3D Probe - Tri-balls type accuracy & DIY

[RCaffin]

I have attached a paper from Renishaw which mentions this with the key word 'trilobular'.
It is also mentioned in US Patent 5,146,691 awarded to Renishaw.
You might find this paper of interest as well:
https://www.precisionballs.com/Micro...Components.php
and this one too:
http://www.measurement.sk/2010/S1/Salah.pdf

It is commonly said that down in the ultra range, halving the error costs 10x more.

Cheers
Roger

[Mecanix]

Thanks Roger. I'm yet to arrive the pre-travel modulus calc (like Renishaw did) however I'm gauging the required rate based of this baseline data (based off Bob's probing velocity & impact).
Adding several types of pre-load rounds up those figures quite nicely and sort of map into a 6 petals shape. I'm yet to be calling this final on the validation side, there are a few tricks I need to give a shot before e.g. conical spring and spring-pre-loaded ball. Failed greatly on the disc spring though so this one can be ruled out. If both the conical and preloaded ball doesn't add any advantages over a cylindrical spring, I'll just call it a day and get a standard rated spring manufactured.

Note that if you reduce the space between the balls the required force is also reduced, although this may affect repeatability and/or create an instability during re-positioning. I've settled for 5mm with R2 balls & pins here.

[maxspongebob]

This looks odd. Why is the polar chart not perfectly symmetrical? I wouldn't expect it to be so uneven.
I would expect that absolute least amount of force to be required when it is applied axially to each of the 3 pins.

Just wondering.

[Mecanix]

This looks odd. Why is the polar chart not perfectly symmetrical? I wouldn't expect it to be so uneven.
Just wondering.

'Facets' and angular set tolerances of CAD i.e. compromise between precision and solving time.

Not a real-world (perfect) assessment but gives a designer an understanding of the mechanism behavior prior real-life testing. Essentially saving you from going back and forth 10 thousand times from & to the prototyping hard-bench. In a case like this the analyst should carry out several solves with varying tols and 'average' the output data, I-did-not-bother in this particular case instead ran it twice with a higher tol only to confirm the prior solved data.

e.g. facets set at 0.001mm, coulomb velocity and friction to 0.01, and angular tol at 1deg (for instance) would take 15~30min per solve X 36 (more or less 15~18hrs lol). You'd get your perfectly symmetric map however non-sense just to find out the highest F point, which is what I was looking for the next step. Plenty enough for me "0.01N".

Sorry for that irrelevant software and solver tutorial (lol), but hope this helps none the less.

[maxspongebob]

All good stuff. I am learning from your build. Thanks.

[Mecanix]

All good stuff. I am learning from your build. Thanks.

Pleasure. I'm impressed you can read and understand my frenchglish.
I'll see if I can get future data on that probe slightly less compressed resolution-wise next time, or do my best to explain why it is not at least (my bad).

[RCaffin]

Why not a 'perfect' trilobe? Two main reasons:
1) approximations and finite steps in the theoretical analysis (you are not looking for an absolutely perfect solution)
2) micron-level differences in sizes and positions in the mechanical world (plus surface irregularities)

"The difference between theory and practice is greater in practice than in theory."
anon

Cheers
Roger

[Sterob]

Hows our new Probe coming along?

[Mecanix]

Hows our new Probe coming along?

By worlds-size the best tool in the shop. Ultra compact, programmable, accurate and rather stable. Like it, a lot.
Plugged in bluetooth comm with a 4.2V lipo just for the kick of it, works great. Thinking about integrating an AM/FM radio with a sml speaker next

[dharmic]

That is a thing of beauty, Mecanix. Impressively quick build time too - usually this kind of project takes me months to tickle along in the background. Well done and thanks for sharing!

[bill south]

Great Work Mecanix and ( I must add zone associates).
Projects like this usually take months to complete and at times never get done. I am very impressed with the dept of research you and the associates have put into this project. If you are going to market the plans or even a finished device shoot me a PM. I think I want to be in line for one.
I watched such a unit being built as a "reasonably priced unit" but in the end, I guess the creator and myself had a very different idea what reasonable is! He built a beautiful unit but after spending months watching, the final product was not within my "hobby" means. But I'm always looking to upgrade my tooling and very much enjoyed your build.
Thanks for posting and very impressed.
Bill

[Sterob]

Yes,...well done!

[Mecanix]

Yes,...well done!

Knew you'd fancy seeing our new probe Thanks



@dharmic, @Bill, thanks guys. Glad to hear you liked the built as much as I did. They are good fun to make after-all so time flies when you dive into this. Happy I've made it before self-deadline (Oct 1st) which is about the time we plan on digitizing a few parts with the probe you see on the pic above.

