| DIY Electrostatic speakers Audio Transformer??? - Click HERE for Original Thread |
| pmscompany |
Hi, mmm my english is very bad. What audio transformer, is possible uses for basical Diy Electrostatic speaker.
I lieve in Chile, mmm not possible buy a audio transformer in online. Chile have a industries work transformers... What is a carracteristics for buy a similar or order or work a similar.
Thanks ! |
|
|
| Calvin |
Hi,
if You are using ESL-panels just from app 200Hz upwards, You may try with standard toroids. In Europe we have 230V/50Hz power supply. A couple of 230V/6V toroid transformers gives an transformation factor of 1:76 (in practise it will be a bit lower, around 1:68 because of losses). I use a couple of 230V/6V on a 50VA-core, connected in series on the 230V-side and parallel on the 6V-side with a panel of 25cmx125cm and 1.2nF. Bandwidth is greater than 20kHz and up till today they work perfectly. You might even get lower in frequency by using trannies with bigger core, but an appropriate panel would become quit huge. You can connect more trannies in series (230V-side, 6V-Side always in parallel) to get higher transformation-factors. A quad of trannies would give app 1:150.
jauu
Calvin |
|
|
| pmscompany |
Thanks, I have a toroidal 220v--->6v (Chile have 220v at 50Hz). My transformer have 2 in and 2 out ??? What is the mode of conection, for ex. http://sound.westhost.com/p105-f4.gif 3 point out, my toroidal transformer have a 2 out. What is the mode of conection to EHT and Stantors.
U have diagram of conection ???
Saludos |
|
|
| Bazukaz |
Hi,
I think it should look like this.
You need two transformers for each panel. Note : between EHT "-" and "membrane" there must be ~20 Megaohm resistor.
I have bought a pair of 80W 220:12V trannies. They start to saturate at input voltage of 13V at 50 Hz. I think that slightly higher crossover point may be better(i am thinking about 250-300Hz , if i split 12V secs into 2x6V) , because it would keep the core saturation lower.
Regards,
Lukas. |
|
|
| pmscompany |
Thanks, I buy in a week a 4 toroidal transformers 220 to 6v have a very good price...
A question, What is a max. power (watts RMS) for speaker ???
Is possible uses a passive crossover for cut a frecuence (in 220 to 440 Hz depending a result). And uses other spekers for compensation mid bass and bass. Total a hibrides speakers 2 ways (ESL for mid and high's and woofer for mid bass and bass).
Or is necesary uses two amps??? Bi-amp ???
Salu2 |
|
|
| Bazukaz |
Hi,
Max. power depends on panel and transformer design.~100W RMS amplifier should be ok.
I think an active filter with at least 12db/oct rollof is better than passive. In this case you would need separate amplifiers for both ESL and dynamic loadspeakers. One more thing : add ~0.7 - 1 ohm resistor between amp's output and transformers.This will make life for you amp a lot easier :).
Regards,
Lukas. |
|
|
| pmscompany |
100 w RMS good power for listening music in a house!!!
I speak via e-mail with Rod Eliot the ESP. He uses a tradicional 100v 20w transformer for PA spk.
http://sound.westhost.com/project1053.htm
The combination of transformer have a ratio 100:1 v/s toroidal 230/6v 1:76, (desgin for a MINI ESL PANELS).... Umm what transformer recomend (toroidal or PA ) for my ESL's and what dimention, recomend for cut my`s grilles and work a spk's.
I buy a chinenses speakers for a woofer. 8u$ for woofer. Good sound, 8" diametre, good magnet... Have a neopren paint cone, for low distortion and water protection. Flex edge foams (spume material)... Good sound for a very low cost !!!
I test a next week in Anecoique camera in my University. For add small factor and responce, For correct calculation the cabinet.
The woofer have solo indications of a sensibility (89 db a1m), 8 ohm impedance and 150 watt max. (mmmmm i test mmmm 70 watts rms max, 130-140 program, 150 peak, hehehee) the horrible is a spec a speker. The chinenses have horrible writing a spec. The spec not correspond a reality.
saludos |
|
|
| Calvin |
Hi,
this is an example what You could do.
