Head-Fi.org › Forums › Equipment Forums › Headphone Amps (full-size) › The Single Power ES-1/2 repair and restoration thread
New Posts  All Forums:Forum Nav:

The Single Power ES-1/2 repair and restoration thread

post #1 of 157
Thread Starter 
Warning: High Voltage!! This, along with all other electrostatic amps, use high voltage which is very dangerous and potentially fatal. In this particular amp there are voltages close to 500v with more then enough current to do permanent damage so a novice should never try to work on it. Even when left off for a while the capacitors can still hold a lot of energy so they should be checked before working on the amp/psu.

1: Introduction

In this thread I, with the help of other qualified members of the community, will address the numerous issues which are common to all the Single Power ES-1/2 amps. With Mikhail being evasive as ever and not repairing the mess he made for him self, it is up to the owners to look elsewhere for service and this thread is intended to guide any techs which will be working on the amps, by listing known issues and ways to fix them.

The main focus of this thread will be on my particular ES-1 which was generously donated to me by our very own Afrikane. I'd like to thank Kane for this kind gesture one more time and giving me a project that will eat up all my spare time for the conceivable future. More on that later though. This amp (which had a list price of just under 10k$ as can be seen in the spec sheet) had never functioned as it should, even after being sent in for servicing so Kane got fed up, sold off his electrostatic headphones and put the amp in storage. Earlier this month he put up a post in the Stax thread here where he expressed his wish to give the amp away to whom ever could have use for it. I volunteered to rip it apart for everybody to see and a few weeks later I had it here. This is how it looked when it arrived:

After unpacking this heavy beast (20kg+) I was struck with just how poor quality the chassis is (two metal plates which are cut and bent and help together with 4 screws) and the painted finish is very cheap. Now the amp isn't new so a few scratches are normal but the amp also had a rather large one on top of the amp which had been touched up, badly. Most of the screws also show varying degrees of rust. Enough about that, time to open up the amp chassis:

Now if you haven't downloaded and read the spec sheet linked to above I would go and do so. All the items highlighted in yellow are supposed to be included in this amp. Amongst them is a 549$ surcharge for V-caps which you should be able to see in the pictures. I sure couldn't find any and instead I found some generic orange drop coupling caps which on top of that only rated at 400V. The minimum for this amp is 600v... There is also supposed to be about 429$ worth of silver wire inside the amp but all I can find is cheap, solid core PVC coated silver plated wire. What is also missing are the Cardas GRFA RCA sockets which cost 45$. This would have been easy to check but nobody apparently did...

It's not all bad news since the 395$ 4 channel gold stepped attenuators are there but they are just bog standard Goldpoint Mini-V units. The Black Gate cathode resistors are also there but paying 185$ for them is a bit excessive given how much these 100uf/25v F units cost back in 2006 when this amp was made. The digital tube biasing system is there too but I will post a seperate article on that later on.

The over all build quality of the amp section is rather poor IMO. It is filled with the cheapest standoffs known to man and they are even used as part of the signal wiring. All resistors are standard stuff and the tube sockets are absolutely horrid. The soldering is very poor in places and no thought has been given to proper layout with all types of wiring mixed together. Running signal, filament and HV lines together is never a good idea.

Next step was to open up the PSU and this is what greeted me:

Now open up that spec sheet and see what we have here. Here we are supposed to see 375$ worth of Black Gate caps plus some 589$ worth of V-caps. None to be found and only some basic quality Nichicon 400v caps.

Some of the items paid for were there though such as those heavy Plitron transformers (4.5kg's each) and there were quad regulators used, two for the HV supply and two for the DC filament supply. Those were all bolted to the chassis but as I later found out, there wasn't any thermal compound bonding the two together.

The transformers are a stock Plitron unit (see a picture of the label here) and they are designed so that the cables exit from the bottom of the solid metal chassis. They therefore sit on aluminum brackets supported by a few 10mm standoffs. This has proven to be inadequte as the washer securing the transformer to the bracket is badly bent on both of them and so is the bracket. I sadly forgot to take a picture of the bracket while I was in there though.

