• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

DIY desktop speakers the ASR way - help needed!

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
All I want from a desktop speaker is these things, and I am struggling to find anything that satisfies the requirements:
  1. Sits on a desk, either side of a computer monitor
  2. Much less visually dominant than an LSR305p. Think JBL 104.
  3. Even frequency response when seated normally - say a 1m equilateral triangle for speakers and listener, with ears 50cm above desk height
  4. The FR not to change much when I move around at the desk a bit, say 50cm sphere around main listening position. If the usable area is bigger then even better!
  5. Whatever distortion there is to be consistent, not jumping up and down with FR
  6. Some bass extension (60Hz f3 ideally)
  7. Moderate volume capability (86dBSPL/1m)
I spend a lot of time at my desk doing lots of things that are not music production, and really I am struggling to find anything that meets those requirements.

#3 rules out all normal 2-ways. #2 rules out all the bigger coaxials. #4-7 as a combination probably rule out the iLoud Micro MTM.

So we find ourselves in the DIY forum :)

I'm thinking that in theory this would work:
  • Line array of BMR drivers
  • Opposing 6" low profile sub drivers on either side
  • DSP to turn the line array into a Constant Beamwidth Transducer like the Dayton CBT24, only a lot smaller
A pair of Tang Band woofers (W6-2253S) and 6 to 8 Tectonic BMRs (TEBM36S05) in a sealed box 80mm wide, 200mm deep and 450mm tall with some DSP help should hit all the requirements above. No idea how to do 8 channels of DSP, 2x100W of bass amplification and 6x10W of mid/high amplification per channel though.

Assuming all the DSP and amplification is external, is this nuts or do I just need to make some better driver choices?
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
Here's a quick render of the front and the right woofer in Speakerboxlite. They need to be offset slightly to keep the width to 80mm. The BMRs are 65mm square and would run up the front. I've assumed a panel thickness of 5mm. SBL actually shows a 50Hz F3 is possible (excursion limited) if total system loudness is limited to 96dBSPL/1m, so that's way above requirements!

Screen Shot 2021-01-07 at 3.46.57 pm.png
 

Sharpi31

Active Member
Forum Donor
Joined
May 20, 2020
Messages
125
Likes
305
I like ambitious projects, but 16 channels of DSP and 16 channels of amplification is going to be expensive. There will also be a lot of modelling & measurement required (additional cost?) and a large time commitment. I’d be doing a quick cost estimate and then look at what the same money will get you from Genelec etc. Then decide if you think your design will be better.......
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
Thanks @Sharpi31. Lets ignore the time cost for now, DIY is never about saving time!

The materials cost (drivers, cabinets, cables, connectors) is about 500USD for a pair. The electronics is indeed going to be quite a lot when done as a prototype, but probably quite cheap if it can ever be integrated - it only needs to do delay and volume and a couple of bands of PEQ, then amplification.

But in return we get things no normal speaker including can deliver:
  • CBT design means no desktop reflection issues - in fact it uses them to advantage
  • BMRs give almost 180 degrees of even dispersion to 15KHz
  • Very small desktop footprint
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
For comparison the cheapest coaxial Genelecs I can buy in Australia are USD7000 for a pair for the 8341A SAM, and those are pro devices that would have desktop reflection issues if just sat either side of a screen.
 

Sharpi31

Active Member
Forum Donor
Joined
May 20, 2020
Messages
125
Likes
305
It’s a very interesting project, and I like the idea of a desktop CBT array, but I don’t think I’m qualified to offer any real guidance :)

In case it helps, I think you may be able to use a Polycom C16 processor here.... (for development, if not final implementation) these were designed for conference room audio but have 16 outputs plus EQ/delay etc. A lot of businesses have been decom’ing these lately and the C16 often sells for $50 or so on popular auction sites. Maybe a low cost way of getting started?
 

hex168

Senior Member
Joined
May 29, 2020
Messages
396
Likes
338
Rick Craig did a desktop CBT with the delay done by a physical curve in the cabinet. Don't know if that approach would be easier or harder for you.

The BMR you selected would meet your stated output requirements, but you'd get a lot more out of the TEBM46C20N-4B if you preferred to go that route:
https://www.parts-express.com/tectonic-tebm46c20n-4b-bmr-3-full-range-speaker-4-ohm--297-2157

And, while I'd have more fun learning from your ambitious project, could you meet your needs with a pair of these?
https://www.parts-express.com/dayto...ver-with-3-4-silk-dome-tweeter-8-ohm--295-384
Or these:
https://www.parts-express.com/pedocs/specs/294-6048--bc-4cxn36-spec-sheet.pdf
Or these:
https://www.madisoundspeakerstore.com/coaxial/sb-acoustics-sb12pfcr25-4-coax-4-coaxial-4-ohms-round/
 

MrPeabody

Addicted to Fun and Learning
Joined
Dec 19, 2020
Messages
657
Likes
942
Location
USA
All I want from a desktop speaker is these things, and I am struggling to find anything that satisfies the requirements:
  1. Sits on a desk, either side of a computer monitor
  2. Much less visually dominant than an LSR305p. Think JBL 104.
  3. Even frequency response when seated normally - say a 1m equilateral triangle for speakers and listener, with ears 50cm above desk height
  4. The FR not to change much when I move around at the desk a bit, say 50cm sphere around main listening position. If the usable area is bigger then even better!
  5. Whatever distortion there is to be consistent, not jumping up and down with FR
  6. Some bass extension (60Hz f3 ideally)
  7. Moderate volume capability (86dBSPL/1m)
I spend a lot of time at my desk doing lots of things that are not music production, and really I am struggling to find anything that meets those requirements.

