Thanks for linking this post! I had found it just a few days ago during my endless online searches =)
Are you going to take up the mini-CBT project again?
Manuele
If I remember correctly, I modeled them in NX11 or Inventor.
The shape was designed to be as stiff as possible (even if I'm considering a MK2 with different internal geometry).
I simulated the eigenvalues response in Comsol but haven't yet measured the resonances. The cabinet are quite deaf to the "nock test".
Manuele
Lately, several advances have been made on the manufacturing side.
So I decided to 3D print some PLA washers that I wanted to use together with a thin layer of vibration-insulated material.
In short, the boxes will be connected in the following way: rubber disc and the PLA disc will be glued on one of the boxes, while the other box will be stuck on the PLA disc using as a plug.
This would avoid piercing all the boxes to pass through a threaded bar (which by the way should be made of non-magnetic material so as not to affect the driver) and therefore all the resulting problems in terms of alignment and air tightness.
Another way would be to only use PLA discs without rubber, but use a layer of rubber in the remaining contact surfaces.
What do you think?
Thanks,
Manuele
P.S. I invested some dollars for 16AWG OFC speaker cable for the internal connection, and a big 12 AWG OFC wire for the whole speaker connection. All of these for 10$ overall
- All boxes have been sanded, grouted, sanded again.
- Two layers of filler primer were sprayed.
- Surfaces prepared for final painting.
- Brass threaded inserts were installed in the boxes to screw the drivers.
- The interface discs between the various boxes have been 3D printed.
So I decided to 3D print some PLA washers that I wanted to use together with a thin layer of vibration-insulated material.
In short, the boxes will be connected in the following way: rubber disc and the PLA disc will be glued on one of the boxes, while the other box will be stuck on the PLA disc using as a plug.
This would avoid piercing all the boxes to pass through a threaded bar (which by the way should be made of non-magnetic material so as not to affect the driver) and therefore all the resulting problems in terms of alignment and air tightness.
Another way would be to only use PLA discs without rubber, but use a layer of rubber in the remaining contact surfaces.
What do you think?
Thanks,
Manuele
P.S. I invested some dollars for 16AWG OFC speaker cable for the internal connection, and a big 12 AWG OFC wire for the whole speaker connection. All of these for 10$ overall
The first attempt will be the classical 3 x series/ 3x parallel and I’ll take some measurements on axis and off axis at different height and distance.
After this outcome, I’ll evaluate if another possible scheme will be better.
According to all the simulations I made, the results don’t change a lot….I have realised a standard simulation with the weird shape of my speakers can’t provide accurate results so it will be much better to proceed step by step using the actual measurements ( as you advices as well 😃).
Stay tuned! 🤣
Manuele
After this outcome, I’ll evaluate if another possible scheme will be better.
According to all the simulations I made, the results don’t change a lot….I have realised a standard simulation with the weird shape of my speakers can’t provide accurate results so it will be much better to proceed step by step using the actual measurements ( as you advices as well 😃).
Stay tuned! 🤣
Manuele
I'd recommend following this thread: https://www.diyaudio.com/community/...heir-behavior-through-simple-modeling.388279/
Some update: finally the external finish of all 18 cabinets has been completed.
The speakers will have an "industrial" look so we decided to paint them in "aged copper" supported by a frame made of metal tubes (perhaps filled with "playground" sand...).
Probably the base of the frame will house the subwoofers (one for each speaker) which will also have the function of counterweight.
The various enclosures will be aligned by means of the PLA washers (manually adapted to fit precisely inside the appropriate slot, already tested).
Epoxy will be used to glue the washers to the enclosures and a (light) layer of silicone will be spread over the flat areas to dampen vibrations.
The internal volume will be filled with Polyfill but I have read some interesting things about the use of melamine as damping material...Any experience with this materials? Any advice?
Unfortunately, over the weekend, I spent a lot of time testing and repairing the threaded inserts and the PLA washers so I was only able to paint the enclosures but I couldn't glue anything...hope to do something today .
I can't wait to start measuring, maybe outside, and start doing something to improve the sound as much as possible.
If I can I would like to create a Spinorama, but I don't know if I will have the patience to do it
Manuele
The speakers will have an "industrial" look so we decided to paint them in "aged copper" supported by a frame made of metal tubes (perhaps filled with "playground" sand...).
Probably the base of the frame will house the subwoofers (one for each speaker) which will also have the function of counterweight.
The various enclosures will be aligned by means of the PLA washers (manually adapted to fit precisely inside the appropriate slot, already tested).
Epoxy will be used to glue the washers to the enclosures and a (light) layer of silicone will be spread over the flat areas to dampen vibrations.
The internal volume will be filled with Polyfill but I have read some interesting things about the use of melamine as damping material...Any experience with this materials? Any advice?
Unfortunately, over the weekend, I spent a lot of time testing and repairing the threaded inserts and the PLA washers so I was only able to paint the enclosures but I couldn't glue anything...hope to do something today
.I can't wait to start measuring, maybe outside, and start doing something to improve the sound as much as possible.