The panel is housed in a aluminium profile frame. Below the panel will be the housing for the electronics (the bass You see is omitted). The Bass-tower is a stacked 8-driver dipole bass with 6.5"-drivers.
The crossing over will be done active and 2 amps will be housed in the electronics cabinet too. So this design is fully active.
But You can try to crossover passively equally well.
Here´s something I tried.
A) and B): pure panels response
C) Filter response
D) and E): Panel +Filter response
F): Response of panel + different Filters (pure, Panel+ 50µF, panel +50µF+1mH/1Ohm, panel + 50µF+1mH/5Ohm). The comparison to ML´s Prodigy panel shows two differences. The higher Fs of the Prodigy, which is possible because they use a membrane 4 times as thick as I use and the earlier rolloff at the uper end, where my lighter merane shines. Apart from beeing double as wide as my panel and featuring a thicker membrane, both panels are very similar.
Placing the crossover frequency of the filter right at a point between the panels ground resonance fs (which is ~170Hz in this case) and ~1kHz (where the acoustical short ends) and let it swing over a bit (Fc ~350Hz, high Q) as You can see in the filters response curve, You get a very linear response down to the panels fs with the best possible efficiency. Thats the way MartinLogan did it (comparison ML-Prodigy and my panel). Though this solution looks marvellous on paper and it works in practise, I find that the sonic footprint of the panels resonance is audible. Too the long decay in step response and raising distortion below 300Hz of the panel is of no good. Still You may try this first, as it is a simple solution and squeezes the max out of the panel.
Don´t wonder about the falling frequency response. Measured in nearfield, this response is Your ideal target!
I´d suggest though, that You filter the panel such way:
- linearize Impedance --> use a 15-22Ohms high Wattage resistor in parallel, or a specialized RLC-net.
- use a Notch filter to compensate for the ground resonance
- adjust high-frequency response wth a small valued high wattage resistor
- design the HP filter as 1st or 2nd order with an fc min. double the panels fs(here it is ~550Hz electrically giving -6dB@350Hz acoustically)
It could look like this (keep in mind that the response is measured at a distance of 2m now, not nearfield!)
This filter gives equally high efficiency as the simpler filter, but lessens the stress on the panel, as well as the trannies since its crossover-freq is higher. The panel now only works in a range where its distortion is extremly low (well below 0.1%@95dB) and You reach ear-shattering levels of SPL.
Compared to the first simpler filter the response at distance is smooother and better...You see, how important it is to get rid of the ground resonance. While the response in nearfileld looks perfect it isn´t any more on distance.
Adding to this a comparably good bass is not an easy task, but will end at a state-of-the-art speaker. I strongly recommend to use a dipole bass with such panels.
jauu
Calvin |
|
|
| bear |
It might also pay to read Peter Walker's articles on ESLs - he discusses matching the panel with the transformer and its effect on freq response vs. sensitivity.
Also to review Strickland's patent on the same subject...
_-_-bear :Pawprint: |
|
|
| pmscompany |
| quote: | Originally posted by Calvin
Hi,
0hhh exellent work !!! What transformer used in u ESL, what recomend toroidal 1:76 (u ex.) or ex. the PA 100v transformers 1:100..
What mic u used for test Frecuence responce and what computer program for grafics ????
Saludos y muchas gracias
this is an example what You could do.
The panel is housed in a aluminium profile frame. Below the panel will be the housing for the electronics (the bass You see is omitted). The Bass-tower is a stacked 8-driver dipole bass with 6.5"-drivers.
The crossing over will be done active and 2 amps will be housed in the electronics cabinet too. So this design is fully active.
But You can try to crossover passively equally well.
Here´s something I tried.
A) and B): pure panels response
C) Filter response
D) and E): Panel +Filter response
F): Response of panel + different Filters (pure, Panel+ 50µF, panel +50µF+1mH/1Ohm, panel + 50µF+1mH/5Ohm). The comparison to ML´s Prodigy panel shows two differences. The higher Fs of the Prodigy, which is possible because they use a membrane 4 times as thick as I use and the earlier rolloff at the uper end, where my lighter merane shines. Apart from beeing double as wide as my panel and featuring a thicker membrane, both panels are very similar.
