The Headcase Stax thread

[Duggeh]

If I had a DIY fetish, I'd have at minimum, a semi-on at that.

I will get around to building my own drivers one day. Gotta be 3 or 4 years hence though.

[Smeggy]

Spritzer, try not to fry yourself with these mad contraptions

So, when are you making the other one and adding earpads? hehe.

[spritzer]

Spritzer, try not to fry yourself with these mad contraptions

So, when are you making the other one and adding earpads? hehe.

It's perfectly safe unless you touch it... It creeks and crackles so I'll scrap it but the charging time is very low and the coating presents no problems for the Stax bias supplies. I'd rate that as a success.

I do have the "itch" to build some speakers as I was reading about Quads most impressive project, spherical ESL design with stacked diaphragms tied to a time delay. That means it throws sound into a room and the drivers can be smaller as they make up for the surface area by stacking. These were meant as replacements for the ESL63 but the tooling was very expensive and the build would require more precision then the Quad plant could muster so they scrapped them.

[Dusty Chalk]

So would the sound still be directional, or are we heading into omnidirectional territory? It's late, so I'm having trouble thinking it through, the spherical driver would push it in the omnidirectional territory, but I don't know what the stacked diaphragms would do, counter it, promote it, or what? Is that related at all to their staggered signal delay technology? (I suspect not.) They were actually stacking drivers, like a pseudo-line-source? So it's a pseudo-line-source with a slight omnidirectional bent, like the Gallo Reference 3.x, except with an electrostatic basis?

[spritzer]

So would the sound still be directional, or are we heading into omnidirectional territory? It's late, so I'm having trouble thinking it through, the spherical driver would push it in the omnidirectional territory, but I don't know what the stacked diaphragms would do, counter it, promote it, or what? Is that related at all to their staggered signal delay technology? (I suspect not.) They were actually stacking drivers, like a pseudo-line-source? So it's a pseudo-line-source with a slight omnidirectional bent, like the Gallo Reference 3.x, except with an electrostatic basis?

Peter Walker stacked the drivers in the same way Koss did with the Model 1 speaker i.e. multiple diaphragms one in front of the other. Sony also did the same thing with the extremely rare SS-R10 speakers but only Quad employed time delays (similar to the ones that drive the dispersion rings in the ESL63) to time the release and make a spherical wavefront. Whats amazing is that all the time delay stuff is analog and done with lots and lots of copper wire so this could be done today with DSP relatively easily.

All electrostatics are directional unless the sound if fed through an acoustical lens like Beveridge for instance. The spherical wavefront would make them a bit less directional (like the ESL63 with its pseudo point source) but directionality is always an issue. The reason for the design was to make a smaller version of the ESL63 that would fit into more homes as the 63 is a pretty large and ugly beast. I'm not so sure a 60cm long sphere would have been that much better...

[Smeggy]

I once read a big article about Peter and the stacked, time-delay panels in the Quads. Very good read and a very clever guy. I fell in love with the original Quad 30 years ago and have very fond memories of it even though I never owned a set myself. Shame they were so ugly. No WAF whatsoever!

[Dusty Chalk]

...multiple diaphragms one in front of the other...

Oh. That's actually kind of bizarre, but by lowering the amount each driver had to move, you might also improve power handling and dynamics (the Quads' fatal weakness, IMHO).

Whats amazing is that all the time delay stuff is analog and done with lots and lots of copper wire so this could be done today with DSP relatively easily.

Well, yes and no -- I don't think there's much call for doing DSP at speaker level signals, it's usually done at line level signals, which means either going all the way and making an active speaker (not unheard of, since Quad make some fairly respectable amplifiers), or some sort of pre-processor. Also, believe it or not, at the delay levels we're talking about, we'd need some serious DSP processing power -- it's perfectly fine to do it in the analog realm. And it's not just using long copper wire -- there are circuits that can do small analog time delays (I think -- I suspect that's what's in the Sunfire subs).

Thanks for the info, that would actually be something I'd be interested in, I think. I was also thinking ribbon (for an omnidirectional tweeter), though I haven't seen a curved one, and am not sufficiently familiar with the technology to know if it would work or not.

[spritzer]

I once read a big article about Peter and the stacked, time-delay panels in the Quads. Very good read and a very clever guy. I fell in love with the original Quad 30 years ago and have very fond memories of it even though I never owned a set myself. Shame they were so ugly. No WAF whatsoever!

The basic model (all brown and British looking) looks like crap but some of the restored models look great. I've also been told that the new modernized version with 3um film is quite a bit better then the original which is no small feat. It also needs to be placed on stands to sound anywhere near its potential.

Oh. That's actually kind of bizarre, but by lowering the amount each driver had to move, you might also improve power handling and dynamics (the Quads' fatal weakness, IMHO).

The multiple diaphragms are to increase dynamics and bass output so they often used different diaphragm thickness depending on what they wanted each layer to do. A 12um diaphragms sound much thicker then a 3um film and if you take the tension into account you can tune the speaker relatively easily. Just to clarify the diaphragms all have + and - stators so you don't stack diaphragms one on top of the other. That wouldn't work.

