DIY/T2 Electrostatic Amp Thread - General

[mulveling]

A T2 build is an undertaking unlike any other amplifier - electrostatic or otherwise. It's an order of magnitude more complex and difficult, even before you get into issues concerning parts obsolescence. There are insanely high internal voltages, plus massive amounts of heat to be dissipated - and the high voltage transistors start to behave very badly if heat isn't dissipated fast enough. A T2's gain factor is also comparable to that of an MC phono stage (1000x gain!!) - and though this bit is not unique to the T2 among e-stat amps, so much gain at & after the small-signal tubes (6922) always makes noise floor a real challenge. Sadly, 6922/6DJ8 types also have a reputation for being fussy and going bad/noisy quite often.

There's good reason there was never (and will never be) an off-the-shelf T2 offering. It can only be done on commission by very skilled & experienced builders, or as a very advanced DIY effort. We are very, very close to not having any commissioned build options for T2, because it is so difficult, costly, and time-consuming.

Here are some prerequisites for those interested in a T2:

1. You are embarking on a journey. Abandon the usual instant-gratification expectations!
2. If you're commissioning one, you need to be flexible on timelines and have good patience.
3. You have to expect there is a chance challenges or problems may arise. Though it's not your fault, these can require a significant amount of additional time to resolve.

Totally not @oneguy's fault though, and I hope he receives a solid resolution sooner rather than later!

 

[oneguy]

@mulveling it’s looking like sooner

 

[KDS315]

A T2 build is an undertaking unlike any other amplifier - electrostatic or otherwise. It's an order of magnitude more complex and difficult, even before you get into issues concerning parts obsolescence. There are insanely high internal voltages, plus massive amounts of heat to be dissipated - and the high voltage transistors start to behave very badly if heat isn't dissipated fast enough. A T2's gain factor is also comparable to that of an MC phono stage (1000x gain!!) - and though this bit is not unique to the T2 among e-stat amps, so much gain at & after the small-signal tubes (6922) always makes noise floor a real challenge. Sadly, 6922/6DJ8 types also have a reputation for being fussy and going bad/noisy quite often.

There's good reason there was never (and will never be) an off-the-shelf T2 offering. It can only be done on commission by very skilled & experienced builders, or as a very advanced DIY effort. We are very, very close to not having any commissioned build options for T2, because it is so difficult, costly, and time-consuming.

Here are some prerequisites for those interested in a T2:

1. You are embarking on a journey. Abandon the usual instant-gratification expectations!
2. If you're commissioning one, you need to be flexible on timelines and have good patience.
3. You have to expect there is a chance challenges or problems may arise. Though it's not your fault, these can require a significant amount of additional time to resolve.

Totally not @oneguy's fault though, and I hope he receives a solid resolution sooner rather than later!

Honestly I never really understood why today one would want 60dB (1000x) amplification, as most ppl nowadays use a PRE-Amp anyway with high (XLR) output voltage (my Primare PRE does up to 12V actually), so that makes no sense to divide it down to say 350mV or even 100mV and then amplify it up again. Just a two-stage amplification would be enough and not (noisy) 3-stage! A friend of mine even built a one-stage amplifier for STAX which needs about 7Vrms to be driven well and that PRE does that easily. Sounds amazing clear and very black silence! All my STAX-amps (or amps for STAX) I bypass the volume potentiometer and set volume ONLY with my PRE Amp.

 

[mbrennwa]

I never really understood why today one would want 60dB (1000x) amplification as most use a PRE-Amp anyway with high (XLR) output voltage (my Primare PRE does up to 12V actually), so that makes no sense to divide it down to say 350mV or even 100mV and the amplify it up again.

Couldn't agree more.
I guess many of the proven designs (like the Blue Hawaii) date back to times where source components with low output were still a thing, so the gain was good to have back then.

 

[justin w.]

Honestly I never really understood why today one would want 60dB (1000x) amplification, as most ppl nowadays use a PRE-Amp anyway with high (XLR) output voltage (my Primare PRE does up to 12V actually), so that makes no sense to divide it down to say 350mV or even 100mV and then amplify it up again. Just a two-stage amplification would be enough and not (noisy) 3-stage! A friend of mine even built a one-stage amplifier for STAX which needs about 7Vrms to be driven well and that PRE does that easily. Sounds amazing clear and very black silence! All my STAX-amps (or amps for STAX) I bypass the volume potentiometer and set volume ONLY with my PRE Amp.

