Hello!
Here is a center channel project enclosure. It is a d'Appolito WMTMW that will have a 18dB/octave active crossover between woofers and midranges, and a 6dB/octave passive crossover between midranges and the tweeter.
The 4 woofers are Dayton RS 150-4, 1 compound assembly in the back of each T shaped side. The 2 midranges are Bohlender Graebener Neo8, going on either side of the tweeter. The tweeter is a home built Neil Davis type Heil AMT.
http://www.diyaudio.com/forums/planars-exotics/149226-neil-davis-heil-amt.html
The 3 front facing drivers are angled down 5 degrees, because this speaker will sit above a screen. The AMT will also be able to be adjusted fore and aft with a jack screw.
The front and sides are 3/4" 11-ply oak, 1/4" MDF, silicone tile adhesive, and 1/16" sheet lead, outside to inside. The back is 1" MDF, silicone adhesive, and lead. The main load-bearing beam is the only continuous vertical plane from left to right. The woofers are attached to that. It is also 1" MDF, as are the interior cross braces. The inside of the midrange and bass enclosures are slathered with latex acoustic compound. Hope it is dead enough!
Problems to come include:
- a measured 10dB peak at 7kHz in the AMT, which it turns out is not unusual for closed-back AMTs;
- maybe not enough room in the back for passive crossover components, especially for things like the notch filter that will have to be used for that peak in the AMT.
- this thing is getting heavy! I have to bolt it into brick above a fireplace where our display sits, preferably in the mortar and not the brick itself.
- I like to use rubber well nuts and foam strips to mount drivers, but this time braces and fasteners are in the way, so I must use screws in pilot holes. Any suggestions on snubbers, dampeners, or nylon washers with collars would be appreciated.
Thanks!
Paul
Here is a center channel project enclosure. It is a d'Appolito WMTMW that will have a 18dB/octave active crossover between woofers and midranges, and a 6dB/octave passive crossover between midranges and the tweeter.
The 4 woofers are Dayton RS 150-4, 1 compound assembly in the back of each T shaped side. The 2 midranges are Bohlender Graebener Neo8, going on either side of the tweeter. The tweeter is a home built Neil Davis type Heil AMT.
http://www.diyaudio.com/forums/planars-exotics/149226-neil-davis-heil-amt.html
The 3 front facing drivers are angled down 5 degrees, because this speaker will sit above a screen. The AMT will also be able to be adjusted fore and aft with a jack screw.
The front and sides are 3/4" 11-ply oak, 1/4" MDF, silicone tile adhesive, and 1/16" sheet lead, outside to inside. The back is 1" MDF, silicone adhesive, and lead. The main load-bearing beam is the only continuous vertical plane from left to right. The woofers are attached to that. It is also 1" MDF, as are the interior cross braces. The inside of the midrange and bass enclosures are slathered with latex acoustic compound. Hope it is dead enough!
Problems to come include:
- a measured 10dB peak at 7kHz in the AMT, which it turns out is not unusual for closed-back AMTs;
- maybe not enough room in the back for passive crossover components, especially for things like the notch filter that will have to be used for that peak in the AMT.
- this thing is getting heavy! I have to bolt it into brick above a fireplace where our display sits, preferably in the mortar and not the brick itself.
- I like to use rubber well nuts and foam strips to mount drivers, but this time braces and fasteners are in the way, so I must use screws in pilot holes. Any suggestions on snubbers, dampeners, or nylon washers with collars would be appreciated.
Thanks!
Paul
Attachments
Last edited:
Now for the NEO8s
BG NEO8s are now in. Nothing fancy on the mounting hardware, just course thread screws turned until clamped and then backed off 1/4 turn. 3/16" foam weather strip tape, cut out for the screw holes and the pins that protrude from the backside of the mounting edge. There are a few dents visible from trial fitting the Davis Heil AMT that will have to be touched up...
BG NEO8s are now in. Nothing fancy on the mounting hardware, just course thread screws turned until clamped and then backed off 1/4 turn. 3/16" foam weather strip tape, cut out for the screw holes and the pins that protrude from the backside of the mounting edge. There are a few dents visible from trial fitting the Davis Heil AMT that will have to be touched up...