If I can I would like to create a Spinorama, but I don't know if I will have the patience to do it
Manuele
Hello,
One "enclosure" is finally completed!
I spent some time for align every single box but I did it with satisfactory results. First of all I glued together only 3 boxes, to be accurate and test the adhesion of the glue used for the purpose (recommended for similar applications).
The bonding between the modules is very strong. The silicon layer seems working both in terms of adhesive and vibration damping as well. I will confirm my feelings after some measurements of the resonances as well as the directivity during the week-end. I need to check how to capture the impedance response...
Hope me some good results!
I started to work on cabling as well; I would like to leave all the wires out of the cabinet to highlight the "industrial style" of this speakers.
I was thinking to solder the speaker cables (16AWG OFC wires) directly onto the TC9s. I don't really the actual difference between soldered vs crimped connection in terms of signal losses. I found diverse opinions on this forum as well as other sources but from an engineering point of view it seems really a low importance point to take care about...
Another much discussed topic would seem to be the damping material. On my first test I used the Polyfill (quite a lot!) to fill completely the interior of the cabinet. I don't know if it's the best solution, but it's definitely the cheapest!
Any thoughts @wesayso ? I have low experience on this topic so every comment is more than welcome!
Next steps in the pipeline:
Manuele
One "enclosure" is finally completed!
I spent some time for align every single box but I did it with satisfactory results. First of all I glued together only 3 boxes, to be accurate and test the adhesion of the glue used for the purpose (recommended for similar applications).
The bonding between the modules is very strong. The silicon layer seems working both in terms of adhesive and vibration damping as well. I will confirm my feelings after some measurements of the resonances as well as the directivity during the week-end. I need to check how to capture the impedance response...
Hope me some good results!
I started to work on cabling as well; I would like to leave all the wires out of the cabinet to highlight the "industrial style" of this speakers.
I was thinking to solder the speaker cables (16AWG OFC wires) directly onto the TC9s. I don't really the actual difference between soldered vs crimped connection in terms of signal losses. I found diverse opinions on this forum as well as other sources but from an engineering point of view it seems really a low importance point to take care about...
Another much discussed topic would seem to be the damping material. On my first test I used the Polyfill (quite a lot!) to fill completely the interior of the cabinet. I don't know if it's the best solution, but it's definitely the cheapest!
Any thoughts @wesayso ? I have low experience on this topic so every comment is more than welcome!
Next steps in the pipeline:
- Figure out whether to crimp or solder the drivers in a 3x3 (series / parallel) configuration
- Check out the outdoor and in-room performances in a wide range of angles and distance from the speaker
- In case of unsatisfactory performance, change the wiring and repeat the measurements
- Finalize the wiring/ crossover
- Build the frame from metal pipes
- ENJOY THE SOUND!
- Start thinking about how EQ / DPS to improve the in-room response
- Start thinking about the next project... perhaps a nice pair of 3D printed subwoofers similarly shaped or start a brand new 25-boxes array (VERY LONG TERM PROJECT)
Manuele
I've used polyfill once, I wasn't impressed by it's ability to absorb anything... maybe I did something wrong .
What worked for me is the itchy and nasty fiberglass. But as it is pretty nasty (and very cheap) I used wool felt
to seal it off from the driver. The felt + fiberglass worked better than each of the materials on their own. I've
used the felt on the walls of the enclosure too, which worked well to get rid of slight impedance bumps. You'll
read that damping material does it's best job in the middle of the enclosure. While that certainly is true, it's quite
telling how felt on the wall does something that the damping material in the middle cannot get rid of. That's the
reason of running these impedance tests for me. I hate going in blind. I like to confirm what does what.
A lot can be seen while taking impedance measurements of the driver(s) in the enclosure. It's much like putting a
microphone in there. If there is a resonance, you'll see it in the impedance trace. Popping up like little bumps that
are not present if you run an impedance test of a driver in free air. You can also see if you put too much damping
material inside the enclosure. As the impedance peak will stop moving down and to the left of the screen and move
back up and to the right again.
Impedance tests can be made with REW and a simple and cheap impedance rig. I made mine from a (free) resistor
and a couple of old mini-jacks to fit my old laptop. This is all that is needed...
One more note on that itchy fiberglass stuffing, it's the fluffy yellow type, used to insulate walls in houses etc.
It is by far the best broadband material I have tested. Meaning it works across a wide range of frequencies.
For mid-range the Angel hair (from Twaron) worked very well, but that is super pricey. It's like long hair wool,
but even thinner strands that easily absorb energy (by turning it into heat). A little goes a long way.
I've read good things about Melamine, but never compared it myself.
.What worked for me is the itchy and nasty fiberglass. But as it is pretty nasty (and very cheap) I used wool felt
to seal it off from the driver. The felt + fiberglass worked better than each of the materials on their own. I've
used the felt on the walls of the enclosure too, which worked well to get rid of slight impedance bumps. You'll
read that damping material does it's best job in the middle of the enclosure. While that certainly is true, it's quite
telling how felt on the wall does something that the damping material in the middle cannot get rid of. That's the
reason of running these impedance tests for me. I hate going in blind. I like to confirm what does what.
