If you're looking for other options, Github allows hosting .exe files without any issues. There's also a release management system.
I don't think it actually requires hosting the source code together with the executable, though it's typical.
-
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!
Small 2-way speakers with linear on-axis and power response characteristics (Scan Speak and SB Acoustics drivers). H&V off-axis measurements included
- Thread starter XMechanik
- Start date
I don't think it actually requires hosting the source code together with the executable, though it's typical.
Maiky76
Addicted to Fun and Learning
It is a very good practice to look at the distribution of the modes in the box.
I am not familiar with the tool you are using (Vcad) so I don’t know if it takes into account tangential modes
i.e. modes with multiple bounces before looping back in phase) but the 4 colors look like
H/D/L + something else (port itself?), I might be wrong though.
Anyway as long as you have modes that are not all overlapping and high enough in frequency this should not be a problem with a decent internal damping material.
For example:
https://p-cdn.rockwool.com/syssitea...oard-insulation-techdata.pdf?f=20230123192603
Above 500Hz and even more so 1000Hz, the damping material should provide enough absorption to curb the box resonances contribution as long as it is applied correctly i.e. not only on the wall (where is does not do its best) but in the middle of the box as @Xmechanic did.
Low voltage impedance measurements of the LF driver in the box with no Xover with and without the damping will help to visualize the impact/issues/improvements:
example of impedance with resonances:
https://www.audiosciencereview.com/...s/boston-acoustics-a-25-speaker-review.17625/
You need to settle on a port diameter that you can actually make.
Standard PVC tubes have discrete diameters so it may be tricky to find one that can work.
The easy thing is indeed to use 1x or 2xMBR35 or 1xMBR50 so you can tune the port to suit your room/taste.
So make sure to procure a correct damping material and make sure that the first few resonances are not overlapping.
The Rockwhool is rigide so you can cut and it will be easy to fit even with funny shapes and fill part of the “center” of the box.
Have a close look at the excursion of the driver with power (WinISD can show maxSPL).
A digression
If one wants to maximize power handling and SPL let’s say for a 2.1 or Home Theater configuration then
Box: 4.6L net (with actual resistor of the filter that would need to be changed)
FB~83.5Hz 2xMBR35 (14.5cm, 19.2cm2) or 1xMBR35 (cut down to 6.4cm after starting from a longer value ;-), 9.6cm2)
NOT optional Active 2nd order High-pass filter set to 95Hz with a Q of 0.62 (Home Theater amplifier?) can be simulated with WinISD (SOS HP filter)
This will provide close to optimal power handing / max SPL with no port issue. All the LF (below 80Hz) duties would then be fulfilled by one or multiple SWs.
96dB (above 120Hz) would barely exceed 1mm excursion on the LF driver, well within its capabilities.
In fact, the system becomes thermally limited i.e. the driver will burn before bottoming/breaking with about 101dB for 40W.
One would need to be careful not to damage the speaker because there will be no warning up until the driver is burnt.
Indeed it's H/D/L + port.
I actually hoped to use an entirely ready port tube; as this one has a tapering diameter I used (Y+Z)*0.5 for calculations:
Though the MBR-50 might be a safer choice.
To keep excursion at bay I considered including a 2nd order high-pass filter at 35Hz or 30Hz; either included in the crossover or via the EQ (this is at 5.1V/96dB SPL):
Will also keep temperature in mind; though it seems like it should be easy enough to measure?
Voice Coil Temperature
It's great to know this can be validated experimentally, thank you!
Earlier in the thread, multiple people mentioned that rounding the edges could help with diffraction; do you perhaps know what radius would be necessary to make a difference?
Due to the woofer, there's only space for 17mm if it's a constant radius, maybe a little more if it's decreasing towards the side (there are cutter bits like that for woodwork).
Would that make a difference, and which option (constant/variable radii) would be better?
Inertiaman
Active Member
- Joined
- Sep 25, 2023
- Messages
- 264
- Likes
- 611
Can the crossover experts please help me finalize my component list?
Among the infinite design options, three viable crossovers have been shown on ASR for the Mechano23. There are also choices within each design to consider (electrolytic cap or not, for example). So maybe @XMechanik , @Rick Sykora , @kimmosto and others can comment or recommend.