Placing the crossover frequency of the filter right at a point between the panels ground resonance fs (which is ~170Hz in this case) and ~1kHz (where the acoustical short ends) and let it swing over a bit (Fc ~350Hz, high Q) as You can see in the filters response curve, You get a very linear response down to the panels fs with the best possible efficiency. Thats the way MartinLogan did it (comparison ML-Prodigy and my panel). Though this solution looks marvellous on paper and it works in practise, I find that the sonic footprint of the panels resonance is audible. Too the long decay in step response and raising distortion below 300Hz of the panel is of no good. Still You may try this first, as it is a simple solution and squeezes the max out of the panel.
Don´t wonder about the falling frequency response. Measured in nearfield, this response is Your ideal target!
I´d suggest though, that You filter the panel such way:
- linearize Impedance --> use a 15-22Ohms high Wattage resistor in parallel, or a specialized RLC-net.
- use a Notch filter to compensate for the ground resonance
- adjust high-frequency response wth a small valued high wattage resistor
- design the HP filter as 1st or 2nd order with an fc min. double the panels fs(here it is ~550Hz electrically giving -6dB@350Hz acoustically)
It could look like this (keep in mind that the response is measured at a distance of 2m now, not nearfield!)
This filter gives equally high efficiency as the simpler filter, but lessens the stress on the panel, as well as the trannies since its crossover-freq is higher. The panel now only works in a range where its distortion is extremly low (well below 0.1%@95dB) and You reach ear-shattering levels of SPL.
Compared to the first simpler filter the response at distance is smooother and better...You see, how important it is to get rid of the ground resonance. While the response in nearfileld looks perfect it isn´t any more on distance.
Adding to this a comparably good bass is not an easy task, but will end at a state-of-the-art speaker. I strongly recommend to use a dipole bass with such panels.
jauu
Calvin |
|
|
|
| pmscompany |
u have a link the Peter Walker's Article ???
| quote: | Originally posted by bear
It might also pay to read Peter Walker's articles on ESLs - he discusses matching the panel with the transformer and its effect on freq response vs. sensitivity.
Also to review Strickland's patent on the same subject...
_-_-bear :Pawprint: |
|
|
|
| bear |
I have no idea if it is online or not... he was the founder of Quad.
His series of articles are ca 1950s in Wireless World. A British magazine.
I found them in the library of an engineering school. Check ur library system, schools, etc.
_-_-bear :Pawprint: |
|
|
| bear |
Oh, fwiw, the usual trick is to use backwards connected tube output transformers. AC mains xfmrs typically have awful HF response.
You can always test the AC mains xfrmrs, of course.
Caveat Emptor!
_-_-bear :Pawprint: |
|
|
| pmscompany |
OK BEAR, i seach in a libraries...
| quote: | Originally posted by bear
I have no idea if it is online or not... he was the founder of Quad.
His series of articles are ca 1950s in Wireless World. A British magazine.
I found them in the library of an engineering school. Check ur library system, schools, etc.
_-_-bear :Pawprint: |
|
|
|
| pmscompany |
The fture test with a Solid state amp. (my behringer ep2500 is a clon the QSC RMX-2450) i used in events productions. Or test with my old yamaha integrated amp (25x2). In a week i buy a trasformer (is possible i buy a taroidal 1:86 with a respect a 1:100 PA (toroidal a rigide, for low electrical parasite sound, menor ratio ummm).
Calvin a recoment a work a pannels the 125x25 cms umm, good dimentions for my first ESL Speakers !!!
My New's woofer i test in a week in a anechoic camera the my university... My univesity have a very small anechoic camera, the low cut out is a 250 Hz... Real anechoic.
What a uses of the Mylar, what a mylar recomend for buy ???
