Hi there,
I have been meaning to build a OB for quite some time, but never found the guts and time to do it. Various designs were in my head, and I finally took the step to actually build one of them as a prototype. So beware, this thing was never intended to be crafted perfectly, it just had to resemble the design enough for me to: (1) See how my wife likes it and (2) see if the design would actually sound good.
Well, first about the units used. I've had them for years already, just waiting for this project:
- tweeter: B&G Neo3 PDR. I think this one won't need any introduction... It's great, and I even have to spares
- mids: Peerless HDS 180 Exclusive, two per side. These are fine units form Peerless, with pearl colored basket and phase plug. They behave nicely and distortion figures are quite good.
- Woofer: the elusive 15" Peerless SLS 400. I'm probably one of the few to have a pair of these, which is in itself a problem, since I'd rather have two pairs, but sadly, you can't get them anymore and back then I only had the budget for two. They do have quite a low Qts for use in an OB, but nothing a little boost won't be able to fix.
Filtering will be done using a miniDSP 4x10 HD. Amplification is done with eight L25D IRS2092 Class-D amps, powered by a 1KW SMPS. Currently I only have two channels wired up, one powering my old speaker, the other the SLS 400. Mid's and tweeter are currently hooked up to a small T-amp. It works for now, until I can build a nice casing for the L25D amps. With four amps per side I can be flexible, I could bridge two channels to power the big SLS 400, or power each mid with one amp. Currently the mids are in parallel tough, so I only use three channels.
Anyway, enough about that now. Let's talk about the design! Most OB's are fairly wide designs with the woofers flat on a big slab material, for obvious reasons. Well, I did lot's of drawings of these, and found I didn't really like the looks that much. I will be building one of these however, just to find out the difference in sound quality. But that's for later on. For now, I wanted to build something more elegant an narrow, but still retain at least some of the low end performance. So I came op with the idea to put the SLS 400 in a kind of ripol/N-dipole/slot loaded OB. This way, I could have a sleek design and still retain good performance. The only thing I still needed to to is, get rid of the ugly box on the back side. I solved by making it rounded on the back. I think It looks pretty cool like this:
The rounded backside is made up of slices of material (MDF). The experimental unit I build doesn't have this, I used a much simpler (and less MDF consuming) method. If I choose to go with this design, I will have it made this way however. Also, in the Sketchup design I cut some bits off of the NEO faceplate. I did not do that with the experimental unit. So the mids are a but further apart. I build the thing last weekend, and hooked it up yesterday. Currently it looks like this:
While building I immediately noticed some shortcommings: Both the baffle and the outer sides of the rear end are probably to narrow. Hooking the thing up confirmed this. I'd probably need to double them up to lessen vibration.
Currently I've set up some a half baked filtering scheme, with some dipole correction, and filters at 120 and 2200 Hz. I didn't measure the system yet, but the first music tracks already showed some promise Bass performance is actually quite good. There are some resonances however, but that is expected, and measurements will tell me if it can be fixed or not. mid's and high's are already very nice, but it's clear that it still needs a lot of tuning. I'll probably also brace the sides of the backend with some MDF strips to make them stiffer.
I hope to have time today to do some measurements and tweak the miniDSP. If so, I'll post some of the graphs.
I have been meaning to build a OB for quite some time, but never found the guts and time to do it. Various designs were in my head, and I finally took the step to actually build one of them as a prototype. So beware, this thing was never intended to be crafted perfectly, it just had to resemble the design enough for me to: (1) See how my wife likes it and (2) see if the design would actually sound good.
Well, first about the units used. I've had them for years already, just waiting for this project:
- tweeter: B&G Neo3 PDR. I think this one won't need any introduction... It's great, and I even have to spares
- mids: Peerless HDS 180 Exclusive, two per side. These are fine units form Peerless, with pearl colored basket and phase plug. They behave nicely and distortion figures are quite good.
- Woofer: the elusive 15" Peerless SLS 400. I'm probably one of the few to have a pair of these, which is in itself a problem, since I'd rather have two pairs, but sadly, you can't get them anymore and back then I only had the budget for two. They do have quite a low Qts for use in an OB, but nothing a little boost won't be able to fix.
