, Mouser and B&D for KGSS and the BH. Add in a case vendor, wire vendor, output jack vendor, a attenuator vendor and FPE, so that makes 8, max. (Since this is a beginner project, we need to assume nothing is on hand, fair?)

So, how many vendors did you use for your exstata build?

Edited July 25, 2010 by luvdunhill

[nattonrice]

Bdent are great.

Tony goes out of his way to help you get stuff. Has a bunch of suppliers in Japan he use to source stuff upon request which is cool.

[pomme de terre]

Hey guys, I've got my mind set to get a set of Stax. Now I'll admit I haven't read through all of this thread, or the Stax thread at the other place... but just how bad is that upper midrange peak that the SR-404 has? Worse than the peak Grados have?

[mypasswordis]

My opinion is that the bass hump is worse than the upper midrange/low treble peak (if you mention etch I may have to cut a bitch). I'm also of the opinion that the 404 isn't nearly as bad as a lot of people make it out to be, though.

[pomme de terre]

Does the SR-202 have the same problem?

I'm trying to decide between a used SR-202 kit right now, or splurging more for a used SR-404 kit. So which would you guys recommend... or neither?