My project (1GE)
Moderator: rztom
- Speed Freak
- - - - - -
- Posts: 930
- Joined: Sun Oct 28, 2012 6:38 pm
- Location: Voitsberg, Austria
Re: My project (1GE)
It`s winter and cold outside... Due to some other "small" topics like finding and moving to a new flat, being father of a wonderful little daughter and a new position in the company there was not much time for the RD project but I was always thinking about how to make it more easy to get the parts into digital format (especially 3D formed sheetmetal or cast parts).
As always... winter is a hard time for the wallet.
As I know that 3D scans are expensive and I have a lot of parts I calculated what it would cost... Ah, forget it... It`s just because I wanted a 3D scanner at home
My first 3D scan
Yes... it`s a "screenshot" with the phone
I still need to learn a lot how to operate the scanner. It`s not like scanning a sheet of paper. There is much more know how necessary to get good results.
But from first view the scan looks not that bad but I need to check the data at work with a better laptop (not enough RAM at home)
I hope the accuracy is good enough but this sheetmetal part is something like the worst case for the scanner because you need some single scans and there are not so much reference points to match the scans when flipping it from one side to the other.
As always... winter is a hard time for the wallet.
As I know that 3D scans are expensive and I have a lot of parts I calculated what it would cost... Ah, forget it... It`s just because I wanted a 3D scanner at home
My first 3D scan
Yes... it`s a "screenshot" with the phone
I still need to learn a lot how to operate the scanner. It`s not like scanning a sheet of paper. There is much more know how necessary to get good results.
But from first view the scan looks not that bad but I need to check the data at work with a better laptop (not enough RAM at home)
I hope the accuracy is good enough but this sheetmetal part is something like the worst case for the scanner because you need some single scans and there are not so much reference points to match the scans when flipping it from one side to the other.
My bikes:
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
Re: My project (1GE)
Wery good !
tried to scan something with a cheap XYZprinting 3D scannr, but result was dfinitely bad.
which scanner do you use ?
tried to scan something with a cheap XYZprinting 3D scannr, but result was dfinitely bad.
which scanner do you use ?
Alex
- Speed Freak
- - - - - -
- Posts: 930
- Joined: Sun Oct 28, 2012 6:38 pm
- Location: Voitsberg, Austria
Re: My project (1GE)
My decision was to use an Shining Einscan SP.
Middle price range but according to videos and reviews more or less the best you can get for acceptable money.
I would have access to a 3d scanner at work but i don't want to ask the operators all the time for small parts (they dont want that people without training are working with the scanner... it's more a ferrari price range scanner...)
Software looks a bit old fashioned but it works quite well for plug&play scanning.
I'm missing some options to manually optimize the part after scanning. The automatic optimization works good but not perfect.
Common problem of all 3d scanners is that they don't like dark, glossy or transparent parts.
For this part I used some powder spray to make it white.
Good scan result but you can't touch the part without destroying the white coating.
It's also not possible to put the reference point stickers onto the powder, so the alignment of the single scans needs to be done manually and is not as good as with the reference point stickers. Every scan where surfaces are double scanned leads to a worse result.
Here I need a different solution, maybe spray paint for parts which will be sandblasted anyway...
If everything is optimized it should be possible to scan parts up to the size of a motorcycle frame which is the maximum I expect to scan.
After detail review of the scan I'm very happy with the result. Yesterday at night it looked like it has some alignment errors but when I compared it with the real part I found all errors on the real part including the fact that the part is twisted a bit from a crash...
And this scan was more a tryout at 11pm, coating the part in darkness in front of the flat...
Edit: Some pircures
Twisted part
Some errors - holes and dimples - because of missing areas in the scan either because I just did not make another scan or the coating was damaged (automatic closing of the software)
Here a small detail - there are some letters "DOT", maybe 0,2mm deep => they were covered by the powder but they are still visible on the scan.
The 2 bent corners at the hole are also there on the real part.
This result is definitely OK for making propper 3D parts without using other measuring tools too much.
Accuracy - I tried to scan a 2nd oversize piston (new, without coating)
I came to a diameter of 56,85mm on the scan - measurement shows 56,9mm.
Middle price range but according to videos and reviews more or less the best you can get for acceptable money.
