Mechano23 Open-source DIY Speaker Review

[kemmler3D]

Except that these similar speakers scored even higher without a wave guide https://audiosciencereview.com/foru...sierra-1-v2-speaker-review.53350/post-1929112

It appears to me that the overall design is what matters not so much the individual elements.

Yes, but like I said, let's not get carried away with the preference score.

Ascend does amazing work and they rightly have a good reputation here. But if you look at the directivity graphs, the Mechano does a bit better. The Ascend has that funky bit between 2-4khz, which they do account for in the crossover, but they can't avoid some wiggles. I think they have both done the best crossovers they can do with the parts they have, but clearly the waveguide does its job in the Mechano.

 

[Rick Sykora]

pumping station one.

speaker flat packs are very difficult, labour intensive, expensive to ship, and has relatively big risk of damage during shipping. if there is more people interested i am willing to help. i am not sure how much material and shipping would cost. even parts express discontinued their flatpacks.

they used to sell bigger 0.56 cu ft box if i remember for 43$ per box so it wouldn make sense to cost more than 90$ shipped for a pair.

if there is more people interested probably we can make it work

Parts Express and its main flat pack supplier had a falling out. They have a new supplier and have slowly started to ship replacement flat packs. We used the .56 ft3 box for Directiva r1 and PE should have a replacement very soon. That said, is a much larger box than was used for the Mechano23.

There is a separate thread tracking the PE flat pack status.

 

[kemmler3D]

For anyone that has worked in a large company, there won’t be much mystery. I imagine the speaker design process goes like this:

1. Marketing sets a target timeline, price and consumer segment.
2. Finance sets a target cost of goods.
3. Engineering sets a target R&D time and strict deadline.
4. Production sets limitations on what it can/cannot do.
5. A bright engineer or group of engineers develops a flat measuring speaker with great dispersion, and are super proud and excited to demonstrate it to senior management.
6. Upon demonstrating it, a senior management person with power states “Sounds good but not exciting enough. Can we make it more exciting so it grabs people when they hear it?”.
7. The engineers can barely contain their frustration but bite their lips (they have mouths to feed at home) and agree to dial in more “excitement”.
8. Product goes to market with a strong lift in 5kHz+ and some bass boost too.
9. Amir measures it which shows the speaker’s deviation from ideal.
10. ASR users jump in thinking the engineers are idiots because a DIY could have done a better job for 1/10th the price.
11. The speaker go on to sell hundreds of thousands of units and are loved by their owners (B&W comes to mind).

I don’t work in hifi but in my big company, this is often how things go down. What perfectionists like us want and what the consumer wants are often very different things. I do hope that consumers eventually come around and want a nice linear speaker but it takes time to appreciate that sound (look at all the people who complain that Dirac has ruined their bass and sound).

Having worked at a SMALL speaker company, I would say the engineers might ask for more guidance on the voicing up front, as they know that they'd rather not repeat steps 5-7 too many times. At which point Marketing (or a product manager) might suggest making the speaker sound like Bose, Sony, or B&W, since those brands get good reviews in the mainstream and sell hundreds of thousands of units. Then the engineers do their job perfectly (or as close as they can with BOM and size budgets) and everyone goes home on time.

Your point of voicing being marketing-driven instead of adhering to a specific idea of high fidelity is absolutely correct, however. As someone who used to work on these things, the goal is to build a unit that will sell well in its chosen niche. Often that niche is not totally interested in true high fidelity. You need to impress a big youtuber on first listen, most of your customers have never and will never read ASR.

"Showroom sound" for consumer-grade speakers is also very much a thing. While real hi-fi shops have dropped off over the years, a lot of consumers are buying their speakers in the aisles of Best Buy, Costco, etc. The speaker needs to demo well in that environment.

And, speaking of the above, you can mix your tracks specifically for the speaker when you do an on-shelf demo, so take any in-store listening test with a big grain of salt...

