using 3 diamond pins as locating for optics?

[rhmeng]

I saw a callout for how 3 diamond pins can be used to center two parts relative to each other Link (also attached screen shot). I am trying to wrap my head around how using this 3 diamond pin design is any better/more accurate than simply using a pin and slot design.

My application is attaching two lens housing parts to each other, both of which have optics, and i want to keep the optical axis of the parts aligned. So i was planning on using 3 diamond pins when I saw the example above. Now that I am creating the drawings I am wondering how to dimension/locate/tolerance the three diamond pins, and leaning towards just using a pin/pin & hole/slot or pin/diamond pin & hole/hole.

Also was looking at using bushings to press fit as the holes which have tighter tolerances than I think you can hold in an Aluminum part.

Any input is appreciated.

 

[JNieman]

You don't want to press fit pins into bushings (Especially not COTS items that might not be intended to be paired.) unless you're talking about simply using a different material pressed/threaded into he aluminum and then drill/mill/reaming that material in-place.

There's no problem holding tight tolerances on aluminum parts unless thermal properties become an issue during usage - but bushing don't solve that.

Tolerances depend on what their usage needs to be. Personally, I think it becomes a mostly straight forward hole pattern positional-tolerance scheme. There are several examples in the ASME Y14.5 standard. It's the same geometry as a bolt flange except you will likely have much more stringent requirements.

 

[rhmeng]

@JNieman I was not meaning that I wanted to press fit pins into bushings. I was talking about on one part press fitting the pins, and then on the mating part press fitting in bushings that would then mate with the pins for alignment.

I was hoping someone had some insight on the three diamond pin locating method vs hole and slot. I think it is as you say just a hole pattern scheme. Thanks for the info.

 

[3DDave]

It is easier to use a reamer to get a precision diameter than to make a precision width slot. Additionally, a round pin in a slot will only have line contact while a diamond pin may do better.

With 3 pins, the mating of these parts is overconstrained. This means that there will have to be more slop to maintain clearance or there will have to be deformation to assemble it. Typically one would use one round pin for location and a diamond pin for orientation.

 

[JNieman]

@rhmeng, Sorry, I misunderstood your bushing implication. Yes, I'd absolute bush aluminum holes that were receiving diamond pins, if aluminum was the best material otherwise.

@3DDave, precision width slots are not difficult to employ when you buy a slotted bushing Let someone else make the tough part and, as you say, just make a precision hole.

I have never used the 3-diamond-pin system, but the best I can tell... it's only benefit is "playing the numbers" regarding tolerance stack up. It's only way to make a "more precise" mate-up without using more-precise components, is to use more of them and average out the outliers. With two holes, you have a chance of one being at min, the other at max, and ending up with a less-satisfactory location that's still in tolerance. But using one more feature of comparable precision, you can get it to 'settle in' to a more precise location. I'm not explaining my thought very well, I can tell... Hopefully that wasn't too muddy.

This is also possibly where the people who adhere to the "no value is any better than any other value when they are within the acceptable range" will balk and call me a heathen.

Also, I don't think it's over-constrained. I think two non-perpendicular diamond pins would not fully constrain the part on their own- especially at a 120-degree orientation.

 

[pmarc]

rhmeng,
Is there any chance to use a tightly toleranced pilot diameter in one part that would mate with a tightly toleranced pocket/bore in the other part?

 

[rhmeng]

pmarc,
not sure if i follow. Are you saying use a pilot diameter and then hold the two parts together and then drill them both at the same time? Is that what you are getting at? How would two holes mate with eachother?

And Im with JNieman that I think that two non perpendicular diamond pins would not fully define a part, as I think there would be a lot of slop in the rotation. But then adding a third diamond pin that is not perpendicular to either of those pins would still allow rotation, i think...

 

[pmarc]

rhmeng,
I was thinking about something similar to the adapter/crankshaft connection from this illustration:

http://files.engineering.com/getfil...-64ba-439f-9aca-01beff43157e&file=capture.PNG

I simply asked if in your case this kind of configuration was possible.

 

[3DDave]

JNieman,

If you create a free-body diagram you'll see that two diamond pins produce an instantaneous center (even at 120 degrees apart), which controls the location of the part, making it nominally fully constrained. The third pin will ideally share this same center, but variations will cause there to be 3 separate instantaneous centers in the fabricated parts. Since there can't be more than one center for a part to rotate about, the part will either require enough clearance to encompass all three or the part will deform to fit.

The lens mounts I'm familiar with use a precision pilot diameter to center the lens and a pin in a slot for rotation control. Since rotation doesn't affect axi-symmetric optics, the tolerance on the pin/slot only needs to be good enough to align the electrical and other mechanical connections (focus drive/shake-reduction alignment). Earlier than that, lenses used threaded attachments, which are self-centering, but have arbitrary orientation depending on where the threads start.

One other system I've seen used radial slots and precision keys to account for variable materials expansion; like putting the diamond pins at right angles to those depicted. This configuration forces the instantaneous centers to be far apart, constraining rotation while also fixing location. They were match machined so there was no need for clearance to make up for location variation allowing 3 of them to balance side-loads.

 

[JNieman]

3DDave - I agree that 2 diamond pins, even at 120d relation, COULD fully constrain a part in that plane. The only lack of full constraint I referred to was due to size differences in mating features.

I believe you are not accounting for the minute difference in actual-diameter between your typical bushing and pin size, whereas I am taking that into account. It is a small amount, but when talking about precise locations in certain contexts, it's enough. Diamond pin diameters' actual size are usually ~.0002"-.0005" under nominal, with a +0/-.0003 tolerance. The corresponding bushings are typically .0002"-.0015" over size. So you could still have a part 'jiggle' with only two diamond pins in bushed holes. It may not seem like a lot, and in most situations that'd be right. But occasionally it's crucial to reduce repeating-location within less than a couple ten-thousandths.

All that said, we use the typical pin+slot or round+diamond all the time, with much success.