Because an amp that makes you turn the knob 80%+ up is going to get horrific reviews

But even with the higher gain, I have seen volume knobs maxed out in loud CanJam rooms

 

[mbrennwa]

You care about reviews? And deaf listeners?

 

[KDS315]

I was only talking about AMPs used with PRE-Amps, not ones directly attached to a DAC or Vinyl-PRE-Amp.

 

[justin w.]

You care about reviews? And deaf listeners?

There is a video on YouTube, very well produced, that reviews the Blue Hawaii SE and the reviewer complains they had to turn the volume up too high. It's a DACT attenuator, so 3 o'clock on that is equivalent to about 12 to 1 on the Alps RK50.

I wish I had been contacted by the reviewer because I could have explained a lot of things about the amp in their video.

 

[VandyMan]

There is a video on YouTube, very well produced, that reviews the Blue Hawaii SE and the reviewer complains they had to turn the volume up too high. It's a DACT attenuator, so 3 o'clock on that is equivalent to about 12 to 1 on the Alps RK50.

I wish I had been contacted by the reviewer because I could have explained a lot of things about the amp in their video.

A lot of audiophiles say that non-electrostatic amps sounds best roughly between 9:00 and 1:00. Too low, and you can get a channel imbalance. Too high, and you get distortion. My understanding is that this does not apply to electrostatic amps, but I have no idea why.

 

[oneguy]

A lot of audiophiles say that non-electrostatic amps sounds best roughly between 9:00 and 1:00. Too low, and you can get a channel imbalance. Too high, and you get distortion. My understanding is that this does not apply to electrostatic amps, but I have no idea why.

I definitely have heard imbalance at low volume settings on estat amps.

 

[justin w.]

A lot of audiophiles say that non-electrostatic amps sounds best roughly between 9:00 and 1:00. Too low, and you can get a channel imbalance. Too high, and you get distortion. My understanding is that this does not apply to electrostatic amps, but I have no idea why.

Channel imbalance at low volumes is common to normal with volume pots, and few good ones are still for sale since digital volume took over, or stepped attenuators/relays in high-end. the test I use is to set it for 8 or 9 and if the imbalance is less than 1dB we use it. For the RK27 it comes out to about 50% used. From what I've found I don't think most companies are testing them at all - even the original Sennheiser Orpheus ($30k in today's money) used an RK27. The Alps RK50 is one of the few that don't need sorting, and costs nearly 100x as much as a RK27

but every volume control is the same in that they go from completely off to completely on, only the in between is different, and many want to ramp up the volume quickly to create the perception of more power, as long as channel balance is maintained. Cars do the same thing by setting the sensitivity on their brake pedals differently depending on the target market. With the attenuator Kerry is working on, we're definitely going to try to find the curve that seems most appealing and natural, and channel imbalance won't be an issue since its relays.

BTW, I'd assume the reason people say this doesn't apply to electrostatic amps is because there's a narrow range of sensitivity between all the available headphones, compared to IEMs/dynamic/planars. This makes it easier to design an amp where the lower end (with potential imbalance) probably won't see much use.

 

[VandyMan]

BTW, I'd assume the reason people say this doesn't apply to electrostatic amps is because there's a narrow range of sensitivity between all the available headphones, compared to IEMs/dynamic/planars

The reason that I say it is because you advised me to turn my BHSE volume to max when using the volume control from my DAC that feeds it. That would never be the advice for a normal amp since it leads to the worst distortion performance. If the BHSE does have worse distortion performance at max volume, I don't understand why I was given that advice or why you object to the reviewer who said he had to turn the volume control too far.

 

[mulveling]

The reason that I say it is because you advised me to turn my BHSE volume to max when using the volume control from my DAC that feeds it. That would never be the advice for a normal amp since it leads to the worst distortion performance. If the BHSE does have worse distortion performance at max volume, I don't understand why I was given that advice or why you object to the reviewer who said he had to turn the volume control too far.

You get distortion from running the signal so high its peaks approach (or even exceed) the output stage's voltage rails (+/-350V to 500V). Which is something that would never even come close to happening with most dynamic amps (they typically have massive headroom to spare) - but this can be a problem on e-stats for listeners who play loud (very LOUD).

