Cracked pump housing - practice for reverse engineering.
Damaged Yamaha Waverunner Impeller. The edges of the real impeller are bent and also rough from hitting rocks.
Very good @rsilvers
Thanks for sharing !
This was my first try doing reverse engineering. My Yamaha Waverunner impeller was damaged by rocks, and the ends of the blades are bent.
I made a scan of it using scanner spray made from 91% isopropyl alcohol and corn starch. 99% is better because it dries faster, but I didn’t have any.
I did two scans - one of the top, and one of the bottom, and then used the Mac software to merge them.
Once I had the scan, I loaded it into Fusion 360. I then build a CAD model within the scan, using it as a reference to get the blade angle to match. Of course my CAD model didn’t try to copy the bent blades, and instead modeled what they would be like if they were not bent.
The impeller has threads in it, and I looked up the thread size, and modeled those into the CAD. I also took caliper measurements of the wrench flats and snapped them to the exact metric size for such a metric nut.
I made two prints on a Bambu printer (I was product tester and consultant for Bambu leading up their Kickstarter launch).
The photo of the three imepllers - the first is the original metal one. The middle one is a direct print from the RevoPoint Inspire scan with no changed at all, so it also has the bent blades. The third one is a print of the CAD model that I made from the scan reference. In theory, I could make a lost-PLA casting of it, and then put it on a lathe or mill, bore and threat it, and it would work. Or have a metal print made.
In reality, I ordered a new Yamaha impeller for $130 and installed it, as this was just practice. The scanner worked perfectly for this.
A great example of what can be done, and you’re right, sometimes it’s just easier to buy new
I am still impressed with your Impeller scan , not easy scan job to do , but I love my INSPIRE , it do great job above its original specifications .
Did a comparison of Inspire vs iPhone photogrammetry. The Inspire is cleaner, better for curved areas, and almost certainly more dimensionally accurate.
The blueish-tinted scan is from a RevoPoint 3D Inspire. The yellowish-tinted scan is an iPhone using 3D Scanner App in Photomode, from 250 photos and processed on an Mac using 3D Scanner App. The object was sprayed with Zinc Oxide to make it more white.
What was the time taken for each scan, iPhone and inspire, it would be interesting to know how long it took for each result.
Each scan is very different. Someone would have to take a bunch of objects and make a long video comparing scanning each of them both ways to really know.
I used an iPhone 13 Pro Max with 3D Scanner App and did photogrammetry with 250 photos. The mesh is in yellow. The blueish mesh was a RevoScan Inspire. Both scans were of the same object which has been sprayed with zinc oxide. This is a gauge block that my calipers says is 50.81x76.4mm.
The photogrammetry was off by 3.69%, average of X and Y. The Inspire was off by 0.65%, average of X and Y. To me this means that the Inspire is good enough to use without scaling, but the photogrammetry would need adjustment. Moreover, the photogrammetry misread the surface texture due to variations in the brightness of the zinc oxide coating. This didn’t fool the Inspire.