Attachments
Now For the AMT Tweeter...
This tweeter is going in tonight.
Good old duct tape! Solder, tape and staples.
This driver is going to be a handfull. I'll post near field measurements as soon as I can.
By the way, normal speaker builders, use a BG NEO3, don't screw around with an AMT. But I'm going retro.
This tweeter is going in tonight.
Good old duct tape! Solder, tape and staples.
This driver is going to be a handfull. I'll post near field measurements as soon as I can.
By the way, normal speaker builders, use a BG NEO3, don't screw around with an AMT. But I'm going retro.
Dual NEO8 Mid Crossover
First, the good news. I used WinPCD 1.2.0 rev1 to design a low pass crossover. (electronic XO and another amp for the woofers) Mitey Mike test mic 1 meter in front of speaker in the listening room, fed into HolmImpluse, for frequency response. Dayton Audio DATS measuring the impedance of both NEO8s in series. .frd file from HolmImpulse and .zma file from DATS loaded into WinPCD.
It is labeled "Woofer." Blue is as measured, red is what I came up with. Not bad at all considering there is no felt around the edges.
First, the good news. I used WinPCD 1.2.0 rev1 to design a low pass crossover. (electronic XO and another amp for the woofers) Mitey Mike test mic 1 meter in front of speaker in the listening room, fed into HolmImpluse, for frequency response. Dayton Audio DATS measuring the impedance of both NEO8s in series. .frd file from HolmImpulse and .zma file from DATS loaded into WinPCD.
It is labeled "Woofer." Blue is as measured, red is what I came up with. Not bad at all considering there is no felt around the edges.
AMT Crossover
The AMT, on the other hand, has a broad 12 dB peak between 5K and 11K. I was able to monitor the whole system sum response to see what changes in the crossover would do. Blue is raw, red is my failed attempt.
The SPL below is summed, tweeter reversed polarity, 18dB/Octave. The peak I left in the AMT at 7kHz shows in the summed response and 10kHz the response is down a good 5dB.
I suppose I could increase the high-pass slope to 24dB/Octave to cut off the 7kHz peak. The parallel RLC notch filters in PCD are not wide enough to neatly flatten that peak.
The impedance on this driver is very flat. Guys, is there a passive component combination I can use just in front of the AMT to tame that peak?
The AMT, on the other hand, has a broad 12 dB peak between 5K and 11K. I was able to monitor the whole system sum response to see what changes in the crossover would do. Blue is raw, red is my failed attempt.
The SPL below is summed, tweeter reversed polarity, 18dB/Octave. The peak I left in the AMT at 7kHz shows in the summed response and 10kHz the response is down a good 5dB.
I suppose I could increase the high-pass slope to 24dB/Octave to cut off the 7kHz peak. The parallel RLC notch filters in PCD are not wide enough to neatly flatten that peak.
The impedance on this driver is very flat. Guys, is there a passive component combination I can use just in front of the AMT to tame that peak?
Smoke and Flames Coming Out of Heil AMT
Well I wish, but it's too late now, I have to make this stupid thing fit! By now I could have bought a SEAS Excel or a Revelator with the money I've spent making this AMT, but it does look cool.
Using the three most useful tools I have bought for the money in years, DATS from Parts Express for impedance, HolmImpulse for frequency response, and WinPCD for crossover design, I've come up with a curve:
Going down in frequency from 20Khz, the high pass starts on its way just before 11Khz, at 24dB/Octave, so the big blue bump fights the filter until 7K, then drops like a stone below that.
(dead attachment, see post below!)
WinPCD comes with a schematic diagram with variable fields for the values of R, L & C that feed back into the program calculations. I was able to substitute values that I had available from my previous hand-calculated crossover and that were inexpensive from Madisound and Parts Express, and then see what effect if any this had on the response.
