Printer diaries: No need to scream? | Shorts

UPDATE, UPCOMING PLANS: I'm going to get a proper 775 motor for this, update and replace the impeller (no more rubba-rubba-rubba), and do another round of testing with an actual PSI gauge and maybe even a tach. Stay tuned!... (Screamin' pressure test video: • 3D Printed Regenerativ... )
Background: I had an exciting afternoon today as, after a 40-minute run at 60% PWM on cooling, my Shark dust buster ubercooler made a loud bang and a cloud of smoke while I was checking it for overheating. After that it ran a lot better, but I still had to take it out and figure out what happened because electronics don't generally let out the magic smoke and then work better. Turned out the smoothing cap exploded quite violently. ANYWAY, I took the chance to hook up the prototype 3D printed regenerative compressor that you might recall from the screamin' demon pressure test video and find out what % of unleashed hell it actually needs to do an equivalent job versus my normal, very quiet cooling solution. Turns out about 3-5% hell would be PLENTY for my day-to-day cooling needs 😮😮😮
At this speed, if you discount the rubbing of the unbalanced impeller (which I will fix), basically all the noise is the motor whining unhappily on the low PWMs. If I can find a solid motor solution that's better sized for this RPM range, I think it could be near silent. I was bummed about the noise level on this thing (and yes, the sound in this video isn't pretty), but I'm excited about its potential again, now.
Also, one big difference from the dust buster cooler is that the motor is external to the airflow, so it's not pumping heat into it.
I think we might be on to something, here, after all...
You can get STLs for this compressor here: www.printables.com/model/2543...

[@stefanguiton]

The fact that this works at low RPM makes it even more impressive. No jet engine speeds needed which would ruin 3d printed blades

[@jamespray]

That's what I'm thinking, too. Nobody gets very far with 3D printed centrifugal compressors because they don't build boost without tens of thousands of RPMs (if not more) and you just can't do that with plastic. And nobody gets very far with 3D printed roots or screw compressors because you just can't print at the required tolerances. This approach has its own limitations, but it seems VERY well suited for good performance within the limits of what FDM 3D printing can do, which is creating nearly unmachineable geometry on the cheap with close tolerances but (ideally...) non-contact moving parts.

[@stefanguiton]

@@jamespray For sure! I am almost wondering whether you could achieve better tolerances by intentially over-running the blades at a burst of slightly higher rpm, essentially sanding itself down. Then clearing the dust and running it at a lower rpm in normal operation. I have been palnning in building a 3D printed blower fan for a while and its refreshing to see your approach, ill test it out.

[@jamespray]

@@stefanguiton That could work, sure! But the impeller would have to be running flat, first, and it's not (the shaft is tilted slightly, so the impeller disc wobbles). I think the issue is that the lock collars I used don't have well-centered bores. I am planning a redesign of the impeller with a separate hub so it can be adjusted for tilt and centering problems. I have found a motor type that looks easy to work with (775) and should be able to do some more testing soon-ish.