Saludos
| quote: | Originally posted by bear
Oh, fwiw, the usual trick is to use backwards connected tube output transformers. AC mains xfmrs typically have awful HF response.
You can always test the AC mains xfrmrs, of course.
Caveat Emptor!
_-_-bear :Pawprint: |
|
|
|
| Calvin |
Hi,
toroids work much better than standard cores, but You hardly can wind them Yourself any more. The coupling betwen the windings is very good and they offer low stray-inductance and reasonable interwinding-capacity. They are rather low on the inductance scale, but that doesn´t count as long as You´re desired freqency range is well above 100Hz. The two parameters which are define the lower and upper frequency-range in first instance are primary inductance (low end) and stray-inductance (upper end). If You need to go low use bigger cores (100VA), wheras 30VA to 50VA is normally enough. I use a pair of 50VA 230V/6V types and got -apart from slightly less transformation factor- very similar results to an costly but specialized Amplimo 1:75 100VA . Sonically I couldn´t hear any difference at all. You can have toroids wound for Your purposes as well as with standard cores ifYou need higher primary inductance and lower interwinding capacity, but even with my highly capacitive panels the bandwidth of the standard toroids is above 20kHz.
jauu
Calvin |
|
|
| pmscompany |
| NO NO NO !!!!! Not toroides in stores.... What other transformer recomend ??? 0 possibilites of work a toroides !!! |
|
|
| Calvin |
Hi,
if You can´t get Your fingers on toroids, I´m afraid You´ll have to take specialized and therefore costly audio-trannies.
Those use double C, or EI-cores. There are many winders around, but not many that wind a good ESL-trannie.
This is a comparison of a standard(!) power toroid against a quad of connected serial-parallel connected audio trannies.
As You can easily see the toroid outperforms the EI-type. The EI-type doesn´t even reach 20kHz bandwidth with the 1.2nF of the panel, whereas even with 2 panels in parallel (2.4nF) the toroid reaches well beyond 20kHz.
Since I became suspicious about those ´specialized´ trannies I tested a sixpac of standard EI-type power trannies and got exactly the same results but for a much lower price! :mad:
Paralleling the trannies helps in reducing the stray inductance, therefore extending bandwidth. Still so, I´d use EI-types just for panels with low capacitance. For highly capacitive panels toroids are a easier, better and cheaper way.
jauu
Calvin |
|
|
| Bazukaz |
Hi,
I would like to ask what DC current should be considered high for these toroids. The amplifier i have has a DC offset. With a 0,7 ohm resistor in series with toroid's secondary(12v) there is 20 mA of DC current. Is it too much for a toroid ?
Regards,
Lukas. |
|
|
| I_Forgot |
If your amp has a DC offset it is broken and should be fixed instead of designing speakers around the offset.
I_F |
|
|
| Bazukaz |
Well , the DC offset is not large( 28 mV). It is because the amplifier has no DC blocking caps (a DIY gainclone). For standard speakers , 28mV offset has no effect , but with a low-resistance transformer , this may cause some problems. Just i was interested if toroids care about 20-30mA DC current.
Regards,
Lukas. |
|
|
| Capaciti |
Hi bazukaz,
you adressed in issue which is becoming quite important.
1. on standard Class-A, AB or B amps it isnt an issue, since DC up to 0,1 Volt wont increase saturation problems.
2. But in our days Class-D is becoming more interesting, especially for ESL, since they perfectly match the critical load of an ESL. Class-D amps show DC up to 1 Volt and this is critical for toroids having no air gap in the core material.
3. I preffered toroids a long time, but now i will re-investigate other transformers to avoid saturation, if connected to Class-D
4. saturation is very critical, since saturated transformer might induce twice or even tripple voltage in the secondaries, which might "kill" your insulation of stators immediately.
regards, Capaciti |
|
|
| Calvin |
Hi,
You might do that with just a pair of toroids, but You won´t get good results with standard EI- or UI-core trannies.
You will probabely need more trannies. Connecting 6 to 8 of these can be sufficient for low capacitance panels only.