I would have access to a 3d scanner at work but i don't want to ask the operators all the time for small parts (they dont want that people without training are working with the scanner... it's more a ferrari price range scanner...)
Software looks a bit old fashioned but it works quite well for plug&play scanning.
I'm missing some options to manually optimize the part after scanning. The automatic optimization works good but not perfect.
Common problem of all 3d scanners is that they don't like dark, glossy or transparent parts.
For this part I used some powder spray to make it white.
Good scan result but you can't touch the part without destroying the white coating.
It's also not possible to put the reference point stickers onto the powder, so the alignment of the single scans needs to be done manually and is not as good as with the reference point stickers. Every scan where surfaces are double scanned leads to a worse result.
Here I need a different solution, maybe spray paint for parts which will be sandblasted anyway...
If everything is optimized it should be possible to scan parts up to the size of a motorcycle frame which is the maximum I expect to scan.
After detail review of the scan I'm very happy with the result. Yesterday at night it looked like it has some alignment errors but when I compared it with the real part I found all errors on the real part including the fact that the part is twisted a bit from a crash...
And this scan was more a tryout at 11pm, coating the part in darkness in front of the flat...
Edit: Some pircures
Twisted part
Some errors - holes and dimples - because of missing areas in the scan either because I just did not make another scan or the coating was damaged (automatic closing of the software)
Here a small detail - there are some letters "DOT", maybe 0,2mm deep => they were covered by the powder but they are still visible on the scan.
The 2 bent corners at the hole are also there on the real part.
This result is definitely OK for making propper 3D parts without using other measuring tools too much.
Accuracy - I tried to scan a 2nd oversize piston (new, without coating)
I came to a diameter of 56,85mm on the scan - measurement shows 56,9mm.
My bikes:
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
-
- - - - - -
- Posts: 609
- Joined: Thu May 24, 2018 5:22 pm
Re: My project (1GE)
amazing stuff that can be done (not by me) in the 21st century.
Re: My project (1GE)
I guess I'm a bit old fashioned. I assumed you scanned it with a touch probe. That was all the rage 20 years ago. I didn't realize they made affordable laser scanning systems now days. Accuracy within .05mm is pretty darn impressive. Have you used Onshape for editing? I've been meaning to give it a try.
- Speed Freak
- - - - - -
- Posts: 930
- Joined: Sun Oct 28, 2012 6:38 pm
- Location: Voitsberg, Austria
Re: My project (1GE)
It's not a laser scanner. It uses basically a standard projector which is projecting a black/white light pattern onto the Part. 2 cameras are taking pictures of it and the software calculates a point cloud.
Then it is clear why it's best to have a white, not glossy part and scan it in a dark room.
The cameras are like eyes. Scanning into pockets is a bit limited as it only scans surfaces which are visible for both cameras.
I don't know Onshape but I guess it's just another CAD software.
At work I have the 3 most used CAD systems. I guess it's not necessary to learn another one.
Basic problem with this scan is that it's just a point cloud and it's not a parametric model.
There is special software available to get a parametric model from it but it's also not an automatic process.
I think it's the best option to simply get the scan into CAD and build up a parametric model based in the scan (points can be used as references).
Special details can be manually measured with calipers or other tools, general stuff which is not so easy to measure can be taken from the scan.
I guess the accuracy of the whole scan as shown on the pictures is in the range of 0,2mm.
Then it is clear why it's best to have a white, not glossy part and scan it in a dark room.
The cameras are like eyes. Scanning into pockets is a bit limited as it only scans surfaces which are visible for both cameras.
I don't know Onshape but I guess it's just another CAD software.
At work I have the 3 most used CAD systems. I guess it's not necessary to learn another one.
Basic problem with this scan is that it's just a point cloud and it's not a parametric model.
There is special software available to get a parametric model from it but it's also not an automatic process.
I think it's the best option to simply get the scan into CAD and build up a parametric model based in the scan (points can be used as references).
Special details can be manually measured with calipers or other tools, general stuff which is not so easy to measure can be taken from the scan.
I guess the accuracy of the whole scan as shown on the pictures is in the range of 0,2mm.
My bikes:
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
Re: My project (1GE)
At work our inspection dept. has something like you're describing. I had no idea how it worked.
