Type of Fixture/Jig required 1

[nevizamm]

Hi Guys,

Basically what I am trying to do is come up with a solution of a fixture/jig which can hold different plastic parts which are basically toy helmets, all of which have different shape and sizes (i.e. some are oval, some are perfectly circular, etc.) It is rather unfeasible to manufacture a fixture for every shape since there are more than thirty types. Moreover these helmets have to be placed in a jig with the top part facing the jig such that the measurement of the inner diameter can be measured using a vision camera. I was thinking about something similar to a collet. The only problem being that the orientation of the inner diameter of each helmet has to be perfectly horizontal for the readings to be accurate.

Thanks for your time and help.

Neville.

 

[MintJulep]

Basically, three threaded rods with basically hemispherical ends threaded into basically a base-plate and a spirit level.

 

[nevizamm]

Thanks, I thought of something like that but a spirit level isn't ideal for an automatation

 

[MikeHalloran]

Vacuum cup facing up.
Robot with gripper to pick up from previous station,
place in cup, orient, wait for measurement,
and transfer to next station.
Computer to drive robot.
Vision camera to allow computer to see orientation.



Mike Halloran
Pembroke Pines, FL, USA

 

[dhengr]

Nevizamm:
Can you put three small vent holes in the top of the helmets? These also act as your control points for the camera, which locates the three points in the picture and calculates/measures based on the holes and adjusts the rest of the picture accordingly, in space, and calcs. your dimensions. Then, Mike’s robotic system just holds the helmet in position while the picture is taken, no need for the suction cut, and passes the helmet on, or rejects it.

 

[1gibson]

Can you take the measurements through a pane of glass?

 

[tygerdawg]

Neville, can you provide any pictures?

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[3DDave]

I second on 1gibson.

Either a system to flip the helmets opening down onto glass with the camera underneath or an actuated glass that sets onto an opening-up helmet to force it to be square to an overhead camera.

 

[racookpe1978]

What is your throughput time?
How long do you have to set each one into the jig, align it as close as requried, and and then remove it?
Destructive test, or must the rig be non-harmful on each helmet tested?
What accuracy is required?
How often do you change test setup? Once an hour? Each helmet is different? Or do you run a single helmet for several hours/days/weeks, then change to the next?

 

[nevizamm]

Hi Guys thanks for your feedback. I didn't reply that quickly because I am from a different timezone.

First of all I added some pictures of the parts which I was talking about earlier. What can be seen from these helmets is that the inner diameter does not lie completely horizontally to the plane if put with the hole(inner diameter) facing down. This is why I have to opt for some kind of robotic arm to orient it in the required orientation. Moreover the technician will be collecting the parts from the injection molding section and feeds them directly to the system.

Regarding to what racookpe1978 asked me, there are no fixed parameters although the main purpose of the system is to reduce the cycle time. This should certainly be a non destructive test though. The accuracy of the system is only 0.1 mm. These measurements are taken every approximate 4 hours where each sample consists of only 6 parts.

Has anyone ever seen a turntable equipped with grippers? Because what I was thinking, instead of using a robot, is to construct a turntable with 6 grippers which are able to hold the part in the correct orientation when the measurements are taken. The helmets are fed in the right orientation by the operator by attaching them into a cylinder (which has the same diameter as the that of the helmet) and the griper collects them from these cylinders. I may be completely wrong about this but any feedback is truly appreciated.

Thanks again.

Neville.

 

[nevizamm]

This is another picture of the parts as I couldn't attach two simultaneously. Thanks

 

[nevizamm]

This is a sketch of what I was trying to explain.

 

[3DDave]

If you are stuck at optical measurement, have you seen universal grippers? They are made with particulants in a plastic bag that are conformed to a part, then a vacuum is used to draw a compressive load on the particulants locking them in position. You could have the tech put the helmet onto a holder that orients the part as needed, shoves it into the universal gripper which grabs it, then withdraw the holder leaving the helmet in position for measurement. Literally search for Robot Universal Gripper.

Instead of measuring the diameters could you measure the snap-force to push the helmets onto a head simulator? I think this force is what governs the suitability of the helmet for use. It shouldn't be too tight or too loose. If it is oval then the measurement may fail a part that works just fine.

 

[racookpe1978]

That is why I was asking above how fast the measurements have to get processed.

I like the idea of a air-vacuum gripper to grab the helmet, but then it should drop the helmet on a inclined flat plastic sheet (1/4 inch thick or so). With a "dome" or half-dome inside the helmet positioning area, the flat plate ensures the correct angle, and the dome gets it in a uniform x,y alignment. Then the laser can scan the helmet.

(If the inside needs to be measured, then you put the helmet inside an outer half-dome with the same type of alignment plate. )

 

[1gibson]

You could prototype a universal gripper pretty quickly with a shop vac, balloon, and coffee grounds, there should be plenty of hits if you search for it.