PURIFI finally did a fully purified passive speaker design! The SPK 16 prototype is here - with a PTT tweeter

[ayane]

It's called humor

 

[Lars Risbo]

peak forces: at structural resonances, the internal forces are multiplied by the Q factor of the driving force. This is where buzzing and rattling potentially occurs when the mechanics get nonlinear

 

[ayane]

Sir.

Just as I think I’m getting my head round the whole concept of my final, superb, speakers and eventually even getting them to be active … you go and throw in some obvious witchcraft.

If you're curious about something better than a two-way passive system, you might want to hold your breath for the Ascilab S8C.

 

[AudioJester]

This is awesome, I might give it a try.

Whats the issue with considering an active version? Makes a lot of sense for diy enthusiasts, as does a 3D printed cabinet!

 

[Audionaut]

It's called humor

I'm a German without a built-in sense of humor.

 

[Audionaut]

the density would need to be enhanced with a filler to match MDF. So you'd have to get some plaster, pour it into a space in the wall of the cabinet after assembling it, and hope that step goes well

Try using fiber-reinforced castable concrete. You can buy it ready-made, but you can also mix it yourself. Polypropylene fiber and glass fiber for reinforcing concrete are available as ready-to-use additives.
If you're using concrete, you really only need your 3D filaments for the outer layer, so you can adjust the thickness accordingly.
The structural integrity will then come mainly from the fiber-reinforced concrete.

I don't think plaster is a good idea.

 

[Kaladin]

An active version sounds like a terrible idea. Plaster reportedly works well but you need to mix it with a lot of pva glue which makes it not cheap anymore.

 

[Audionaut]

Whats the issue with considering an active version?

In principle, nothing at all.
However, Lars has pointed out on several occasions that the Purifi should still be wired with passive components to reduce distortion; you should take this into account and consult with Purifi.

 

[Roland68]

Try using fiber-reinforced castable concrete. You can buy it ready-made, but you can also mix it yourself. Polypropylene fiber and glass fiber for reinforcing concrete are available as ready-to-use additives.
If you're using concrete, you really only need your 3D filaments for the outer layer, so you can adjust the thickness accordingly.
The structural integrity will then come mainly from the fiber-reinforced concrete.

I don't think plaster is a good idea.

An active version sounds like a terrible idea. Plaster reportedly works well but you need to mix it with a lot of pva glue which makes it not cheap anymore.

Both are terrible ideas, especially for such a high-quality loudspeaker.
Over the past 30 years, I've seen many loudspeaker projects using plaster and concrete, but none of them were particularly good.
Both materials shrink as they dry, making a clean bond with the enclosure impossible. Furthermore, complete drying takes weeks.
The thin material thickness of approximately 8-12 mm makes plaster and concrete prone to cracking, especially with the low-frequency vibrations of the bass drivers.
Bonding these materials with PETG simply doesn't work.

For true cohesion, there are only a few adhesives available, such as PVC-U. Epoxy resin is not a good idea for bonding PETG to PETG, primarily due to vibrations. A fairly good adhesion can be achieved with high-quality mounting adhesives and RTV-1 adhesive silicones.

 

[Mister.Mole]

With the active version having the option to go cardioid. Wouldn't that ask for another compartment or is it fine if the delayed transducer playing the cancellation signal is in the same volume as the front mounted transducer?

 

[Kaladin]

Both are terrible ideas, especially for such a high-quality loudspeaker.
Over the past 30 years, I've seen many loudspeaker projects using plaster and concrete, but none of them were particularly good.
Both materials shrink as they dry, making a clean bond with the enclosure impossible. Furthermore, complete drying takes weeks.
The thin material thickness of approximately 8-12 mm makes plaster and concrete prone to cracking, especially with the low-frequency vibrations of the bass drivers.
Bonding these materials with PETG simply doesn't work.

For true cohesion, there are only a few adhesives available, such as PVC-U. Epoxy resin is not a good idea for bonding PETG to PETG, primarily due to vibrations. A fairly good adhesion can be achieved with high-quality mounting adhesives and RTV-1 adhesive silicones.

I use 3D gloop and it works extremely well for bonding petg to petg. No idea what it's made out but maybe you have an idea?

 

[Roland68]

I use 3D gloop and it works extremely well for bonding petg to petg. No idea what it's made out but maybe you have an idea?

This is a solvent-based adhesive based on methylene chloride, methyl acetate, and heavy benzene (petroleum) with small amounts of a polymer.

Not exactly healthy. I would advise against using it in living spaces and recommend ensuring good ventilation during application.

I always advise consulting the safety data sheets.

 

[JanRSmit]

Cyanoacrylate works fine as well, proper degrease is needed, of course.

