spritzer

Poor or simply broken. Chewing up tubes might just be a design choice...

Happy Birthday!!!

Happy Birthday!!

The Chinese one is a KGSS, not a KGSSHV. It runs at +/-350V when a KGSSHV runs 400V or higher. Not the sole distinction though... Going by the pics the Chinese amp will need some mods to perform better than a nice little desktop unit.

The simplest answer is that the WES and BHSE have nothing in common, at all. Sure they both drive electrostatic phones and have similar tubes but it ends there, same with the Electra. The goal of both Woo and EC is to price their amps at the same level or higher to fool people to think they might be in the same class which they clearly aren't. Let's just view what a WES with all the upgrades and updates looks like: This is what 8000$ will buy you. Now for some impressions of the design, the PSU in this thing is the usual Woo joke. Utterly overloaded tube rectifiers which don't matter due to the bridge present. No regulation what so ever except for the circuit which powers the LED's and front end heaters. This power supply is very similar to what Stax do so why would anybody pay a premium for this? Now for the amp section, here is where it gets from bad to "we will kill your headphones". First off, a HE90 should never be used with a WES or a WEE. Never in a million years should the twain meet as both will destroy the phones. It's easy to see in the amp pic, see the small PCB next to the output sockets? Yeah, those aren't close to be the correct values though for Stax they are closer so not as big of a deal. So with actual amp the worst idea are those output chokes as there should be only two of them. Simply connecting them as a single push-pull unit doesn't work, they need to be on the same core. Inductors for Stax are bad idea though given the electrostatic load and its behaviour but a push-pull choke at least works. This doesn't which is why the amp sounds like shit, hell it sounds like it is broken. You can also clearly see that there are four isolated amp sections here with no adjustment so how can one phase know what the other one is doing? It can't so the natural drift found in all circuits is , quite literally, out of control. The amp should also be wired differently to use the benefits of matched triodes within the same envelope. The phase splitter is also an absolute dog and should never have been there. How come the ESX uses all the same tubes, far less components and doesn't need one? Other stuff such as series connected heaters for the EL34's are a problem. These are not computer/TV tubes and were never designed to be series connected. So the bottom line is this, it's an 8K$ amp that measures and sounds worse than a SRM-323S and has less power. Also worse components and build quality. My initial reaction to hearing a WES at CJ'10 was to ask Jack if it war broken. That sums it up nicely enough...

Same size IRC as the dual fet we put on the KGSSHV board. Almost impossible to solder...

The choice of RCA's on each cup is just bizzare but those earpads do interest me.

Yup, the WES is a clusterfuck and the Coriolan/Egmont also. Let's not go into the ES-1/ES-2 again...

Some good news as we can all expect some more measurements of electrostatic headphones. Kevin and I weren't happy with the lack of a good amp for Tyll to measure on so being a bit loopy we decided to build one for him and donate to the cause. It fell to me to build it and it had to be something special, with some unique features but also a known standard. A souped up KGSS felt like the best choice as it's a well known design and probably the most popular high end Stax amp out there. Justin built a pile of them, it's been a popular DIY project, I'vv and the Stax SRM-717 is pretty much a KGSS so an easy reference point. I've also built a dozen or so of them so plenty of parts in stock. Here it is sitting on top of it's younger sibling. It is clear that this isn't a bog standard KGSS since Tyll as a reviewer needs flexibility. All the outputs are driven off the same amp section but the three major bias voltages are taken care off via their dedicated sockets. The bias supplies are all either generated by fully regulated power supplies or derived via a voltage divider off a fully regulated line. Now the fourth output is a bit special as it is a small variable DC supply driving an EMCO DC-DC converter and goes up to 1kV. There are jacks on the side to plug in multimeter probes and the pot to adjust it. The ballast resistor can also easily be swapped out for any value required by what ever unit is up for testing. I'll ship it with a 4M7 unit rated to 2000V which works just fine for most sets. The LED's for power indication are mounted underneath the chassis to make the front a bit cleaner. I looked into transparent feet but found nothing that I liked. So here are the guts. The PSU is a design derived from the KGSSHV but simplified and limited to single 500V caps. I run it at +/-415V which is about as high as that Edcor transformer will give me. On the KGSS amp board the 2SA1156's are replaced with 2SA1486's and it uses 2SC4686A's as output/gain devices. I had to back the current off a bit as the sinks were reaching 95°C at the stock KGSS setup but even at 8.5mA it is a very powerful amplifier. Vanishingly low output capacitance and no protection resistors means a very linear output. I left the 5k1 output resistors off as they might interfere with a non Stax based design. Other than that it is a pretty standard fair for my builds, XLR inputs and loopouts, quad Alpha pot which is easy enough to bypass by simply maxing it out. The AC input also features a voltage selector which made hooking up three transformers a bit easier. Needless to say this amp sounds awesome and has no problem driving even the tricky SR-007A (SZ2) to perfection. Tyll is happy and and so are we. The "God amp". Who comes up with this shit as it is a pretty lackluster piece of engineering. Same with the Electra which nobody cares about any more. How come nobody can build an all tube electrostatic amp properly?

