Good morning everybody,
I have read with much interest the paper by Jan Didden on a diy direct drive amplifier for QUAD ESL 63 in: https://audioxpress.com/article/you-can-diy-edd-a-direct-drive-amplifier-for-electrostatic-speakers.
I am and have been a lover of QUAD electrostatic speakers for the last 40years and think that the ESL 63 (properly maintained), are the best of them all ,including their successors , but the building of a direct drive amplifier is beyond my limited abilities . However, the search for the best amplifier for QUAD electrostatics still remains (imho)
In particular I would appreciate your opinion and direct experience on driving them with high power class D amplifiers such as Purifi or Hypex n-core cx.
Many thanks
Bruno
PS: I have both the 63s and 2905s and, at present, drive them with both tubes (Quicksiver V4_) and ss amplifiers (Spectral studio 160) .In my opinion tubes are a little ...better ,but only a very little...
I have read with much interest the paper by Jan Didden on a diy direct drive amplifier for QUAD ESL 63 in: https://audioxpress.com/article/you-can-diy-edd-a-direct-drive-amplifier-for-electrostatic-speakers.
I am and have been a lover of QUAD electrostatic speakers for the last 40years and think that the ESL 63 (properly maintained), are the best of them all ,including their successors , but the building of a direct drive amplifier is beyond my limited abilities . However, the search for the best amplifier for QUAD electrostatics still remains (imho)
In particular I would appreciate your opinion and direct experience on driving them with high power class D amplifiers such as Purifi or Hypex n-core cx.
Many thanks
Bruno
PS: I have both the 63s and 2905s and, at present, drive them with both tubes (Quicksiver V4_) and ss amplifiers (Spectral studio 160) .In my opinion tubes are a little ...better ,but only a very little...
Before I designed the direct drive amp I have driven them with a Purifi class D as well as with my own class AB amp.
The Purifi worked very well; the limiting factor is the ESL step-up transformer as documented in my article.
But it is just as with class A(B) - there are crappy class D amps around that should be avoided.
Quality has its price - but the quality remains after the price is forgotten ;-).
Jan
The Purifi worked very well; the limiting factor is the ESL step-up transformer as documented in my article.
But it is just as with class A(B) - there are crappy class D amps around that should be avoided.
Quality has its price - but the quality remains after the price is forgotten ;-).
Jan
My place to go with the 63 is a total sum of only € 3900 for a complete set of pre-amp and 2 mono blocks from Schiit
https://www.schiit.com/products/kara
https://www.schiit.com/products/tyr
Class D ... Naaaaah or should I say shhhhhhhhhhhhhh. Not for me. Tubes, same, not for me.
Class D amps have high numbers, not high power. A 1000W in Class D says nothing.
But most important enjoy your music.
https://www.schiit.com/products/kara
https://www.schiit.com/products/tyr
Class D ... Naaaaah or should I say shhhhhhhhhhhhhh. Not for me. Tubes, same, not for me.
Class D amps have high numbers, not high power. A 1000W in Class D says nothing.
But most important enjoy your music.
I use the 250W Ncore amps with my ESLs and am very happy with the results. Not sure I can tell the difference with good class AB. I have no experience with tube amps. The ESLs are floor to ceiling RC transmission lines - see https://www.diyaudio.com/community/threads/another-segmented-esl.234975/. The panels have about 1200pF total capacitance but the RC transmission line configuration reduces effective capacitance to less than 100 pF above 20 kHz, so the loading effect on the transformer is much reduced.
Jan's comment about transformers is notable. There are three basic ESL designs requiring very different transformers -
Single segment ESLs have large capacitance, > 1000pF, limiting the transformer step up ratio to about 60. Also the impedances on the primary side must be low resistance and low inductance, so faults tend to release smoke.
Multi-segment RC transmission line ESLs have a mixed R,C impedance with effective capacitance at 20 kHz less than 100 pF, so step-up ratios up to 150 practical.
Multi-segment LC transmission line (ESL-63) have nominally resistive impedance, as Quad designs show, transformers up to > 200 are practical.
See posts under https://www.diyaudio.com/community/threads/esl-transformer.55950/ for more detail.
Jan's comment about transformers is notable. There are three basic ESL designs requiring very different transformers -
Single segment ESLs have large capacitance, > 1000pF, limiting the transformer step up ratio to about 60. Also the impedances on the primary side must be low resistance and low inductance, so faults tend to release smoke.
Multi-segment RC transmission line ESLs have a mixed R,C impedance with effective capacitance at 20 kHz less than 100 pF, so step-up ratios up to 150 practical.
Multi-segment LC transmission line (ESL-63) have nominally resistive impedance, as Quad designs show, transformers up to > 200 are practical.
See posts under https://www.diyaudio.com/community/threads/esl-transformer.55950/ for more detail.
I have used a variety of class D amps with the newer quads over the years. In the last few years, I have been using nCore amps on them. My main system with a pair of 2912's uses a Cambridge audio EVO150 that has an nCore amp in it (NC252MP I believe). It works extremely well. That EVO150 is a really good sounding and lovely machine and drives the quads with authority. My shop isn't class D, but is also a world class amp: The benchmark AHB2 driven by their DAC3HGC. Both of these electronics chains are extremely low distortion and transparent and sound virtually identical to each other.
