Hi everyone. I recently decided to try my hand at designing and building my own loudspeakers. I am about to assemble the cabinets for my design.
My first design is inspired by the Harbeth 40.3, hence the name. I had initially planned on making the Visaton Nimrod kit, but lack of subjective reviews made me hesitant. I built the Aviatrix MTM last month, and it sounded horrid on my desk.
Anyway, lets talk about the design:
HxWxD: 22"x12"10.5"
Material: 19mm Baltic Birch, Rear panel and bracing MDF.
Volume:~30L
Box: Dual Ported 5x20 cm or sealed (still unsure). 45 degree bevel for diffraction
Arrangement: TMW
Crossover: Active 3-way, using tinysine 7800B. Xo at 200hz (1st order electric), 2khz (2nd order LR). Replaced later with either passive XO or custom DSP & amp option. Time alignment and BSC using DSP.
Usage: Apartment Living room, mostly HT. Flexibility for bedroom type music listening in the future. Desired bass extension at least 40hz. SPL 85 dB at 3 meters. No subwoofer as that is staying in my bedroom and I don't need another 🙂.
It is a 3-way loudspeaker using the Visaton GF200, SB12CACS25-4, and SB26ADC with WG148 waveguide and adapter. I went for the SB12 because it was on sale at madisound and the SB ceramic line seems to be popular although I'm using the 4" in lieu of the 5. I plan on 3D printing a cylindrical enclosure with spherical back for the mid. About 1L of internal volume. I was able to get a pair of SB26ADCs on ebay for $80 and I snapped the GF200s from PE when I had still planned on making the Nimrod. I did a basic test by putting a sb12 and GRS RT1.0 in a carboard box and a GF200 in a second larger box (volume was only about 15L). It sounded better than my current SVS ultra bookshelves even in the ad hoc cabinet in which it was placed. MTW sounded better than TMW although that I suspect was due to the ribbon being above ear level in my test room.
Some questions I have:
Are there any glaring issues with this design based upon what I've detailed?
How do I do about designing bracing? Just put a bunch of fenestrated panels in and hope for the best?
Is there any benefit to using BB for the rear panel? A removable rear panel of MDF would allow me to iterate on the design with regards to ports/minor changes in volume. It would also be cheaper.
The WO24p-4 went on sale for almost the same price on madisound so I also snapped up a pair of those. Not sure if I should go with that instead of the GF200. Virtuixcad shows I can get much deeper extension with the WO24, but I'm not sure it matters given that it was not part of my design goal. I also have seen people say that the WO24 sounds good in midbass. Not sure how applicable that is given the low XO I'm using.
Harbeth uses rear mounted recessed drivers. I figured I could line up the acoustic center of the mid and woofer this way, without using a stepped baffle. Is this possible? How do I make a good chamfer?
I'm going to a workshop tomorrow to get the baffles and bracing CNC'd. I could get baffles for the GF200 and WO24 done for the same cost as 1 pair but that would mean wasting wood. It would also mean not having being able to test whether I prefer TMW or MTW.
Thank you for your help!
P.S. Yes, I do have a measurement mic.
My first design is inspired by the Harbeth 40.3, hence the name. I had initially planned on making the Visaton Nimrod kit, but lack of subjective reviews made me hesitant. I built the Aviatrix MTM last month, and it sounded horrid on my desk.
Anyway, lets talk about the design:
HxWxD: 22"x12"10.5"
Material: 19mm Baltic Birch, Rear panel and bracing MDF.
Volume:~30L
Box: Dual Ported 5x20 cm or sealed (still unsure). 45 degree bevel for diffraction
Arrangement: TMW
Crossover: Active 3-way, using tinysine 7800B. Xo at 200hz (1st order electric), 2khz (2nd order LR). Replaced later with either passive XO or custom DSP & amp option. Time alignment and BSC using DSP.
Usage: Apartment Living room, mostly HT. Flexibility for bedroom type music listening in the future. Desired bass extension at least 40hz. SPL 85 dB at 3 meters. No subwoofer as that is staying in my bedroom and I don't need another 🙂.
It is a 3-way loudspeaker using the Visaton GF200, SB12CACS25-4, and SB26ADC with WG148 waveguide and adapter. I went for the SB12 because it was on sale at madisound and the SB ceramic line seems to be popular although I'm using the 4" in lieu of the 5. I plan on 3D printing a cylindrical enclosure with spherical back for the mid. About 1L of internal volume. I was able to get a pair of SB26ADCs on ebay for $80 and I snapped the GF200s from PE when I had still planned on making the Nimrod. I did a basic test by putting a sb12 and GRS RT1.0 in a carboard box and a GF200 in a second larger box (volume was only about 15L). It sounded better than my current SVS ultra bookshelves even in the ad hoc cabinet in which it was placed. MTW sounded better than TMW although that I suspect was due to the ribbon being above ear level in my test room.
