RC10Talk - The Net's Largest Vintage R/C Community

For the small detail of the drivers face that's really good , hand scanning is really tricky , you need to keep the scanner as nearly the same distance from the object as possible , I was scanning people's heads so much bigger but I considered a semicircular frame that could rotate round the person from above and you could slide the scanner around the semicircle so you could do a circle below the chin height , on a level with the ear and finally from just over looking down with all three passes at exactly the same distance .

yeah, I noticed that hand scanning is really tricky... I've quickly tried the turn table as well, but the scanner loose track too easily due to the uniformity of the rear of the helmet... Hope the support will change something to this...

I did another session of scan but with my lettle printed support


And that improved the form tracking dramatically... I've kept the face I showed earlier, and did 3 additional cloud of points with a full rotation at different heights (it took me more time to put the proper setup in place, and to understand how to work within the application - I did this with my phone and the post processing was done on my laptop)... And the result is already great. I'll try to print that stl later on to compare, but that looks already good.


I'm curious to see how good it is dimensionnally...

As the resin printer was busy, I did a cut of the STL in Bambu studio and printed the scan in two parts...

Here is the result. There is still a bit of work as the base of the helmet is not clean enough currently (at least for me so I feel there is more work.


I'll also have to deal with the driver figure body... which may be both easier and more complex to scan...

I printed the scan in resin to see how much it is different...


The result is better defined (as expected)... I still need to work the base in CAD, and we'll try to scan the body later on... Normally, the feature tracking should work better on the body so we'll see.

After my day of work, I went for another scanner session...

This time I used high accuracy in the capture (it was in standard accuracy last time). I started by capturing the body. For this I used the black support (black means the scanner will not capture it), as the body itself should have enough features to allow proper tracking easily....


I use some packages as support for the scanner or the turn table to take the capture at different height and orientations. I also put the body in different position directly on the turn table to add additional captures...

Then I merged all the point clouds based on the features, and about 5 minutes later, I had a very nice body...


I redid the same for the helmet, but with the white support with slightly better results on the scan itself than for my previous capture...


I think I can still improve the support, so the bottom of the helmet is even better... But that's for next time...

Gone for some test prints...


I'll put a comparison between the original parts and the printed parts later on after the printed/cleaned part are fully dry and cured...

Thats plenty good enough

yes, plenty good enough already!
I'm still asking myself if there is way to get better definition. I'll tweek the edition, as I believe there is a bit too much of smoothing in the automated editing... In short there is room for improvement... Even though it's probably good enough, considering that most of the reproduction I saw relying on a silicon mould are not really giving much sharp edges... But clearly, this is on the sharp edges that the scanned file could be improved...

I'm really puzzled by resin printing some times : I have around 1/8 print fails with standrad resin.

The kind of fails I can't explain : bed level is good because unchanged, next print with the same file is working. At this stage, I believe my bed might not be as clean as it should, but generally, ending a print, removing printed parts from the bed and start next print would work.

Funnily, most of the times, the first print of a serie is always the one failing (the one after I cleaned the bed since last serie). Knowing this, I generally check my prints few minutes after the beginning of a print, to avoid loosing too much time and material... Let's see what happens for the test prints I launched today...

The test print to check the last scans was done successfully...

I rescan the head of the Super Wheelie differently, with a slightly better result again...


The Big Bear driver figure scan is actually much better than what I thought it would be... I still wish to find another one in parts so it is easier to scan...


Last the Hunter/Galaxy driver is in two parts, like the Super Wheelie : body and head... And the same process I used for the Super Wheelie head was used.


I'm questioning myself about putting those on shapeways or cults3D to make them easily accessible...

As they are then Thingiverse or cults3d would be a nice thing to do , they are really so close to perfect though that I'd be tempted to try and rework them , sharpen the detail of the seat belts and buckles .
Maybe cut up the model in meshmixer to just export a part of the model to cut down on the number of triangles then expert it to fusion to rework the belts and buckles , finally put it back together in meshmixer . It would take time but it could be perfect !

Dadio wrote: ↑Sun May 19, 2024 4:20 pm As they are then Thingiverse or cults3d would be a nice thing to do , they are really so close to perfect though that I'd be tempted to try and rework them , sharpen the detail of the seat belts and buckles .
Maybe cut up the model in meshmixer to just export a part of the model to cut down on the number of triangles then expert it to fusion to rework the belts and buckles , finally put it back together in meshmixer . It would take time but it could be perfect !

At this stage, I did not played yet with the post processing to see if I can get those element sharper. The result is already good, and just using a fine pointer, it's easy to mark the lines and the details in a much better way (I tried on one of the Super Wheelie printed heads), and that would get a perfect result from the print. I got this suggestion by a friend who is a former 1/43 prototypist, and he said that this type of work should give proper results... That said, I'm still looking for a better result directly out of the print...

This morning was resin session end.

I already learnt how to do better with flex resin as it is usually very long... With a standard resin, which is way more fluid, cleaning goes like a charm... 15minutes and it's done...


And as usual, resin prints are so much better than anything else regarding the finish...


I may actually do more resin prints in the future, as with standard resins, this is less cumbersome...

If you're not changing resins and don't have the printer near a window you don't even need to drain the vat unless you're going several months with it not in use.