Hi,
I am a somewhat experienced speaker builder, building/designing several sets of conventional loudspeakers (including my most recent set that uses an active crossover).
Been getting the speaker building itch again and decided I would like to try a set of ESL's.
Just started doing my research on this and as such been reading Ken Seibert's and Jazzman's sites on their projects.
I'm looking to do the wire stator version. I would like build a full range version (or at the least get good, flat response down to 80hz) but not sure what size panels I would need to achieve that.
Just wondering if anyone has any advice for me as a newbie to ESL's?
I am a somewhat experienced speaker builder, building/designing several sets of conventional loudspeakers (including my most recent set that uses an active crossover).
Been getting the speaker building itch again and decided I would like to try a set of ESL's.
Just started doing my research on this and as such been reading Ken Seibert's and Jazzman's sites on their projects.
I'm looking to do the wire stator version. I would like build a full range version (or at the least get good, flat response down to 80hz) but not sure what size panels I would need to achieve that.
Just wondering if anyone has any advice for me as a newbie to ESL's?
With ESL's if we look at well established manufacturers like Martin Logan
and Quad ,they all curve the panel area.
Listening to ESL's that lack curving is very difficult and almost ridiculous.
You end up with extreme beaming, but with a curved panel excellent reproduction.
You should try to hear Quad ESL's and Martin Logans, to ascertain how
curving the panel is an essential requirement.
http://www.quadesl.com/pdf/quad_book.pdf
and a great video showing what is involved with
building them at Martin Logan https://www.youtube.com/user/MartinLoganSpeakers
Cheers / Chris
the panel
Hi Bengel
I suggest searching the DIYaudio site for segmented ESLs.
Segmented ESLs have the manufacturing advantage of flat ESLs, a broad listening area of curved ESLs (actually better, better too than many conventional speakers), and without the distortion of curved ESLs. They are ideally built as a line source - tall but relatively small footprint.
They can be made easily as a wire ESL, or using PCB - there are examples of both here.
You should also check out the paper by Baxandall there is a copy at Links & info - The wire electrostatic Loudspeaker page it is probably the best resource for ESL designers. There is also the paper by White (me) on segmented ESLs available on the same site Its a bit terse for DIY, but it has the basic design formulae for the segmented line-source ESL. Amongst the DIY searches you should also find an excel simulator that you can use to check the design.
For a full range ESL, you should expect a reasonably large area perhaps 0.5 m^2 or bigger - see formula for area vs cutoff frequency in the paper. The diaphragm resonance is also used to boost the bass - much like a bass reflex. I can flesh out details once you are underway.
good luck
I suggest searching the DIYaudio site for segmented ESLs.
Segmented ESLs have the manufacturing advantage of flat ESLs, a broad listening area of curved ESLs (actually better, better too than many conventional speakers), and without the distortion of curved ESLs. They are ideally built as a line source - tall but relatively small footprint.
They can be made easily as a wire ESL, or using PCB - there are examples of both here.
You should also check out the paper by Baxandall there is a copy at Links & info - The wire electrostatic Loudspeaker page it is probably the best resource for ESL designers. There is also the paper by White (me) on segmented ESLs available on the same site Its a bit terse for DIY, but it has the basic design formulae for the segmented line-source ESL. Amongst the DIY searches you should also find an excel simulator that you can use to check the design.
For a full range ESL, you should expect a reasonably large area perhaps 0.5 m^2 or bigger - see formula for area vs cutoff frequency in the paper. The diaphragm resonance is also used to boost the bass - much like a bass reflex. I can flesh out details once you are underway.
good luck
Definitely looking at how to deal with the beaming. Interestingly I have a small pair of magnepans and they beam like mirrors.
So trying to decide if I should do segmented or curved. I have access to sheet metal roller that will do something like 24 inches long...so I could do that.
Right now I am trying to understand why you would stick spacers to the diaphragm. Seems like adding mass to it would be counter productive.....Though I do understand that they keep the diaphragm from slamming into the stators
So trying to decide if I should do segmented or curved. I have access to sheet metal roller that will do something like 24 inches long...so I could do that.
Right now I am trying to understand why you would stick spacers to the diaphragm. Seems like adding mass to it would be counter productive.....Though I do understand that they keep the diaphragm from slamming into the stators
The spacers attach to the stators and support the diaphragm but, since they are stationary, they add no mass at all to the diaphragm.