1) Original Crossover - as described in both this thread and the review thread. Some questions:
a) what is the practical consequence of using an electrolytic for C2 (27uf)? Just a theoretical deterioration in 20 years?
b) the specific components from Solen, listed in Amir's initial review post, are a bit of a grab bag. Many caps are 160V but others are 630V. I get the impression none of these were used in the actual speaker build (parts were sourced from Europe, right, so no Solen?), but rather its simply a quick selection of components that met the uF values, without any regard for voltage? Is there any justified concern, when comparing x-over designs/costs, to consider the 160V Solen caps inferior or problematic vs typical 400V caps (or even 250V) through other sources?
2) Alternative crossover from @kimmosto as shown in post #151 of the review thread.
a) should I consider an electrolytic cap for C2 (8.2uf) for an apples to apples comparison of x-over cost? (or inversely, use film cap in x-over #1?)
b) the 910mH coil L1 (.470 ohm) is difficult to find. Is there any practical reason not to use a 900mH coil (also .470 ohm) which is easily obtained?
c) I assume 10W versions are fine for the resistors?
3) Alternative crossover from @Rick Sykora as shown in post # 117 of this thread
a) once again, should an electrolytic cap be considered for the 33uF cap? and/or the 15uf? The large caps in this design keep the total cost relatively high despite the much lower component count
b) again, shall I assume 10W resistors are appropriate?
b) I haven't seen many comments/reactions to this option . . . the graphs look pretty good (although, like option #2, it doesn't trap the woofer resonance) . . . are there negative consequences I'm missing?
Ultimately I'm trying to decide which of these options to pursue, and also clarify specific component choices where they are in question. My personal objective is to minimize cost while approaching diminishing returns in quality/performance. I suspect others may be making similar assessments, and although the outcome may depend on regional sourcing or personal preference, some discussion of the options could benefit many here.
Among the infinite design options, three viable crossovers have been shown on ASR for the Mechano23. There are also choices within each design to consider (electrolytic cap or not, for example). So maybe @XMechanik , @Rick Sykora , @kimmosto and others can comment or recommend.
1) Original Crossover - as described in both this thread and the review thread. Some questions:
a) what is the practical consequence of using an electrolytic for C2 (27uf)? Just a theoretical deterioration in 20 years?
b) the specific components from Solen, listed in Amir's initial review post, are a bit of a grab bag. Many caps are 160V but others are 630V. I get the impression none of these were used in the actual speaker build (parts were sourced from Europe, right, so no Solen?), but rather its simply a quick selection of components that met the uF values, without any regard for voltage? Is there any justified concern, when comparing x-over designs/costs, to consider the 160V Solen caps inferior or problematic vs typical 400V caps (or even 250V) through other sources?
2) Alternative crossover from @kimmosto as shown in post #151 of the review thread.
a) should I consider an electrolytic cap for C2 (8.2uf) for an apples to apples comparison of x-over cost? (or inversely, use film cap in x-over #1?)
b) the 910mH coil L1 (.470 ohm) is difficult to find. Is there any practical reason not to use a 900mH coil (also .470 ohm) which is easily obtained?
c) I assume 10W versions are fine for the resistors?
3) Alternative crossover from @Rick Sykora as shown in post # 117 of this thread
a) once again, should an electrolytic cap be considered for the 33uF cap? and/or the 15uf? The large caps in this design keep the total cost relatively high despite the much lower component count
b) again, shall I assume 10W resistors are appropriate?
b) I haven't seen many comments/reactions to this option . . . the graphs look pretty good (although, like option #2, it doesn't trap the woofer resonance) . . . are there negative consequences I'm missing?
Ultimately I'm trying to decide which of these options to pursue, and also clarify specific component choices where they are in question. My personal objective is to minimize cost while approaching diminishing returns in quality/performance. I suspect others may be making similar assessments, and although the outcome may depend on regional sourcing or personal preference, some discussion of the options could benefit many here.
kimmosto
Active Member
- Joined
- Sep 12, 2019
- Messages
- 272
- Likes
- 672
Purpose of that filter was to show (locally) that less AI and components can be considered. I have not given permission to publish that filter on this forum. In addition, it was using older measurement data/processing so just forget it.
jaakkopetteri
Active Member
- Joined
- Apr 10, 2022
- Messages
- 260
- Likes
- 177
Curious that you would post it "locally" on a completely public forum accompanied by some high-handed mentions of "common sense vs AI" and now that it's actually exposed to criticism, you downplay it with some copyright thing and it not being serious.
Yes.
Probably fine for most purposes. If you drive them hard, might push rating to 20-25 watts. Looking at options, prob cheaper to parallel some 24-25 ohm resistors.
