hello everybody !
sorry to be so obsessional about waveguides
my idea of project about a fullrange in a waveguide is still going on, i've ordered a couple of 3" TB driver to do some tests with them. but i'm not yet convinced by this solution, i would like to do a more classical three way, but waveguiding the medium is causing design problems
this is why i would like to do a waveguide like this one : Genelec 11034b
and i'm wondering about how i should build such a waveguide. The organical shape of it is pretty complex. I'm ready to do things like fiberglass and such, but would like to keep the thing totally DIY.
help about design and building techniques would be much appreciated
sorry to be so obsessional about waveguides
my idea of project about a fullrange in a waveguide is still going on, i've ordered a couple of 3" TB driver to do some tests with them. but i'm not yet convinced by this solution, i would like to do a more classical three way, but waveguiding the medium is causing design problems
this is why i would like to do a waveguide like this one : Genelec 11034b
and i'm wondering about how i should build such a waveguide. The organical shape of it is pretty complex. I'm ready to do things like fiberglass and such, but would like to keep the thing totally DIY.
help about design and building techniques would be much appreciated
you could cheat and buy one form genelec (just the waveguide casting as a spare part, not the whole monitor) and then take a mould from it, the make a fibreglass copy from that mould.
failing that I think you'll need to make both waveguides individually, take negative moulds of them and then hack them together and smooth the transition by hand. then make fibreglass waveguides from that mould/plug.
failing that I think you'll need to make both waveguides individually, take negative moulds of them and then hack them together and smooth the transition by hand. then make fibreglass waveguides from that mould/plug.
I would probably do the math and generate an AutoCad 3D output file that I could take to a CNC shop to route out of solid surface Corian counter top material. Earl Geddes has published several papers on waveguides as well as a book with the basic math.
Most full range, or almost full range designs favor horn loading. Tracix horns often win for best sound. A waveguide will not extend the fequency range or efficiency as much as the horn loading.
Most full range, or almost full range designs favor horn loading. Tracix horns often win for best sound. A waveguide will not extend the fequency range or efficiency as much as the horn loading.
Thanks for the link to the Genelec w/waveguide; very slick.
The polar diagram
http://www.genelec.com/pdf/OM1034b.pdf
doesn't look as consistent as I'd like to see.
The polar diagram
http://www.genelec.com/pdf/OM1034b.pdf
doesn't look as consistent as I'd like to see.
Hi guys
For what its worth, the every first one of these below was made using burlap to make the “waveguide”.
Perhaps you can use spandex to do the same thing.
Stretch the fabric across a framework, uniformly stretched a bit.
With something round, the size you need, press in the fabric to the depth you need and clamp in position.
Get some thin epoxy resin from the hobby shop and brush on a few coats until the fabric is ridged enough to deal with. Fill in the rear area with “great stuff” or equivalent polyurethane expanding foam.
Be sure to drill a number of big holes on the back side to let the excess out as it expands.
Figure it will take a couple trys until you get the hang of it but then it will be easy once you figure out what you need for your application.
Good luck
Tom Danley
http://www.danleysoundlabs.com/SH100.html
For what its worth, the every first one of these below was made using burlap to make the “waveguide”.
Perhaps you can use spandex to do the same thing.
Stretch the fabric across a framework, uniformly stretched a bit.
With something round, the size you need, press in the fabric to the depth you need and clamp in position.
Get some thin epoxy resin from the hobby shop and brush on a few coats until the fabric is ridged enough to deal with. Fill in the rear area with “great stuff” or equivalent polyurethane expanding foam.
Be sure to drill a number of big holes on the back side to let the excess out as it expands.
Figure it will take a couple trys until you get the hang of it but then it will be easy once you figure out what you need for your application.
Good luck
Tom Danley
http://www.danleysoundlabs.com/SH100.html
hello again !
puggie :this spare part is a big idea, i'm gonna call genelec, but i suspect the price is gonna be high, if there ever is a price, i don't know how they do that kind of stuff
about making two waveguides and fitting them, i have thought about it , but it seems to me like a very time consuming method.
linesource: naturally this would be the best solution . But now i have no idea such a thing would cost (cnc shop, provided i can handle the CAD). do you have experience with cnc shops ? Also the calculation i can't handle. Theses are not basic waveguides, they are oval shaped, shall i say , 2/3 of two oval shaped waveguides, i have no idea about how to calculate that, and at least the papers of geddes i found on the subject on the internet can't help me with that. Guess i will have to build and measure, and that is why i fear the CNC
noah kats: you are welcome
what do you mean by consistent ?
the genelec waveguides appears to be particularly shallow, maybe this is what you are refering to ?
