Iowa DIY 2026 - ElectRelics

[technodanvan]

This project is kind of a rush since I’m trying to get it done by IowaDIY in the middle of October. I’m waist deep in parts and it’s a struggle to move forward on everything but I spent some significant time working this out. On a phone right now so details are slim, but here goes. This will hopefully be a largish bookshelf that pushes my 3D printers to their capacity.

Tweeter: Peerless DA25TX

Midrange: TBD - FaitalPRO 5FE125, Anarchy 558, MAC-05, and Satori MW13TX are all contenders. Not sure I’ll have time or filament to make test baffles for all, but I’m definitely going to test the FaitalPRO and the Satori.

Woofer (rear firing, separate chamber): Anarchy 708 with a Dayton Audio RSS210-PR in a downfiring situation.

Lots of questions, very little time to get things done. CAD work and progress photos below. Essentially I’m making two trapezoidal enclosures. One shallow enclosure around 4-5 liters for the tweeter and mid. One ~10 liter enclosure for the woofer, PR, and hopefully the crossover.

[JonMarsh]

Interesting progress, but I know from experience that getting things done on a tight schedule is never easy...

[technodanvan]

Indeed, this will be pretty tight, especially if some of the crossover parts are difficult to source. Here's hoping I have most of what will eventually be needed.

The color of the Satori matches really well with the baffle, IMO. It's less black than the 5FE125 and is close to the shade of gray selected. Baffle for that one is printing now.

[technodanvan]

Rear baffle is pretty good, minor improvements need to be made but this will work for testing I think. Printing the woofer cabinet now. It's a 3-day print on that one so hopefully I got it right the first time (and hopefully the print doesn't fail!).

Edit: Not sure what color of bolts to use (black or silver stainless steel). I was planning on black for the 9 bolts that hold the baffle on and mirroring that for the little guys, but I think black would look weird against the red input plate.



Edit:

Satori baffle and surround are complete as well. Big fan of the appearance.

[technodanvan]

The 3-day print that was the woofer cabinet sides turned out okay. Something is going on with my seals, for some reason they only fit one direction and I'm not sure if it's the seal or the cabinet. Seal is easy enough to modify but it shouldn't need to be. Very confusing. Anyway, the cabinet itself needs to be updated a bit so I may have two more very long prints to endure. I'll probably still use this for testing though as I need to get moving on that.

This is not going to be a small speaker. However, with the PR resting there and nothing bolted together, the cabinet feels quite light compared to what I imagine a Baltic birch version of the same would weigh, so that's nice.

[Steve Manning]

Looking very cool Danny

[technodanvan]

I've started measuring a bit to get my head wrapped around what I have here. Turns out the 5" bolts I ordered to hold the front baffle, mid chamber, and woofer compartment together were about an inch too short for the job, but they were sufficient for the front baffle and mid chamber so I started there. The 6" bolts will be here before next weekend, so I probably can't test the woofer until then. My first test failed miserably and I was confused until I tried a second and third tweeter. Turns out that very last unpaired tweeter I got from PE a couple years ago was DOA, so that's kind of a bummer. Of course now I have a dead one to tear apart and (perhaps) think about waveguide development for the others.

But that is a future project.

Initial Satori and Peerless combination is shown below. Keep in mind this is a rudimentary setup in a pretty small room - I'm basically just proving my setup is working - and no gating has been applied here either. Clearly a room issue at ~400 Hz, I'm guessing it's the distance to the concrete floor (which I have no rug or anything placed upon. Probable room issue between 200-300 Hz as well, that would correlate roughly to the distance to the nearest wall. Input voltage was static (but not set specifically at 2.83v), but shows we won't need to attenuate the tweeter much to get it in line with the mid.

I'm trying to determine how I could maybe get this setup outside.



After this, I got to thinking about how to measure the woofer since it is rear firing. My first thought is to leave the test setup the way it is - about a meter or so away from the speaker, and have the front baffle pointed at the mic. This would emulate how the speaker would actually be listened to, which seems appropriate?

Now, I'm sure the combination of REW and VituixCAD would allow me to measure the woofer directly (woofer pointed at the mic) but I don't know how to do that being the novice that I am.

Thoughts JonMarsh?

[JonMarsh]

Hey Danny, looks like you're making some good progress with the cabinet fabrication!

Now, about measurements, or what I typically do- for example, I have a concept project on hold derived from the Diapson, originally a candidate for the Satori MW19TX, but then I must have accidentally sniffed some glue or had a glass of wine, and got it in my head that perhaps the PuriFi PTT8.0NAB (which I have some of still) would be an even more powerful candidate as regards output and distortion, and might handle the frequency response needs adequately combined with a DA32TX and it's extended low end.