Are you able to click on the part to get the nearest point data? In other words is it easy to work with the information it gives you?
I use Solidworks at work but I never seem to find time for my personal jobs. I'd like to find a CAD program that I can use at home without spending a fortune.
Are you able to click on the part to get the nearest point data? In other words is it easy to work with the information it gives you?
I use Solidworks at work but I never seem to find time for my personal jobs. I'd like to find a CAD program that I can use at home without spending a fortune.
- Speed Freak
- - - - - -
- Posts: 930
- Joined: Sun Oct 28, 2012 6:38 pm
- Location: Voitsberg, Austria
Re: My project (1GE)
If you go a bit more into detail and turn on the edges you can see the mesh, every corner is one of the points created by the scanner.
You can select every single point in the CAD software to put a plane on 3 points or use 3 points projected onto a plane to draw a circle (which was not the best point selection in the example... The circle has 10mm diameter in reality)
With professional software for reverse engineering you can select a region (where you as user know that it should be a flat surface) of the part and it interpolates automatically all points in this region and puts the plane there.
But if you take 3 points which are more or less at each corner of the part the accuracy is sufficient.
I`m always searching for some base dimensions/base planes of a part and then I build up the part from there using constraints like parallelism etc.
Often the quality of the hardware part is not that good, so you can have the best scan in the world but on the hardware is smply no flat surface.
It`s a bit experience to know how the parts are initially dimensioned on the factory drawing. If you are used to create technical drawings it`s often easy to find the right corner to measure.
Here I tried to scan the "ideal" part.
It`s white, no gloss and not transparent and it fits to the automatic turtable where the alignment of the single scans works much better.
Software optimization setting was set to "middle" as the software seems to be a bit unstable using "high" - maybe just too less RAM (8GB).
Yes, this is a scan, not a picture
You can select every single point in the CAD software to put a plane on 3 points or use 3 points projected onto a plane to draw a circle (which was not the best point selection in the example... The circle has 10mm diameter in reality)
With professional software for reverse engineering you can select a region (where you as user know that it should be a flat surface) of the part and it interpolates automatically all points in this region and puts the plane there.
But if you take 3 points which are more or less at each corner of the part the accuracy is sufficient.
I`m always searching for some base dimensions/base planes of a part and then I build up the part from there using constraints like parallelism etc.
Often the quality of the hardware part is not that good, so you can have the best scan in the world but on the hardware is smply no flat surface.
It`s a bit experience to know how the parts are initially dimensioned on the factory drawing. If you are used to create technical drawings it`s often easy to find the right corner to measure.
Here I tried to scan the "ideal" part.
It`s white, no gloss and not transparent and it fits to the automatic turtable where the alignment of the single scans works much better.
Software optimization setting was set to "middle" as the software seems to be a bit unstable using "high" - maybe just too less RAM (8GB).
Yes, this is a scan, not a picture
My bikes:
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
Re: My project (1GE)
Really interesting stuff - I love your technical insight into things - I’ve still got to convert the crankcase scan I did a year or two back into a CAD drawing -sadly even with 24gb of RAM my CAD program hangs. My computer is too old to install more memory.
I’d love to redraw a crankcase that would take TZ250 cylinders and get it CNC’d and matched to the lower case - probably cost a fortune though ?
Cheers
Paul
I’d love to redraw a crankcase that would take TZ250 cylinders and get it CNC’d and matched to the lower case - probably cost a fortune though ?
Cheers
Paul
Re: My project (1GE)
Hi BR Michael, Did you find out about item 38? as needing one myself. regards VinSpeed Freak wrote: ↑Sun Jan 08, 2017 5:16 pm More parts:
Some more details needed on the upper air filter case and the "cover" (no.37) is missing.
Can someone send me a picture and dimensions of following parts?
38 - 97701-50116
39 - 92901-05200
The washer is missing on my bike and I don`t know if the screw is original (cross head, 25mm long with plastic thread outer diameter 5mm and inner diameter ~3,5mm)
Re: My project (1GE)
Really neat stuff! Thanks for sharing!!
