My project (1GE)

[Smoker]

When you put the individual CAD drawings onto the same diagram, how do you know the exact position of each part?

Do you have to scan the complete bike to get the location?

[Speed Freak]

Most of the positions you can see on the picture are just estimated.
If I have the positions, means I have the counterparts in 3D it's just positioning like in reality.

Trying to scan the main parts of the bike in near future.
Frame, rear frame and front stay are the most important parts for positioning the other parts.

But I have close to one month holidays now and did not manage to draw a single screw...

R1 needed a carb rebuild to be ready for 2 track days, another bike is reactivated after 2 years of storage and of course wife and daughter also got some family time.

[Speed Freak]

[Smoker]

Man, I love this stuff!

How long before you can put the wheel CAD into the assembled bike CAD?

It's going to start looking like an RZ.

[Speed Freak]

The wheel isn`t that complex. Only some elements which has to be mutliplied 3 or 6 times.
Biger problem is that the scan data is quite big (1GB), so even my CAD workstation is very slow with this data opened.

I think I need to reduce the quality of the scan result to get it down to 300-400MB.
Or a different way would be to scan only 1/3 of the wheel which is enough to draw it in CAD.
The full scan looks nice but it`s somehow an overkill...

But anyway, I need to find time to draw all the scanned objects.
Oil pump housing - ~80% finished
Headlight - ~5% finished
Upper crankcase - ~30% finished

[Speed Freak]

95%

[Smoker]

Only 95%?

Looks amazingly good. Better than any actual part I've seen.

If you get the frame and wheels into the CAD assembly, I'll nominate you for Greatest Of All Time.

And, you know it needs to be done.

[Speed Freak]

Some threads are missing, balls for closing some passages are missing and the oil inlet spigot is also missing. I think this evening it will be finished.

That's a common problem of CAD, it looks finished quite early if you want, but from technical side it will not work.
If you know this fact you can play with it.
If you want to show your boss (who is not involved in the technical process) that you are the fastest engineer in the world - show him the model including all radius.
If you want to get more time for finishing the model just hide them and show the model with sharp edges...
Of course you have to talk about the high technical complexity then which requires a lot of time...

[Jeff B]

Just curious, have you done this with any other bikes?

[Speed Freak]

No and I will never do this with another bike.
This is a "once in my life" thing without any practical sense as already discussed in the beginning.

Maybe if it's finished I will go a different way creating something new instead of just drawing existing parts.

[Speed Freak]

Only the spring is not right...

[Speed Freak]

Again very happy about the accuracy of the scanner.
It`s a bit tricky to get such a big part scanned without aligning issues, even with marker spots.
Especially from center part to the outer ring - only small surfaces available for marker spots (7mm diameter) and you need to get 3, better 4 marker spots between 2 scans to align it.
The software can align the single scans by it`s own, but the wheel is too symmetrical for that and the accuracy is not sufficient for a scan which is made out of 50-100 single "pictures"

The scanner has one issue, the detection of the markers is very poor in my opinion.
It`s very sensitive to the brightness. The markers are bright white spots with a black ring around.
The sand blasted wheel is a bit darker, so I can`t put the setting to high brightness, otherwise the scanner doesn`t get the 3D geometry of the wheel.
Most of the marker spots are overexposed then and it doesn`t catch them as markers... Then you can only delete the latest scan and try to find a new position of the scan that it catches at least 3 common points.

But I`m getting better to find the right areas for the spots, right amount and good scan angles.
And I have some other tricks now
I`m testing a lot because I want to scan the frame parts. They are a nightmare for alignment.
No big surfaces to place markers on them, long slim pipes where a small alignment error is multiplied several times from one end to the other.

But the result is quite good...
Diameter of the wheel bearing is 42mm - just took 3 points from the scan and draw a circle => 41,97mm

I`m only missing an option to reduce the amount of geometry information in areas where I don`t need it.
On a flat surface the point cloud can be simplified without quality loss but if you reduce the point amount at edges it leads to a radius instead of a sharp edge.
I have no software which is able to reduce only information in certain areas, they all reduce the information of the whole part.

The CAD file has 3GB... Not funny to handle such big data amounts.

[Speed Freak]

Starts to look like a bike...

[lost1750GTV]

I came to this thread late, as a more recent purchaser and having specific search queries .
Yesterday I read this from the beginning - and on page 10 there is a great exploded view of the petcock, and mention of the glues holding fuel fittings together.
I expect I would have saved time/energy resolving my various leaks if I had read the earlier posts earlier.

So just wanted to add that your efforts are another great resource.

[Speed Freak]

Inner core finished.


Valve hole, threads and some small details missing.
The end is near
At least for this part...

For the rear wheel I will be a bit faster as I can reuse some of the front wheel geometry.
Also the core is not that complex.