Bengel,
I concur with Golfnut on segmentation… it works beautifully. My current segmented welding rod panels sound great but were very time consuming to build, since I had to splice all 540 rods to get them to 48” length, and the plastic supporting grids are not nearly as visually appealing as my old perf metal panels were.
I’m now planning another pair of segmented panels using insulated copper wire supported by a wooden lattice that should be quite visually appealing. And I really like segmentation. My current panels even have a multi-pole rotary switch in the loop that allows me to select either narrow dispersion mode (resistor networks jumped over) or wide dispersion mode (resistor networks in). You certainly can’t do that with perf-metal panels!
Most of the time, I leave them in wide dispersion mode, even for solo listening.
I’ve not actually built any curved panels so I won’t comment on those.
Good luck with your project!
Charlie (Jazzman)
I concur with Golfnut on segmentation… it works beautifully. My current segmented welding rod panels sound great but were very time consuming to build, since I had to splice all 540 rods to get them to 48” length, and the plastic supporting grids are not nearly as visually appealing as my old perf metal panels were.
I’m now planning another pair of segmented panels using insulated copper wire supported by a wooden lattice that should be quite visually appealing. And I really like segmentation. My current panels even have a multi-pole rotary switch in the loop that allows me to select either narrow dispersion mode (resistor networks jumped over) or wide dispersion mode (resistor networks in). You certainly can’t do that with perf-metal panels!
Most of the time, I leave them in wide dispersion mode, even for solo listening.
I’ve not actually built any curved panels so I won’t comment on those.
Good luck with your project!
Charlie (Jazzman)
I think I would limit myself to 36 inches tall unless I could find some straight metal rod longer. Splicing 500+ rods would drive me crazy 
I was looking at this stuff for the supporting grids, which I think looks better, it is hexagonal holes..... any issues you can see with it? ParaHex
Also, I think a hybrid is only thing practical for my 1st project. Definitely will be using active crossovers, for that I will use this kit... I like the analog-ness of them: Linkwitz-Riley Electronic Crossover
I could use some pointers on designing the transmission line "box" on the bottom. I've never designed a transmission line speaker before (just standard sealed and ported).
Thanks...
I was looking at this stuff for the supporting grids, which I think looks better, it is hexagonal holes..... any issues you can see with it? ParaHex
Also, I think a hybrid is only thing practical for my 1st project. Definitely will be using active crossovers, for that I will use this kit... I like the analog-ness of them: Linkwitz-Riley Electronic Crossover
I could use some pointers on designing the transmission line "box" on the bottom. I've never designed a transmission line speaker before (just standard sealed and ported).
Thanks...
I think I would limit myself to 36 inches tall unless I could find some straight metal rod longer. Splicing 500+ rods would drive me crazy
I was looking at this stuff for the supporting grids, which I think looks better, it is hexagonal holes..... any issues you can see with it? ParaHex
Also, I think a hybrid is only thing practical for my 1st project. Definitely will be using active crossovers, for that I will use this kit... I like the analog-ness of them: Linkwitz-Riley Electronic Crossover
I could use some pointers on designing the transmission line "box" on the bottom. I've never designed a transmission line speaker before (just standard sealed and ported).
Thanks...
The plastic hex grid should work fine. And so too the 24db/octave LR crossover.
Modeling a transmission line leads to insanity. But if you're up for the math you might try Martin J. King's TL calculator. King's math was over my head so I just used the generic guidelines in Sanders' ESL Cookbook; which are basically as follows:
- Line length should be about 1/4 wavelength of the lowest desired frequency
- For folded lines, the wall behind the woofer should be curved such that standing waves cannot rebound to the woofer cone.
- Line section area at the woofer should be at least 125% of the woofer's piston area (s/d) and should taper to 100% of s/d at the terminus.
- Stuffing density should be about 0.5 lbs/ft3.
- In order to best blend with the ESL's near-zero Q, choose a woofer with low Q, low inductance and a strong magnet--low inductance being most important.
I was thinking about using these drivers (I have 4 and love them). Looking for good response down to 50hz or so then let my sub take over from there. .... certainly have low inductance but Qms is a bit high than I would like.
Efficiency is terrible also but may match the ESL
http://www.glasswolf.net/papers/Extremis6.8Datasheet.pdf
Efficiency is terrible also but may match the ESL
http://www.glasswolf.net/papers/Extremis6.8Datasheet.pdf
I am a total noob to ESL's so what do you mean by this? I am planning to do a segemented ESL BTW, so they would have to be connected together, I would think.