Tom Danley That is a solution that looks very interresting. I could do the framework with cutted wood. how did you do yours if i can ask ?
thanks you very much anyway, i had a similar idea in mind, about doing a framework, but was stuck into the material i should drop on it. I first thought bout paper mache
puggie :this spare part is a big idea, i'm gonna call genelec, but i suspect the price is gonna be high, if there ever is a price, i don't know how they do that kind of stuff
about making two waveguides and fitting them, i have thought about it , but it seems to me like a very time consuming method.
linesource: naturally this would be the best solution . But now i have no idea such a thing would cost (cnc shop, provided i can handle the CAD). do you have experience with cnc shops ? Also the calculation i can't handle. Theses are not basic waveguides, they are oval shaped, shall i say , 2/3 of two oval shaped waveguides, i have no idea about how to calculate that, and at least the papers of geddes i found on the subject on the internet can't help me with that. Guess i will have to build and measure, and that is why i fear the CNC
noah kats: you are welcome
what do you mean by consistent ?the genelec waveguides appears to be particularly shallow, maybe this is what you are refering to ?
Tom Danley That is a solution that looks very interresting. I could do the framework with cutted wood. how did you do yours if i can ask ?
thanks you very much anyway, i had a similar idea in mind, about doing a framework, but was stuck into the material i should drop on it. I first thought bout paper mache
Hi Tom,
Very slick prototyping method!
Can you tell me what coax you use for those?
Is it suitable for high quality HT? I want to make surround speakers with cylindrical enclosures with an 8" coax on the front and an 8" PR on the back.
[edit] I just noticed there's a 2nd page on the spec's. That's pretty darn smooth response.
Do you think filling the waveguide with 30 ppi foam a la Geddes would smooth out those jaggies? They seem like they might be just a bit beyond insignificant.
Thanks
Very slick prototyping method!
Can you tell me what coax you use for those?
Is it suitable for high quality HT? I want to make surround speakers with cylindrical enclosures with an 8" coax on the front and an 8" PR on the back.
[edit] I just noticed there's a 2nd page on the spec's. That's pretty darn smooth response.
Do you think filling the waveguide with 30 ppi foam a la Geddes would smooth out those jaggies? They seem like they might be just a bit beyond insignificant.
Thanks
Hi Nemophyle, Noah
I knew what shape I needed on the outside (square), I knew I needed ports in the corners.
I cut out the center of a board, with a broad radius in the corners, roughly the shape of the flat part of the horn in the photo. I stretched the burlap across the opening and stapled it down on the back sides.
I pressed in the center of the cloth with an aluminum tube. Initially, it was too taught to stretch as far as I needed so I got it wet. After it was in position and dry, I painted on a couple coats of epoxy to make it stiff. After it was hard, I took the cylinder out and cut the opening.
I would try Spandex, it stretches a great deal, easier than burlap (in one direction a little more than the other) and has a nice smooth surface. I would try the same thing, use the built in tension in the fabric to form the curves you need. Then, stiffen it with epoxy into a solid part you can sand and paint.
Good luck, don’t be afraid to do a trial one (I made two) first to get the feel of the fabric, its cheap!
Noah, it would be wrong to tell you what driver it is although there is an Italian company with a short name that makes one VERY similar.
With any coax driver, there is a discontinuity between the compression driver and cone body, which acts as a horn. This driver had the “smallest issues” of all that I looked at for this job.
The SH100 was arranged to continue that horn to a dimension which allows it to have forward directivity down to a lower than average frequency, allowing it to be placed back against a wall without significant reflection
The discontinuity is still present at the transition point however and foam would hot fix it.
I suppose the traditional fix would be to measure at 1 meter instead of 2 and apply more smoothing to the measurement before publishing.
To make the acoustic phase response it has, the seam between the ranges had to be right, radiated power constant (more or less) but the that discontinuity shows up as small changes in on axis level / radiation pattern.
I can send you files we have just received that let you see the speaker radiation in 3-d, in several ways when you download the CLF free viewer. (files are not linked at the web site yet).
If you e-mail me at danleynospamlabs@comcast.net and (remove the no meat by product reference) I will send you the CLF files for the SH100 and SH50.
As you can see, both the SH50 and 100 are free from lobes and nulls normally associated with multiway loudspeakers at crossover.
The arrangement on the seven drivers in the SH50 show no interaction between frequency ranges and is the one, which can reproduce a square wave from about 220 Hz to 2600Hz. (posted some screen shots here a few weeks ago)
The SH100 can reproduce a square wave too over a narrower range.
FWIW, The response curve for the SH50 currently on the web site needs to be changed (has a reflection, time setting is off) but it hasn’t been done yet and theirs still snow on the ground here.