Some questions I have:
Are there any glaring issues with this design based upon what I've detailed?
How do I do about designing bracing? Just put a bunch of fenestrated panels in and hope for the best?
Is there any benefit to using BB for the rear panel? A removable rear panel of MDF would allow me to iterate on the design with regards to ports/minor changes in volume. It would also be cheaper.
The WO24p-4 went on sale for almost the same price on madisound so I also snapped up a pair of those. Not sure if I should go with that instead of the GF200. Virtuixcad shows I can get much deeper extension with the WO24, but I'm not sure it matters given that it was not part of my design goal. I also have seen people say that the WO24 sounds good in midbass. Not sure how applicable that is given the low XO I'm using.
Harbeth uses rear mounted recessed drivers. I figured I could line up the acoustic center of the mid and woofer this way, without using a stepped baffle. Is this possible? How do I make a good chamfer?
I'm going to a workshop tomorrow to get the baffles and bracing CNC'd. I could get baffles for the GF200 and WO24 done for the same cost as 1 pair but that would mean wasting wood. It would also mean not having being able to test whether I prefer TMW or MTW.
Thank you for your help!
P.S. Yes, I do have a measurement mic.
I have only enough experience to give you a couple tidbits.
I have been quite underwhelmed with the performance of any Visaton driver. That satori will blow the visaton away. Also, go with large woofers in a 3 way. I just finished a 8"/5"/Planar T build and realized I would have been better off with a larger woofer because you can cross an 8" all the way up to around 900-1000hz. If you're crossing around 300-400 then might as well go bigger to get more extension.
I have experience 3d printing enclosure as this is how I did it before getting into more multi way stuff.
Use PETG Filament. It has the least "ring" of any of the filaments I tried. Sticking some butyl rubber on the outside and inside of your mid enclosure will lower the frequency of the filament even lower. What you want is the ring to be in a frequency that is lower than your crossover. If you have the space, do not make it a cone or a cylinder. I found some research papers on mid enclosures after the fact and found out that the best thing to use is a wierd shape like a floppy goatskin wine bag. This screws up the back wave frequencies which is what you want. See attached picture (hard to explain in words).
I have been quite underwhelmed with the performance of any Visaton driver. That satori will blow the visaton away. Also, go with large woofers in a 3 way. I just finished a 8"/5"/Planar T build and realized I would have been better off with a larger woofer because you can cross an 8" all the way up to around 900-1000hz. If you're crossing around 300-400 then might as well go bigger to get more extension.
I have experience 3d printing enclosure as this is how I did it before getting into more multi way stuff.
Use PETG Filament. It has the least "ring" of any of the filaments I tried. Sticking some butyl rubber on the outside and inside of your mid enclosure will lower the frequency of the filament even lower. What you want is the ring to be in a frequency that is lower than your crossover. If you have the space, do not make it a cone or a cylinder. I found some research papers on mid enclosures after the fact and found out that the best thing to use is a wierd shape like a floppy goatskin wine bag. This screws up the back wave frequencies which is what you want. See attached picture (hard to explain in words).
Attachments
Thanks for the response! I was leaning toward returning the Visatons. I went for it kinda because it wasn't too expensive, and was used in a proven design, but the WO24P on sale seems like a no brainer. Holding them side by side it's significantly more substantial. The rationale behind using an 8 vs larger was:
1. Price
2. No need for greater output/extension
3. Flexibility with crossover (I see many people stating not X frequency is a crucial frequency and you shouldn't crossover there. X, of course is anything between 100hz and 4000hz.)
4. Similarity to a lot of designs I've heard good things about like Linton, mission 770 etc.
5. It allows me to use 1st order XO if I decide to go passive. Also means that they are both minimum phase and linear, so no faffing about with alignment.
If the mid enclosure should be deformed as such, would it make sense to just melt it with a heat gun and do some bending?
1. Price
2. No need for greater output/extension
3. Flexibility with crossover (I see many people stating not X frequency is a crucial frequency and you shouldn't crossover there. X, of course is anything between 100hz and 4000hz.)
4. Similarity to a lot of designs I've heard good things about like Linton, mission 770 etc.
5. It allows me to use 1st order XO if I decide to go passive. Also means that they are both minimum phase and linear, so no faffing about with alignment.
If the mid enclosure should be deformed as such, would it make sense to just melt it with a heat gun and do some bending?
I understand your rationale. It a makes sense to me. Return the Visatons. Go with the Satori, you'll be happy you did when it's all said and done.
Do you use REW to take measurements? If so, you can see where the drivers begin to distort and base your XO points on that to limit distortion.
