First ESLs, any advice?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Been reading a lot about esls recently and I'm considering building a pair of small tweeter panels to start with. I've got a few ideas I'd like to bounce around, bare with me as I'm thinking about trying things out without spending much/any cash... :D

Thinking about an A4 size or less panel, any advice on shape? I notice people seem to use really rather thin/tall sections of large full range panels for "mid/top", this is for better dispersion characteristics I understand, is there any rule of thumb for this? Target usable frequency range is (I guess) 2k and up, but if I could go lower (and what affects how low a panel this small is useful for) would be interesting for testing purposes.

Things I've found around the house to get the proverbial juices going:

- Cooking/roasting 'bags'. Have a vague memory of reading about some people using them for diaphragm material, have taken the digital callipers to them and I make them 12.5um (16 sheets come in at 0.2mm). They are polyester, probably already heat treated in some way (they'r for cooking things in) no heat shrinking to tension is possible. Is this to thick for a treble panel? Any UK members have a source of mylar/hostaphan?

- Looking at the front of my microwave I noticed that the grill that allows you to see in but completes the faraday cage is of what I assume to be a usable hole size/spacing... any thoughts? I can get hold of discarded microwaves with relative ease (gotta love local dumps), which also must contain step-up transformers, something to think about.

- Have some quite fine (for what it is) chicken wire mesh, it has a hole size of roughly 5.5mm and a wire diameter of about 0.65mm (in a square lattice if your unfamiliar with the term). I assume this is far to open for my purposes but would a mesh screen of different dimentions be suitable, anyone made/got reference to a panel with similar stators? I've found a relatively cheap source of mesh with a 2.5mm spacing with a 0.7mm wire diameter (60% open).

- I have 6 mains toroidal transformers (cheap from maplin years ago, old stock, they were not cheap to start with) that are single 230V primary, 12V-0-12V secondary, 50VA. Cant find any specs on them with google. Using two of these 'backwards' would give me a step up ratio of around 1:40 correct? Are these usable for my application?

- I have a roll of 3M '9485' double sided tape type adhesive lying around, found a recommendation for it on 'The audio circuit' esl diy pages, anyone have experience with this/any other type of 'tape' adhesive for fixing diaphragms?

Any input would be appreciated!
Thanks,
Theo
 
On the Kings Sound page scan down to the Hummingbird tweeter fotr size performance idea. A 70 volt line transformer wired backwards will runnthese that all they use (size wise).
As for diaphragm material you can contact Ray harlan and purchase 1.45 micron Hostaphan RE or 0.9 micton DuPont Mylar C for SOTA diaphragms. You can probably find 1.5 micron polyester in film/foil capacitors just find one the right width and cut the ends off and unwind the cap.
You can obtain diaphragm coating from Matts at MT Audio.
For stators I would use double build 30 gage magnet wire check out Superior Essex or use Kynar wire wrap wire on some suitable frame.
Scotch Grip #4693 is a 100% gripping contact adhesive that will never slip or let go. You might also experiment with locally available water based contact adhesives. If you decide to use your transfer tape remember that heat and pressure are required if you want it to set up quickly and not allow the diaphragm to creep over time.
I would suggest that you spend the money and buy the proper parts to build your tweeters. This will result in a better chance of a positive outcome to your project. Your time is worth money and a negative outcome is not very insperational to a second attempt.
Hope this helps and is of some use.



http://www.kingsaudio.com.hk/products.html#hummingbird
http://www.indoorspecialties.com/
http://user.tninet.se/~vhw129w/mt_audio_design/
http://www.superioressex.com/upload...tion/Europe/Essex_Nexans/round data sheet.pdf
 
<i>Thinking about an A4 size or less panel, any advice on shape? </i>

Yes, my advice is to play with the ESL simulator and see what panel size does to frequency response and dispersion. You can play a bit with the dimensions of the panel and the off-axis position as well.

High and small is a good combination because you have the improved dispersion of a small panel and the larger diaphragm area of a large panel. Of course vertical dispersion is not so good.
 
A4 ESL panel why not!!

I encourage you to try, my prototype for testing purpose is about 8 inchs wide by 4 inchs tall, and i assure you that the sound is great.

I test frequency response with a tone generator and i can produce loud 400 Hz with it. So if you plan to use it as a tweeter i'm sure you can us the grill of microwave succes fully. I dont recommend strech grill since there are to many sharp edge, and DS spacing will not be consistent. For diaphragm you can use windows treatment film for isolation. For coating i use graphite powder, in very small quantity.

You can use double tack foam tape for spacer but it wont hold i tension of the film and it will shear if the film has not something to hold to.
You can use a ridgid frame to tape the film to or simply put adhesive tape that come with the windows treatment on the other side of the grill on the edge and you wrap the grill so tension will be transfer to the ridgid grill, make sure you sand the edge of the grill so the film will not be cut.
For the transformers since you have toroidal type i suggest that you buy some fine bobbin wire and bobbin your primary to have better ratio 1/70.
Make somes experiments it's fun.
Bye;)
 
Many thanks, excellent links there moray james, I particularly like the mylar supplier, 10ft of 1.5um film for $9... wonder what shipping is like.

arend-jan, brilliant simulator, didnt realise one existed! Not sure I understand the two lines. I dont seem to be able to get any kind of proper freq response in the 'voltage drive' line only 'current drive', I thought almost all amplifiers were voltage drive, that current driven amps could be constructed but are no means common place....