#3 rules out all normal 2-ways. #2 rules out all the bigger coaxials. #4-7 as a combination probably rule out the iLoud Micro MTM.

So we find ourselves in the DIY forum :)

I'm thinking that in theory this would work:
  • Line array of BMR drivers
  • Opposing 6" low profile sub drivers on either side
  • DSP to turn the line array into a Constant Beamwidth Transducer like the Dayton CBT24, only a lot smaller
A pair of Tang Band woofers (W6-2253S) and 6 to 8 Tectonic BMRs (TEBM36S05) in a sealed box 80mm wide, 200mm deep and 450mm tall with some DSP help should hit all the requirements above. No idea how to do 8 channels of DSP, 2x100W of bass amplification and 6x10W of mid/high amplification per channel though.

Assuming all the DSP and amplification is external, is this nuts or do I just need to make some better driver choices?


I'm a little confused by some of the requirements. You say you don't want the speaker to be visually dominant, but then propose what amounts to a tall tower sitting on the desk. The proposed enclosure will have the advantage of having a narrow front, which is important for a desktop speaker, but it will be visually conspicuous.

What confuses me more is that you say that you want the FR to remain stable as you move around, which is certainly a good requirement to identify, but then you reject a small concentric because you think it will violate the rule about not being visually dominant, which is not consistent with the preference for a desktop speaker 18" tall. You say that the rule for stability of FR rules out a small 2-way, but don't explain exactly why this would be, and then suggest a line array, but don't explain exactly how this would be the solution to the problem of stability of FR. This leaves me puzzled. The line array is going to narrow the vertical dispersion unless you use a varying time delay scheme to achieve phase differences from the middle element to the ends of the column. The complexity of a solution to any problem needs to be commensurate to the problem itself, and something like this would an extreme violation of that principle. The only issue with the small 2-way would be the suck-out at the crossover point when you are off-axis in the vertical polar plane. As long as the woofer and tweeter are adequately close together, this shouldn't be a problem unless you sit vertically off-axis, i.e., your ears are higher than the speakers, but this can be corrected by placing the small 2-way speakers on the desk and tilting them back. If this solution wouldn't be good enough, then a small concentric is the manifest solution, but if the small concentric will be wider than you'd like, then a substitute solution, that will give much of the benefit of the concentric but without the width, would be an MTM arrangement using two small midrange drivers and a tweeter with a very small flange, so that the distance separating the two midrange drivers will be as small as it could be made. You could take this a step further and put two midrange drivers above the tweeter and two below, but in this case I would be inclined to place an additional 1st-order low-pass on the top and bottom midrange drivers. There are two reasons for doing this, first, the top and bottom midrange drivers will be too far from the tweeter, and second, they will be too far from each other. By using an additional low-pass on the top and bottom midrange drivers, so that they only operate below 700 Hz or thereabouts, they will not interfere with each other in the frequency range from about 1 kHz to wherever the crossover to the tweeter is, and will not interfere with the tweeter in the vicinity of the crossover point. This would defeat the line array effect, which I think makes more sense. I think some thought needs to be given to whether, aside from consideration of packaging and conspicuity, the concentric driver or the line array is preferable. You first mentioned the desirability of the concentric and then rejected it for reasons that weren't very clear, then went straight to the line array, which is almost the opposite.

Of course you could still use a side mounted woofer, maybe even a pair.

If I weren't inclined to build something, I would consider the small Elac coaxial. (UB51?)
 
Last edited:

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
721
Likes
359
I like the idea, interested to see the result! I am thinking how about MTM in the front and a flat woofer at the sides? Do you really want CBT or normal speakers are fine too?
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
So when I say "The ASR way" I mean using science to solve the problem, which I think CBTs do for desktop monitors. The question is are they practical in this application?

The desktop CBT by Rick Craig (@Selah Audio) is really close, but wow the curvature is huge! Here's the PE Tech Talk thread on Rick's build: Post 59. Don Keele's research shows that using a curve to do the delay is better than time delaying a flat array, but I _think_ this is due to narrowing dispersion of conventional drivers at high frequencies. It does save a lot on DSP though, so I'd need to do the math and see what the curvature would need to be like for a 400 or 450mm tall array.