Almost forgot, make that mid enclosure at least 4mm thick. The thicker it is, the less it will ring. At this thickness you won't be able to melt the shape. You're printing it anyways so just let the gcode do the work. If you find organic modeling like that out of your skill set I can do it for you. I make aerodynamic parts as my side business so I make those types of shapes a lot in CAD and so it's quite easy for me. It was not easy to learn.
Do you use REW to take measurements? If so, you can see where the drivers begin to distort and base your XO points on that to limit distortion.
Almost forgot, make that mid enclosure at least 4mm thick. The thicker it is, the less it will ring. At this thickness you won't be able to melt the shape. You're printing it anyways so just let the gcode do the work. If you find organic modeling like that out of your skill set I can do it for you. I make aerodynamic parts as my side business so I make those types of shapes a lot in CAD and so it's quite easy for me. It was not easy to learn.
Your design isn't following the pecularities of the Harbeth speaker which limit it's technical performance but increase it's attractiveness to retro enthusiasts. I would suggest a different name.
You haven't chosen your drivers so that they are good fit for the speaker configuration and work well together. Instead you have bought what was on offer which can lead to an OK speaker but almost certainly won't lead to a good one. It's not necessarily a poor move though if the intention is to learn and develop but likely would be if the intention is to stop and keep the speakers.
The best thing to do would be to learn about the physics of bracing and hence understand why DIYers have a tendency to brace the hell out of everything whereas professionals designing speakers with the highest technical performance only brace where required and give more attention to other aspects of cabinet design.
BB is a nicer material to work with but the pros and cons w.r.t. to technical performance will depend on the details of the speaker design. If you intend to evolve your design then I would suggest starting with what is easiest.
Why not learn from using both? Try changing the crossover frequency and slope to improve the performance. The Harbeth 12" driver crosses at a significantly higher frequency and you are using an 8" driver which the Harbeth uses for the midrange and crosses at a few kHz!
It's an option but uncommon because to install and remove the drivers requires rear access and hence a removable panel. For a first speaker you intend to mess about with it seems a reasonable choice. If you are going to be swapping midwoofers a removable front may be preferable to a removable rear.
Good because this is pretty much essential if you are going to identify, understand and improve your first design which, like most people's, is almost certainly going to have some disappointing aspects you didn't know about or fully consider. Mine certainly did. Good luck.
I had both baffles CNC'd shortly after posting. I started cutting the other panels with a circular saw and quickly came to the realization that a table saw would be needed. Currently working on getting that done. I also realized that the sub-10" wo24 does not really give the Harbeth look I expected. The project is an attempt to blend interior design with good audio design. Or to be more specific learn which WAF parameters are acoustic, and which are not. @andy19191 I will probably call these the Debutantes
In the meantime, I've been using them as open baffle and have been impressed with the sound so far although tweaks are needed. Currently crossing at 300hz and 3khz butterworth first order. I found that: A. Crossing at 150hz is a bit unnecessary and does not provide any noticeable improvements. B. The poor 4" midrange has significant excursion at higher volumes. C. The WO24p is still well behaved 3 octaves up which allows for a first order XO.
Non-gated measurements at 1m in room show a 6 dB rise in the mid bass centered around 400hz, a small dip around 1k and then falling SPL past 3khz. I'm thinking that the midbass hump is a room effect and the 1k dip is a surround resonance from the woofer. The falling hf response is due to the horn gain from the waveguide. I think a high shelf filter is warranted, but I have not yet determined how to implement in sigma studio.
I also picked up some bliemsa m74S for a good price and apparently, they will fit in the cut-out in the baffle with minimal alteration. Will probably compare head-to-head with the current midrange.
@Bmsluite I might take you up on that offer to design the mid enclosure.
In the meantime, I've been using them as open baffle and have been impressed with the sound so far although tweaks are needed. Currently crossing at 300hz and 3khz butterworth first order. I found that: A. Crossing at 150hz is a bit unnecessary and does not provide any noticeable improvements. B. The poor 4" midrange has significant excursion at higher volumes. C. The WO24p is still well behaved 3 octaves up which allows for a first order XO.
Non-gated measurements at 1m in room show a 6 dB rise in the mid bass centered around 400hz, a small dip around 1k and then falling SPL past 3khz. I'm thinking that the midbass hump is a room effect and the 1k dip is a surround resonance from the woofer. The falling hf response is due to the horn gain from the waveguide. I think a high shelf filter is warranted, but I have not yet determined how to implement in sigma studio.
I also picked up some bliemsa m74S for a good price and apparently, they will fit in the cut-out in the baffle with minimal alteration. Will probably compare head-to-head with the current midrange.
@Bmsluite I might take you up on that offer to design the mid enclosure.