Electrostatic speakers?

[Jazzman53]

Mostly due to not having heard one that I find convincing, I still look on hybrid stats with suspicion...

Mind you, I haven't auditioned a set in a few decades!

Back in the day, I enjoyed the ML CLS & CLS-II but not any of the Hybrid models...

I have to agree that most hybrid ESLs never quite sound like a single, seamless speaker.

The only commercial hybrid ESLs I know of are from Sanders, Janzen and Martin Logan. Of those; I've only personally heard the older Martin Logan models, which in my opinion had good bass extension and punch but lacked seamless integration between the woofer and ESL panel.

The only full range ESLs I've hard are the Quad 57's and 63's. Their bass extended surprisingly low but wasn't as punchy as hybrid designs and was also a bit boomy and ill-defined at the bottom end, which I attributed to the diaphragm's drum-head resonance.

I've been an ESL designer and builder since 2008, and to-date I've built 20 pairs, in several configurations.

The challenge in full range ESL is mitigating the nasty-loud drum-head resonance. In a hybrid the stat panel doesn't have to play down low where the drum-head resonance dominates, but its challenge is achieving a seamless blend between the woofer and panel.

I wanted something different than the commercial ESLs I had heard so I took a different path; initially favoring hybrids with transmission-line bass, and now hybrids with segmented wire-stators and dipolar bass.

My latest and final hybrid ESL build sounds like a single-driver with the cleanest bass I've ever heard.

The key to seamless integration was using a low-Q [i.e well damped] woofer with low-inductance [i.e. minimal back-emf for faster transient response] mounted on an open baffle to match the ESL's dipolar radiation pattern. The low-Q woofer sacrifices deep bass extension to prioritize a seamless blending with the ESL, so subs are required for the bottom end (below 60Hz).

Building my speaker is more work than any sane person would attempt but I freely share my CAD drawings and all build info with other DIY'ers.

If interested; here's my website: http://jazzman-esl-page.blogspot.com

 

[SIY]

i.e. minimal back-emf for faster transient response

If you have a low-pass crossover before the woofer, the transient response isn't particularly relevant. Damping absolutely is, though. I have my current subs set for second order Q=0.5 alignment and they seem to blend in very nicely with my Quads using a very steep lowpass on the woofers, a similarly steep acoustic highpass on the Quads, and a delay to align their acoustic centers.

 

[Newman]

The very use of the term 'faster transient response' for subwoofers, and also the second idea of linking it to back-EMF, reveal false notions of what causes speakers to sound like they do.

 

[SIY]

The other issue may be polar pattern. I'm building a version of the ripole that @Jazzman53 suggested and I'll be curious to see if that's an improvement. I should also mention that I've set up a DBA and that might also have a lot to do with the quality of integration.

 

[egellings]

The very use of the term 'faster transient response' for subwoofers, and also the second idea of linking it to back-EMF, reveal false notions of what causes speakers to sound like they do.

A woofer doesn't have to be fast, since it's fed by a low pass filter set to a low cutoff frequency. I can jump up and down that fast with a beer in my hand.

 

[617]

A woofer doesn't have to be fast, since it's fed by a low pass filter set to a low cutoff frequency. I can jump up and down that fast with a beer in my hand.

It's still pretty fast depending on excursion but you're right.

I'd pin this on directivity and room interaction differences in the 80-200 region.

 

[Jazzman53]

The very use of the term 'faster transient response' for subwoofers, and also the second idea of linking it to back-EMF, reveal false notions of what causes speakers to sound like they do.

"Fast" (or "faster") was not the best choice of words here but I do think transient response (in addition to a well damped suspension) is important for blending a woofer in a hybrid ESL.

My inference that that the mid-bass woofer's back-EMF affects it's transient response follows from the back-EMF opposing the drive signal, and since a low-inductance voice coil generates less back-EMF, it follows that lower inductance gives better transient response.

I neglected to mention the importance of the woofer having a well designed shorting ring (or rings) to minimize its inductance, as opposed to merely having fewer turns in the voice coil, giving lower drive force.

In fact, for integrating a woofer in a hybrid ESL, I prioritize its voice coil inductance over [even] its moving mass, for best transient response in the crossover band.

 

[AdamG]

The other issue may be polar pattern. I'm building a version of the ripole that @Jazzman53 suggested and I'll be curious to see if that's an improvement. I should also mention that I've set up a DBA and that might also have a lot to do with the quality of integration.

Pictures and progress reports are requested.

 

[SIY]

Pictures and progress reports are requested.

AudioXpress gets first crack, but if they don't want to publish it, I'll happily post it here.

 

[NTK]

"Fast" (or "faster") was not the best choice of words here but I do think transient response (in addition to a well damped suspension) is important for blending a woofer in a hybrid ESL.