The problem is that the drivers are even harder to make, let alone to fix if something breaks down. Koss had major issues with their speakers and even sunk so low as to add a tweeter. There are a lot of different forces at play here so they need to be controlled. I did make a prototype of a dual diaphragm headphone driver and while it worked it would need months to perfect to find the right combo of diaphragm thickness and how taught they should be.

Well, yes and no -- I don't think there's much call for doing DSP at speaker level signals, it's usually done at line level signals, which means either going all the way and making an active speaker (not unheard of, since Quad make some fairly respectable amplifiers), or some sort of pre-processor. Also, believe it or not, at the delay levels we're talking about, we'd need some serious DSP processing power -- it's perfectly fine to do it in the analog realm. And it's not just using long copper wire -- there are circuits that can do small analog time delays (I think -- I suspect that's what's in the Sunfire subs).

Thanks for the info, that would actually be something I'd be interested in, I think. I was also thinking ribbon (for an omnidirectional tweeter), though I haven't seen a curved one, and am not sufficiently familiar with the technology to know if it would work or not.

It would be impossible (price wise) to do the time delay at line level as that would require lots of amps (4 per speaker) all with their own transformer and protective circuits. All the panels need to be fed their own signal so this would get big and heavy very quickly. It's too bad that nothing like this gets made so we are stuck with Martin Logan or yet another rehash of the ESL63.

Stax also used resistor to control the dispersion of the "F" speaker line so that the flat panel would produce an arc wave form for a wider dispersion. This would be similar to what Peter Walker was planning to do with the flat panels on the sphere speaker. Curved speakers are a bad idea and I just don't get why M-L has stuck with it.

[Dusty Chalk]

Curved speakers are a bad idea and I just don't get why M-L has stuck with it.

They actually have had varying bits of luck with that (speaking of interviews, there was one with one of the founders), so the short answer is, because that's what their target market wants. But then, at this point, it's what their target market wants because that's what they make successfully. And by "varying bits of luck", I mean, both good and bad. The bad has led to their rather tight production line, so the good hasn't been simply luck, but manufactured opportunity. (Can you tell I've become a big Martin Logan fan? Pretty much after reading that interview.)

Curved stators make perfect sense. You're just stretching materials/textiles (both mylar and the metal bits) over a frame, there's no reason the frame has to be flat.

[spritzer]

They actually have had varying bits of luck with that (speaking of interviews, there was one with one of the founders), so the short answer is, because that's what their target market wants. But then, at this point, it's what their target market wants because that's what they make successfully. And by "varying bits of luck", I mean, both good and bad. The bad has led to their rather tight production line, so the good hasn't been simply luck, but manufactured opportunity. (Can you tell I've become a big Martin Logan fan? Pretty much after reading that interview.)

It does make perfect sense for them as their customers aren't after quality, they just want something flashy. M-L is going leaps and bounds backwards by decreasing the panel size and letting it flex freely. They are also firmly in the 80's with the diaphragm thickness. Basically made to look pretty but with little to no regard for sound quality. They even have a part of the panel overlap the bass unit to have something nice and flat to reflect the backwave.

Curved stators make perfect sense. You're just stretching materials/textiles (both mylar and the metal bits) over a frame, there's no reason the frame has to be flat.

It makes no sense at all really. The diaphragm is only clamped on the sides so this creates an uneven range of motion while not really delivering on the promise of a wide sweet spot. Diffused is just that, diffused. A flat panel works as a piston while a curved unit doesn't have same simple movement. If you want a wide dispersion angle then the multi faceted way used by Sound Lab is the only proper way.

[laxx]

I've always wanted to have a listen at ML Summits. They look nice and I think it's easier to get those approved by significant others compared to alot of other speakers.

[spritzer]

I've always wanted to have a listen at ML Summits. They look nice and I think it's easier to get those approved by significant others compared to alot of other speakers.

That's really their function in life i.e. WAF. They could be modified though to be much better but it would damage the looks

[Dusty Chalk]

The diaphragm is only clamped on the sides so this creates an uneven range of motion while not really delivering on the promise of a wide sweet spot. Diffused is just that, diffused. A flat panel works as a piston while a curved unit doesn't have same simple movement. If you want a wide dispersion angle then the multi faceted way used by Sound Lab is the only proper way.

So it's okay to criticize Martin Logan for using a curved surface, but not Quad? Methinketh you doth protest too much. I mean, there's curmudgeonly (which I enjoy) and ... then there's just plain cantankerous (which gets old).

[spritzer]

So it's okay to criticize Martin Logan for using a curved surface, but not Quad? Methinketh you doth protest too much. I mean, there's curmudgeonly (which I enjoy) and ... then there's just plain cantankerous (which gets old).

Quad never made any ESL speakers with a curved panel. If you would read the white papers Peter Walker posted on ESL design in 1955 he strongly opposes all tricks like curved panels and hybrids for the simple reason that they don't sound like music. He even said the same thing about his own older corner ribbon speaker. Something like it produces bass and treble but it doesn't sound like music. The ESL63 was designed to mimic a point source for a wider dispersion angle and that was simply done by manipulating the waveform.