Mandrake5, very encouraging to hear that tiny panels can sound good and go that low... What materials did you use to construct it, this 'windows treatment' stuff you mentioned I guess, do you have a link? not a product I've ever heard of before.
 
check out your local UK suppliers/distributors....

of model airpland building supplies. They use thin film down to about two micron for ultralite model skins. Postage will be minimum as the small cardboard roll it will be shipped in is what you are paying for. You can experiment with 5 to 1 dilute PVA adhesive (white glue for wood) for diaphragm coating or dilute acrylic floor polish. Both these can be misted onto the diaphragm exactly where you want it with a small cosmetic style atomiser bottle (perfume).
you can experiment with the plastic bags they provide for produce at the store but the real deal is the best and the most stable. The two films that Harlan sells are the very best films that money can buy (DuPont Mylar C and Hostaphan RE) and you can get them in small quantity. Perfect for a tweeter, I would go for the 0,9 micron DuPont because it will give you a wider bandwidth. You won't hear the high stuff but you will catch downward harmonics. If you build a small simple stretching frame like the one that Sheldon Stokes shows on his web page you can build it large enough so you can stretch one piece of film of a size to make both tweeters at the same time. This way both tweeters will have identical resonant frequencies.
 
Theo404 said:


arend-jan, brilliant simulator, didnt realise one existed! Not sure I understand the two lines. I dont seem to be able to get any kind of proper freq response in the 'voltage drive' line only 'current drive', I thought almost all amplifiers were voltage drive, that current driven amps could be constructed but are no means common place....

You can create current drive by adding a large resistor in series with the stators (play with series resistor value). Most panels are in between voltage drive and current drive. They use resistors to segment the panel and shape directivity and frequency response.

At a large distance, a voltage driven ESL will have a rising response of 6db/octave with frequency. This is because the impedance is capacitive, hence current increases with frequency. But only if the source driving it is a perfect voltage source. In the real world there are a number of factors that can deminish this effect: diaphragm mass, behaviour of the step-up transformer and close(r) distance to the speaker. A current driven panel will have a theoretical flat response.

Take for example a panel of 60cm high, 4cm wide at 3meters. This looks pretty nice. With segmentation and perhaps some frequency response shaping in the drive circuitry you should be able to get a nice flat response from the voltage drive.

If you make a speaker (almost) has heigh as the living room then you speak of a line source. Sound waves can now only expand horizontally and not vertically. This changes the behaviour to a slope of +3db/oct for voltage drive and -3db/oct for current drive. You can simulate the behaviour by making the panel really high, say10 meters (it will take a while to calculate). But the interference pattern will be for a 10m high speaker not for a 2m.

Really small panels have very low capacitance, which means you need very high voltage to get any current trough them. So you need a higher step-up ratio.

Final remark is that this is all theory. In the real world there are a lot of factors that change the behaviour, like resonances, diaphragm mass, step-up transformer, losses, the room etc etc. The list is long.

For a tweeter, I would definately go with high and small, something like 60cm x 5cm or so. It's a proven concept.
 
As an exemple Cadence Avita...

Hi Theo404
Here is an exemple of what a small panel can do as a tweeter.
The crossover frequency is about 1.8 kHz like a standard 2 way
system. But i'm telling you that you will have proper bass frenquency as low as 400Hz. And if you use panels of the dimension of microwave door, it sure is possible.

The windows treatment that i mention is a film used to isolate thermicaly in winter period windows of different size, even patio size.

Here in Canada temperature drop well below minus 20 or 30 Celcius, so all good hardware store sells those films, in flat boxes about 10 dollars depending of how many windows you can treat. There is a roll of double side tack tape very thin film that glue the film to the frame of the window. After you use a conventional hair dryer to heat the film to a drum tight tension film, nearly invisible film that trap air between the inside side of the window and the interior air mass. Tickness is about 6 to 12 microns so it's similar to some Hostaphan film tickness, that represent very tough film with a little less high frequency but it's not a problem. You can get graphite in hardware store to as lock lubricants.
Bye
 

Attachments

  • avita cadence.jpg
    avita cadence.jpg
    30.9 KB · Views: 645
arend-jan said:



At a large distance, a voltage driven ESL will have a rising response of 6db/octave with frequency. This is because the impedance is capacitive, hence current increases with frequency.

If you make a speaker (almost) has heigh as the living room then you speak of a line source. Sound waves can now only expand horizontally and not vertically. This changes the behaviour to a slope of +3db/oct for voltage drive and -3db/oct for current drive.



Hi


According theory, the excursions of an esl are lineair to voltage. So a headphone ESL should be perfect.
I assume the rising in output of ESLs in a room is only caused by the way sounds travels from ESL to your ears (e.g. beaming, interference) and has nothing to do with electrical effects.

Am I right here?
 
I think it has more to do with the front wave of the panel meeting with the back wave and squishing around the sides and cancelling, which happens less as frequency increases, up to a frequency determined by panel dimensions.

I didn't know a line source changed the frequency response rolloff. Isn't the "3dB per" a 3dB per each doubling of listening distance from the speaker? Or both?
 
MJ Dijkstra said:



Hi


According theory, the excursions of an esl are lineair to voltage. So a headphone ESL should be perfect.
I assume the rising in output of ESLs in a room is only caused by the way sounds travels from ESL to your ears (e.g. beaming, interference) and has nothing to do with electrical effects.

Am I right here?

You are 100% correct.

This is also nicely shown by the simulator when you enter a distance of 0.01m.
 
Hi Theo404,

just some thoughts...

try to get the low cutoff frequency as low as possible; a lot of the magic in music goes on around 1kHz (voices etc.).

When your panel has a low aspect ratio it may help to support the diaphragm with a few dots of silicone in the middle.

Keep your first prototype simple to avoid disappointments and eventual cancellation of the project... Once you have that first one working you can try harder designs with renewed motivation :)

Be very careful with microwave oven transformers, they can easily kill you! Besides their transformation ratio is probably too low for ESL use.

Have fun,
-Kenneth
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.