Juspur said:
Well based on those statements I'd say you're able to retain some objectivity. And that actually allows true progress.
From what I've personally experienced there are far too few folks in the audio game who are willing to admit that maybe that cash outlay (whether big or small) they just made had no effect on the sound, or even a detrimental effect. Generally they -swear- it's better, but then one day the items suddenly and mysteriously disappear from their systems without any explanation.
Re: More Freq response considerations
Here's another option that you might like then:
http://www.htguide.com/forum/showthread.php4?t=29620&highlight=duos
The basket on these speakers is great, thin cast metal legs. I doubt you'll find another series of drivers with better FR from the backside...
Hmm, perhaps then, what Juspur's speaker could use to imrpove it would be a bipole or dipole tweeter?
Jack Caldwell said:
Here's another option that you might like then:
http://www.htguide.com/forum/showthread.php4?t=29620&highlight=duos
The basket on these speakers is great, thin cast metal legs. I doubt you'll find another series of drivers with better FR from the backside...
Hmm, perhaps then, what Juspur's speaker could use to imrpove it would be a bipole or dipole tweeter?
Walsh Mode FR
I recently measured several drivers mounted in Walsh mode and have a couple of conclusions to share:
1) Front radiator vs Walsch mode
For Walsch mode. the upper end response seems directly linked to the le inductance of the voice coil.
With a driver working in direct front radiating response, there is a "rising" response due to the rigid piston. To get this to work as a"flat response" driver, a low pass filter is needed. in practice, many driver designers use the voice coil inductance to achieve this goal.
BUT, the Walsch mode does NOT seem to exhibit these characteristics, because the "piston" is at a 60 to 90 degree incidence angle. SO, the Le inductance that yields flat response in direct radiator mode, now gives a 1st order type rolloff at a frequency mostly determined by the Le vs Re of the voice coil.
2) There seems to be some minor "horn load" effect in the midrange area, due to the size and proximity of the magnet coupling with a shallow cone. This usually seems big enough to require EQ.
Supporting evidence:
- Tried the Walsch mode with the Peerless 6.5" Nomex drivers(Le=1.4mH) and determining they were rolling off 1st order at about 1.2KHz up to 5Khz where the rolloff goes 3rd order.
- Then did the same using Vifa XG18 wood cone drivers 6.5", (Le=0.54mH) they started roling off at about 3.2KHz
- Then did the same with SBacoustics 6.5" (Le=0,15mH) and these go all the way up past 10kHz (albeit with some major peaking)
All three exhibit some midrange boost from 700Hz to about 2kHz which I am inclined to attribute to partial "horn load" created by the shallow cones.
BTW the "horn load" would be much less of a problem if we had big tall cones like the original Walsch drivers, only using modern materials and magnet structures with ultra low inductance like ScanSpeak or the SB Acoustics drivers.
SO, for "Walsch mode"use of drivers it does appear as if the upper FR is strictly related to the ratio of voice coil inductance vs the resistance.
Depending on the tweeters used, and on the type of crossover, it appears one could pretty reliably predict the rolloff of the driver and design that into the filter.
So there you have it, a mystery that seems to be resolved...
I recently measured several drivers mounted in Walsh mode and have a couple of conclusions to share:
1) Front radiator vs Walsch mode
For Walsch mode. the upper end response seems directly linked to the le inductance of the voice coil.
With a driver working in direct front radiating response, there is a "rising" response due to the rigid piston. To get this to work as a"flat response" driver, a low pass filter is needed. in practice, many driver designers use the voice coil inductance to achieve this goal.
BUT, the Walsch mode does NOT seem to exhibit these characteristics, because the "piston" is at a 60 to 90 degree incidence angle. SO, the Le inductance that yields flat response in direct radiator mode, now gives a 1st order type rolloff at a frequency mostly determined by the Le vs Re of the voice coil.