Now, I'm a Fuzzmeasure from RHØDE Test boy for a variety of reasons, including all the cool acoustics measurements it can do that are not just about speaker design, and from the way it approaches setup and control. To each his own... it has an awful lot of flexible control and functionality baked in, and taking good advantage of Apple API's, so yeah, this is only available on a Mac.

So, when I get started, after deciding on an enclosure design and specific driver, the next step for me is generating what I call a QNF, "Quasi Near Field" measurement, which doesn't quite follow the classical description; I usually use measurement at a distance of about the diameter of the driver radiating cone, something around 4-5" and keep that fixed distance for things like dome midranges and tweeters- the intent to be close enough so that baffle step response and room effects are greatly attenuated. The intent is to get a good idea of the driver response with the rear volume of the design, and as close to large flat baffle free field, by making the direct sound far stronger than any reflection or outside cabinet effects.

I have a range of preferred impulse width measurement conversions from the sine chirp, and a preferred sine chirp duration. For nominal full range measurements, I use an impulse window time of 100 msec; and that's what I usually use for QNF. On top of that one has time offset controls, amplitude sampling filtering (from 1/48th octave down to 1/3 octave) and a variety of impulse filter functions on the step response, like rectangular, Hamming, Half Hamming, Hanning, Half Hanning, Bingham, etc. You get the idea.

And I usually do non near field measurements at several distances, like 0.5 to 0.7 m, 1m, 2 m, so that I can see tradeoffs in the cabinet and driver response versus the specific test or listening room and it's interactions. I also like Fuzzmeasure because of the flexibility in combining primary SPL and distortion harmonics. Mainly I focus on 2nd and 3rd, to get an idea of the balance of odd (level sensitive) and even (polarity sensitive) performance. (there are a lot of drivers in the past that have much higher 2nd order distortion than 3rd, and this contributes to more IM and congestion when playing music- ScanSpeak, I'm looking at you... )

Note, while Fuzzmeasure can generate absolute acoustic level plots inputting calibration data, the basic response plot is based in dB around the response of the A/D system in the hardware one is using, and so is a guide to setup and avoiding test distortion from the gear- I usually shoot to keep maximum levels no higher than -10 db. And most tests are done with 2.83 VRMS at all times, monitored on a digital oscilloscope connected to the DUT terminals.

So, for the PTT8.0 in an old Madisound cabinet, sitting on a tall stool on my test angle turntable, at 0.7 meters, this is the result:



This shows baffle effects and to a lesser extent room issues.


Here is the QNF measurement, no change to the setup except for the microphone position and reducing the mic gain:





These are with a newer microphone than I've been using in the past for a long time; it's an Audix TM1, which seems to have a lower noise floor and returns better results on distortion tests on high performance drivers like the PuriFi parts. My other microphone is an Earthworks M30. The test controller preamp is a MOTU M4. The system preamp is a 2nd generation Schitt Freya with the Nexus balanced current mode amplification; the current version of that is called a Kara, as Freyas now all have solid state plus vacuum tubes.

Note, considering the sensitivity of the PuriFi PTT8.0, the output distortion for a speaker is VERY low...

[JonMarsh]

A little more info- this is what one of my test carts looks like, the one with the Schitt preamp and Cambridge Audio Azur 840W stereo power amp, which itself is very clean, and has a low noise floor. However, it currently has a Mac Studio desktop computer and LG monitor, instead of the 10 year old laptop.

[technodanvan]

As always, thanks Jon for your insight. Very much appreciated.

Unfortunately, I (sort of) figured out why the gaskets aren’t lining up perfectly. Somehow the cabinet is not entirely symmetrical, it’s off by at least 1 mm, maybe more in some areas. I’m unsure where the error began which makes it difficult to fix without remodeling the whole thing in CAD. So I’m at a crossroads - do I remodel the whole thing (probably a full day+ worth of effort) or just live with it as-is?

As much as it'll be annoying to do, I'm leaning towards the remodel. It wasn't obvious, but now I can't unsee the asymmetry.

[Steve Manning]

As always, thanks Jon for your insight. Very much appreciated.

Unfortunately, I (sort of) figured out why the gaskets aren’t lining up perfectly. Somehow the cabinet is not entirely symmetrical, it’s off by at least 1 mm, maybe more in some areas. I’m unsure where the error began which makes it difficult to fix without remodeling the whole thing in CAD. So I’m at a crossroads - do I remodel the whole thing (probably a full day+ worth of effort) or just live with it as-is?