I´d really try to get my hands on toroids.
jauu
Calvin |
|
|
| pmscompany |
I dissambled my audio Compressor (behringer 4600 http://www.behringer.com/MDX4600/index.cfm?lang=ENG ).
In the interior have a toroidal tranformer (220, 110 to 22 volt at 400mA)... I contact with a technical services the behringer, for $$$ cost of transformer, or disponibility)... Is possible a used a two 22 volt toroidal transformer ???
- Saludos |
|
|
| Bazukaz |
Hi,
You can use 22:220 V transformer as well. However , this would only yield a 1:20 step-up ratio with 2 transformers.You could split 22V secondary into 4x5.5V secondaries and connect them in paralell.This would give approx 1:80 transform ratio.
Regards,
Lukas. |
|
|
| I_Forgot |
Any unpotted toroid power transformer is very easy to modify for use in driving an ESL. If the secondary voltage (the low voltage winding) is too high you can either remove turns or remove the whole winding completely and replace it with one of your own.
How do you know how many turns to use on that new low voltage winding? Wind a few turns on the transformer, then plug the HV side into the power line and measure the voltage on the newly added winding. Too low, add a few more turns and measure again or just calculate the volts per turn and then add the exact number of turns needed for the desired voltage change ratio.
I_F |
|
|
| SillyCone |
I dont know if the following have been discussed before.
That is, to build a dedicated Esl amplifier with ordinary output
transistors or fets but including the stepup transformer
in the feedback circuit?
This should eliminate much of the distorsion and nonlinearity
produced by the stepup transformer.
Or? |
|
|
| Capaciti |
Hi Sillycone,
i do not know, which stepups you utilize, but an appropriate designed one shows distortions 10 times lower than the ESL membran itself.
@ pmscompany
There is nothing to say against the efforts to get cheap toroid power trannies running at ESL's, but nobody should wonder if those "playtools" distort and sound like a piezo-tweeter.
capaciti |
|
|
| SillyCone |
Hello Capaciti,
It must be an advantage to include the stepup transformer
in the feedback chain, as you are not only reducing
distortion, you are also extending bandwidth in both directions.
Also, if a toroid power transformer is used as stepup in this Esl
dedicated Amp, maybe the grainy and harsh elements in the sound emerging from the cheap iron core can be reduced or
even eliminated.
I have not seen this idea realized anywhere, so there's surely
a catch that I haven't discovered.
I about to experiment with my old Nad amplifier and a pair
of 230:6 volt toroids to see what happens. |
|
|
| Bazukaz |
Hi,
The high-frequency limit of power toroids is usually > 20kHz.
Including it in feedback chain would not allow to improve the low frequency bandwidth anyway.
Cheap power toroids can work well , i think , but there may be a big difference between manufacturers.They seem to saturate quite easily.
As far as i know, transformer's distortion is dependant upon drive source impedance.So , by making the source impedance the lowest possible , we should lower trafo's distorion as well. Am i right ?
Regards,
Lukas. |
|
|
| bear |
| quote: | Originally posted by I_Forgot
Any unpotted toroid power transformer is very easy to modify for use in driving an ESL. If the secondary voltage (the low voltage winding) is too high you can either remove turns or remove the whole winding completely and replace it with one of your own.
How do you know how many turns to use on that new low voltage winding? Wind a few turns on the transformer, then plug the HV side into the power line and measure the voltage on the newly added winding. Too low, add a few more turns and measure again or just calculate the volts per turn and then add the exact number of turns needed for the desired voltage change ratio.
I_F |
oh geez...
Do NOT PLUG THE TRANFORMER UNDER TEST INTO THE AC MAINS!!!
Is it necessary to spell this out?
The alternative is to use, for example and ease of calculation, a 10volt low current transformer as the voltage source to test unknown transformers. A stupid AC wall wart will do, be it 6, 9 or 12 volts AC type - use a calculator to determine the ratio.
_-_-bear :Pawprint: |
|
|
| bear |
| quote: | Originally posted by Bazukaz
Hi,
The high-frequency limit of power toroids is usually > 20kHz.
Including it in feedback chain would not allow to improve the low frequency bandwidth anyway.
