Small 2-way speakers with linear on-axis and power response characteristics (Scan Speak and SB Acoustics drivers). H&V off-axis measurements included

[HighFutility]

Can anyone recommend something to seal the bond between the drivers and the MDF?

 

[carbidetooth]

Can anyone recommend something to seal the bond between the drivers and the MDF?

I like this foam tape.

McMaster-Carr

McMaster-Carr is the complete source for your plant with over 595,000 products. 98% of products ordered ship from stock and deliver same or next day.

www.mcmaster.com

 

[HighFutility]

Nope i used CNC
Btw. i created pcb mount if you have 3d printer and some M3 5mm heat inserts (3mf in mechano23-xover-mount.zip). I recommend to use foam sound dampening touching the pcb to eliminate vibration transmitted to the pcb.

I already put them together without sound sound dampening (waiting for order) and the sounds good but vocals are somehow very directional (almost hollow like). Its little bit strange. Maybe it will help sound dampening or i made mistake somewhere.

I just quickly tested mine out before veneering them. I had the exact same thoughts. The vocals sounded hollow like. Im going to check my PCB and make sure I input everything correctly.

 

[celstark]

I've hit many, but not all of the pages here... Has anyone tried a 3D print of the cabinet? I've modeled one up, tweaking the dimensions slightly to a) fit on my printer in 2 parts and b) accommodate some more extensive bracing and a thicker front baffle.

 

[somebodyelse]

I've hit many, but not all of the pages here... Has anyone tried a 3D print of the cabinet? I've modeled one up, tweaking the dimensions slightly to a) fit on my printer in 2 parts and b) accommodate some more extensive bracing and a thicker front baffle

There are a couple of other threads about 3D printed cabinets, including links to some tests people have done. IIRC at infill >50% you get similar performance to various woods. Another option is to print an inner and outer shell and fill the gap with a plaster/PVA mix. You can either print the shells separately with spacers, or use an infill you can pour through, like gyroid with the right density. A similar technique should work with sand or similar, but I haven't seen measurements for that.

 

[xPakrikx]

I've hit many, but not all of the pages here... Has anyone tried a 3D print of the cabinet? I've modeled one up, tweaking the dimensions slightly to a) fit on my printer in 2 parts and b) accommodate some more extensive bracing and a thicker front baffle.

Yes, at first I was interested but I didn't have the time or the filament to spare. I can actually print full cabinet on my voron 2.4 350. In future i want do some experiments with infill like making actual internal volume just with inflill, so no walls, or optimize infill density is some places/patterns. But for this i need to learn simulations. But about your question, yes 3d printed cabinet could work just fine, maybe you can fill walls with plaster+pva glue mix.

 

[celstark]

Thanks @somebodyelse and @xPakrikx -- I've seen a number of these and have been working with a test box to test out a few materials / shapes / infills, etc. I'm still quite new to the CAD aspects, so that box I drew up was pretty basic. With printing, we have so many options for the design that I was hoping someone more skilled might have done up an enclosure. Without going back to the crossover, we'd need to keep the volume and driver spacing set. The design seems aimed at minimizing the baffle, but that would still give us a lot of options in an 8L box. Someone with more skills at 3D CAD could really do something fun.

 

[wineds]

I looked at this some time ago and put together a quick and dirty model less baffle. Slicing it with 7 walls and 40% infill and about 5 top/bottom walls I recall it used more than 1kg of filament. This means maybe using 3kg rolls which complicates things a bit as some of the more exotic filaments may not be available in that case.

 

[Avi56]

Thanks. I contacted the PCB vendor (JLCPCB) and they said a public account was not possible. If anyone knows otherwise I'm all ears. However its not that hard to order yourself. Just setup an account and upload the RAR file. With fastest shipping I had them in just over 4 days. I have uploaded the files and link to JLCPCB ordering instructions to github here :

Mechano23-xover/README.md at main · wineds/Mechano23-xover

Mechano23 xover. Contribute to wineds/Mechano23-xover development by creating an account on GitHub.

github.com

View attachment 398180

I’m a new member building the Mechano23 with wineds’ PCB. I have a question about the nomenclature of the capacitors. Is C3 27uF (per wineds) or 18 uF (per XMechanik’s first post)? Similarly, what are the values for C2 and C4 on the PCB? The nomenclature for the rest of the components are consistent. Thanks.