@MrPeabody thanks for the feedback! The CBT design delays and shades the drivers in the array, creating a virtual curved line array which actually exploits the desk reflections. On a desk even an MTM setup or a coaxial suffers from either quite tricky reflection issues or a narrow sweet spot. Coaxials do solve for most polar issues when freestanding, but can't remove desk issues unfortunately :-(

As for perceived size, 8cm wide speakers next to an existing 27" monitor would be much less obtrusive for me than the alternatives. The JBL 104 is 15cm wide and 25cm tall, so presents a 375cm^2 face to the user. 8cm x 45cm is 360cm^s.
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
Here's a 2018 writeup from noaudiophile.com on a later iteration of Rick's desktop CBTs (about 1/3 of the way through the page, or search for 'desktopia'): Lone Star 2018. A BMR array wouldn't have the lobing issues he heard on these, but would need the side woofers.
 

tomchr

Addicted to Fun and Learning
Audio Company
Forum Donor
Joined
Nov 5, 2018
Messages
984
Likes
2,607
Location
Calgary, Canada
Have you considered some of Frugalphile's plans? My "casual listening while working" desktop speakers are mMar-Ken6 with the Alpair 6P. The mMar-Ken6 was originally intended for the Mark Audio Alpair 6 (metal cone), but if you remove the spacer in the port it works well with the paper cone version of the Alpair 6 (called 6P).

I don't know if that meets your criteria for "the ASR way" but it's an option.

Tom
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
Thanks Tom, those look interesting, though a bit big as they are around the size of the LSR305 I think. How far away do you sit from them?

I guess what I mean by "the ASR way" is looking beyond received wisdom for solutions that solve specific problems well. The Evidence based speaker designs thread is a nice example. Desktop speakers seems like an area which is in need of that approach.
 

tomchr

Addicted to Fun and Learning
Audio Company
Forum Donor
Joined
Nov 5, 2018
Messages
984
Likes
2,607
Location
Calgary, Canada
17x18x30 cm give/take. I would put that in the "small" category. Not tiny. They're larger than PC speakers but you wanted an f3 of 60 Hz... I doubt they go that deep, actually. They might do an f10 of 60 Hz. I think I measured them at one point, but it doesn't look like I stored the measurements.

I'm usually 1-1.5 m away. They sit on a shelf behind my monitor. Or, rather they used to. I used to have my workstation in the basement lab but have since then moved upstairs. I've attached a picture of my old workstation.

Tom
 

Attachments

  • DSCF4268.jpg
    DSCF4268.jpg
    165.6 KB · Views: 215
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
Nice setup! I'm looking to work something out that will be on-desk though, not on a shelf or brackets. I really want to stop fighting the desk and do something that works with it. Coaxials are better than two-way (MT or MTM) speakers in most aspects for desktop use, but the bounce off the desk is still harsh and they have their own issues with response and diffraction effects.
 

tomchr

Addicted to Fun and Learning
Audio Company
Forum Donor
Joined
Nov 5, 2018
Messages
984
Likes
2,607
Location
Calgary, Canada
I think you'll have a hard time reaching 60 Hz f3 with a desktop speaker. You'll have to trade on that or put a mumble box (aka subwoofer) under the desk. Maybe a compact studio monitor...?

Tom
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
I think you'll have a hard time reaching 60 Hz f3 with a desktop speaker. You'll have to trade on that or put a mumble box (aka subwoofer) under the desk. Maybe a compact studio monitor...?
Tom

Oh I'm assuming this whole project is active with DSP. A pair of Tang Band woofers (W6-2253S) per speaker in a small box is only 5dB down at 60Hz so not a lot of correction needed anyway.

Screen Shot 2021-01-16 at 4.05.53 pm.pngScreen Shot 2021-01-16 at 4.10.30 pm.png
 

tomchr

Addicted to Fun and Learning
Audio Company
Forum Donor
Joined
Nov 5, 2018
Messages
984
Likes
2,607
Location
Calgary, Canada
Isn't that a fairly large box, though?

Tom
 
OP
M

mcdn

Addicted to Fun and Learning
Forum Donor
Joined
Mar 7, 2020
Messages
559
Likes
780
It is a bit tall (450mm) and deep (200mm) but it's only 80mm wide. Total box volume is 7.2L vs 9.1L for the mMar-Ken6, and now you mention it the interior volume would probably be about the same at 4.5L as I've assumed thinner walls.

I suppose for me the form factor makes a big difference. For speakers that sit next to a computer monitor on a desk I don't feel the height is an issue as it's not like anything would go above them - like a skyscraper, once you've committed to the footprint you night as well go up, and my ears are 50cm above the desk anyway. Likewise with depth within reason.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
@mcdn the concept is solid (indeed I've considered building something similar myself a number of times). The only significant downside I see to such a design vs a conventional monitor is that the apparent source of the sound will be somewhere down just below the level of the desk.

Personally I would curve the array to generate the required delays, and use passive networks (resistors) to implement the shading, with the only active XO point being between the woofer and the array. This will reduce costs significantly (without any cost in performance).

I'm not sure about the driver selections you've made there. Have you measured the BMRs already to verify their performance? And what's your intended XO frequency?
 
Top Bottom