2) There seems to be some minor "horn load" effect in the midrange area, due to the size and proximity of the magnet coupling with a shallow cone. This usually seems big enough to require EQ.
Supporting evidence:
- Tried the Walsch mode with the Peerless 6.5" Nomex drivers(Le=1.4mH) and determining they were rolling off 1st order at about 1.2KHz up to 5Khz where the rolloff goes 3rd order.
- Then did the same using Vifa XG18 wood cone drivers 6.5", (Le=0.54mH) they started roling off at about 3.2KHz
- Then did the same with SBacoustics 6.5" (Le=0,15mH) and these go all the way up past 10kHz (albeit with some major peaking)
All three exhibit some midrange boost from 700Hz to about 2kHz which I am inclined to attribute to partial "horn load" created by the shallow cones.
BTW the "horn load" would be much less of a problem if we had big tall cones like the original Walsch drivers, only using modern materials and magnet structures with ultra low inductance like ScanSpeak or the SB Acoustics drivers.
SO, for "Walsch mode"use of drivers it does appear as if the upper FR is strictly related to the ratio of voice coil inductance vs the resistance.
Depending on the tweeters used, and on the type of crossover, it appears one could pretty reliably predict the rolloff of the driver and design that into the filter.
So there you have it, a mystery that seems to be resolved...
Jusper, i'm really inrigued by the simplicity of this design for surround duty in a hifi HT application but with maybe a tad more low end extension and power handling......what do you think of trying the 8" version of the Nomex and would you suggest the nomex cone or the Poly as the poly doesn't exhibit the peaky response around 6k you mentioned? For surround duty. i've contemplated a different approach to the tweeters, maybe 4 neo domes each mounted on the corners sliced at a 45 degree angle with each opposing pair wired out of phase with an adjustable Lpad of course. When modelling the bass response of the 8" version, i can get solid extension to 35hz in only 45L.....do you think the Walsh alignment will produce close to the same results? I would like to build these as 12" cube towers lined with 10" acrylic cast tube and dampened between the outside wall of the tube and the wood enclosure.....maybe rigid foam standoffs and then sand fill the voids?
8" peerless?
in answer to your recent question about 8" driver in Walsh mode:
Given the same internal box volume, the bass response would be identical for Walsh type mounting with the magnet out of the box when compared to conventional front firing.... only remember with Walsh style mounting you are actually getting a little more box volume rather than less.
This WILL make a difference in your bass response if you do not take the volume difference into account.
With a sealed box it's probably not a biggie, but with a vented cabinet, this could change the alignment by a fair amount, especially if you are using low Q drivers where the box has a small volume.
I second Critofur's comment about not going much beyond 2KHz if you use the Peerless, due to the fairly high inductance of the voice coil.
Another more expensive choice if you want to cross higher would be the scanspeak 8".... $$$
For tweeters, you might want to look into the ScanSpeak Air Circ tweeter or the Heil type AMT's from Mark & Daniel....especially if you want an Omni dispersion for the highs you should look into the DM2a with a diffuser.... check my other post "New Heil type driver" on this site for details on how to do this.
Using the DM2a would give you a single driver solution with no beaming or interference effects, and it can be very successfully crossed in at 2kHz
in answer to your recent question about 8" driver in Walsh mode:
Given the same internal box volume, the bass response would be identical for Walsh type mounting with the magnet out of the box when compared to conventional front firing.... only remember with Walsh style mounting you are actually getting a little more box volume rather than less.
This WILL make a difference in your bass response if you do not take the volume difference into account.
With a sealed box it's probably not a biggie, but with a vented cabinet, this could change the alignment by a fair amount, especially if you are using low Q drivers where the box has a small volume.
I second Critofur's comment about not going much beyond 2KHz if you use the Peerless, due to the fairly high inductance of the voice coil.