Filtering will be done using a miniDSP 4x10 HD. Amplification is done with eight L25D IRS2092 Class-D amps, powered by a 1KW SMPS. Currently I only have two channels wired up, one powering my old speaker, the other the SLS 400. Mid's and tweeter are currently hooked up to a small T-amp. It works for now, until I can build a nice casing for the L25D amps. With four amps per side I can be flexible, I could bridge two channels to power the big SLS 400, or power each mid with one amp. Currently the mids are in parallel tough, so I only use three channels.
Anyway, enough about that now. Let's talk about the design! Most OB's are fairly wide designs with the woofers flat on a big slab material, for obvious reasons. Well, I did lot's of drawings of these, and found I didn't really like the looks that much. I will be building one of these however, just to find out the difference in sound quality. But that's for later on. For now, I wanted to build something more elegant an narrow, but still retain at least some of the low end performance. So I came op with the idea to put the SLS 400 in a kind of ripol/N-dipole/slot loaded OB. This way, I could have a sleek design and still retain good performance. The only thing I still needed to to is, get rid of the ugly box on the back side. I solved by making it rounded on the back. I think It looks pretty cool like this:
The rounded backside is made up of slices of material (MDF). The experimental unit I build doesn't have this, I used a much simpler (and less MDF consuming) method. If I choose to go with this design, I will have it made this way however. Also, in the Sketchup design I cut some bits off of the NEO faceplate. I did not do that with the experimental unit. So the mids are a but further apart. I build the thing last weekend, and hooked it up yesterday. Currently it looks like this:
While building I immediately noticed some shortcommings: Both the baffle and the outer sides of the rear end are probably to narrow. Hooking the thing up confirmed this. I'd probably need to double them up to lessen vibration.
Currently I've set up some a half baked filtering scheme, with some dipole correction, and filters at 120 and 2200 Hz. I didn't measure the system yet, but the first music tracks already showed some promise
Bass performance is actually quite good. There are some resonances however, but that is expected, and measurements will tell me if it can be fixed or not. mid's and high's are already very nice, but it's clear that it still needs a lot of tuning. I'll probably also brace the sides of the backend with some MDF strips to make them stiffer.I hope to have time today to do some measurements and tweak the miniDSP. If so, I'll post some of the graphs.
Attachments
Nice project!
Could you explain your rationale behind keeping the mounting frame of the Neo3? Wouldn't it be better to discard it and get the HDS drivers closer together?
Rudolf
I wouldn't double up the baffle. How about a "spine" in the back which fixes the magnets of the HDS drivers.
Could you explain your rationale behind keeping the mounting frame of the Neo3? Wouldn't it be better to discard it and get the HDS drivers closer together?
Rudolf
Nice looking design!
Be aware that the Neo3 front plate also works as some kind of waveguide, that improves the frequency response and dispersion somewhat. Taking it off will change the response. But in the other hand, by taking it off you will get more symmetry in front/rear radiation....
To isolate the mids and tweeter from the woofer vibration you could build the mid/tweeter frame as a separate unit that is decoupled from the woofer frame. I did this on one of my designs, and it worked well.
Be aware that the Neo3 front plate also works as some kind of waveguide, that improves the frequency response and dispersion somewhat. Taking it off will change the response. But in the other hand, by taking it off you will get more symmetry in front/rear radiation....
To isolate the mids and tweeter from the woofer vibration you could build the mid/tweeter frame as a separate unit that is decoupled from the woofer frame. I did this on one of my designs, and it worked well.
I'm curious the advantage of having a wider baffle around the tweeter, in theory and in listening? I ask cause I'm about to build something like this myself. Linkwitz has gone super skinny on his latest which makes me think its probably a good idea? (though perhaps something less circular/symmetrical?)
Also 4real, I'd be curious to see where you measure your ripole resonance at. I'm going ripole with 12s and really hoping to get up to 200hz 4th/8th order without problems.....
Cheers
B
Also 4real, I'd be curious to see where you measure your ripole resonance at. I'm going ripole with 12s and really hoping to get up to 200hz 4th/8th order without problems.....
Cheers
B
The final design will probably have a symetrical front and back side, both having a bit of a wave guide (the baffle is 18 or 22 mm in thinkness). Different response can be fixed easyly.
Yes, I've seen this done all over the place. I guess it could even be done without killing the design.
