Does anybody have any good techniques for inspecting some small, flimsy silicone gaskets? Depending upon how you set it down, it's dimensions change. How do we verify that the dimensions called out on the drawing are correct if it is so flimsy?

Currently we are using a horizontal optical comparitor and placing the part on a piece of upright acrylic. The weight of the part deforms itself when 'stuck' on the glass vertically and we would like to find a way to avoid any deformation for quick and precise inspection.

We are thinking about making a go/no go template (an overlay on the display for the critical dimensions, but we would like something more concrete for the various radii this part has. Any suggestions are appreciated.

An interesting problem, and one for which I have no first hand experience. But I'll try to answer anyway! :D

My first thought is that if you can't measure it, does it really matter much? :rolleyes: Perhaps you can just test if it works for the application it is designed for - if it fits, its good enough!

Another thought would be to weigh the piece. A bad weight would tell you the piece is a bad size. However, the reverse is not true - the piece could be too big on the OD but too thin and still fall in the correct range.

If you really want to test the dimension, it seems you need a way to hold the piece and measure the piece without putting any stress on it. The optical comparator approach should perform the measurements without stress. Holding it without any stress might be tougher. Two ideas that spring to mind include

air suspension: like an miniature air hockey table, only with closer-spaced holes and clear materials. Set the whole thing on top of the comparator. One problem is that the part may tend ot move around. (or perhaps slowly shut off the air flow until it settles to the surface.)
fluid suspension. you could float the part on a liquid. A saturated salt water solution has a slightly higher density (1.20) than silicone (1.17), so the part should float. Or glycerin has a higher density yet (1.26), so the part would float better.Tim F

A vertical optical comparator is an option, but a pricy one.

A very similiar option is a toolmakers microscope, and you can get a radius eyepiece as an option with most of them. This may fit better into your price range. Also it allows the addition of assorted thread eyepieces as needed.

What about mounting between 2 acrylic plates?

Is this part molded or stamped?

If stamped, we use a "mylar template" to verify the tooling, which tells us the tooling is ok. Our customers have approved this method, when no other easy way seems to work. Other than making a fixture resembling the actual component the part is used in or a go/no go gage. Do I dare mention, we yet to have a problem with fit, form or function of any parts, using this method. :notme:

This is just an opinion........

I like the idea of making a go/no-go gage based on the how the part will fit in use.

It's pointless to measure what is not measureable.

I run across this with o-ring and quad ring seals.

The part is molded and the challenge is that every time you set it down, the dimensions change slightly. So a vertical comparator and toolmaker's microscope would give us the same problems.

I like the template / go-no go approach and we have discussed this. How can this be implemented practically into a quality system since we have not done it before? We have been doing inspecting our parts (mostly machined) using calipers, micrometers and the like, so this is a new approach and I would want to make sure it was done correctly.

Would someone mind further elaborating?

The part is molded and the challenge is that every time you set it down, the dimensions change slightly. So a vertical comparator and toolmaker's microscope would give us the same problems.

I like the template / go-no go approach and we have discussed this. How can this be implemented practically into a quality system since we have not done it before? We have been doing inspecting our parts (mostly machined) using calipers, micrometers and the like, so this is a new approach and I would want to make sure it was done correctly.

Would someone mind further elaborating?

Could I get a picture of the part and maybe leader lines showing the dimensions you are concerned about or the drawing? Hopefully I'm not the only one, but working in the abstract sometimes causes me to miss obvious t hings.

For inspecting small flimsy parts, I use a Micro-Vu video measuring system. As far as the part moving, you may want to study the print. If it is drawn to ASME Y14.5M-1994 standards, it may specify if inspection should be carried out in "free state variation" or a "restrained condition".

The part is molded and the challenge is that every time you set it down, the dimensions change slightly. So a vertical comparator and toolmaker's microscope would give us the same problems.

I like the template / go-no go approach and we have discussed this. How can this be implemented practically into a quality system since we have not done it before? We have been doing inspecting our parts (mostly machined) using calipers, micrometers and the like, so this is a new approach and I would want to make sure it was done correctly.

Would someone mind further elaborating?

how many parts does your system call for inspecting now?
Can you get a sample the part(s) it sits on/between from your customer?

For inspecting small flimsy parts, I use a Micro-Vu video measuring system. As far as the part moving, you may want to study the print. If it is drawn to ASME Y14.5M-1994 standards, it may specify if inspection should be carried out in "free state variation" or a "restrained condition".

We are designing the part, putting the requirements on the drawing and determining the inspection method. (We are a small operation.) So we could put that it is to be inspected in a restrained condition (ie. within a template).

how many parts does your system call for inspecting now?
Can you get a sample the part(s) it sits on/between from your customer?

Our sampling plan calls out for 35 out of 1000 parts. We have the parts that they are assembled into and have tested them and found that they work.

Attached is a section of the part. (I could not show the whole part due to proprietary constraints.) We are looking to verify the radii called out and that is what changes when the parts are moved around.

We are designing the part, putting the requirements on the drawing and determining the inspection method. (We are a small operation.) So we could put that it is to be inspected in a restrained condition (ie. within a template).



Our sampling plan calls out for 35 out of 1000 parts. We have the parts that they are assembled into and have tested them and found that they work.

Attached is a section of the part. (I could not show the whole part due to proprietary constraints.) We are looking to verify the radii called out and that is what changes when the parts are moved around.

that certainly does not look like it lends itself to measuement.

If I was in your shoes I'd pursue the go/no-go option. Lay it on the customer provided part and if it fits then it's a go.

If you really want to measure you can always look into something like a Smartscope that does video measurement - no concerns about shadows like with a comparitor. Thing is you'll probably end up laying the piece on the mating part anyway to get the lay right.
But Smartscope are, like, $40k.

I can inspect all those dimensions on the Micro-Vu. PM me if you need further details.

I appreciate all of this input. I will probably go with some type of template or go/nogo.

Thanks to you all.

I can inspect all those dimensions on the Micro-Vu. PM me if you need further details.

off topic -
Crash Not: do you use the video measuring system? Automated or manual?
How is it priced? I was pricing these systems about 6 months ago but did not get a quote or demo on the Micro-Vu.

Is this part molded or stamped?

If stamped, we use a "mylar template" to verify the tooling, which tells us the tooling is ok. Our customers have approved this method, when no other easy way seems to work. Other than making a fixture resembling the actual component the part is used in or a go/no go gage. Do I dare mention, we yet to have a problem with fit, form or function of any parts, using this method. :notme:

This is just an opinion........

In my experience, this is the best and least expensive method. You can do a one-time correlation between the hard tooling and sample parts, determine that they work in the assembly, and then monitor the tooling. You can, perhaps, just produce a 2D template that you can lay the part down on and line it up to determine whether the part can be formed to fit the template

Added in edit: If you do something like this, the drawing should specify the method, so that there aren't any peeing matches over block tolerances in future.

I use an automated Micro-Vu, work envelope of 24x18x12 in. Cost was about 42K in 1998.

In my experience, this is the best and least expensive method. You can do a one-time correlation between the hard tooling and sample parts, determine that they work in the assembly, and then monitor the tooling. You can, perhaps, just produce a 2D template that you can lay the part down on and line it up to determine whether the part can be formed to fit the template

Added in edit: If you do something like this, the drawing should specify the method, so that there aren't any peeing matches over block tolerances in future.

Thanks, this makes it much clearer for me.