Another more expensive choice if you want to cross higher would be the scanspeak 8".... $$$
For tweeters, you might want to look into the ScanSpeak Air Circ tweeter or the Heil type AMT's from Mark & Daniel....especially if you want an Omni dispersion for the highs you should look into the DM2a with a diffuser.... check my other post "New Heil type driver" on this site for details on how to do this.
Using the DM2a would give you a single driver solution with no beaming or interference effects, and it can be very successfully crossed in at 2kHz
More expensive drivers are not nesc. better, particularly for use as a "Walsh" driver.
The Peerless drivers DO have shorting rings, and, pretty smooth response as far as non-polypropylene drivers go.
Now - some drivers such as some of the ScanSpeak Revelators, maybe some Usher woofers, etc, MAY allow simpler higher crossovers, may have lower distortion. But, until you're a real expert at this, you won't know what the back side FR looks like until you measure it. You can count on it always being different than the "front" side FR measurements.
Regardless of the woofer's freq. response, I still would generally say you don't want to crossover much above 2Khz, one reason is the distance between the woofer cone and the tweeter.
If you're lucky enough to find a woofer that plays smoothly much above 2 Khz from the backside, I would certainly TRY listening with the tweeter crossed as high as possible, as a learning experiment, if nothing else. Perhaps some day there will once again be a driver that works "full range" in the "Walsh" mode.
Some of those Peerless Nomex woofers are a great value, and I'd have to hear the completed speaker design side-by-side before I would assume that some other more expensive driver would make a "better" speaker.
For that 8" woofer - don't worry about the difference in bass/volume dispacement between "Walsh" mode and "conventional" mode mounting - the displacement of the magnet is not going to be that signifigant considering the VAS of that woofer...
The Peerless drivers DO have shorting rings, and, pretty smooth response as far as non-polypropylene drivers go.
Now - some drivers such as some of the ScanSpeak Revelators, maybe some Usher woofers, etc, MAY allow simpler higher crossovers, may have lower distortion. But, until you're a real expert at this, you won't know what the back side FR looks like until you measure it. You can count on it always being different than the "front" side FR measurements.
Regardless of the woofer's freq. response, I still would generally say you don't want to crossover much above 2Khz, one reason is the distance between the woofer cone and the tweeter.
If you're lucky enough to find a woofer that plays smoothly much above 2 Khz from the backside, I would certainly TRY listening with the tweeter crossed as high as possible, as a learning experiment, if nothing else. Perhaps some day there will once again be a driver that works "full range" in the "Walsh" mode.
Some of those Peerless Nomex woofers are a great value, and I'd have to hear the completed speaker design side-by-side before I would assume that some other more expensive driver would make a "better" speaker.
For that 8" woofer - don't worry about the difference in bass/volume dispacement between "Walsh" mode and "conventional" mode mounting - the displacement of the magnet is not going to be that signifigant considering the VAS of that woofer...
Thanx for the replies. I'm gonna start out with just the woofers for now and take some measurements on a horizontal axis as the purpose for these is strictly omni-surround duty. Simple MDF towers for now tuned to 37hz. We'll see what the response turns up.....i'm hoping to press it to 3.5khz. Considering the proposed 4 corner tweeter array, this will result in a mild horn loading of the woofer into the enclosure...anyone see a problem with this? BTW i'm considering the Aurasound NT1's.
