Quote Originally Posted by mike@aravind.org View Post
Feedback from the hospital added to my notes and questions..... My text is in black, and Rajasekar's feedback is in red....

[AW: My responses are added in square brackets.]


==================================

What parts will we scan? Most likely “machine tool” type parts, no larger than 6 inches in any dimension, and I would expect not much smaller than an inch or so. I can picture machining a prototype part, and once it’s finished, scanning it to allow us to make more parts the same way.

[If you've machined it with a CNC machine, you've probably had to create a 3D model to base the toolpaths on. This model would be more accurate than any scan you made from the finished part.]

Metal and plastic parts (opaque transperant,translucent)(medical devices), 6 inches in any dimension ok- less than or slightly more than one inch also needed.
Inner profile and shape also to be scanned. What 3d modelling/cad software needed to join different surfaces of a part after scanning.(stiching), what will be the prerequisites of the technical staff to do this job?

[If there are interior volumes that need to be scanned, no optical scanner will do that. If your hospital has access to a CAT scanner, that might work. Otherwise, you'd need to fill the volume with something, remove the outside (by dissolving it in acid?), scan the resulting form, and subtract it from the scan of the exterior. Scanners are often sold with software that is used for integrating the various sub-scans that they make into a solid printable object, but if not, one good one is Wrap, from Geomagic (3DS). The staffers assigned to this part of the job would have to learn how to use it.]


==================================

How much accuracy and resolution do we need... I will try to get figures that relate to our needs. For accuracy, I’m pretty sure 50 microns will not be good enough and 5 microns is more than we need. I will ask Rajasekar, along with a question about the required resolution.
5 microns may be ok

[The Capture Mini can resolve details down to .034 mm (34 microns). There may be scanners that can do better, but they will be correspondingly more expensive. It may be that you're expecting more from a scanner than is reasonable; if you need extreme accuracy in your prints (or toolpaths), you won't be printing directly off your scan data; you'll need to go through a reverse engineering step, using a program that constructs clean CAD data from the scans.]

==================================

How fast do we need it to scan? I don’t see that as a main concern. We are only going to be scanning “static objects”. If scanning more slowly gives us better results in a longer time, I think that will be acceptable.
yes

==================================

Data storage: After the scanning is completed, I assume the data will be stored on our computer system, possibly for use in re-creating the object in a 3D printer, possibly with modifications. I already know of parts such as a mobile phone “case” that we may scan, so that we can print more of them.


Sometimes scanned data copied to 3d modelling software like solidworks/pro-e, do the required changes if required and converted to a file format suitable for 3d printing


[Things like that work better if you go through the reverse-engineering step I talked about above; generally scanners produce "point clouds" which normal CAD programs aren't equipped to deal with.]

==================================

I assume the scanner will be set up in an “office” environment, or a separate room kept more or less like a laboratory. Not knowing more about scanners, I’m not sure what else we might have to do, other than cover it when not in use. I’m assuming normal office lighting won’t interfere with the scanner. If it did, we would change the lighting to something more appropriate.

Normal office environment

[That shouldn't be a problem; scanning is a fairly clean process.]

==================================

I assume the above answers will allow us to calculate the cost range for the scanner. I don’t see us trying to make something fit into a budget, but I do see us trying to keep the cost reasonable by not including features or capability that we don’t expect to need.

[Look seriously at the tolerances you're calling out - if you really need to hold 5 microns, be prepared to pay for them.]

==================================

We have an engineering department. As a group, I assume they will learn how to use the scanner. As with anything else, learning how to use the scanner effectively will take time, but I don’t understand enough to say much more. I assumed we would put a part in front of the scanner, start the scan, and after a certain amount of time, the scan would complete. As with a camera, there are probably a lot of settings to adjust before the scan – but I don’t know anything about these, let alone how difficult they are to learn.
Yes

[Scanners often come with a turntable that semi-automates the scanning process, but anything not in the line of sight will manifest itself as a hole in the scan mesh. If you want more or less whole objects, you'll generally have to scan at least twice with the turntable, and then register, align and merge the two sets of scans to make a solid object.]

==================================


Well, that's the best I've been able to come up with to suggest what we think are our requirements. I have a feeling that we may need to re-evaluate some of these requirements, as I think the cost of what I've described is going to be more than will be approved. I read that an accuracy of 5 microns would be very costly. Other than that, everything seems reasonable to me, which is scary as I don't know enough yet to say that. Maybe I should say it sounds in line with what I've been reading over the past few weeks.

There will be one other requirement - I should have listed it long ago. The printer is going to be use in South India, and it would be best if we were to purchase it from a company that provides support in that part of the world. I guess that can come later on - the first step is to find which 3D printers can do these things. Thanks!!!!!!!!!
[I think you're doing the right thing by asking questions and getting information on the process before spending your hospital's money. The 3D printer is just one piece of a rather complex puzzle you're trying to solve, but it sounds like one that's worth solving. Have you investigated how other health-care institutions are dealing with these issues?]

Andrew Werby
www.computersculpture.com