 

[Groove01]

I’m a new member building the Mechano23 with wineds’ PCB. I have a question about the nomenclature of the capacitors. Is C3 27uF (per wineds) or 18 uF (per XMechanik’s first post)? Similarly, what are the values for C2 and C4 on the PCB? The nomenclature for the rest of the components are consistent. Thanks.

I had the same question. It’s not xmechaniks first post. I did the same thing - lol. Luckily I caught it early.

Wined provided me with this schematic and is what he based the board on. This is what I used.

I also cross referenced with the pic he posted of his x-over build.

 

[wineds]

Thanks groove1. That's correct. I've since updated GitHub and added that schematic to clarify. The PCB was designed based on the schematic.

 

[Avi56]

I’m still confused. The schematic shows seven capacitors but the PCB has six?

 

[wineds]

Ignore c7. Schematic originally showed two capacitors (c6 + c7) in series to yield a total of 2uf. I'll add a BOM to the GitHub when I get a chance.

 

[XMechanik]

I added a note to the first post that component numbering in VituixCad and XMachina schematics may differ.

The XMachina schematic was included there as a curiosity, I was a bit surprised that optimizer managed to create an E24 component where needed with only the E12 list available. (C6 and C7).

 

[wineds]

I added a note to the first post that component numbering in VituixCad and XMachina schematics may differ.

The XMachina schematic was included there as a curiosity, I was a bit surprised that optimizer managed to create an E24 component where needed with only the E12 list available. (C6 and C7).

Thanks. I've uploaded a JPG of the KiCAD schematic and also a simple excel BOM here :

GitHub - wineds/Mechano23-xover: Mechano23 xover

Mechano23 xover. Contribute to wineds/Mechano23-xover development by creating an account on GitHub.

github.com

 

[agnostic1er]

Hello to all,

I'm a former french speaker diyer, new to this forum, so please be indulgent toward my english, thanks.
I have a question: In this Mechano23 thread I don't understand why Vituix graphs show a crossover frequency of 3.5kHz while Amir's measured graph show à 2.5kHz Fx.
Did I misinterpret this last one? Does the near-field measurement not be able to render the actual Fx?

 

[XMechanik]

I have a question: In this Mechano23 thread I don't understand why Vituix graphs show a crossover frequency of 3.5kHz while Amir's measured graph show à 2.5kHz Fx.
Did I misinterpret this last one? Does the near-field measurement not be able to render the actual Fx?

Interesting point. Below is my attempt to explain this effect, not necessarily completely correct or complete so please share your interpretations or remarks.
Measurement in the far field reflects the radiated signal and to radiate a signal of the same level for a higher frequency, a smaller amplitude of the same cone is needed. Therefore measurement in the near field has a certain slope, which causes the intersection points not to occur in the same place as for measurement in the far field.
The near field measurement is not a scaled far field measurement. In the near field, the part of the signal that is later returned to the source (reactance effects) is also measured, which may add up to the difference in crossover frequency, but helps to highlight problems associated with resonances (and this is the main point of performing measurements in the near field).

 

[agnostic1er]

@XMechanik,

Thanks for your reply.

It seems to me that your explanation reflects the consequence of the near-field measurement which is free from the baffle effect.
In my opinion, if Amir would have presented the complete woofer response raised by 6dB compared to that of the tweeter, the Fx would have been the one visible with a far-field measurement i.e. 3.5kHz. Perhaps he did not choose it this way so as to minimize the showing of the spurious bass-boost relative to this type of measurement.
Do you agree with my analysis?

 

[Ktacos]

if Amir would have presented the complete woofer response raised by 6dB compared to that of the tweeter

You can't just add gain, you need to simulate the baffle diffraction profile of the baffle. Amir could probably do it easily, but it would most likely not be entirely accurate and fall in line with klippel level accuracy.

Here's what the klippel response looks like simming the baffle diffraction, using amirs measurements.

 

[XMechanik]

There is also an impact of the difference in the source diameters. In this case it could be estimated as 20*log(26mm/110mm)=12.5dB and this is more or less the difference between the tweeter and woofer graphs on the near field characteristics at 3.5kHz.