The power rating for the SH50 isn’t determined yet.
Rather than a 2 or 24 hr “death test” rating, the speaker is driven with an iec pink noise spectrum, until the speakers output simply departs 6 dB from linear, given the power.
This is a lower but more realistic rating for music etc.
Got to run,
Tom
I knew what shape I needed on the outside (square), I knew I needed ports in the corners.
I cut out the center of a board, with a broad radius in the corners, roughly the shape of the flat part of the horn in the photo. I stretched the burlap across the opening and stapled it down on the back sides.
I pressed in the center of the cloth with an aluminum tube. Initially, it was too taught to stretch as far as I needed so I got it wet. After it was in position and dry, I painted on a couple coats of epoxy to make it stiff. After it was hard, I took the cylinder out and cut the opening.
I would try Spandex, it stretches a great deal, easier than burlap (in one direction a little more than the other) and has a nice smooth surface. I would try the same thing, use the built in tension in the fabric to form the curves you need. Then, stiffen it with epoxy into a solid part you can sand and paint.
Good luck, don’t be afraid to do a trial one (I made two) first to get the feel of the fabric, its cheap!
Noah, it would be wrong to tell you what driver it is although there is an Italian company with a short name that makes one VERY similar.
With any coax driver, there is a discontinuity between the compression driver and cone body, which acts as a horn. This driver had the “smallest issues” of all that I looked at for this job.
The SH100 was arranged to continue that horn to a dimension which allows it to have forward directivity down to a lower than average frequency, allowing it to be placed back against a wall without significant reflection
The discontinuity is still present at the transition point however and foam would hot fix it.
I suppose the traditional fix would be to measure at 1 meter instead of 2 and apply more smoothing to the measurement before publishing.
To make the acoustic phase response it has, the seam between the ranges had to be right, radiated power constant (more or less) but the that discontinuity shows up as small changes in on axis level / radiation pattern.
I can send you files we have just received that let you see the speaker radiation in 3-d, in several ways when you download the CLF free viewer. (files are not linked at the web site yet).
If you e-mail me at danleynospamlabs@comcast.net and (remove the no meat by product reference) I will send you the CLF files for the SH100 and SH50.
As you can see, both the SH50 and 100 are free from lobes and nulls normally associated with multiway loudspeakers at crossover.
The arrangement on the seven drivers in the SH50 show no interaction between frequency ranges and is the one, which can reproduce a square wave from about 220 Hz to 2600Hz. (posted some screen shots here a few weeks ago)
The SH100 can reproduce a square wave too over a narrower range.
FWIW, The response curve for the SH50 currently on the web site needs to be changed (has a reflection, time setting is off) but it hasn’t been done yet and theirs still snow on the ground here.
The power rating for the SH50 isn’t determined yet.
Rather than a 2 or 24 hr “death test” rating, the speaker is driven with an iec pink noise spectrum, until the speakers output simply departs 6 dB from linear, given the power.
This is a lower but more realistic rating for music etc.
Got to run,
Tom
Tom,
I understand, no problem. I was wondering if it was the B&C (since you're not using it I presume it's OK to mention).
One question everyone wonders about is if the moving cone causes modulation of the tweeter output.
Do you know if it does or not?
What do you think of this crazy idea: I have a couple dozen nice (4 mm xmax) little Focal 4C212 that I was going to build line arrays with.
What about putting 4 ea in a 2x2 array *behind* a, say, 8" waveguide, crossed at about 1.5 kHz, using very steep slopes and offset correction with a digital XO.
The 4" could be tilted back towards the centerline to match the average waveguide angle and to make a bigger hole in the center for the tweeter.
The big question is, what does the polar response look like?
1.5kHz wavelength is about 9", so I don't think the cancellation among the 4" is an issue, but what kind of polar pattern do you get from the "smoke ring" of radiation from around the waveguide?
Thanks
I understand, no problem. I was wondering if it was the B&C (since you're not using it I presume it's OK to mention).
One question everyone wonders about is if the moving cone causes modulation of the tweeter output.
Do you know if it does or not?
What do you think of this crazy idea: I have a couple dozen nice (4 mm xmax) little Focal 4C212 that I was going to build line arrays with.
What about putting 4 ea in a 2x2 array *behind* a, say, 8" waveguide, crossed at about 1.5 kHz, using very steep slopes and offset correction with a digital XO.
The 4" could be tilted back towards the centerline to match the average waveguide angle and to make a bigger hole in the center for the tweeter.
The big question is, what does the polar response look like?
1.5kHz wavelength is about 9", so I don't think the cancellation among the 4" is an issue, but what kind of polar pattern do you get from the "smoke ring" of radiation from around the waveguide?
Thanks
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.