Once upon a F cabinet
Juspur, I love your design and the engineering that went with it. I wish my cabinet making skills were nearly as good as yours. One of my first successful Walsh alignment creations came back in '94 out of necessity. I burnt the left side tweeter in my Ohm F2's and couldn't afford repair, so I made a mounting board out of 3/4" plywood and placed 10" Oaktron poly woofers upside down in the original F cabinets and hot glued an ESS Heil monopole "tweeter" to the top of each one. When you run a pink noise signal you can time align a hi frequency driver very well by listening to the sound become coherent. I surrounded the backs with glass wool, kinda like wall insulation but much more dense, inside the original F covers, as I had learned from the F's of potential (and severe) reflection waves, both destructive and constructive. This did not really solve that issue but smoothed out the midrange substantially. Big peaks around 2500 Hz using the original ESS x-overs. So I wound some air core inductors (no L pad, did it by counting windings, ugh.) and used a cap to cross over the heils at approxiately 4200 Hz. They looked a mess, almost as bad as if you pull apart a current Ohm driver but with the brown covers you couldn't see. So to make it short, very powerful sound with more transparent highs than the F2's although the mids needed some EQ. I agree that poly woofers work much better with a walsh alignment and virtually any cone angle will produce a very coherent sound stage. However, as someone mentioned earlier, not all woofers or mid bass drivers work well this way! I lucked out. OK,I'm done with this for now. I am running some 3 1/2" drivers this way with some auto tweeters from Radio Shack into 4 liter polypropylene bottles. I might post a pic or 2 and describe them if ayone is interested in my $30 baby walsh-like creations. Yes, they do sound very good, actually.
Juspur, I love your design and the engineering that went with it. I wish my cabinet making skills were nearly as good as yours. One of my first successful Walsh alignment creations came back in '94 out of necessity. I burnt the left side tweeter in my Ohm F2's and couldn't afford repair, so I made a mounting board out of 3/4" plywood and placed 10" Oaktron poly woofers upside down in the original F cabinets and hot glued an ESS Heil monopole "tweeter" to the top of each one. When you run a pink noise signal you can time align a hi frequency driver very well by listening to the sound become coherent. I surrounded the backs with glass wool, kinda like wall insulation but much more dense, inside the original F covers, as I had learned from the F's of potential (and severe) reflection waves, both destructive and constructive. This did not really solve that issue but smoothed out the midrange substantially. Big peaks around 2500 Hz using the original ESS x-overs. So I wound some air core inductors (no L pad, did it by counting windings, ugh.) and used a cap to cross over the heils at approxiately 4200 Hz. They looked a mess, almost as bad as if you pull apart a current Ohm driver but with the brown covers you couldn't see. So to make it short, very powerful sound with more transparent highs than the F2's although the mids needed some EQ. I agree that poly woofers work much better with a walsh alignment and virtually any cone angle will produce a very coherent sound stage. However, as someone mentioned earlier, not all woofers or mid bass drivers work well this way! I lucked out. OK,I'm done with this for now. I am running some 3 1/2" drivers this way with some auto tweeters from Radio Shack into 4 liter polypropylene bottles. I might post a pic or 2 and describe them if ayone is interested in my $30 baby walsh-like creations. Yes, they do sound very good, actually.
Hi Spikedart - and Welcome to diyAudio!!
I see we share common interest in things ESS and Walsh so you might find this thread of interest.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=100390&perpage=25&pagenumber=2
I used to work for DEC over in Springfield and my Bro still lives in MA so if I get up that way maybe we can get together for some chowder and compare notes.
I see we share common interest in things ESS and Walsh so you might find this thread of interest.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=100390&perpage=25&pagenumber=2
I used to work for DEC over in Springfield and my Bro still lives in MA so if I get up that way maybe we can get together for some chowder and compare notes.
Hi there
My webhost went down - however the initial pic's are submitted below. Enjoy
My webhost went down - however the initial pic's are submitted below. Enjoy
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Hi Renron
A_men said just the same.
Yes, my webhost suddenly gave up business. I've resubmitted the 4
original pictures to diyaudio, but for some strange reason, they've
been scaled down (making the drawing difficult to see).
Well, I'll try again
A_men said just the same.
Yes, my webhost suddenly gave up business. I've resubmitted the 4
original pictures to diyaudio, but for some strange reason, they've
been scaled down (making the drawing difficult to see).
Well, I'll try again
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Another main driver worth considering?