As can be see from the pics then the build quality here is utterly terrible. Just about everything is wired with the same PVC coated black wire which makes troubleshooting that much harder. This chassis design is also a horrible choice here since the whole PSU is built into the bottom with all the connectors, LED's and switches on the top part. There was quite a bit of slack in the wires (which were never fixed in place in any way to the chassis) but this is a bad way to do things.

In the last pic you can see the "bias supply" for the headphones which is just a string of 1M resistors, one each for the HE90 and Stax PRO outputs. This is nothing but fail but there is a reason why Mikhail did this which shows that he knew the amp had major design issues.

In the next post I will talk about the known issues with the ES-1/2 amps and what can be done to fix them. In the third post I will document the ongoing journey of my amp where I transform it into what it should have been in the first place.

Stay tuned...
post #2 of 157
Thread Starter 
2: Known issues with the design, it's origins and how to fix them

Let's just get right to it and here are the schematics for the ES-1/2 amps, as compiled from my amp and a number of others. There is some variation in parts used (and their values) but the basic circuit stays the same. First off is the amp section and then the PSU:

If you've spent any time looking at the schematics of electrostatic amps then the circuit will seem very familiar and that's because it is. This is a very familiar site to many of us and pops up in most searches on the subject. Anyone with even no understanding of electronics can see that the ES-1 schematic is very similar, the same in most regards. This amp circuit, often called the Stax SRX, was given away for free by Stax in 1968 in the Japanese instructions manual for the then brand new SR-3 earspeakers. I happen to have that instructions manual and here is a shot of the schematic there:

It is a sound circuit but Mikhail did a poor job upgrading it. A shockingly poor job in fact. The biggest issue is the single filament supply which is grounded on top of that. Look at the Stax schematic and it's clear that the first two tube stages should share a grounded filament stage (AC or DC doesn't matter) but the last stage has to be left floating on its own. That's because the high cathode voltages will run most tubes out of spec. All tubes have a rated voltage gap which shouldn't be crossed at any time and that is the filament-cathode limit. Run the tubes past it and the tube will start to misbehave though it is hard to predict what will actually happen, since it varies from tube to tube. To make this a non issue then the filament is left floating. Luckily on most of the amps there are two filament windings which can supply enough current to drive the tubes so this is an easy fix. The next one isn't but that will have to wait...

2.1: How to fix the amp Added Dec 2009

It's been a while but I finally had some time to work on the ES-1 over the holidays. What I had to do was to completely strip down the amp section, have the top parts of both chassis repainted (simply too much work to gut the PSU as well so the bottom parts will stay black) and then rebuild it properly. Just about nothing from the original amp will be reused, the HE90 socket, attenuators, XLR inputs are just about the only things not scrapped. It may seem drastic but there are very few parts in an ES-1 so why not go crazy and get the really good stuff? First off, here is the schematic of the fixed amp courtesy of Dr. Gilmore.

As you can see it is pretty much the same but there are some crucial changes. The extra 5K pot on the input stage is there to decrease distortion and the two 50k pots on the output stage do work now and can limit the DC offset down to zero. For this you do need a new PSU but we'll get to that later. It is possible to simply modify a point-to-point wired amp according to the schematic above and the same should be true about the PCB versions. It will involve cutting and rerouting wires/traces but there is really no need for a complete rebuild.

As for the main PSU I chose to use the Blue Hawaii unit designed by Dr. Gilmore. This is a PSU design I've used for years and I had both PCB's and ample supply of parts for it. It's also easy to work with and it takes a lot to blow it up, always a good quality to have in a HT PSU. The design can be easily found over on Headwize but I'll include it here:

I built it just like that except the capacitors I used are 500uf/500V JJ units. Since the transformers are rated at 340VAC don't even think about using anything less then 500V caps, at least in the filter posistion (340V*1.41=~480VDC).

In the schematic above there was now a need for a new C- PSU giving use roughly -450V (-440V was enough for me but if you want to use 3d21a's then you need the extra voltage). The simplest solution here is to simply take a 50V supply and have it floating on top of the -400V supply. Lucky for use the Plitron transformers used by Mikhail have a 40V/100ma bias winding sitting unused which is perfect for this. You could simply wire up a bridge and some capacitors (hell I tested it with 3*9V batteries connected in series but this is not recommended as the batteries are now floating at 400V) but I really prefer a regulated PSU. The ideal regulator for this job is the TL783 and the PSU is just lifted directly from the data sheet.