My inference that that the mid-bass woofer's back-EMF affects it's transient response follows from the back-EMF opposing the drive signal, and since a low-inductance voice coil generates less back-EMF, it follows that lower inductance gives better transient response.

I neglected to mention the importance of the woofer having a well designed shorting ring (or rings) to minimize its inductance, as opposed to merely having fewer turns in the voice coil, giving lower drive force.

In fact, for integrating a woofer in a hybrid ESL, I prioritize its voice coil inductance over [even] its moving mass, for best transient response in the crossover band.

Voice coil inductance is one of the contributing factors that limits the bandwidth of a dynamic driver. If the bandwidth is sufficient, and woofers don't need much, it is not an issue. Back EMF is the product of the Bl factor (force factor) and the cone/voice-coil velocity (see equation 10.2.4: motional emf). Voice coil inductance comes into play by limiting the voice coil current with frequency, which limits the force generated by the motor (force = Bl × current), which drives the motion of the cone.

For woofers, the effect from voice coil inductance is so small that it is neglected in most analyses. See below (source). With an inductance of 0.3 mH of a typical woofer, at 120 Hz, the impedance from inductance is only 0.23j Ω.

The effect of moving mass on the woofer is sensitivity (efficiency), not "speed". See link.

 

[Jazzman53]

AudioXpress gets first crack, but if they don't want to publish it, I'll happily post it here.

Great idea! AudioXpress is always looking for quality project articles, and I'm deferring dibs to you for a RiPol article.

BTW; per AudioXpress article guidelines, there would be no conflict if you post your project here first. I know this because two of my earlier ESL builds were posted on several forums, and I later published the same projects in the April 2012 & September 2017 editions of AudioXpress-- I even copied entire paragraphs verbatim from my website and forum posts into both AudioXpress articles. I checked with the AudioExpress editor about this first, and was told that forum posts aren't copywrited, and copying verbiage that I wrote wasn't plagiarizism.

 

[SIY]

Great idea! AudioXpress is always looking for quality project articles, and I'm deferring dibs to you for a RiPol article.

BTW; per AudioXpress article guidelines, there would be no conflict if you post your project here first. I know this because two of my earlier ESL builds were posted on several forums, and I later published the same projects in the April 2012 & September 2017 editions of AudioXpress-- I even copied entire paragraphs verbatim from my website and forum posts into both AudioXpress articles. I checked with the AudioExpress editor about this first, and was told that forum posts aren't copywrited, and copying verbiage that I wrote wasn't plagiarizism.

I read those articles and they are terrific. I wish I had the ability to build panels the way you do.

 

[misterdog]

I use a Dipole (Lab12C) sub fed from a MiniDSP plate amp which gives me full control, crossover frequency, volume control etc. from my laptop.

This compliments my Quad 989 ESL's. which are fed full range (Benchmark AHB2 ).

I have taken all the panels out of the flimsy Quad enclosures and fitted them to a bespoke steel frame with the panel centre at ear height.
The increase in rigidity and removal of the safety grid and dust covers proved to be a transformation.

After the refurbishment, I had to lower the resistance on my phono stage as there was so much additional high frequency information coming through, I think this also affected the rest of the frequency range, ie. more usable bass.
The sub just adds to the lowest octave, and 'activates' the room, as people expect from a sound system.

 

[sq225917]

Probably worth a mention that I live near mrdog and know his system well. It's quite simply the best sounding home hifi rig I've heard, and I get about more than most as I build and sell a pretty expensive phonostage, so I'm exposed to a variety of high end rigs quite often. In fact I'm off to hear one tonight.

With big stats it's all about room integration.

 

[617]

I use a Dipole (Lab12C) sub fed from a MiniDSP plate amp which gives me full control, crossover frequency, volume control etc. from my laptop.

This compliments my Quad 989 ESL's. which are fed full range (Benchmark AHB2 ).

I have taken all the panels out of the flimsy Quad enclosures and fitted them to a bespoke steel frame with the panel centre at ear height.
The increase in rigidity and removal of the safety grid and dust covers proved to be a transformation.

After the refurbishment, I had to lower the resistance on my phono stage as there was so much additional high frequency information coming through, I think this also affected the rest of the frequency range, ie. more usable bass.
The sub just adds to the lowest octave, and 'activates' the room, as people expect from a sound system.

Have you taken detailed anechoic measurements of the 989 by any chance? I was always very impressed by how innovative these units were, truly one-of-a-kind, but nobody has ever measured them.

 

[Blumlein 88]

Have you taken detailed anechoic measurements of the 989 by any chance? I was always very impressed by how innovative these units were, truly one-of-a-kind, but nobody has ever measured them.