As much as it'll be annoying to do, I'm leaning towards the remodel. It wasn't obvious, but now I can't unsee the asymmetry.

Danny, I don't know if your printers require similar treatment, but I have to calibrate my drives on my CNC to make sure when I tell it to move a specific distance, that's what they actually move. I also have to "tram" the spindle to make sure it's perpendicular to the table surface. Your CAD files could be fine, but the machine is not moving the correct amount?

[technodanvan]

Unfortunately I noticed it in the CAD files first, so it's definitely there. Parts that don't rely on these particular dimensions (such as all the round cutouts) are fine, so I'm pretty sure the printers are not contributing to the error.

[Hdale85]

Just accept the loss and remodel it, you'll end up doing it anyways lol. We're all far too much of perfectionists to let a problem we know we could fix go lol.

[technodanvan]

Agreed!

So I’ve mostly completed the new 3D model and have several parts printed. Before doing that I asked around on the Bambu forums to see what could be done about the warping I was getting on the corners of the larger parts. They were very helpful! I don’t know if any one of these changes did it, but the combination certainly helped significantly.

1. Dry the filament for 24 hours. This was a step I normally do for just 6-12 hours for the nicer projects I do with PETG (usually not at all for PLA given how dry it is here in southern NV). I skipped this step entirely for the prototype prints, but this time I gave it the full 24 hours.
2. Increase the bed temperature from the stock value of 70C. I chose 75C, no idea if it’s doing anything…but maybe?
3. Turn DOWN the fans to keep heat in the chamber relatively high. I could also choose to control the temperature directly but that didn’t seem to be necessary - usually that’s only for the advanced filaments that need very high temps.
4. Design fix #1 - Chamfer the entirety of the part that would otherwise make contact with the bed. I used a 2mm chamfer and made it a feature. This I had done before even asking at Bambu and it definitely helped a little. Someone here might have made this suggestion.
5. Design fix #2 - Hard, sharp corners (as those seen on the bottom of the picture) are subject to the worst warping due to high stress on small points. The fix is to usually just round off the corners. I did this on the top, but for the bottom I chopped off the corner then rounded the remaining edges.
6. Design fix #3 - Long, uninterrupted stretches that make contact with the bed are subject to the most stress. To alleviate this issue, you need to break up those stretches. I made some small, shallow circles on the backside of the part to do this. If the part were much more solid (i.e. greater than the 25% infill I used) It may be prudent to have the cylinder go nearly the whole way through the part and hide it with exterior skin on each side. This was unnecessary for my situation.
7. Design fix #4 (NOT IMPLEMENTED) - One could also ‘ridge’ the edges of the design with tiny curves. I didn’t care for how that looked so I did not do this.

Here’s how the back of the rear baffle looks. I haven’t actually printed this part yet but the front baffle and mid chamber printed flawlessly and used the same techniques.



I also did a couple things differently on this build.

• Previously the bolt holes were inset by a least 12-16mm compared to how they are above, where they are relatively close to the outer edge. This was because they actually used bolts to hold everything together and you need to be able to access the bolt head. This time I’m using threaded rod that extends all the way through the speaker. I used M8 heat inserts on the back of the baffle (you can’t see them from the front), screwed M8 threaded rod into it, and on the backside there will be small steel thumbscrews (not like those phenolic ones shown below) that’ll cap everything off. There’s a little more to it on the bottom where you can’t use threaded rod due to the placement of the passive radiator, but in general this is the way forward. I have short threaded rod and temporary knobs holding together the mid cabinet for further testing. Once the new woofer compartment is done I’ll use the longer threaded rod in all positions.
• You can also see in both images four T-slot insets. This is to aid with alignment because I no longer have a groove for the gaskets. Gaskets are just flat now.The T-slots on the woofer compartment are for braces which will just extend past the cabinet a few millimeters through the gasket and into the baffle.
• Something that I should have noted in the first build is the audio from the woofer vibrated the panel ‘shared’ between the mid compartment and woofer compartment. The panel vibrates (and presumably makes some level of noise) and audio also passes through the panel itself. That’s obviously no good, so I’m taking a few steps to try and alleviate that issue. Hence the Siless-branded stuff shown below (it’s also inside the mid compartment). Because of this I’m not sure where exactly the crossover will now go, but we’ll see how big and complex it gets and go from there. It may be cool to create a whole ‘base’ for the speaker that the crossover fits in and the speaker rests upon. If time (and filament!) allows, of course.

Shown below is the back of the mid compartment with temporary ‘short’ (~6") threaded rod and knobs holding it together.