 

[Yevhen]

@amirm What is the speaker stand height in this estimated room response? I don's see a floor bounce dip around 200Hz, which I think is typical for the bookelf models.

Here is my simulation for 2 identical woofers placed 30cm (green) and 60cm (blue) from the floor.

 

[Bogda89]

Parts Express and its main flat pack supplier had a falling out. They have a new supplier and have started to ship replacement flat packs. We used the .56 ft3 box for Directiva r1 and PE should have a replacement very soon. That said, is a much larger box than was used for the Mechano23.

There is a separate thread tracking the PE flat pack status.

parts express flat packs are very reasonably priced especially compared to some other flatpacks. and quality is very good. last time i bought 2 sheets of mdf in home depot and it was junk compared to the quality of parts express flat pack mdf. I guess most people who are interested in diy have a garage and some tools, but not in my case. even cutting front baffle was mission imposible until i joined makerspace and got access to all the tools you can think off.

probably in a week or two i will build this kit and post some pictures and first impressions.

 

[olieb]

Yes, but like I said, let's not get carried away with the preference score.

Ascend does amazing work and they rightly have a good reputation here. But if you look at the directivity graphs, the Mechano does a bit better. The Ascend has that funky bit between 2-4khz, which they do account for in the crossover, but they can't avoid some wiggles. I think they have both done the best crossovers they can do with the parts they have, but clearly the waveguide does its job in the Mechano.

If you know how to do it and work around the problems of naked domes you can get very good results.
But considering that a waveguide is only a piece of plastic (an extra inductor in the crossover will probably cost more) it is astonishing how many tweeters still don't have one (built in) or two to go with them. Is it all about the tooling costs?

 

[amirm]

@amirm What is the speaker stand height in this estimated room response?

The paper doesn't say. But I suspect you can figure it out from this:

The Vertical Reflections
Not surprisingly the floor reflection has the tightest
distribution of all the early arrivals. On average it
arrives at the listener 1.8 ms after the direct sound
and occurs at an angle of 32.5 degree below the main
axis.

 

[Scgorg]

What is the speaker stand height in this estimated room response? I don's see a floor bounce dip around 200Hz, which I think is typical for the bookelf models.

The estimated in-room response does not include destructive room effects. It basically takes the spectrum of various curves and then adds them as if they were all in phase. Specifically, it takes 3 curves and applies weighting to them. Listening window (12%), early reflections (44%), and sound power (44%). For this reason the estimated in-room response is not as good of a predictor of frequency response in the sparsely modal and boundary-interference dominated regions, since effects like room modes, destructive floor bounce, and destructive front wall reflections are not taken into account.

 

[ocinn]

Truly an amazing swan-song for the high value/budget passive speaker era, as the industry trends towards DSP/Active to force budget designs into submission.

Huge kudos to the designer for pulling this off at this price point. Round the edges of the baffle and implement a proper xover to a sub and you’ll have a fantastic system for an unbelievably low cost.

 

[amirm]

The estimated in-room response does not include destructive room effects. It basically takes the spectrum of various curves and then adds them as if they were all in phase. Specifically, it takes 3 curves and applies weighting to them. Listening window (12%), early reflections (44%), and sound power (44%). For this reason the estimated in-room response is not as good of a predictor of frequency response in the sparsely modal and boundary-interference dominated regions, since effects like room modes, destructive floor bounce, and destructive front wall reflections are not taken into account.

Well, if you mic a few inches in a room, you too can get very different measurements. The idea is to have an average that correlates well with listening tests. For that, PIR works fairly well although I tend to put more emphasis on on-axis response than PIR.

 

[Yevhen]

Well, if you mic a few inches in a room, you too can get very different measurements. The idea is to have an average that correlates well with listening tests. For that, PIR works fairly well although I tend to put more emphasis on on-axis response than PIR.

Thanks. Did you hear that floor bounce? In my simulations that includes the actual room, speakers with the low woofer have no dip around 150-200Hz

Fuller sound in upper bass? Flat upper bass response in normal listening position (2-3m)

Weak upper bass because of the floor bounce?