The curve above is with these available components. That 3dB bump between 5 and 6K, if it survives into real life, will be audible. There is a 60 ohm double point impedance peak around there too, so notching that will have to wait until I install the components and then run the DATS device and software so I can do a parallel impedance filter. And the sacred odd-order MTM d'Appolito slope is out the window, of course, so there will be nulls off-axis. Time to run HolmImpulse and hear just what I've got. One saving grace is, most of the music spectrum will be from the NEO8s.
Reader, if you are about to build your first system, check out "Introduction to designing crossovers without measurement." I've been screwing with this speaker for a year and a half. I've also burned up three tweeters and an expensive Focal woofer in my hobby, and those acrid, rubbing voice coils haunt me. It's about the crossover, unless you are avoiding it entirely.
Well I wish, but it's too late now, I have to make this stupid thing fit! By now I could have bought a SEAS Excel or a Revelator with the money I've spent making this AMT, but it does look cool.
Using the three most useful tools I have bought for the money in years, DATS from Parts Express for impedance, HolmImpulse for frequency response, and WinPCD for crossover design, I've come up with a curve:
Going down in frequency from 20Khz, the high pass starts on its way just before 11Khz, at 24dB/Octave, so the big blue bump fights the filter until 7K, then drops like a stone below that.
(dead attachment, see post below!)
WinPCD comes with a schematic diagram with variable fields for the values of R, L & C that feed back into the program calculations. I was able to substitute values that I had available from my previous hand-calculated crossover and that were inexpensive from Madisound and Parts Express, and then see what effect if any this had on the response.
The curve above is with these available components. That 3dB bump between 5 and 6K, if it survives into real life, will be audible. There is a 60 ohm double point impedance peak around there too, so notching that will have to wait until I install the components and then run the DATS device and software so I can do a parallel impedance filter. And the sacred odd-order MTM d'Appolito slope is out the window, of course, so there will be nulls off-axis. Time to run HolmImpulse and hear just what I've got. One saving grace is, most of the music spectrum will be from the NEO8s.
Reader, if you are about to build your first system, check out "Introduction to designing crossovers without measurement." I've been screwing with this speaker for a year and a half. I've also burned up three tweeters and an expensive Focal woofer in my hobby, and those acrid, rubbing voice coils haunt me. It's about the crossover, unless you are avoiding it entirely.
Last edited:
From Airbrushed Theory to Dirty Practice
All,
I bought a first round of caps coils and resistors based on the previous efforts, and then ran Holm Impulse. The 3 db peak I was concerned about between 5 and 6k is in reality a broad 5db plateau between 5 and 9kHz. There is also a gentle peak centered at 650 Hz. In order to get the hot AMT down in level, I had to put 75 ohms in series with it. This made an impedance curve rising to 70 ohms at 20kHz.
This gave me an opportunity to look at the 2 NEO8s and 1 AMT with the crossover as if they were a single driver. Another impedance run into the whole collection so far using DATS, and another frequency response run using HolmImpulse. Again, loading the .zma file from DATS and the .frd file from HolmImpulse into WinPCD for one more session.
With that rising impedance, I made a Zobel that practically flattened the impedance at 9 ohms. This allowed me to add two notch filters for the two peaks. The designed curve has two impedance peaks, which I will leave in.
The target frequency response is much better. This is still without any felt diffraction damping, or any adjustment of the AMT in or out.
I've bought and installed the additional caps coils and resistors. The wife hates the sound of frequency sweeps, so I can't do that yet, but preliminary tests are very good!
Again I must sing praises of a good test mic, HolmImpulse, DATS, and WinPCD in helping me interatively design this messy crossover and tame that AMT.
All,
I bought a first round of caps coils and resistors based on the previous efforts, and then ran Holm Impulse. The 3 db peak I was concerned about between 5 and 6k is in reality a broad 5db plateau between 5 and 9kHz. There is also a gentle peak centered at 650 Hz. In order to get the hot AMT down in level, I had to put 75 ohms in series with it. This made an impedance curve rising to 70 ohms at 20kHz.
This gave me an opportunity to look at the 2 NEO8s and 1 AMT with the crossover as if they were a single driver. Another impedance run into the whole collection so far using DATS, and another frequency response run using HolmImpulse. Again, loading the .zma file from DATS and the .frd file from HolmImpulse into WinPCD for one more session.
