[Staxnuts]

Which one you prefer and why?
 
Direct connection freaks me a little bit, because  if something goes wrong you're at risk of getting full B+ right next to your ears. And you need a separate high-voltage begative supply so that the plates can rest at around 0V during normal operation.
 
A transformer is a component more in the signal path, and a costly one at that, but at least you can be sure your head will be isolated. A supplemental B- supply won't be needed anymore. Another benefit is that a unit with a CT primary can also act as plate load in a diff amp (that's what i use almost exclusively). This make for a simpler to build.
 
I think transformer coupling sounds very good anyways. Would DC coupling really be superior, considering the extra components needed?
I am teared between a will for a minimalist DC coupled path on one side, vs ease of building on the other side.
 
Opinions and especially listening impressions of both, are welcome!
Joris.

 

[FrankCooter]

A quick look around the electrostatic amp world will reveal that all  commercial amps (that I know of) and almost all the DIY designs use some sort of differential topology and either direct or capacitor-coupled outputs. 
 
There are a lot of reasons for this:
 
1) Transformers are bulky, heavy, and consume a lot of amp real estate .
 
2) Even at their best, they will never achieve the technical measurements of a good direct-coupled design or even a top-tiered cap-coupled design
 
3) There are no top quality "off-the-shelf" electrostatic output transformers available. If there were, they would probably cost as much as the total component cost for a non-transformer amp. Within their ratings, you can use you can use just about any high level phase-splitting transformer as an output device. If you want a purpose built design, you'll have to have them custom built. Mine are Electra-Print.
 
4) Pure speculation, but I think part of the reason transformers get a bad rap is because of the underbuilt Stax SRD boxes.  At their original price point, they weren't going to use  world-class Tango  or Tamura components. And for a tube amp at least, running any step-down transformer straight into a step-up transformer results in compounding the weaknesses of both transformers.
 
Despite the above weaknesses, I believe that a well executed transformer-coupled amp can be a very viable and competitive alternative to any of the traditional electrostatic amp topologies. It will certainly be different, and will open a whole lot of design alternatives that are not practical with any other topologies. ( Tube regulated power supplies, something besides a long-tailed pair as an input stage, exotic output tubes ,etc.).  I have built 3 or 4 transformer-coupled amps and have been surprised and pleased with the results. I love EL-34's, but what if you want to use something different? I'm a single-ended DHT fan. The only way to get a DHT electrostatic amp is to build it yourself.  And by far the most practical way to incorporate DHT's into an electrostatic amp is to employ an output transformer. I'm currently running a three stage  76/6v6/845 design with Stax 007's. I haven't conducted any measurements beyond basic sweeps, but it probably measures OK by the standards of other amps of this type. By the standards of traditional top-tier electrostatic amps , I'm sure it's not even in the ball park. Like most modern audio components, electrostatic headphones have always impressed me with their precision, speed, imaging, and detail. Yet at the same time I've found them to be somewhat lacking in body, tonality, warmth. Call it distortion, "non-neutral", "fun", etc., but it can be very rewarding to build an amp that errs a bit in the opposite direction. I think it's  no accident that most Stax enthusiasts musical tastes seem to run towards jazz and small scale classical. I'm an analog classic rock kind of guy. Most of my listening is done at the end of a long work day with a beer in my hand. The little extra warmth of the DHT/transformer combination is appreciated.

 

[jcx]

SE DHT - do you mean a gapped xfmr with DC current? - makes an already hard job really expensive in material
 
I don't see anything saying you can't push-pull DHT - but that does largely cancel even order tube distortion
 
 
and why would Classic Rock lack enough distortion in the recording? - that's what Marshalls, Fenders on the stage are for
 
 
 
once you have enough V to punch through dry skin's normal insulation it doesn't help all that much that its xfmr coupled or not - if there's enough current it can kill, has to be treated with the same respect as direct coupled
 
the most dangerous is current through the heart triggering fibrillation
 
Tetanization - "can't let go" may be a problem too - if you can get the ~20-50 mA current range that causes it

 

[Staxnuts]

SE DHT to electrostats - that must sound sweet. I agree that the trannies must have been a challenge to design.
 
Given the nature of the cans naturally points to a balanced topology which is in fact push-pull - with all that it implies in terms of even order harmonics cancellation. That alone probably contributes to what people call "push-pull sound". Personnally i don't dislike it, as i find advantages to both topologies.
 
Concerning the HV safety question, as jcx pointed i will avoid currents greater than 20mA at the cans. My current design uses 6SN7 at around 10 mA, about 300V p-p swing, but that's AC - it crosses at 0v at times. What concerns me is the possibility of DC, and the tranny would take care of it. Still looking for high-level 1:1 interstages candidates. The choice is rather limited, and the price is high.
 
OTOH, there are tons of DC coupled amp out there - have you heard of any accident happening? Sometimes i wonder if i'm not worrying too much.
 
Choices, choices, choices - the beauty of DIY!
 
Thanks for the replies!
Joris

 

[jcx]

AC can be as dangerous or more than DC - even DC danger depends on DC current direction through thoracic cavity, for both AC and DC time and current matter, less current over longer time - by few seconds even ES headphone amp peak output currents could be fatal
 
once the V either AC or DC exceeds 200 V it becomes more likely to punch through dry skin - even lower V when you are wet, or have poked yourself with the conductor
 
a output xfmr does block some failure modes from giving large output V such as internal parts shorting to rails - but it isn't protection against the large intended Vac  output
 
 
the 1st line of safety is the insulation of the connector, cable and headphone - these should be in good condition - only modded by the very experienced
 
ideally the connector should be shrouded - a poor choice by Stax since it is possible to touch the pins as you insert/remove the plug