Scanning Vintage Cars

[Matt Young]

Before I got into surveying I used to work in Vintage car restoration, mainly Bentleys from the 1920's/30's.

I will be scanning an ash frame for a vintage bentley, with a view to creating 2D or 3D drawings that can be used to re-create the frame. I know it will be possible to get a detailed 3D scan of the frame, but I am still unsure as to how the drawings will be done. If anyone has any serious Ideas as to how this could be done then suggestions are welcome. The real problem is going to be getting all the joints done correctly.

I may even contract the work out to someone with Polyworks or similar, as it is too expensive for me to buy right now.

Matt

[dirdim]

Hi Matt -

Perhaps this will help. I just uploaded a PowerPoint presentation that we gave at the SME (Society of Manufacturing Engineers) annual RAPID Conference in Orlando, FL, USA in 2008. This is a case study of a project we did to scan a vintage Jeep undercarriage frame. It was completely 'shot', rusted and distorted. The purpose was for the car club to re-manufacture new Jeep frames. So besides scanning it, we also reverse engineered it into an accurate CAD model as if it were brand new. This presentation walks you thru the process, the tools & software we used, and some of the behind the scenes look at reverse engineering into solid CAD models.

Click here to see the PowerPoint Case Study of the Jeep Frame Reverse Engineering Project.

190-jeep frame reproduction.jpg

[Matt Young]

Micheal,

I have had a go with the Rapid form software and it seems to be quite a powerful tool. Its very impressive that you were able to reproduce the parts on the Jeep chassis to such an accurate level of detail.

The Vintage ash frames that we will be looking at will be almost as complex to re-produce. The example shown is similar to the type of frame we are going to scan. the thing that will be hardest to do will be to find all the joints in the framework and be able to model them as individual parts, to produce a set of kit part drawings.

boattail-ash-frame.jpg

I may post up some sample data when the scans are done to see if anyone can help to come up with a solution.

Thanks

Matt

[Matt Young]

Last week, with the help of Steve Ramsey and the new Scanstation C10 we scanned the frame of a vintage Bentley.

To scan the entire frame took one hour and included nine scans. The speed of the C10 and the sharper looking data made this job very easy, I am still amazed at how easy it is to use compared to the previous ScanStations. I have written procedures for using all the previous Leica scanners and their are hundreds of pages, I could put the procedures for a C10 on two A4 pages.

I will be producing some section drawings of the frame to aid in the construction of new frames.

The attached link is to a sample of the data that I am offering to anyone that would like to try and create a solid CAD model.

Frame Sample.jpg

Ultimately I would like to create a set of parts for the frame that can be re-produced using CNC and would use company with the best model to do this.

http://www.warnerlandsurveys.net/FTP/Frame-Sample.zip

Matt Young

[derijones]

Hi
Looks like an ace project. Getting it in to Solidworks is the obvious one for sorting out the joints as this would give you the tools to be able to get the clearances right and model the geometry of each part correctly.
We do kits for boats and while welded joints are far simpler than the presumably tenon joints on this, we've looked at doing large scale kits for timber frame houses and the modelling in Solidworks is pretty simple to get parts that can be used for CNC with correct tolerances for tenon offsets etc.
Looking at the data you've put up, my immediate concern is that you're not going to get the required accuracy without taking off a series of check points as well - it's pretty fuzzy for that level of modelling as there's obviously some details on the faces of the frames - holes in the bottom etc.
I'd suggest getting a series of key points for the corners of the key joints and maybe 4 points along the edge of the frames to give clean control curves to model to. You'd also need to either dismantle it or give some detailed info on how the jointing structure works - which bit tenons in to which as I'd imagine it's an area where a lot of "suck it an see" has gone in to figuring out the strongest configuration and I know the grain constraints can play a big part in getting the strongest joint.
Unfortunately I'm off scanning boats next week, but if you want to drop me an email to discuss further - feel free.
Are you getting in to the wheeling of the skin panels as well - that's another "black art" area that you could get some interesting results from with scanning. It's another pet project of mine for the boat building stuff - same principals, just 10mm steel plate instead of skinny aluminium!
Keep us updated on the project - it's a fascinating one!
Cheers
Deri

[Matt Young]

Hi Deri,

That is one cool looking boat model! some really complex geometry going on there.

Although the joints on the Bently frame look complicated, they are not actually that bad. I have some detailed photo's of the car and I will try and find the ones that match the data sample, so hopefully they will help. The important thing with this particular car is that the finished shape of the frame looks right. When the cars were origionally made in the 20's they were all slightly different and never perfectly symetrical. the most popular car now is the 1924 Le Mans Team car that Tim berkin drove in the race, and this car is basically a replica. The ash frame when completed is covered with leather and the joints are hopefully not seen for another 85 years.

I have thought about solidworks but I didnt know if the scan data could be used with it. I guess that having some key points highlighted on the frame would help in this, and that is something we could do with the high order Sokia instruments that we use.

I used to use a wheeling machine to make Cycle wings, Bonnets, body panels on these cars for quite few years and it was a skill taught to me in my second apprentiship. I used to really enjoy the work and although I enjoy being a survey engineer now, I miss building Bentleys.

I know what you mean about the data being fuzzy, unfortunately the ScanStations are not designed for close range scanning but bearing that in mind, I think that the C10 did pretty good.

The Berkin Blower

800px-1929_Bentley_front_34_right.jpg

Gallery_Birkin_LeMans30.jpg

[Joe Parsons]

Hi Matt, don't supose you could up the .ptx file for the sample section or any other file types you've got going.
Regards Joe

[Matt Young]

Hi Joe,

http://www.warnerlandsurveys.net/FTP/Frame-Sample.zip

This is the link for the PTS file.

Matt

[Joe Parsons]

Hi Matt, I've had a go at that data and knocked out a NURBS surface it's a bit rough though:
Joe

Bentley Screenshot 1.jpg

Bentley Screenshot 2.jpg

Bentley Screenshot 3.jpg

[stevenramsey]

That's amazing Joe. I helped Matt do the scanning so now that the C10 is not designed for this type of work.

Could you explain to the forum what your process to get to this point is as I'm sure others would be as interested as Matt and myself.