A lot can be seen while taking impedance measurements of the driver(s) in the enclosure. It's much like putting a
microphone in there. If there is a resonance, you'll see it in the impedance trace. Popping up like little bumps that
are not present if you run an impedance test of a driver in free air. You can also see if you put too much damping
material inside the enclosure. As the impedance peak will stop moving down and to the left of the screen and move
back up and to the right again.
Impedance tests can be made with REW and a simple and cheap impedance rig. I made mine from a (free) resistor
and a couple of old mini-jacks to fit my old laptop. This is all that is needed...
One more note on that itchy fiberglass stuffing, it's the fluffy yellow type, used to insulate walls in houses etc.
It is by far the best broadband material I have tested. Meaning it works across a wide range of frequencies.
For mid-range the Angel hair (from Twaron) worked very well, but that is super pricey. It's like long hair wool,
but even thinner strands that easily absorb energy (by turning it into heat). A little goes a long way.
I've read good things about Melamine, but never compared it myself.
Last edited:
Hello,
I finally completed one speaker and made several measurements both in-room and outdoor.
My first set of measurements are taken with the speaker without any kind of filling materials and the speaker only attached with 2 of 4 screws for a quick first setup. No damping layer between drivers and boxes as well...very rough in few words.
I measured only "on-axis" but at three different heights: lower driver (27" / 69 cm), center (40" / 102 cm) and max height of the speaker (53" / 135cm ).
I checked the frequency response for several distances, from nearfield (12" / 30cm) up to 4.5m (180") to verify the near field to far field transition.
This is my measurement setup (very improvised!). You can see that the mic is very close to a "vertical barrier" that can perhaps alter the responses. I'll do in a more appropriate setup using a long tube to avoid any kind of reflections.
Find below some measurements:
Some thoughts:
Next steps in the pipeline:
Manuele
I finally completed one speaker and made several measurements both in-room and outdoor.
My first set of measurements are taken with the speaker without any kind of filling materials and the speaker only attached with 2 of 4 screws for a quick first setup. No damping layer between drivers and boxes as well...very rough in few words
.I measured only "on-axis" but at three different heights: lower driver (27" / 69 cm), center (40" / 102 cm) and max height of the speaker (53" / 135cm ).
I checked the frequency response for several distances, from nearfield (12" / 30cm) up to 4.5m (180") to verify the near field to far field transition.
This is my measurement setup (very improvised!). You can see that the mic is very close to a "vertical barrier" that can perhaps alter the responses. I'll do in a more appropriate setup using a long tube to avoid any kind of reflections.
Find below some measurements:
Some thoughts:
- The measurements made indoor are quite similar to the ones made outdoor (after filtering the impulse).
- The peaks and dips at 800Hz and 1,5 kHz and 2 kHz are recurrent for all the measurements. They look weird to me, perhaps is my measurement setup (very artisanal... the microphone can capture several reflection from the 3-foor stand) or something that I still need to understand (some idea?). They appear in the simulations as well, perhaps it's the driver itself but they seem to me too sharp...
- The last measurements are not completely coherent because of the different in distance, but very similar so this indicate the far field has already be reached.
- Apart from the high frequency dips in the 0.3m and 1m measurements, the response seem quite good to my eyes (unsmoothed), much better than the simulated values.
Next steps in the pipeline:
- Assembly the second speaker with a metal frame
- Understand which damping materials to use for my speaker
- Crimp or solder?
- Re-measure with a variety of horizontal and vertical responses.
- Start thinking about how EQ / DPS to improve the in-room response
- Start thinking about the next project... perhaps a nice pair of 3D printed subwoofers similarly shaped or start a brand new 25-boxes array (VERY LONG TERM PROJECT)
Manuele
Weird peaks...i do analytical chemistry by trade...and we see these often. Its positive and negative peak coliding close to each other. Like you have one reinforcing and one negating driver contributing at the same frequency. One comes in positive phase, one in negative phase, thus you get up and down sharply. Line array is multiple point sources interacting with each other. Hence the peaks. I am sure you would get these at different frequency if you move mic up or down, or closer vs further. I am no expert, but it seems ok to me. It will smooth out in the room full of reflections.
Great work so far.
Great work so far.
I think the results look pretty normal and comply with the sims. Which would make me rethink running these
arrays without any filter.
I think the 'needle like' directivity we get at 3.5 meter really is not that useful...
Unless all the people in the household are exactly the same length.
Even though, checking plus/minus 10 cm looks kind of ok-ish (within +/- 1 dB up to 10 K):
arrays without any filter.
I think the 'needle like' directivity we get at 3.5 meter really is not that useful...
Unless all the people in the household are exactly the same length
.Even though, checking plus/minus 10 cm looks kind of ok-ish (within +/- 1 dB up to 10 K):
Actually when I designed these speakers I was inspired by the 9-cell superconducting cavities I work with
.Later I discovered the existence of the Scaena!
Manuele