- Speed Freak
- - - - - -
- Posts: 930
- Joined: Sun Oct 28, 2012 6:38 pm
- Location: Voitsberg, Austria
Re: My project (1GE)
@rd84: I already posted the status of the upper crankcase some time ago. Since then no progress on this part as there was no real opportunity to work focused after normal work.
It doesn`t really make sense to start work for 1-2h, you just think about how to get something done start productive work for 30min and then 2h are gone.
It makes more sense to work 6-8h focused onto the part and keep working on it the next day where you still remember what was done before
For such parts I like to have the equipment of the company (big screen, keybord with number block etc.) but as described it doesn`t make sense to start drawing after 8-10h regular work.
Around 50% of the work is done which took around one week.
It`s a quite complex part if you go into detail.
Try the software 3D-Tool, it`s freeware and I can open the upper crankcase on my close to 15 year old laptop with 2GB RAM. The CAD software crashes when trying to open it.
.stl files are incredible big...
Talked to the Guy in the company yesterday. He is also new to the 3D scanner (combined laser scanner and touch probe) and it might be possible to do one or another more complex part as training for him.
I will try to get the lower crankcase scanned including a detail measurement (0,01mm or better) of all the machinings for the shafts.
There is still the problem to match the bore of the upper case to the bore of the lower case, but if you bolt the 2 cases together and you have a good machinist you can get it done.
@vin: No sorry, I never got any info about this 2 parts. It`s marked as part with missing informations in my parts list.
It doesn`t really make sense to start work for 1-2h, you just think about how to get something done start productive work for 30min and then 2h are gone.
It makes more sense to work 6-8h focused onto the part and keep working on it the next day where you still remember what was done before
For such parts I like to have the equipment of the company (big screen, keybord with number block etc.) but as described it doesn`t make sense to start drawing after 8-10h regular work.
Around 50% of the work is done which took around one week.
It`s a quite complex part if you go into detail.
Try the software 3D-Tool, it`s freeware and I can open the upper crankcase on my close to 15 year old laptop with 2GB RAM. The CAD software crashes when trying to open it.
.stl files are incredible big...
Talked to the Guy in the company yesterday. He is also new to the 3D scanner (combined laser scanner and touch probe) and it might be possible to do one or another more complex part as training for him.
I will try to get the lower crankcase scanned including a detail measurement (0,01mm or better) of all the machinings for the shafts.
There is still the problem to match the bore of the upper case to the bore of the lower case, but if you bolt the 2 cases together and you have a good machinist you can get it done.
@vin: No sorry, I never got any info about this 2 parts. It`s marked as part with missing informations in my parts list.
My bikes:
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
-
- - - - - -
- Posts: 3236
- Joined: Sat Feb 04, 2012 7:36 am
Re: My project (1GE)
What is the question with this screw? It is suppose to be a simple self tapping screw.
Re: My project (1GE)
Hi,
I see you made some progress - I definitely understand that there aren’t enough hours in the day.
I’ll try another software tool - I was surprised when the 24gb of DDR RAM wasn’t enough.
Cheers
Paul
I see you made some progress - I definitely understand that there aren’t enough hours in the day.
I’ll try another software tool - I was surprised when the 24gb of DDR RAM wasn’t enough.
Cheers
Paul
- Speed Freak
- - - - - -
- Posts: 930
- Joined: Sun Oct 28, 2012 6:38 pm
- Location: Voitsberg, Austria
Re: My project (1GE)
@Silverstrom: I have 2 air filter cases but on one the washer is missing and on the second everything is disassembled.
Not really sure if the screw is the one which came with the bike.
It's not my goal to get screws from the local DIY store into CAD
If you have a bike where it looks like this screw was never touched it would be nice if you can take some dimensions and pictures from it. Same for the washer.
@rd84: 24gb should be more than enough.
At work i had 16gb before it was upgraded to 32.
Not really sure if the screw is the one which came with the bike.
It's not my goal to get screws from the local DIY store into CAD
If you have a bike where it looks like this screw was never touched it would be nice if you can take some dimensions and pictures from it. Same for the washer.
@rd84: 24gb should be more than enough.
At work i had 16gb before it was upgraded to 32.
My bikes:
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04
RD500 YPVS 1GE
RD350 YPVS 31K 1985
Honda CBR 1000 RR SC57
Yamaha R1 RN04