Bill, the one you are looking at isn't all cheap... something like 50-60 hrs went into this all together I'm sure (roughly), so not a cheap one-off probe at all lol. Possibly what happened with the one you previously got... that builder must have passed the rate/hrs to you guys which is a bit unfair :/ I actually do plan on running a small batch and already have 6 standby; 3 of those going to the same guy (I think he plans on crashing two lol). I'll add you in if you want. Think in the line of $125/135 machined/anodized/assembled/cabling/packaged. Something like that. Just can't think of anything that could skyrocket the price above that (AM/FM Radio?).

Next Step: Running this probe 25,000+ times on my scans here and watching how it performs. Ultimate test, I'll report back. If all goes well this can ship out to you during Now'19.

Special thanks to Bob(Max), Dharmic, Roger, Britt, jkkmobile and everybody else who've contributed brilliant ideas, support and helped along the built! Cool team, great forum. Very grateful guys, thanks.

I'll keep this thread up to date as often as I can guys.

[maxspongebob]

Just getting back to this forum and I am so happy that you have completed the probe. Fabulous looking device. I am dying to see a video of it in action.

[Mecanix]

Just getting back to this forum and I am so happy that you have completed the probe. Fabulous looking device. I am dying to see a video of it in action.

Yea, done that! Struggling to select a good 3rd party Digi Point Cloud Software lately, although my CAD can interface both modbus or OPC (mechatronics) so I might just port this over that.
3D probe is very much live, ultra sensitive also, fairly happy with the end result. ps: don't ask me to make vids man I really sux at this stuff... a quality camera is possibly what I need maybe lol
Anyway, here's one of 14sec. https://www.dailymotion.com/video/x7lsi5b

Including two screenshots for you also. Sure you'd like to see how things look like under the hood

[bill south]

Thanks for the update Mecanix!
Yes, put me in for 2 of therm if available. I always want a spare on the shelf to make sure the primary is never damaged. Seems I only kill those instrument that I only have 1 available.
Many thanks again.
Bill

[Mecanix]

Yes, put me in for 2 of therm if available.

Bill, start with one and if that works for you then crack on for a spare one (my advice). I plan on making a few so I'll have that spare for you if you really need it anyway.
If however you insist having the spare ready and in hands then, I'll be happy to send two also Thanks btw!

In the Nutther News; I bet some of you are quite curious to know about the accuracy of this type of probe. In fact I was also, and from the beginning... if you read my first post.
Verdict on that bad boy is a whooping max of 4 microns (4um) (repeatable). Reading on the pic is taken at 21dC and on a side that requires the most force to lift that loaded assembly (i.e. worst scenario).
Sure some here will debate that theirs is within 1~2 microns, sure... I believe you, mine too, but only on the lightest points/sides.

Conclusion: "Accuracy between 1~4 microns". For one and with shop's machine-tools within 0.01~0.02mm accuracy (huge max), I call this project a success. Beats my machines by twice!

PS: I also have a HSM production tapping center dialed/squared to roughly 0.003mm (depending on the time of the day lol). I bet this probe would be sufficient for this one also, I might actually try that eventually.

[bill south]

Thanks for the update Mechanix.
I would still like to have two of your probes if available. As stated, I do like having a spare but also, it's a bit of a pay back for all of your effort. I do hope you are making a profit but being a hobbiest, I never think of the cost of something I'm building, but I just want to be satisfied with my effort. Thank for posting and let us know (PM) when you need payment!
Bill

[Mecanix]

Oh wow, cool Bill, thanks. So kind you are almost forcing me to add that AM/FM add-on! I'll see what I can do about that but can't promise
And yes, absolutely, I'm making a profit. Pretty cool in fact, considering this is done from the home shop side (i.e. no heavy bills/insurance/overheads). Counting $20~$25/hr, roughly.
Takes a good 3~4hrs to get one shipped out (machining time/16 M&L operations, anodizing (40min), pcb/20min, flashing, cabling, assembly, testing, packing, shipping). Adds up to about that, 3~4hrs + mats + probe tip = $125~135.

Had a mate giving me a quote for the machining + anodizing last week and it came to about $220/probe (packaged) :/ That's with overheads (employees, tax, and you name it). So at that home shop price of $125 I think we are both getting a good deal on this one I think Bob wants one too, I'll shoot him a PM to dbl-check with him.

PM you when this passes endurance and ready to turn & mill. Should be good to go early Nov. Too kind, thanks again!

[Hanzo]

Hello people!
From an old Bulgarian Rais 3 axis CNC, i succeed to save one set of Hexagon m&h 3D probe, tool setter and data receiver (optical eye).
I talk about of one m&h IRP40.01 3D touch probe, one m&h IRT35.70 tool setter (Z and X axis) and one m&h IRR91.40 data receiver for this two transmitter, all of them produced by Hexagon.
I want to know if someone can use this on another CNC machine without another components from old "donor" machine and I want to know if someone is interested about that components (use PM).
Try to sell them or give up that idea ?
Thanks, have a good day.