Cheap power toroids can work well , i think , but there may be a big difference between manufacturers.They seem to saturate quite easily.
As far as i know, transformer's distortion is dependant upon drive source impedance.So , by making the source impedance the lowest possible , we should lower trafo's distorion as well. Am i right ?
Regards,
Lukas. |
Wrapping a transformer with feedback will flatten the frequency response at the expense of transient response, not a good trade off.
This is what many tube amps were doing, and the result is/was a better freq response spec, but horrid transient response (look at what they do to a square wave). The bad transient response that results is quite audible.
The smaller the power toroid physically, the better the chance it will have something of a decent HF response. The power toroids that I have measured - as small as 225VA class have had horrid response as an audio transformer. Not something I would consider using, ever. Imho, I would not want to use a transformer for audio with a response that only went out to 20kHz. in the first place. Maybe for initial testing, in a pinch.
Use an audio output transformer run in reverse... even a stupid generic guitar amp output transformer will likely outperform any power transformer.
_-_-bear :Pawprint: |
|
|
| Bazukaz |
Hi,
I have made some measurements a few months ago with 250W toroid.Frequency response was > 20kHz. Better than that of an old tube amp. THD was less than 0.1% in the audio band.
I think different trannies may behave different.The steel for toroids is usually laminated grain oriented silicon stell. The problem is that these transformers saturate very easy even from small DC offsets.
Regards,
Lukas. |
|
|
| Bazukaz |
Hi,
As far as i know , the distortion , caused by step-up transformer , is dependant on source impedance.Adding a passive crossover in series with transformer would increase source impedance. Are passive filters used in commercial products , and what are their effects on distortion ?
Lukas. |
|
|
| maudio |
| quote: | Originally posted by Bazukaz
Hi,
As far as i know , the distortion , caused by step-up transformer , is dependant on source impedance.Adding a passive crossover in series with transformer would increase source impedance. Are passive filters used in commercial products , and what are their effects on distortion ?
Lukas. |
Yes they are, I know at least Quad, ML and Audiostatic use some sort of passive network on the primairy side of the trannies, so do most DIY esl's I've seen.
Purpose is a) equalization and b) smoothing the impedance curves, which helps reducing amplifier distortion (not transformer distortion).
Main distortion problems with esl's IMHO are in the amplifier struggling to drive a load that often drops below 1 ohm or so @ 20k :clown: All amps show a raise in distortion when load imps drop. I think this is the main distortion source, distortion in the transformer itself is very low and probably totally neglectible. At least as long as you stay well below core saturation point. Increasing source impedance might increase transformer distortion theoretically but I don't think it will be of any significance, the improvement in amplifier performance will easily compensate for it.
Adding some resistance (either on prim or sec side) is mandatory anyway to dampen the transformer series resonance which is (hopefully) just above 20 k in most fullrange trannies. This peak, when not taken care of, can easily reach 10 or 20 dB and might cause isolation failure in the transformer.
The main problem with including the transformer in the feedback loop is the amount of phaseshift it causes. Therefore it is virtually impossible to place any decent amount feedback over the trannie without causing instability. Possible way around it is using nested feedback loops, I am experimenting with that now.
Even in such a design a correction network is needed to relieve the output stage from driving very low imps, the distortion caused by that can't be cured by feedback! |
|
|
| maudio |
| quote: | Originally posted by Bazukaz
Hi,
I have made some measurements a few months ago with 250W toroid.Frequency response was > 20kHz. Better than that of an old tube amp. THD was less than 0.1% in the audio band.
I think different trannies may behave different.The steel for toroids is usually laminated grain oriented silicon stell. The problem is that these transformers saturate very easy even from small DC offsets.
Regards,
Lukas. |
Did you measure THD at the output of the toroid or did you measure the difference in THD between the signals on the primary and secondary? Only the last method gives you transformer distortion in isolation. |
|
|
| Bazukaz |
Hi,
Thanks for help. Now i know that i'll go with a passive crossover :).