As you can see there is no ground here, you take a wire from the -400VDC supply where the earth would normally go and that's it. There is also no need to use 500V rated parts as the components only care about the delta voltage and not voltage to ground. Here everything is lifted to -400VDC so 63V caps would be fine. I used just what components I had in the parts box so the values reflect that but also what the data sheet recommended. The formula to set the output voltage is in the data sheet so no need to repeat it here.

Since I had one extra 40V/100mA winding left unused in the amp I thought I'd use it to power the two indicator led's on the front panels. Mikhail powered them off the DC filament supply but I'd rather not do it that way so I built a simple PSU with a bridge, small filter cap and a voltage divider to give me roughly 14V. Again I just used parts I had at hand and there are many ways to do this. You could even put in a regulated PSU here too...
post #3 of 157
Thread Starter 
3: The repair and restoration of my amp

When I had opened up the amp it was quite clear that I had a long road a head of me. Virtually none of the parts used to build can be salvaged though the chassis, transformers, He90 connector and attenuators are the most likely to be used after the rebuild. The other parts are simply to cheap or running out of spec to have any use.

When I got the amp it did work (though I never turned it on) but I had to cut a lot of wires in the PSU to be able to remove the transformers and rewire them for 230v. The PSU was also a textbook example of horrible build quality and I would never be comfortable using it. That meant it would be scrapped later on but first I had to do some work inside the amp. My main target was to rewire the filament supply since the original Stax schematic (which this amp is based on) clearly states the need to keep the output tube filaments seperate from the other two stages since the cathode-filament limit will be broken and the supply therefor can't be grounded. Rewiring it was easy enough since I used the stock wiring for most of it and then wired up one of the Speakon sockets just for the filaments. They will be AC straight off the transformer so it's a good idea to keep it away from the other power lines. With only 8 poles available in the stock umbilicals it is sadly not enough for this amp after it has been rebuilt so I found a nice 10 pin system which should do the trick. More on that later when I have actually mounted it.

While I had the amp open I swapped out the main coupling caps for some cheap 0.33uF/630V units I had for safety reasons and I also removed the HE90 socket and some unwanted wiring. Here is how the amp looked after the surgery:

The connector closer to the edge is now for filament use only while we have GND, B+, B- and bias on the other umbilical. Mikhail used two seperate bias lines for the Stax and HE90 sockets which is a waste of wire as it is easy enough to install a voltage divider which will reduce the 580v Stax Pro bias down to 500v which is what the HE90 is supposed to run at.

Having wired up the sockets on the amp side I wanted to make sure that the umbilical were up for the job and that the legendary Single Power build quality on the umbilicals wouldn't end up causing my new PSU to blow up. The cables are easy enough to open, the collar on the back of the plug screws off and the housing for the plug just slides off. Here is what I saw inside:

As you can see, they didn't even bother to put heatshrink on it so that the techflex wouldn't fray. The wires are also rather small (roughly 18AWG) and the bare metal contacts too long so that they are almost touching. Let's look past the fact that this connector was never designed for this job (it's a Pro speaker connector so low voltage, high current) but this is appalling craftsmanship. One of the wires was also loose inside the plastic tubing and free to move around, potentially shorting out something.

With that done it was time to gut the PSU. I had decided to scrap the whole thing and just power the filaments directly off the transformer (so no DC supply needed) and put in one of my old Blue Hawaii power supplies since it might just fit next to the transformers. I've used these supplies since 2003, blown up quite a few , but they are a good workhorse and easy to configure. Removing the old parts was pretty straight forward if a bit fiddly and in the usual fashion with this project, there were a few nasty surprises:

First off is the lack of any thermal compound between the regulators and the chassis which they were mounted to. This costs nothing and can only be construed as sloppy work. When I started to remove the main PSU caps it was clear that the capacitor clamps were to tight and had even damaged a few of them. The one in the picture was just the first one with damage and there were a couple which were worse. This could have ended badly... With all that gone I ended up with this:

The problem with those rather large transformers is that there are only about 9cm of clearence on either side and no way to move them over since they fit through holes in the top cover. Fortunately for me I had a few of the older style BH PCB's with the caps mounted off the board. It would be a tight fit but it did fit in the end:

The caps used were JJ 500uf/500v units and there are four of them. The voltages on the screens are with the PSU unloaded so I'm not running it that high. I did keep one of those cheap tie points to act as a starground and to ground the chassis. It is also used to ground one of the filament windings.