In the past I've tried tripath amps, B&O amps, and the older UcD models with the quads, but it wasn't until the nCores that they really became transparent and world-beaters. It's been interesting to have played with solid state, tubes and these class D amps over the years and as distortion has steadily dropped with the best of them, their sonic signatures fade away and they sound more and more like nothing.
I use the nCore amps on the original quads but have not used the older class D amps on them.
Sheldon
In the past I've tried tripath amps, B&O amps, and the older UcD models with the quads, but it wasn't until the nCores that they really became transparent and world-beaters. It's been interesting to have played with solid state, tubes and these class D amps over the years and as distortion has steadily dropped with the best of them, their sonic signatures fade away and they sound more and more like nothing.
I use the nCore amps on the original quads but have not used the older class D amps on them.
Sheldon
I migrated from nCore to Purifi some time ago and have the impression that the midrange got a bit smoother and, for lack of a better word, more authorative. But subtle, sometimes I doubted if there was a change. I'm using Acourate for xover, speaker and room correction with the ' 63's and the Gradien subs and have also been swapping multichannel DACs. Topping DM7 (a secret well-kept) worked quite well, but I needed AES inputs (Topping is strictly USB) and am now playing with an Okto Research DAC8 Pro. Better than the Topping, but again, subtle. Baby steps, but class D really has come of age.
Yesterday I was in Germany for a preview of a system that will play at the Munich High-End show in May. 3-way system by Joachim Gerhard, Acourate-corrected. All Purifi, drivers and amps. It didn't really sound much better than my own system but of course with a quartet of large Purifi woofers, more dynamic and powerful than the '63 + Gradient. We live in the best of audio-times!
Jan
This is the '63 impedance, measured directly at the stator terminals. Mostly resistive.
Jan
Resurrecting a dormant thread.
I am routinely measuring signals directly at the output of the step-up transformers, using a purpose-designed high voltage 1:1000 attenuator.
The attenuator and its XLR connecting cable have been tuned to be flat to 50kHz within a fraction of a dB loaded with the the AP XLR input impedance. This works very well.
I now would like to use this to measure the step-up output when it is driven by a class D amp. Just wondering how much of the switching residu makes it through the step-ups, knowing that the xformers are high-frequency limited. You get my curiosity ;-)
The AP input autoranging circuits can get confused by high frequency residues, and to fix that AP developed a series of class D measurement filters, and I am the proud owner of an AUX-0025. You pick the signal off the amp load, run it through the AUX-0025 and then into the AP analyzer. Works like a champ.
But to run the signal from my 1:1000 attenuator through the AUX-0025 presents a problem because that filter does not present the same impedance as the AP input would. So I see two solutions:
- add a low output impedance buffer to the high voltage attenuator, which presents the required Zin to the attenuator and that can drive the AUX-0025. One problem: that buffer should be low distortion at audio and be flat to 1MHz or so. Not trivial.
- not worry and just connect the attenuator directly to the analyzer not worrying about any residuals. After all, the step-ups won't let through those hundreds of kHz switching component anyway.
And that thus is my basic question: would there any class D switching residue at the secondary of the step-up transformers?
Just collecting informed views here before setting the goals of the project.
As always grateful for your insights.
Jan
I am routinely measuring signals directly at the output of the step-up transformers, using a purpose-designed high voltage 1:1000 attenuator.
The attenuator and its XLR connecting cable have been tuned to be flat to 50kHz within a fraction of a dB loaded with the the AP XLR input impedance. This works very well.
I now would like to use this to measure the step-up output when it is driven by a class D amp. Just wondering how much of the switching residu makes it through the step-ups, knowing that the xformers are high-frequency limited. You get my curiosity ;-)
The AP input autoranging circuits can get confused by high frequency residues, and to fix that AP developed a series of class D measurement filters, and I am the proud owner of an AUX-0025. You pick the signal off the amp load, run it through the AUX-0025 and then into the AP analyzer. Works like a champ.
But to run the signal from my 1:1000 attenuator through the AUX-0025 presents a problem because that filter does not present the same impedance as the AP input would. So I see two solutions:
- add a low output impedance buffer to the high voltage attenuator, which presents the required Zin to the attenuator and that can drive the AUX-0025. One problem: that buffer should be low distortion at audio and be flat to 1MHz or so. Not trivial.
- not worry and just connect the attenuator directly to the analyzer not worrying about any residuals. After all, the step-ups won't let through those hundreds of kHz switching component anyway.
And that thus is my basic question: would there any class D switching residue at the secondary of the step-up transformers?
Just collecting informed views here before setting the goals of the project.
As always grateful for your insights.
Jan
I listen to the Quads in the traditional meeting / listening and point of the situation with a top of the range Sony CD source with Perreax PFM 2350 pre and power amplifier and it makes them sound divine. The only problem is that at a certain listening level you have to hold back, otherwise you risk knocking them out with a repair estimate of more than 1000 euros.
I always found that I used a lower listening level with the 63s, and had to turn up my cone speakers to get the same resolution of detail.
@jan.didden Have you considered passive impedance matching?
You may have to give it some time. There are a few very knowledgeable users that don't seem to visit as frequently these days.