I am planning to cut the ESL at approx 250Hz. The filter gives some boost at ~ 400-800 Hz to partially compensate for dipole cancellation .
Now one question about woofers : i've read that many recommend dipole woofer arrays. However , i have no space in my living room to place such a thing :). How about sealed/ported boxes ? I have some experience with sealed subwoofers , and like sound of them.
Here is an interesting article :
http://www.musicanddesign.com/roomgain2.html
As i understand , it states that in-room response of dipole is not very different from monopole , except that monopole would compress the air below room's fundamental , while with dipole , this is completely opposite.
Regards,
Lukas. |
|
|
| zenmasterbrian |
I just searched and found this thread, because I wanted to show a neat web site I found, Martin Logan, in Lawrence Kansas.
http://www.us.martinlogan.com/
I have no experience with electrostatic speakers. But they are of interest to me.
I've never been real keen about something that has to be driven by a transformer.
I've imagined the possibility of a high voltage amp that could direct drive and electrostatic speaker.
This might be more feasible if it were MF/HF only, or even HF only?
Such an amp might be solid state. But maybe it would use horizontal output tubes, or radio transmitter tubes, like 6146B??
Just an idea.
What do you think of Martin Logan? |
|
|
| bear |
Zenmasterbrain: this is a time when the advice to search the forum is good. There have been a number of threads on direct drive ESLs and ESL amps... as well as prior discussions of ML speakers.
ML is well known.
They have some issues, not so good LF response, and a very low Z being two. But they sound pretty good.
Bazukaz: no matter what you do with cone speakers they will not sound like an ESL. The main thing with respect to a woofer , and you want a woofer to mate with ur ESL at 250Hz., imho, is to get it to be as low distortion as you can find, and to get it out acoustically as fast as possible where it meets the ESL... that and the Q will be all that matters. Bi-amping may help too, since you can use an amp that has subjective characteristics that may be more appropriate, and select a DF that works well.
A quick look at the article you cited was interesting, but unclear to me... since ESLs need "air" behind them, meaning that they are best typically 3-6 ft from the rear walls, and the rear walls play a role in the ultimate sound too, the usual placement of subs (or woofers) or for that matter speakers is not feasable! In my case I found that it is best to have the acoustic center of the subs slightly in front of my ESLs for best overall system impulse response.
I suppose that dipole woofers have their advocates, but the issue that I see with dipole woofers is that they inevitably need LF boost EQ'ing, and that means both very large amps, and usually extra drivers to increase Sd and output... I think I prefer to deal with extra room gain by reducing (cutting) EQ than the converse.
... but again for an ESL at 250Hz rollover, it will be tough imho to get good midbass sound and at the same time excellent LF extension. Somehow these two things seem to be mutually exclusive when it comes to woofers! (bummer, eh?) So figure on a separate sub if you want to maximize the LF and still get excellent midbass to merge with the ESL...
As always, merely my opinions...
_-_-bear :Pawprint:
PS. of course, if ur in a relatively small room, the issues change yet again... |
|
|
| zenmasterbrian |
ML makes lots of wall mounted speakers, for surrounds.
Direct drive could be done a number of different ways.
I notice that ML makes two center speakers for home theater. But the don't use their electrostatics for the tweeter. They use a 1" dynamic dome driver.
Electrostatic must have issues with the large drive panels. This has always seemed curious to me. |
|
|
| maudio |
| quote: | Originally posted by bear
no matter what you do with cone speakers they will not sound like an ESL. |
Fully agree on that, the only exception in my opinion being dipole woofers crossed over at 100 hz or below.
| quote: | | I suppose that dipole woofers have their advocates, but the issue that I see with dipole woofers is that they inevitably need LF boost EQ'ing, and that means both very large amps, and usually extra drivers to increase Sd and output... |
I guess I'm one of those ;) If you are willing to spend the time, money and effort to build a good dipole woofer design *and* build your esl panel large enough to allow crossover below 100 hz, you won't be dissapointed.
You'll need biamping + active crossovers including eq for the woofer, a notchfilter for the esl's resonance + probably some eq for the esl. Amps for the woofer don's have to be that big, 70W is plenty for my own dipoles.