With the amp up and running it was clear that the 1145$ Digital biasing system was just a toy and I'm more then capable of just using a DMM to set up the amp. It turned out to just a 20$ voltmeter with screen (plenty of them on ebay) with some extra wires and 8 push-to-connect switches. Hardly worth the cash but you can be the judge of that:

3.1 Ongoing journey Added Dec 2009

Other projects have come and gone over the past 6 months so the ES-1 has been mostly neglected but time to change that. The only way forward in my opinion would be to completely gut the amp chassis, add the necessary PSU's and rebuild the amp to the same standard as my other amps. With that in mind I set to work and started to remove the wiring plus all the larger bits such as connectors, attenuators etc. There really aren't that many parts which make up the ES-1 circuit:

The only circuit components missing from that picture are the pots and two 1k5/5W resistors. It's been clear from day 1 that almost everything in the amp was headed for the bin so I I didn't even try to save any of the parts. I may find some use for the 20K/6.5W plate resistors but that's about it. The wiring is utter crap, covered with PVC that melts if you look at it the wrong way but I may use it for the filament wiring since it is all solid core. Here's how the amp chassis looks like when I was done with it.

The nest step was to sand it down and have the two top halves repainted silver to remove any Single Power markings from the amp. That should happen after the holidays. I also added ventilation holes, drilled extra holes for the new trimpots/test points and sorted out the back panel. Since I'm using just one umbilical connector I have two large openings in the back panel plus the old DC input for the useless voltmeter so I just drilled a new hole for the umbilical connector in the middle of the back panel and now each side of the amp will have a RCA/XLR input plus a RCA/XLR loopout with a selector switch. All connectors will also be rear mounted and all the holes countersunk for a cleaner look. I'll post pictures of this when I get the top plates back and can start to reassemble the amp.

There was only one major problem which I've encountered with all of this and that's to do with the tube sockets. The ones Mikhail used are the cheapest crap money can buy and the pins are all heavily oxidized plus they are all loose. The plan was always to discard them and put in some nice teflon sockets but here is where Mikhail screwed me, the openings are 30mm. That means the teflon sockets will just fall through so I either have to have some adapter plates made or find some other sockets. I do have some nice ceramic sockets that fit and it offers up the possibility to use loctal sockets for the first two stages. Decisions, decisions...

Now in the meantime I could sort out the PSU. Like I said in part 2.1 above, I needed a C- supply to properly bias the tubes and a small PSU to power the LED's.

Both power supplies are located on the small perfboard towards the back of the PSU chassis. I'm also going to change the bias supply setup a bit, move the 5M ballast resistor over to the amp chassis so that I can install a voltage divider to give me the 500V needed for the He90. I'll probably add a smoothing cap and some small tricks to eliminate any noise the bias line might pickup in the umbilical cable.

3.2: Rebuilding the amp Added Feb 2010

Now that the amp has been running for a few weeks now and I'm happy with it I thought I'd put the final touch on this saga. I didn't take any pics of the different stages of the amp rebuild simply because I was eager to finish it.

First up was the PSU since that was the least work. I only had to install the led, power switch, iec inlet and umbilical connector to the top of the chassis and then wire them up. I also added a wire mesh to block the second umbilical opening (I use a single 10 pin umbilical and not two 4 pin Speakon connectors like Mikhail did) while providing ventilation. I also drilled holes in the top of the PSU case and the bottom to provide some extra ventilation.

The majority of the work would be on the amp chassis since just about everything had to be wired point to point and there are a lot of components to attach to the chassis plus the standoffs, connectors etc. I was stuck using cheap sockets because of the size of the cutouts in the chassis but I could mount them flush so it appears that the tubes are just plugged directly into the box. Pretty cool... Anyway here are some pics:

As you can see it has changed quite a bit. Properly wired RCA/XLR inputs with separate input switches and loop outs for both types plus full adjustment over the operational parameters of the amp via 4 pots (precision 10 turn wire wound) and 5 test points per channel. It's also been fitted with a Teflon Stax socket from luvdunhill and both outputs now have the proper bias voltage and ballast resistor plus some noise canceling tricks. The first two stages now have loctal sockets but more on that later. Here are some pics of it in action:

Seen here with 4 7n7 tubes and 4 EL34's. With the loctal sockets the amp can be used with 7n7's, 7af7's and 7f7's which are far cheaper then octal variants (6SN7 and 6SL7) and are plentiful. I did build in the ability to use 3d21a's instead of the EL34's (via the old SP adapters) but I haven't tried it yet. For this to work you have to increase the C- supply voltage but the configuration I posted above should do the trick.