Another thing I am curious about is in "traditional" drivers the movement of the cone causes back EMF that the amplifier "dampens" (which it does quite well if you do not have a passive crossover in the mix ).
With the amp audio output going through a step up transformer, I am curious as to what this does to the dampening factor of the amp - i.e. the amount of "control" it has over the back emf from the diaphragm moving.
). With the amp audio output going through a step up transformer, I am curious as to what this does to the dampening factor of the amp - i.e. the amount of "control" it has over the back emf from the diaphragm moving.
Anyone have any advice for the step up transformer(s)? I've read jazzman's&seibert's sites and a long thread with commentary by golfnut (who really knows what he is talking about when it comes to this).
Right now I am looking at this (or its 6v cousin) but not sure I like having to have 4 of these per panel.
AN-0105 - 10VA 5V Transformer - AnTek Products Corp
The pros and cons theory is a bit over my head of electronic knowledge so appreciate any general advice.
I am not looking for a super cheap option but not a super expensive one either (sound quality matters to me).
Right now I am looking at this (or its 6v cousin) but not sure I like having to have 4 of these per panel.
AN-0105 - 10VA 5V Transformer - AnTek Products Corp
The pros and cons theory is a bit over my head of electronic knowledge so appreciate any general advice.
I am not looking for a super cheap option but not a super expensive one either (sound quality matters to me).
If you make two 36" high segmented ESLs, you can stack one above the other to make a single 72" ESL. So there is no need to connect the rods or wires, just run the two panels connected in parallel to the amp. The segmented ESL simulates a line source, so stacking two will make a longer line source and follow theoretical response better (i.e. better bass response).
Do not consider the transformers until you decide what sort of ESL you build. A single segment ESL (e.g. perforated metal plate) favours larger transformers (50VA toroids or bigger) while the segmented ESL favours smaller transformers (10 VA toroids or more conventional EI or C-core type transformers)
Have no fear about using multiple toroids in each ESL. MY ESLs each have 16 toroids in them - slightly pricey solution but works well for segmented speaker.
regards
Rod
16?? holy.... thats gone be more expensive then a proper big trannie>? i bet you get some sick stepup though
Hi Wrinex
I found that 32 small toroids at wholesale rates was cheaper than a set of commercial ESL transformers. Also there are no commercial transformers made for segmented ESLs - they all have the wrong balance of leakage inductance and winding capacitance and would seriously limit the bandwidth - same problem as big toroids.
The number of transformers required also depends on the low frequency required, if I was happy with a response down to 100 Hz instead of 50Hz, I could halve the number. Actually, when segmented speakers are used in hybrids or not for really full range applications, choosing a bunch of 10VA toroids is really a no brainer. You just need to avoid the ones with the twin 110 V windings - capacitance is really high.
The best solution is of course a purpose made transformer - that can be done for less than US$150 for a pair, I understand - that's the approach I would take in the future. Most transformer manufacturers will do them if you provide the design. But that requires a bit more expertise than I had when I started playing this silly game.
regards
I found that 32 small toroids at wholesale rates was cheaper than a set of commercial ESL transformers. Also there are no commercial transformers made for segmented ESLs - they all have the wrong balance of leakage inductance and winding capacitance and would seriously limit the bandwidth - same problem as big toroids.
The number of transformers required also depends on the low frequency required, if I was happy with a response down to 100 Hz instead of 50Hz, I could halve the number. Actually, when segmented speakers are used in hybrids or not for really full range applications, choosing a bunch of 10VA toroids is really a no brainer. You just need to avoid the ones with the twin 110 V windings - capacitance is really high.
The best solution is of course a purpose made transformer - that can be done for less than US$150 for a pair, I understand - that's the approach I would take in the future. Most transformer manufacturers will do them if you provide the design. But that requires a bit more expertise than I had when I started playing this silly game.
regards
That's good advice, I am getting ahead of myself. I get what you're saying about the separate rods.
I just ordered the tig rods, mylar, louver panels etc.....so I am going to start on the stators soon.
So if I wanted response down to 50 is hz..... and the panels 15 inches ish wide....how tall would they have to be? Assuming 6u mylar and .0625 or maybe .0185 spacing.
Also I definitely will be doing a segmented design.....not sure on the configuration of that yet.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.