If that's correct, then our estimated in-room response from Klippel is not very representative...

 

[Yevhen]

The estimated in-room response does not include destructive room effects. It basically takes the spectrum of various curves and then adds them as if they were all in phase. Specifically, it takes 3 curves and applies weighting to them. Listening window (12%), early reflections (44%), and sound power (44%). For this reason the estimated in-room response is not as good of a predictor of frequency response in the sparsely modal and boundary-interference dominated regions, since effects like room modes, destructive floor bounce, and destructive front wall reflections are not taken into account.

Is there any way to include it in Klippel software? According to my experience (and at least in my listening rooms), the two-way speakers with the woofer placed 50-70cm from the floor never sound as full/complete as the 3- or 4-way with the woofer placed close to the floor, even if both have a perfectly flat anechoic response. So I just wonder where that subjective difference is coming from.

 

[Scgorg]

Well, if you mic a few inches in a room, you too can get very different measurements. The idea is to have an average that correlates well with listening tests. For that, PIR works fairly well although I tend to put more emphasis on on-axis response than PIR.

For what it's worth, I do not disagree that the predicted in-room response is useful! I just think that one should be aware of what it is actually showing, as with any measurement. In this case I was merely explaining to user Yevhen why the expected upper bass/lower midrange dip from floor bounce is not in the PIR.

Is there any way to include it in Klippel software? According to my experience (and at least in my listening rooms), the two-way speakers with the woofer placed 50-70cm from the floor never sound as full/complete as the 3- or 4-way with the woofer placed close to the floor, even if both have a perfectly flat anechoic response. So I just wonder where that subjective difference is coming from.

Not sure about the klippel software, as I don't have access to that. With simple assumptions you can calculate the floor bounce frequency and severity based on speaker height, vertical off-axis data, and seated position distance/height.

 

[milosz]

What you are calling a wave guide is more properly called a horn. A wave guide is a kind of transmission line.

 

[bert stoltenborg]

That tweeter has been around since the 1980-s under the Vifa HT-275 moniker https://www.speakerenco.nl/ht275-2. It also had a industrial name, something like the D-27 etc.
You could replace the frontplate with a horn with 3 metal springs clipping it to the magnet. It made the tweeter louder at
the cost of some lower frequencies.
We used it for studio monitors and when Stage Accompany noticed that it wasn't immediately destroyed they designed the F5 and F9 cabinets around those tweeters.
Seems you don't always need a Bliesma or
Purify unit (they are unbelievable, but... ) to make a nice loudspreaker.

 

[beagleman]

For anyone that has worked in a large company, there won’t be much mystery. I imagine the speaker design process goes like this:

1. Marketing sets a target timeline, price and consumer segment.
2. Finance sets a target cost of goods.
3. Engineering sets a target R&D time and strict deadline.
4. Production sets limitations on what it can/cannot do.
5. A bright engineer or group of engineers develops a flat measuring speaker with great dispersion, and are super proud and excited to demonstrate it to senior management.
6. Upon demonstrating it, a senior management person with power states “Sounds good but not exciting enough. Can we make it more exciting so it grabs people when they hear it?”.
7. The engineers can barely contain their frustration but bite their lips (they have mouths to feed at home) and agree to dial in more “excitement”.
8. Product goes to market with a strong lift in 5kHz+ and some bass boost too.
9. Amir measures it which shows the speaker’s deviation from ideal.
10. ASR users jump in thinking the engineers are idiots because a DIY could have done a better job for 1/10th the price.
11. The speaker go on to sell hundreds of thousands of units and are loved by their owners (B&W comes to mind).

I don’t work in hifi but in my big company, this is often how things go down. What perfectionists like us want and what the consumer wants are often very different things. I do hope that consumers eventually come around and want a nice linear speaker but it takes time to appreciate that sound (look at all the people who complain that Dirac has ruined their bass and sound).

Add in the fact that many customers simply prefer "Bad" sound to some degre.....