I've never built a 500V offboard amp so only dealt with isolated packages but zero issues here. I do bend the leads so that they enter the PCB holes at a straight 90° angle so no chance of shorts. This is not fun with the super tiny board...

There is no way to run a stock 323 at 400V without building a new PSU and fitting a new transformer. It isn't regulated so there is no headroom, the B+/B- is just what comes out of the rectifier. Here is a SRM-323A schematic I found somewhere. It's a bit clearer than our own: Very simple circuit and quite clear where the LL was stolen from. The output devices are 800V units so good for +/-400V but any higher you'd have to start swapping out parts. When I look at this, it would be interesting to design a PCB for this as it's so bloody simple to build. It would require matched J74's which is a problem, at least for now. Ramp up the current, swap out the gain stage CCS for a 10M90 solution and it's a cracker of an amp. Far better than either the Exstata or the LL. Edit: The 717 is a slightly modified KGSS but the 727 is a bit different. It can be modified to become a KGSSHV though which is what Tony did. Still limited to +/-350V though.

This certainly isn't BH levels of heat which does make it more user friendly. The "normal" board is still on for production, just finishing up one other design to save on shipping.

Happy Birthday Purk

Happy Birthday Shelly!!!

That would be a good idea. Having 50 useless boards does sting a bit...

They could be hid next to the output devices but since the original article doesn't have them I think you are right. Justin's version didn't have them either As for the sound, it does depend on what output device is used.

There is no mod for the 323S unless you count buying a 727 instead. Raising the voltage would involve higher rail voltages so pretty much everything new...

Stunning work Kerry. Q1 and Q5 are silly small so congrats on being able to hand solder them.

I have KuboTen order them from Ippinkan. Stax refused to sell me some 009 pads though claiming they aren't user replaceable. "That's not a flaw, it's a feature...." The tubes starting to conduct would be my guess which is why output relays would be a good idea. We've wanted somebody in China to make a KGSS for years so that is all good. Some recognition would be nice but mostly so people know what they are buying. The price is certainly attractive and the build quality seems good enough: Cheap pot but at just over 700$ shipped to the US it isn't all that bad. The one I bought shipped yesterday so we'll soon know what Kevin thinks.

Well naturally since somebody could possibly have dissed that piece of shit 45 amp that he's peddling for Eddie Current. You know the one that screams at you when you power it on. Pure quality there... One aspect of that whole discussion is the ludicrous concept of how Stax "voice" the headphones. It's quite clear that Stax have just one goal when designing transducers and that's to make them better, something which doesn't always work out. Ever since day one in 1959 it's been about getting better diaphragms which then in turn means the structure is the issue which then needs to be redone and then the cycle repeats. There have been a few exceptions from this rule such as the SR-Omega but that in turn lead to the 007 which ticked both boxes so not a whole lot of difference. Then the 007 SZ2 which fixed some issues for a better product but generated a whole lot of other problems which the SZ3 set out to fix. Then they dropped in the new diaphragm which made everything even worse... In terms of amp design this could be compared to increasing feedback, it will yield some better specs but over all you are getting something that sounds progressively worse. Not magic to this, it's just that THD is pretty meaningless as the sole basis of amp design. I also know full well that Stax believe each new version is an improvement and in some aspects they are, but some do miss the point a bit. The LNS was a bit too mellow for some people so the SR-404 spiced things up a bit which didn't work too well. They also introduced new earpads then which didn't work all that well either.

Indeed so here is something Stax related. This one has been discussed before: http://item.taobao.com/item.htm?id=36525626565 Well I did some digging and found some pretty poor internal pics but I'm 99% sure that this is indeed a KGSS. Certainly uses Kevin's BH power supply and the layout matches so we've naturally ordered one... I'm sure we can find a far more elegant solution to that. *Goes off to design a low pass filter for Stax*... Since you have a KGSSHV now, does it make you want to rip out your ears with the 009? I'm a year older than you so clearly half deaf by now...

For Dale resistors there are always the CCF60 type which have color bands on them for easy identification. 500V spec too which is nice. Nahh, better ask away so things can be fixed. Fucking up is just part of the process

That would do it.

Couldn't have said it better. He also thinks the HV is bad with the 009's, he should try that SRM-323S... I also love his response about age. Sure, being younger will often lead to better hearing but experience is the key here. Well that and not being an idiot.

There are 4 zeners for the bias supply, 3*150V and one 130V. That would work but I highly recommend doing a proto run on any untested pcb.