[Dusty Chalk]

No, the way this whole conversation started was, you were talking about spherical electrostatic drivers. "Spherical" counts as curved.

[spritzer]

No, the way this whole conversation started was, you were talking about spherical electrostatic drivers. "Spherical" counts as curved.

Ahh no you must have misunderstood me. The drivers were flat but they were mounted in a spherical case, 60cm long, filled with absorbing material and vented at the back. It's similar to that Fujitsu speaker line.

[Dusty Chalk]

You mean the Eclipse? I didn't realize that was Fujitsu.

And: oh. Never mind then.

[spritzer]

You mean the Eclipse? I didn't realize that was Fujitsu.

And: oh. Never mind then.

Yup, they are/were a devision of Fujitsu.

It would be possible to make a spherical or dome ESL driver but the stretching jig would pretty hard to design.

[Dusty Chalk]

Yeah, well after your explanation, I picture more of a multifaceted thing more than a true hemisphere.

[Dusty Chalk]

And on an unrelated note -- a sphere, though less "boxy" than a box, still has a resonance. If it's more of an egg shape, then less so, but if it's truly spherical inside, it's like every side has a parallel surface. A pyramid would be better.

[spritzer]

And on an unrelated note -- a sphere, though less "boxy" than a box, still has a resonance. If it's more of an egg shape, then less so, but if it's truly spherical inside, it's like every side has a parallel surface. A pyramid would be better.

It looked a bit like a Zeppelin with a flat front end and they chose it because while it does resonate it is a very robust shape so cheaper materials could be used. Still the pictures aren't very good as they might be afraid somebody would copy it.

Parallel sides are much more of an issue with dynamic drivers then with electrostatics since ESL dispersion is a figure 8. The backwave needs to be either absorbed or reflected in some other direction as it kills the bass output.

[Dusty Chalk]

It looked a bit like a Zeppelin with a flat front end and they chose it because while it does resonate it is a very robust shape so cheaper materials could be used. Still the pictures aren't very good as they might be afraid somebody would copy it.

I'm not sure if egg-shape falls to my same criticism, though. In fact, I'm pretty sure it doesn't. Imagine being on an egg surface on the inside -- if you look straight across, there are only a couple of places where you're actually looking at a parallel surface (top, bottom, "equator").

Also, one resonance is easier to deal with than three. So it's not like it's a fatal flaw to go with that particular enclosure shape.

Parallel sides are much more of an issue with dynamic drivers then with electrostatics since ESL dispersion is a figure 8. The backwave needs to be either absorbed or reflected in some other direction as it kills the bass output.

I'm not sure your first statement makes sense. A dynamic driver has rear dispersion as well as forward dispersion as well, and is figure-8-ish in shape as well. The only difference (and I won't pretend to know more about this than I do, which is well nigh naught) may be that ESL drivers handle back-EMF differently than dynamic drivers do, which I suspect is the case. As to whether that's "better" or "worse", I don't know. But either way, you still have to worry about the signal reflecting and interfering with the forward signal and causing comb filtering, so I certainly agree with your conclusion.

[spritzer]

I'm not sure if egg-shape falls to my same criticism, though. In fact, I'm pretty sure it doesn't. Imagine being on an egg surface on the inside -- if you look straight across, there are only a couple of places where you're actually looking at a parallel surface (top, bottom, "equator").

Also, one resonance is easier to deal with than three. So it's not like it's a fatal flaw to go with that particular enclosure shape.

The parallel surfaces are an issue if the backwave manages to bounce back through the driver. This will kill the bass and all sound staging so you have tune the shape for a certain wavelength and port the back. I think this would be possible to pull off today as we have much better adsorbtion materials then in the early 80's.

I'm not sure your first statement makes sense. A dynamic driver has rear dispersion as well as forward dispersion as well, and is figure-8-ish in shape as well. The only difference (and I won't pretend to know more about this than I do, which is well nigh naught) may be that ESL drivers handle back-EMF differently than dynamic drivers do, which I suspect is the case. As to whether that's "better" or "worse", I don't know. But either way, you still have to worry about the signal reflecting and interfering with the forward signal and causing comb filtering, so I certainly agree with your conclusion.

I rarely make any sense at all... The parallel sides are both a reflection issue and structural issue. Dynamics produce so much vibration and sound so designers try to tune it out pretty much the same as ESL designers try to handle the backwave. Peter Walker nearly worked himself to death until he was satisfied that a backwave would be acceptable in a domestic setting though he did put the filter on the tweeter panels.

Speaking of ESL design, has anybody seen the new Hi-fi News with a review of the new Siltech Pantheon or what ever it is called. The pair matching is beyond horrible for a 120-130k$ speaker and the ESL tweeters aren't even similar.

[Elephas]

Adjusting SRM-T1W tube bias is not easy.

Just installed a pair of NOS made in Japan Raytheon-brand 6CG7 tubes.

Can I become a "made guy" now?

[spritzer]

New tubes drift like crazy so that would be expected.

I think we have to bring your status up for a vote...