Having worked in audio sales for a few years in the 80s, we had everything from decent entry level stuff to quite serious semi-high end stuff in my old store.

The number of guys that would come in, listen a bit, then go to the Receiver/preamp we had running and instantly find the loudness and bass treble controls and switch the loudness on and bass/treble all the way up, was quite literally Scary!

I would hear often "Now thats how it should sound!"

Those of us that expect neutral and accurate are NOT the entire Hi-Fi world.

 

[Aperiodic]

A sub and a high pass filter for these and the bass limits would probably be eliminated and it would go from excellent to giant killer

 

[Ajax]

Add in the fact that many customers simply prefer "Bad" sound to some degre.....

Having worked in audio sales for a few years in the 80s, we had everything from decent entry level stuff to quite serious semi-high end stuff in my old store.

The number of guys that would come in, listen a bit, then go to the Receiver/preamp we had running and instantly find the loudness and bass treble controls and switch the loudness on and bass/treble all the way up, was quite literally Scary!

I would hear often "Now thats how it should sound!"

Those of us that expect neutral and accurate are NOT the entire Hi-Fi world.

In 1993 I took over a suspended ceiling business that was begun in the 1960's by a large property company, which had become surplus to their needs and who wanted to concentrate on large scale development. It comprised mineral fibre ceiling tiles manufactured in Japan located within an aluminium grid that I made in Australia. My market was Sydney CBD office buildings and my main competitors were two large US companies.

Prior to that time I'd obtained a degree in Engineering (Civil) 1978 and a diploma in Bus Admin (baby MBA) 1985 and had worked in construction and property development for 15 years. I started the ceiling business with no idea about how to run it but I did understand the basics of business and how the property industry worked. I new more than my customers about what they needed and basically ignored them and developed new ceiling systems that provided easy access to the buildings' services so they were less likely to be damaged.

I tracked buildings that were over 15 years old that had not yet been refurbished but soon would as tenants' leases expired. I tracked tenant movements in old buildings, who would exit say 3 floors providing a small refurbishment project. The new tenants would relocate the existing partitions to suit their needs and destroy the exiting ceiling system. The building owner would take the opportunity to install a new system (mine as I'd already approached him), who would then install my products in the remainder of the building to ensure consistency as more tenants came and went, as much as another 50 floors in the one building.

I ended up with 70% of the Sydney market, servicing over 500 office buildings with 400 floors land banked, which I was able to service at great margins. I'd developed a cash cow which I sold to a multi national in 2007, a couple of months prior to the GFC (pure arse).

The moral of the story - become an expert in your field and know more about your market and your customers' needs than they do. Despite the quote to the contrary, your customers are very rarely right. You are the expert not them.

Another example of knowing what your customer needs is the late great Steve Jobs.

IMO Amir is on the right track - he knows much more about how to produce great Audio than most of the manufacturers.

 

[levimax]

If you know how to do it and work around the problems of naked domes you can get very good results.
But considering that a waveguide is only a piece of plastic (an extra inductor in the crossover will probably cost more) it is astonishing how many tweeters still don't have one (built in) or two to go with them. Is it all about the tooling costs?

I think the issue is the diameter required for a really good and effective wave guide. If you look at the response of most "waveguide / horn" tweeters that are around 104 mm diameter (more or less standard size) they tend to get a little uneven between 8Khz and 14 Khz due to some cancellations from the "horn". If you look at Genelec and the like they use a much larger waveguide that is usually part of the baffle and these waveguides are more effective and smoother. I think adding a waveguide "big enough" to be really effective does add a lot of costs, it is also not the size people are used to so probably limits the market. It will be interesting to see if this changes and there are more offerings of larger built in waveguide tweeters.

 

[brunes]

He does! It is called Mechano323 and performance-wise it looks great too. The crossover doesn't look overly complicated for a 3-way speaker.

Would be awesome to test those as well, quite a few 2 way DIY speakers floating around, but 3 way with ribbon tweeter is a bit different.