Now for the important bit, the insides. Given the relative low parts count of the this amp I could really go all out in terms of parts used with some of the best components money can buy. The coupling caps are very important and the sky is really the limit in terms of cost but not many audiophile caps are rated above 630VDC. This is simply not enough for this design as the caps could "see" upwards of 800V or even more so 1KV caps should be the minimum. This really limits choices but my favorite caps just happen to be Mundorf and most of their range is rated at 1200V which is perfect for this amp. I settled on their Supreme line which offers a very neutral sound in a small form factor and isn't as terribly expensive like their more esoteric caps. There are also some Nichicon ES bi-polar caps in there for the 100uf spots but these will be switched out for some polypropylene caps later on.

When it comes to resistors there are a lot more choices but care should be taken to make sure they are rated at a high enough voltage. It is absolutely crucial to use 500VDC+ rated resistor in this amp but one could always series connect to increase the voltage handling. For the large wattage resistors there is really only one choice in current production, the non inductive wire wound Mills resistors. For the EL34 plate resistors I used 50W units bolted to the chassis but for lower noise it would be a good idea to use a few smaller resistors in series/parallel configuration. Should an EL34 fail then the cathode resistors will be hit hard so I used 12W Mills resistors here. The rest of the larger resistor are Kiwame and Riken, both excellent carbon resistors. With the smaller units I used two types, all the 499K resistors are Caddock MS-132V (about 5$ each) but the rest are all PRP audio resistors. I've had great results with these in the past and since I was buying a metric "crap-load" for the T2 project it made sense to use them here as well. With that in mind, here is a pic:

What's missing from the picture is the RCA wiring which I've added now. There are also spots near the inputs to mount input transformers but they aren't really needed so I haven't fitted them yet. The amp is right on the edge to function as a phase splitter so with the wrong tubes there can be a slight channel imbalance with the RCA inputs but the dual volume controls clear that right up. All the wiring is Belden except for the filament wires which are from the stock amp. Showing that Single Power truly is cursed that is the only part of the amp that has given me any trouble.

After a few weeks of use the amp has been noting but rock solid. I've played with 7N7's and 7F7's in the input spots and both sound great. At the same time I haven't messed with the output tubes at all and they haven't drifted more then 1V in this time. Mission accomplished I dare say...
post #4 of 157
I am wondering Singlepower have been around for a few years but no one ever realize the terrible inside of the amp until now?
post #5 of 157
Originally Posted by squall343 View Post
I am wondering Singlepower have been around for a few years but no one ever realize the terrible inside of the amp until now?
Before I purchased my Woo, I was interested in Singlepower, That is until I saw a picture of the rats nest wiring someone posted. The lack of pride in his work was a real turnoff to me. This was over a year ago....
post #6 of 157
Thread Starter 
The issue of terrible build quality had always been been synonymous with Single Power but is has been overlooked based on how the amps sound. He was also always willing to fix them when they break (often ripping the customers off by charging for unwanted upgrades) but now he isn't speaking to his customers and his buisness is all but dead. Mikhail has also made damn sure that all attempts to post threads like this have been shut down by stepping in and offering to service the amps for free and/or with very fast turnaround time. Quite a few owners have been burned by this and their amps are still sitting in that black hole which is SP. Also, this isn't my first thread posting about the issues with ES-1/2 amps (the first one was closed back in October) but now I have an actual unit in my hands for the world to see.
post #7 of 157
Thread Starter 
Updated the third post
post #8 of 157
Spritzer, what is your opinion of this circuit, assuming it is built correctly? Do you expect to have a good amp once your work is complete?

Also, if I recall, this amp is a clone of a Stax unit. Would you recommend this circuit as a DIY project?