 

[JanRSmit]

Why not use pertinax or whatever its name in english , specifically the version with cotton weave, for the cabinet?
The baffle can be constructed with a cld layer. I used a sheet from a swedish firm, need to search my archive. Was specific for fibreglass hulls where the engine mounts are. So ment for quite stiff materials. Mainly stiff viscuous therefore.

 

[Roland68]

Cyanoacrylate works fine as well, proper degrease is needed, of course.

But it's significantly worse; it's very sensitive to impacts and any kind of sustained vibrations.

 

[bmc0]

Both materials shrink as they dry

Pure gypsum plaster expands slightly as it sets and then changes very little dimensionally with moisture content. Portland cement and lime plasters shrink. That said, I agree that filling a 3D-printed shell with plaster is probably not the best idea.

 

[kemmler3D]

Pure gypsum plaster expands slightly as it sets and then changes very little dimensionally with moisture content. Portland cement and lime plasters shrink. That said, I agree that filling a 3D-printed shell with plaster is probably not the best idea.

I was thinking plaster because it is cheap, dense, and can be dimensionally stable. That said, it seems like the right density is achievable with just PETG or whatever filament is used without making the print job totally unreasonable, so I'm not prioritizing a fillable version right now.

My goal is to make the printable version as easy to build as I can, so not asking the user to fill it with liquids is a help. Filling is also a concern because certain cements and plasters can heat up enough while curing to soften the plastic (I've seen 100C mentioned) so that's another variable that could cause failures that can't be controlled up front. I don't think most plaster will do this in the quantities we're looking at, but who knows what someone will decide to buy online...?

I think plaster/ concrete does make sense if you were building something much larger than this. Say a floor stander or a sub. To finish those in a reasonable amount of time you would need to print much lighter, but the need for added mass would only go up.

My current plan is to advise using 2-part epoxy for assembly. I've found it to be pretty reliable, and it's viscous enough that creating a proper seal with it shouldn't be too difficult.

 

[kemmler3D]

why would you prefer to save desk space vs. The inevitable replacing of a plate amp?

If desk space is really at a premium, and you have a lot of confidence in the plate amp lasting a long time, you may find that the risk of amp failure in the long run is worth it. I have been using active nearfield speakers for about 15 years now (with built-in amps) and I haven't had a problem. I have often had a problem with desk space though.

What if you can't get the same plate amp or a plate amp with the same dimensions?
There's an enlightening thread on hypex capacitor failures. The eval boards use cheap caps that are sub 2k hour rated. Even the Nichicon caps are rated at 7k hours. I admit I'm an outlier since I listen to my system for 8+ hours per day. Work. Relaxing. Tinnitus therapy. But even in kit form the Purifi speaker kit plus cabinets are not cheap. For me this would be approximately 3 to 4 years.

Yes, I actually had a cap on a Hypex amp fail on me too, although it was not in a plate amp, the principle applies.

You make many highly intelligent posts and I've learned lots from them. I don't agree with you on this one.

Thanks, I don't think plate amps are the right answer for everyone. I simply like the all-in-one aspect and I am willing to "whistle past the graveyard" of amp failures, but I don't think any of your objections are wrong. As with all things in speakers, there is a trade-off to be made.

 

[bmc0]

I think plaster/ concrete does make sense if you were building something much larger than this.

In addition to the potential lack of bond with the plastic, one other concern I have with a mostly closed, fillable shell is simply getting rid of the free water in a reasonable amount of time. Gypsum plaster isn't highly alkaline like portland cement and will tend to support mold growth if not dried completely and in a timely manner.

The eval boards use cheap caps that are sub 2k hour rated. Even the Nichicon caps are rated at 7k hours.

The hour ratings aren't very useful without temperature ratings. If I'm not remembering wrong, expected life approximately doubles for every 10C decrease in operating temperature. So a cap rated for 8k hours at 85C and one rated for 2k hours at 105C will have a similar expected life if both are operated at, say, 65C.

 

[Luvchampagne]

In addition to the potential lack of bond with the plastic, one other concern I have with a mostly closed, fillable shell is simply getting rid of the free water in a reasonable amount of time. Gypsum plaster isn't highly alkaline like portland cement and will tend to support mold growth if not dried completely and in a timely manner.


The hour ratings aren't very useful without temperature ratings. If I'm not remembering wrong, expected life approximately doubles for every 10C decrease in operating temperature. So a cap rated for 8k hours at 85C and one rated for 2k hours at 105C will have a similar expected life if both are operated at, say, 65C.

You're missing the obvious.
1. Plate amp is enclosed in a speaker. Worse heat wise than in an amp enclosure. Depending on how hard the speaker is being driven, much hotter.
2. Pistonic vibrations within the enclosure especially with the Purifi drivers. It's part of the secret sauce. These affect not only the capacitors but all solder joints.
Add those to your calculations and it's far worse.
Do you disagree? If so, have you looked at the calculations including excess heat and vibrations?