Quad ESL-989 electrostatic loudspeaker Measurements

Sidebar 3: Measurements Although the different kind of interaction between a panel speaker and the room can make it sound louder than its measured sensitivity would suggest, the Quad ESL-989 is still on the low side, at an estimated 83dB(B)/2.83V/m. Given its strict 100W power handling, this...

www.stereophile.com

Not quite up to Klippel data. They appear to be near identical to the Quad ESL-63 in terms of the measurements done.

 

[mhardy6647]

The only commercial hybrid ESLs I know of are from Sanders, Janzen and Martin Logan. Of those; I've only personally heard the older Martin Logan models, which in my opinion had good bass extension and punch but lacked seamless integration between the woofer and ESL panel.

Depending on how one defines "hybrid ESLs", there were numerous others, especially in the olden days.
Bowers & Wilkins, e.g., sort of established themselves with their DM70 (one of the few B&W loudspeakers I've heard that I can actually bear listening to for more than a few minutes).

Blog - Beginnings | Bowers & Wilkins

The early history of the world’s best-known and most acclaimed loudspeaker company is a fascinating tale of obsession for the very best – mixed with the occasional stroke of good fortune.

www.bowerswilkins.com

(borrowed image)

Electrostatic tweeters (nearly "supertweeters" by current reckoning) were all the rage in the 1950s and 1960s, often as add-ons to good loudspeakers lacking adequate HF extension). Here's an example from 1961. Radio Shack apparently did pretty well for themselves selling add-on electrostatic tweeters for the Acoustic Research AR-2. The AR-2 didn't really have "tweeters" as we might think of them today -- only a pair of "cross-firing" 5" cone drivers. Eventually, AR offered AR-2a, adding a dome tweeter to the original AR-2 design.

Radio Shack also offered inexpensive two-way loudspeaker systems for many years based on an 8" woofer and an array of four single-ended electrostatic tweeters (the Electrostat-2 and later -2A). The Electrostat-3 addon (above and also below) may have been a proper push-pull electrostatic and has been reputed by some on teh webz to have been JansZen OEM. Not sure about that, though.

both images above from https://www.radioshackcatalogs.com/flipbook/1961_radioshack_catalog.html

EDIT... and I plumb forgot to mention the Dayton-Wright XG-8 and XG-10 loudspeakers!




Better European and even some American table radios of the 1950s included electrostatic tweeters, too. It was a thing in those days.

 

[Blumlein 88]

I had a pair of the Janzen tweeters like this.

Also one of the Realistic type Electrostatic 2 tweeters. It had another name on them which I don't recall at the moment. They made them for Radio Shack.

 

[mhardy6647]

In a bizarre but perhaps not surprising coincidence I have a box containing what appear to be four NOS replacement tweeter panels for the Electrostat 2/2A/4 that I found one day, many years ago, on the swap pile at the much-beloved Harvard, MA town dump.

There's also a pair of Electrostat-2As here (again, perhaps not surprisingly), although one of the original woofers was (is) toast.
In principle, the single-ended* electrostatic tweeters will generate high levels of distortion -- and they probably do. In practice, they're quite sweet sounding.


(Electrostat-2 catalog blurb from the early '60s)

Oh, there're also two pairs of Electrostat 3 add-on tweeters here. Stuff just kind of accretes here sometimes...

[/url
One of the Electrostat 3s is visible (under a R/S 40-1375 leaf tweeter in a plastic Hammond cabinet) to the right of the FrankenAltec.
_____________________
* i.e., only one driven membrane -- the Quads and most if not all other good electrostatic drivers are push-pull, with both sides' membranes driven.

 

[617]

In a bizarre but perhaps not surprising coincidence I have a box containing what appear to be four NOS replacement tweeter panels for the Electrostat 2/2A/4 that I found one day, many years ago, on the swap pile at the much-beloved Harvard, MA town dump.

View attachment 395001
View attachment 395002
There's also a pair of Electrostat-2As here (again, perhaps not surprisingly), although one of the original woofers was (is) toast.
In principle, the single-ended* electrostatic tweeters will generate high levels of distortion -- and they probably do. In practice, they're quite sweet sounding.


(Electrostat-2 catalog blurb from the early '60s)

Oh, there're also two pairs of Electrostat 3 add-on tweeters here. Stuff just kind of accretes here sometimes...

[/url
One of the Electrostat 3s is visible (under a R/S 40-1375 leaf tweeter in a plastic Hammond cabinet) to the right of the FrankenAltec.
_____________________
* i.e., only one driven membrane -- the Quads and most if not all other good electrostatic drivers are push-pull, with both sides' membranes driven.

You in Harvard? One of my favorite places in the area.