SR009 / WES / BHSE / LL hit the mainstream - Page 33

That would only be the case if all these amps were truly linear, two of them have a vary pronounced sound signature and it's hard to counter it.  No EQ in the world will remove high levels of THD for instance. 

Quote:

Originally Posted by spritzer 

That would only be the case if all these amps were truly linear, two of them have a vary pronounced sound signature and it's hard to counter it.  No EQ in the world will remove high levels of THD for instance. 

I see, thank you sir.  To state the obvious, I'm assuming you're referring to the WES and LL as the two that have a very pronounced sound signature?

Yes, the BHSE is pretty much linear but a bit of a second harmonic distortion was engineered into the circuit.  There is also the load tolerance to think about, output impedance, output capacitance, slew rate etc. that no EQ can compensate for. 

Quote:

Originally Posted by spritzer 

Yes, the BHSE is pretty much linear but a bit of a second harmonic distortion was engineered into the circuit.  

Long time no speak, hope all is well in sunny Reykjavik. I haven't had an electrostatic setup in a while, and thus I've got no skin in this game, but can you please enlighten me how the distortion-by-design plays into the whole wire-with-gain concept, which the BHSE is famous for? I understand the notion of euphonics, but good or bad it can still be argued it's a degree of coloration. This isn't a loaded question, I'm strictly curious.

All engineering is about compromises and choosing what you can live with and what you can't.  It's well documented that a slight second harmonic distortion (or rather even order distortion) sounds pleasing compared to the odd order one normally gets from solid state gear.  So designers such as KG, Nelson Pass, Ayre, Krell all engineer amps with vanishingly low distortion but what little distortion there is, it's mostly second harmonic.  It's still wire with gain or as close to it as is possible.  If you just focus on THD as the big evil you end up with garbage such as the O2 amp and all those horrible amps from the 70's and 80's. 

The bottom line is indeed all about managing the compromises needed and if you can give a little bit of ground here to gain a lot elsewhere that's what I'd do.  I'm not a fan of capacitors in the signal path but I'd take them over badly executed DC coupled stage any time.  Same thing applies to feedback, you can make very nice stages with no NFB but takes a lot of work and it also takes power.  Take something like the SRM-727, at the current power levels is sounds dire but increase the standing power 4 times and it will be far more linear. 

Anyone know of any used LL/BHSE/009 for sale?  I assume once the LL2 is out there will be some LL1 units on the market.  I sort of hope no one comes forward so that I don't follow through with the crazy decision to spend this kind of money, especially since I don't know how much better/different it is than my current rig (LF/HD800/LCD2.2/W3000ANV).

how does this happen with balanced drive? - classically there should be even harmonic cancellation where it matches in +,- outputs

Quote:

Originally Posted by spritzer 

Yes, the BHSE is pretty much linear but a bit of a second harmonic distortion was engineered into the circuit.  There is also the load tolerance to think about, output impedance, output capacitance, slew rate etc. that no EQ can compensate for. 

I fail to see how balanced drive would cancel out harmonic distortion. It's a push pull transduction, I except the distortion to show right up in the diaphragm motion. No?

Or maybe I am confused with the terminology and you're referring to a push-pull amplification where the + and - halves of the signal and separately amplified? But again here, aren't most ES design NOT of such type? 

BH, many other ESL amps are "fully differential" amps with differential input,  opposed polarity, differential outputs, those with Global feedback (like BH) have 2 identical feedback networks/paths

identical circuitry for both +,- polarity outputs that drive each stator, one on each side of  the diaphragm

if the even order distortion in each of the paths to the output are the same then the even terms have the same sign/polarity on both sides of the membrane, so they don't cause any force on it

That's what I don't get JCX, the harmonics are expected to be in phase with the fundamental. The signal on both stators is opposite phase (fundamental & even order harmonics) such that it acts on the diaphragm in push/pull fashion (both fundamental & harmonics). No?

say the gain function for each side of the stator looks like

f(x) = 500x +10( x^2)

now put in x and -x as the differential signal and take the difference of the outputs:

( 500x +10(x^2)) - ( 500(-x) +10((-x)^2) ) = 1000x

even powers always give positive result for both +,-x that makes the +,-x even order distortion terms cancel in the difference output

you do have 10x^2 common mode - which subtracts from the bias V but doesn't push or pull on the diaphragm because it is the same on both sides

the effect of subtracting from the bias V is negligible at audio for typical high resistance diaphragms since the time constant for actually changing diaphragm charge is long

Quote:

Originally Posted by jcx 

say the gain function for each side of the stator looks like

 

f(x) = 500x +10( x^2)

 

now put in x and -x as the differential signal and take the difference of the outputs:

 

( 500x +10(x^2)) - ( 500(-x) +10((-x)^2) ) = 1000x

 

even powers always give positive result for both +,-x that makes the +,-x even order distortion terms cancel in the difference output

 

you do have 10x^2 common mode - which subtracts from the bias V but doesn't push or pull on the diaphragm because it is the same on both sides

 

the effect of subtracting from the bias V is negligible at audio for typical high resistance diaphragms since the time constant for actually changing diaphragm charge is long

jcx, sorry I must be really thick but I still don't get the above. You're calculating the difference between stators and it cancels out the quadratic term in the gain function. What is that supposed to mean in practice? That the effective gain function in regards to push/pull drive is really just 1000x, then why bother with the quadratic term in the first place?

Also, what makes it sensible to calculate the differential output between stators when the diaphragm will independently react to each stator's voltage (one repulsing while the other attracts)?

PS: I had to dig back in the stax thread to read your / Kevin's and Birgir's posts about esl amp topologies and sensible performance specs, utterly informative so please bare with me and the 101 lecture series for the masses ;)

the 2nd order term was included as an example of a amplifier circuit's possible distortion - with the assumption that the distortion matches in both +,- drive sub circuits

with the usual linearizing assumptions the the net force on the charge stuck on the diaphragm is proportional to the difference between the V on the surrounding stators

I suppose there are a conceptual few steps in "the usual linearizing assumptions" getting to superposition, ignoring the bias V and seeing the "bound" charge on the diaphragm just responding to the difference V between the stators - not sure I can explain every step in a post or two

my motivation for the example is to further my question about how in a symmetic, full differential circuit you can "tune" if for a desired amount of 2nd order distortion