About direct drive: After playing with it for a few years I believe it's only practical for panels in the 1000hz and above range, read for instance this or this tread for some more detail and there's plenty more to find. The bottom line: There is a good reason that almost nobody uses dd amps :D
Since crossover frequencies of 1k are utterly useless for combining esl's with dynamic drivers, my advice is: use transformers, they are not as bad as you think, it's often the rest of the chain giving the real headaches
Just my ideas... |
|
|
| zenmasterbrian |
maudio, thanks for finding those threads!
I'm impressed to see such here.
Of course I do understand that it would only be above some minimum frequency.
If that could be 300hz, not 1khz, I would say it could be a go.
How about the 6146B tubes. They are old horizontal output tubes, but they were mostly used for HAM radio transmitters.
They might be able to get the voltage higher.
Notice that ML makes wall mount speakers. Not much air behind them.
Notice that on their two center channel home theater speakers they don't use the esl panel for the tweeter.
There must be phasing or beaming issues in esl speakers because of the large panels.
But suppose you just had a real tiny panel for a tweeter? Maybe that could be direct driven with high voltage devices. |
|
|
| maudio |
| quote: | Originally posted by zenmasterbrian
[B]maudio, thanks for finding those threads!
I'm impressed to see such here.
Of course I do understand that it would only be above some minimum frequency.
If that could be 300hz, not 1khz, I would say it could be a go. |
Might be possible if you manage to increase the sensitivity of your panels. That would require very small membrame-to-stator spacing (< 1.5 mm), very high membrame tension (raising resonance!) and high bias voltage, requiring well insulated wire stators.
Small spacing will also mean reduction in air displacement = need for higher crossover point and/or steeper filtering. Not a good trade-off I think.
| quote: | How about the 6146B tubes. They are old horizontal output tubes, but they were mostly used for HAM radio transmitters.
They might be able to get the voltage higher. |
Be it tubes, fets or bjts, the problem is that with neither of those you can build quality high voltage amps without running into serious dissipation problems above a kV or 2 tt output swing. You'll need current as well :cannotbe:
I truly believe it's not possible to build a good sounding ESL-woofer system with crossovers in the 200-300 hz range. I heard several such designs and none of those could convince me. Lowering crossover point to the 100 hz range is the only way to go.
Still, that's only my experience, if you're willing to compromise here and there you might end up with something you actually like. Don't let other keep you from experimenting;) |
|
|
| zenmasterbrian |
Using 2 6146B tubes, one could probably do 2kv p-p. Maybe one could also have some kind of impedance matching LC circuit that would raise the voltage at the lower frequency end.
Just an idea.
Any ideas why Martin Logan doesn't use the electrostatic unit for the tweeter on their two center channel speakers? Why do they opt for a dome dynamic tweeter?
Why to they opt for the opposite with their Left and Right tower speakers and surround speakers? |
|
|
| Capaciti |
Hi,
the membrane area of the centers is too small to achieve the required efficiency over the entire frequency range. So they use a dome tweeter. Since the ESL isn't used for high frequencies, stepup ratio can be selcted rather high,leading to high efficiency.
Capaciti |
|
|
| Bazukaz |
Well , for a tweeter , you could even probably use a core-less transformer , if you don't like tradditional transformers.
Why do use a direct drive amplifier ? It is difficult to build ;
it dissipates a lot of heat ; is more dangerous than a typical transformer configuration because signal is not separated from ground , etc.
IMO the only practical use of direct drive amplifier is to drive ESL headphones, where above problems do not airise.
Regards,
Lukas. |
|
|
| stokessd |
If you folks want to roll your own Toroid input transformers, you can get pre-wound primaries (secondaries for this application) and you wind your own secondary (primary) in place. I've used these to make multi-tap power transformers for other projects but never tried them as either input or output transformers. I think you may need a special order one with dual primaries, but the price is pretty cheap:
http://www.toroid.com/standard_tran...former_kits.htm
Sheldon |
|
|
|
|
|