Phaseplug-to-phaseplug distance is currently 310 mm (90+130+90 mm). Ideal crossover point is probably somewhere near 1.8 to 2 Khz. I guess I can drive the NEO3 that low with a high enough order filter. I might even go as low as 1.5Khz..
The basic advantage in my case is not have the upper woofer break the MDF, and off
I have seen lots of variants, from a naked NEO, to vert small sides, or even the use of damping material. From what I've gathered it looks like a small baffle or the use of damping on the sides yields best results. I't didn't quite fit my visual design however, so I didn't go with it. I can probably make the tweeter section a bit narrower, specially if would stiffen the baffle by fixing the magnets.
If there is any way to reduce this distance the better. Driving the tweeter that low will induce a significant amount of distortion. Trim those bezels if necessary. Optimum would be 2.2kHz ¼ wave distance to the tweeter center.
Really like your setup
BG crosses that tweeter at 1500
Today, I made some measurements and threw together a fast and simple filter that at least gives me something resembling a flat response.
First off the woofer:
Dipole peek is at about 170Hz, with a second resonance right after that. The somewhat corrected woofer response looks something like this.
Next up, the mids:
And the corrected version:
Tweeter looks like this, corrected (somewhat:
There are still quite some ups and downs here, probably related to the rear end
Lastly, the full response with some more tweaking done:
As you can see, It still needs tweaking, but I doesn't sound half bad at all. Bass resonances are mostly gone, and it really sounds quite nice
So, now, I'll lay back and enjoy the current settings for tonight
First off the woofer:
Dipole peek is at about 170Hz, with a second resonance right after that. The somewhat corrected woofer response looks something like this.
Next up, the mids:
And the corrected version:
Tweeter looks like this, corrected (somewhat:
There are still quite some ups and downs here, probably related to the rear end
Lastly, the full response with some more tweaking done:
As you can see, It still needs tweaking, but I doesn't sound half bad at all. Bass resonances are mostly gone, and it really sounds quite nice
So, now, I'll lay back and enjoy the current settings for tonight
Attachments
Last edited:
Nice effort so far. And you deserve to enjoy it.
But dipole truth is in the polars
. I look forward to see the shortcomings - if any.Rudolf
Thanks guys for the complements
Well Linkwitz helped a lot with his documentation on how to do active correction for dipole. Still, it is really astonishing what you can do in about an hours worth of measuring and tweaking!
Yes, It would be nice if I could make them. Do you know any free tools that work with OSX? REW doesn't seem to support polar plots. I will do some off axis measurements later this week. That should also give you an idea.
Well, this one won't get any paint, since it's only a test model. I'm still not sure about the final version. I'm not sure if I want to use MDF and give it a paint job, or make it out of bamboo, and just clear-coat it. I have some time still... the design is not yet final anyway.
Well Linkwitz helped a lot with his documentation on how to do active correction for dipole. Still, it is really astonishing what you can do in about an hours worth of measuring and tweaking!
Rodulf said:
Yes, It would be nice if I could make them. Do you know any free tools that work with OSX? REW doesn't seem to support polar plots. I will do some off axis measurements later this week. That should also give you an idea.
gainphile said:
Well, this one won't get any paint, since it's only a test model. I'm still not sure about the final version. I'm not sure if I want to use MDF and give it a paint job, or make it out of bamboo, and just clear-coat it. I have some time still... the design is not yet final anyway.
No need to show polar diagrams. I meant to just ask for off-axis measurements up to 90°
.If you can export your impulse response measurements as .wav files, I could turn them into polar plots or sonograms for you using ARTA.
Rudolf
I like this!
I'm interested in horizontal off-axis measurements too! Just do the various measurements and mak an Overlay of them! And what xo frequencies you use, what topology?
I think it should be easy to eq the woofer even more (flat to 30-35?) , but perhaps at the expense of overall "efficiency" Be careful for not boosting low end, because without high-pass it leads to digital clipping in minidsp.
I'm interested in horizontal off-axis measurements too! Just do the various measurements and mak an Overlay of them! And what xo frequencies you use, what topology?
I think it should be easy to eq the woofer even more (flat to 30-35?) , but perhaps at the expense of overall "efficiency" Be careful for not boosting low end, because without high-pass it leads to digital clipping in minidsp.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.