If you're controlling signal level from your source or a preamp and playing at normal levels, that doesn't apply. Putting the e-stat amp's volume pot at MAX effectively shorts it out of the circuit, so you would get the lowest possible distortion (for any given net output level) and more perfect channel matching in that scenario - assuming your source's volume isn't sacrificing bit depth at those levels.

The amplification circuitry doesn't know or care what level your volume pot is at - it only cares about the signal level that gets through on the other side.

The use of 2ch preamp with an e-stat amp's volume at MAX (or even removed) is an interesting approach I should have played with more.

So just to illustrate, a 2.0V RMS input signal has peaks +/- 2.82 Volts. If you run that into a T2 with gain 60dB and volume at MAX, it tries to make +/-2,820 Volt peaks - so far into hard clipping over its +/-500V rails it will be pure distortion - and your headphones/ears/etc will probably be toast. 20dB of pre-attenuation (from a preamp or digital source) gets that down to +/-282 Volts peaks which is very loud but do-able (for BHSE, T2, etc). 26dB and more of attenuation gets you into reasonable level territory. So - if you do take the pre-attenuation approach, just be sure to adopt a usage protocol that doesn't accidentally result in sending a 0dB (un-attenuated) signal level into your headphones (e.g. you accidentally hit a remote button and swap to another live source input that does NOT have attenuation applied).

 

[VandyMan]

You get distortion from running the signal so high its peaks approach (or even exceed) the output stage's voltage rails (+/-350V to 500V). Which is something that would never even come close to happening with most dynamic amps (they typically have massive headroom to spare) - but this can be a problem on e-stats for listeners who play loud (very LOUD).

If you're controlling signal level from your source or a preamp and playing at normal levels, that doesn't apply. Putting the e-stat amp's volume pot at MAX effectively shorts it out of the circuit, so you would get the lowest possible distortion (for any given net output level) and more perfect channel matching in that scenario - assuming your source's volume isn't sacrificing bit depth at those levels.

The amplification circuitry doesn't know or care what level your volume pot is at - it only cares about the signal level that gets through on the other side.

The use of 2ch preamp with an e-stat amp's volume at MAX (or even removed) is an interesting approach I should have played with more.

So just to illustrate, a 2.0V RMS input signal has peaks +/- 2.82 Volts. If you run that into a T2 with gain 60dB and volume at MAX, it tries to make +/-2,820 Volt peaks - so far into hard clipping over its +/-500V rails - and your headphones/ears/etc will probably be toast. 20dB of pre-attenuation (from a preamp or digital source) gets that down to +/-282 Volts peaks which is very loud but do-able (for BHSE, T2, etc). 26dB and more of attenuation gets you into reasonable level territory.

Thank you. That is an excellent explanation.

Putting the e-stat amp's volume pot at MAX effectively shorts it out of the circuit,

If you don't mind, a follow up question about this. For a standard amp, putting it at MAX means maximum gain and usually standard amps produce more distortion at higher gain (correct?). What is it about estat amps volume that causes MAX to mean, if I understand correctly, 0 dB of gain?

The use of 2ch preamp with an e-stat amp's volume at MAX (or even removed) is an interesting approach I should have played with more.

I do this because my Benchmark DAC-3b/HPA-4 has a remote and my BHSE is across the room from my listening seat.

 

[mulveling]

Thank you. That is an excellent explanation.



If you don't mind, a follow up question about this. For a standard amp, putting it at MAX means maximum gain and usually standard amps produce more distortion at higher gain (correct?). What is it about estat amps volume that causes MAX to mean, if I understand correctly, 0 dB of gain?

Most amps and preamps put their active circuitry *after* the passive volume control. So they're really running at "max gain" all the time! The volume control is just there to pre-attenuate the signal to a reasonable level, so you don't overload any amplification stage (or your ears!). As you raise volume the signal can encroach on the overload points of various stages, and distortion necessarily rises. But if you pre-pre-attenuate to compensate, then the net signal level stays the same and distortion doesn't go up.

A few amps & preamps have placed circuitry before the volume control. The danger with that, is you have to be VERY careful to protect the input stage from overload conditions. Lowering the volume control won't save you here, because it's after the input circuit. I don't know why some designers make that choice. I have a Meridian 808 with analog inputs NOT protected by a passive volume - but Meridian solves the problem by providing for adjustable "input sensitivity" on that input, from 0.5V to 2.5V.