The reason I ask is because I think I'll have an Omega 2 Mk. 1 in a week or two () and want to DIY an amp for it. I've had my eye on Dr. Gilmore's AC-coupled all triode amp, but this one also interests me. The Blue Hawaii is a bit too intimidating, and I prefer just using tubes in point-to-point configuration.

I hope this doesn't derail your thread, but your opinion of a correctly built ES-1 would be much appreciated.
post #9 of 157
huge thanks to Spritzer for doing this project to help out the head-fi community! in light of the huge mess of a 10K amp and power supply, in your opinion, how much would such an amp cost WITH the parts given by Michail? how much would the amp cost if done with all the upgrades listed in the spreadsheet? one last question, if you don't mind. Would it be possible for me (or anyone else of this community) to commission you, Spritzer, to build a similar amp? if so, what would be your asking price?

Thanks again, david
post #10 of 157
Jeez what a terrific undertaking this rebuild presents but at the same time a valuable lesson in what it takes to build or rebuild something the right way.

I'll be following this thread with keen interest. Thanks Spritzer for taking on such a job and then sharing it with the rest of us. When all is said and done I vote this project/thread be stickied so others can gain from it that are facing a similar situation.

post #11 of 157
Thread Starter 
Originally Posted by Uncle Erik View Post
Spritzer, what is your opinion of this circuit, assuming it is built correctly? Do you expect to have a good amp once your work is complete?

Also, if I recall, this amp is a clone of a Stax unit. Would you recommend this circuit as a DIY project?

The reason I ask is because I think I'll have an Omega 2 Mk. 1 in a week or two () and want to DIY an amp for it. I've had my eye on Dr. Gilmore's AC-coupled all triode amp, but this one also interests me. The Blue Hawaii is a bit too intimidating, and I prefer just using tubes in point-to-point configuration.

I hope this doesn't derail your thread, but your opinion of a correctly built ES-1 would be much appreciated.
The circuit is a good one and has aged very well. Things have moved on since this design was debuted in 1968 so now we have constant current sources to help deal with the massive impedance swings which something like the SR-007 presents and fully regulated power supplies, even shunt regulated as with the latest designs. Not everybody likes these changes though, preferring the slightly warm and fuzzy sound of vintage tube amps and that's why these amps have such a following.

With the amp fully rebuilt it should be comparable with the Woo WES, using the same output tubes and similar output stages. In it's stock form it only swings about 400V P-P (the Blue Hawaii swings 1500V P-P and the newest incarnation of the KGSS circuit 1900V P-P) so less then the all triode amp but the EL34's have a lot more standing power which makes up for it. Voltage swing is useless without the current to back it up.

Given just how cheaply an amp like this can be built I do think it is a good circuit and it's a much smaller project then something like the BH with its massive heatsinks and sometimes overwhelming complexity.

Originally Posted by leberserkfury View Post
huge thanks to Spritzer for doing this project to help out the head-fi community! in light of the huge mess of a 10K amp and power supply, in your opinion, how much would such an amp cost WITH the parts given by Michail? how much would the amp cost if done with all the upgrades listed in the spreadsheet? one last question, if you don't mind. Would it be possible for me (or anyone else of this community) to commission you, Spritzer, to build a similar amp? if so, what would be your asking price?

Thanks again, david
I don't really want to get into whole mess what the amps should cost as markup varies quite a bit and some part prices are hard to gauge. I'd rather focus on what was paid for yet didn't end up inside the amp.

I'm not an MOT so I don't do work for others.
post #12 of 157
Thread Starter 
Second post updated with schematics
post #13 of 157
the perfect thread for the disaster tourists ...
post #14 of 157
Originally Posted by spritzer View Post
<snip> Mikhail has also made damn sure that all attempts to post threads like this have been shut down. <snip>
Really? I was not aware that Mikhail had such power.
post #15 of 157
Thread Starter 
I think nobody doubts his powers to BS and spin anything in his favor. That's certainly what happened the last time I posted about this, he spun some story for Jude claiming that the design was his own and all the usual BS. You'll be the judge whether that is true or not.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Headphone Amps (full-size)
Head-Fi.org › Forums › Equipment Forums › Headphone Amps (full-size) › The Single Power ES-1/2 repair and restoration thread