With that rising impedance, I made a Zobel that practically flattened the impedance at 9 ohms. This allowed me to add two notch filters for the two peaks. The designed curve has two impedance peaks, which I will leave in.
The target frequency response is much better. This is still without any felt diffraction damping, or any adjustment of the AMT in or out.
I've bought and installed the additional caps coils and resistors. The wife hates the sound of frequency sweeps, so I can't do that yet, but preliminary tests are very good!
Again I must sing praises of a good test mic, HolmImpulse, DATS, and WinPCD in helping me interatively design this messy crossover and tame that AMT.
Done With This
Dear Diary: (or, Dear diyAry?)
Well, the crossover is done, or about as done as I want to bother with. Please see the picture, above in red is the predicted mid and high response from Passive Crossover Designer, below in blue is that measured with Holm Impulse.
Below in red is the whole center channel with active crossover, set to around 250Hz. This is 1 meter out into the room towards the listener. The peak around 40 to 50 Hz is the mantel, and the pictures on it vibrating and moving around. I can press on various parts of the mantel and the shelf holding the center-channel and stop the resonances. Time for some rubber damping and screws.
This speaker sounds very good in mono, because I'm re-cabling the rest of the system and this project has held things up.
Next comes felt, and maybe screwing the tweeter backwards a few millimeters.
The crossover will only fit in a separate tray, but that's okay because I already had to do that for the left and right speakers some time ago.
Dear Diary: (or, Dear diyAry?)
Well, the crossover is done, or about as done as I want to bother with. Please see the picture, above in red is the predicted mid and high response from Passive Crossover Designer, below in blue is that measured with Holm Impulse.
Below in red is the whole center channel with active crossover, set to around 250Hz. This is 1 meter out into the room towards the listener. The peak around 40 to 50 Hz is the mantel, and the pictures on it vibrating and moving around. I can press on various parts of the mantel and the shelf holding the center-channel and stop the resonances. Time for some rubber damping and screws.
This speaker sounds very good in mono, because I'm re-cabling the rest of the system and this project has held things up.
Next comes felt, and maybe screwing the tweeter backwards a few millimeters.
The crossover will only fit in a separate tray, but that's okay because I already had to do that for the left and right speakers some time ago.
Attachments
That is my biggest disappointment in this project. I think I can hear a null around 8K, somewhere off to the either side, and the listening room has chairs on the far right and left for which I wanted a smooth roll off.
What I was aiming for was an overlap of a good two octaves and a first-order crossover no higher than 2/3 wavelength between high and mids, in order to get that d'Appolito no-lobe radiation pattern. I had to go much steeper on the AMT high-pass in order to tame the frequency response, 4th order.
The BG NEO8 drivers were to be "helpers," transitioning to the woofers. Instead, they have given this system a quick, in-phase midrange from 400 Hz to 4kHz. Their low pass is only an inductor, and a notch filter for their upper range peak. They make this speaker.
The AMT can slide in or out 2 cm each way by turning a jack screw, and I want to experiment with different positions to see how measured FR, phase, transient spike and angle of the nulls are affected.
If I had it to do over, I would use the bohlender graebener neo3, as close as possible to the NEO8s, and a third-order crossover. The diy Heil AMT has been an embarrassing love affair.
What I was aiming for was an overlap of a good two octaves and a first-order crossover no higher than 2/3 wavelength between high and mids, in order to get that d'Appolito no-lobe radiation pattern. I had to go much steeper on the AMT high-pass in order to tame the frequency response, 4th order.
The BG NEO8 drivers were to be "helpers," transitioning to the woofers. Instead, they have given this system a quick, in-phase midrange from 400 Hz to 4kHz. Their low pass is only an inductor, and a notch filter for their upper range peak. They make this speaker.
The AMT can slide in or out 2 cm each way by turning a jack screw, and I want to experiment with different positions to see how measured FR, phase, transient spike and angle of the nulls are affected.
If I had it to do over, I would use the bohlender graebener neo3, as close as possible to the NEO8s, and a third-order crossover. The diy Heil AMT has been an embarrassing love affair.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.