Having spoken to Dave at DC Gold Audio, he indicates that his initial (not thorough) test of the 7" driver produced very positive results when walsh aligned. Note the spread cast basket and steep angle of the driver. These are full range drivers usually meant for aquatic use or for other harsh environments. I do not have the driver specs for these but I am sure he will provide them if asked. He quotes:
Size - 7", Power handling - 80W cont / 300 W max, Efficiency @ 1W / 1m - 92 db, Freq. +/-3d - 35Hz - 19kHz. They seem like the ideal test driver for this application. Unfortunately, they are not cheap and if they aren't good as inverted cones, well I suppose if you have a boat...
Web site= not so hot, the actual driver picture can be found on the home page and you see these on the auction sites. I don't have nearly $600 on hand at the moment for this project, but hope to at the start of the year. DC Gold Audio : N7C [N7C] - $ 279.11
Having spoken to Dave at DC Gold Audio, he indicates that his initial (not thorough) test of the 7" driver produced very positive results when walsh aligned. Note the spread cast basket and steep angle of the driver. These are full range drivers usually meant for aquatic use or for other harsh environments. I do not have the driver specs for these but I am sure he will provide them if asked. He quotes:
Size - 7", Power handling - 80W cont / 300 W max, Efficiency @ 1W / 1m - 92 db, Freq. +/-3d - 35Hz - 19kHz. They seem like the ideal test driver for this application. Unfortunately, they are not cheap and if they aren't good as inverted cones, well I suppose if you have a boat...
Web site= not so hot, the actual driver picture can be found on the home page and you see these on the auction sites. I don't have nearly $600 on hand at the moment for this project, but hope to at the start of the year. DC Gold Audio : N7C [N7C] - $ 279.11
The price of the DC Gold N7C drivers is per pair.
Here is some data you might find useful.
I have found that a Zobel consisting of 12 ohms and 3.1uF does a very respectable job of smoothing the top end.
As a woofer these things are capable of some ridiculous bottom end for their size.
In a two way they can be crossed over much higher than a normal 7"driver thus avoiding a x-over point in the critical area of 1-3k.
As a full range driver they do very well also, however, for this purpose the N7R is very worth the extra money.
Here is some data you might find useful.
I have found that a Zobel consisting of 12 ohms and 3.1uF does a very respectable job of smoothing the top end.
As a woofer these things are capable of some ridiculous bottom end for their size.
In a two way they can be crossed over much higher than a normal 7"driver thus avoiding a x-over point in the critical area of 1-3k.
As a full range driver they do very well also, however, for this purpose the N7R is very worth the extra money.
Attachments
Is that trace for real?
Thanks for the post, I am impressed. The physical driver size (elongated) and wide claimed FR seemed to indicate the potential. Nice find. Does the Zobel network reduce the efficiency of this driver? I noted the high R in the trace.
Thanks for the post, I am impressed. The physical driver size (elongated) and wide claimed FR seemed to indicate the potential. Nice find. Does the Zobel network reduce the efficiency of this driver? I noted the high R in the trace.
Juspur,
If you are listening to the back side of the woofer it needs to be wired in reverse or it is playing out of phase. It would be very easy for you to wire both drivers in reverse where the woofer would be in phase and the tweeter would be out. This is a common trick to smooth a tweeter in a 1st order x-over arrangement.
My estimation is that with only a 2.7uF cap your x-over point is going to be in the 7.8khz range. I highly recommend putting a zoebel across the woofer, 8.2uF and 5.6 ohm is a pretty good place to start.
If you are listening to the back side of the woofer it needs to be wired in reverse or it is playing out of phase. It would be very easy for you to wire both drivers in reverse where the woofer would be in phase and the tweeter would be out. This is a common trick to smooth a tweeter in a 1st order x-over arrangement.
My estimation is that with only a 2.7uF cap your x-over point is going to be in the 7.8khz range. I highly recommend putting a zoebel across the woofer, 8.2uF and 5.6 ohm is a pretty good place to start.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.