No, mine has same trap circuit for the woofer resonance, it just shows a bit more with the lower order filter on the woofer. Also shows more as I used different scaling. As mentioned, not likely audible. Here is a comparison to the original using same scaling...
I doubt there is much if any audible difference. but is @XMechanik's design, would go with his advice.
Last edited:
- Thread Starter
- #149
A general note is that since it's an "open design", anyone is free to modify/simplify/complicate or create their own filters, change box tuning, etc. Let's have Mechano23 playground here.
Referring to Rick's x-over version I would add 3 remarks: 1) rising sensitivity means also rising f6 which may not be that good if no sub cooperation is planned, 2) filter architecture has capacitive shunt which may be uncomfortable load for some amps, 3) this filter version creates some riples on PIR and PR characteristics (esp. 3-10kHz region):
Some riples are also present on the original design characteristics but they are noticeably less prominent:
It's hard to say if any there will be any audible difference. On the other hand, it seems to me that adding 3 components does not significantly increase the complexity and costs.
Referring to Rick's x-over version I would add 3 remarks: 1) rising sensitivity means also rising f6 which may not be that good if no sub cooperation is planned, 2) filter architecture has capacitive shunt which may be uncomfortable load for some amps, 3) this filter version creates some riples on PIR and PR characteristics (esp. 3-10kHz region):
Some riples are also present on the original design characteristics but they are noticeably less prominent:
It's hard to say if any there will be any audible difference. On the other hand, it seems to me that adding 3 components does not significantly increase the complexity and costs.
Agree just a different set of tradeoffs and will stick to any real difference is minimal and any audible difference would more likely come from voicing tweaks. I did manage to get some nice linearity from 300 to 1000 and doubt the slightly higher sensitivity will send anyone running for a subwoofer! There are some EPDR swings but just in a couple of spots, so the impedance diff is not a major differentiator even on the sims.
Will echo that a few more components are not a major expense for the hobbyist. Using other terms, would consider the original to be a perfectionist crossover, mine is more the pragmatist one. If it was to be made in larger numbers, parts counts and board space weigh in more, but unless one cannot find the extra parts, would not be a deal breaker for me. 
kimmosto
Active Member
- Joined
- Sep 12, 2019
- Messages
- 272
- Likes
- 672
I can write directly here in case my opinions or proposals are necessary (imo) for community or original author so I don't need you or any other gossipier to copy paste images and words from other forums. One of the main problems is that as a designer of commercial products I cannot have public and honest opinions with any negative flavor about designs/products of other designers/companies. Zero tolerance is very difficult to maintain especially with reviews and (too) many messages on ASR. Spreading messages just helps to block me out.
Last edited:
Nuigurumi
Member
- Joined
- Oct 14, 2022
- Messages
- 20
- Likes
- 10
This might become my first DIY attempt.
How important is it to stick to the specific enclosure wood type and thickness?
Say I want to do 18-24mm MDF instead of 18mm plywood.
Would this transform the speaker from great into 'not so great'?
I imagine resonances would be different, so it's not advisable.
How important is it to stick to the specific enclosure wood type and thickness?
Say I want to do 18-24mm MDF instead of 18mm plywood.
Would this transform the speaker from great into 'not so great'?
I imagine resonances would be different, so it's not advisable.
Quality MDF is fine and may be an improvement. Cutting MDF does create much more dust and should use mask when you do.
The most important issue is to keep front baffle dimensions (esp. width) constant. Resonances can be measured and then dealt with by applying dampening/bracing.
Personally I am considering doing the same, since I intend to finish it in natural wood veneer anyway and I found it instructive to look at the Directiva R1 prototype build - an enclosure with twice the volume, made from 3/4 inch thick MDF with a brace in the middle. So a similar solution should work here; but it's worth noting that the brace itself will take up a non-insignificant amount of volume (~0.1l-0.2l).
- Thread Starter
- #156
Everything crossed is connected. Compare to the VituixCad schematic and the wiring diagram.
Thank you!
- Thread Starter
- #158
For a 40W 4ohm bookshelf speaker max voltage magnitude on terminal would be 17.8V (12.6V RMS). Even assuming that some resonance effects going on filters will double or triple this value there is still a big headroom comparing to 160V. To be honest, I didn't even look at the maximum voltage parameter when selecting sample caps for the list.
The caps used in the design were different from those listed, but this should make no difference as long as you keep the values and tolerances.
Local extraction a must with mdf as well.