A quick update.
This is the difference between having the right software vs the wrong.
The smaller dia hole was from the 1st setup. The larger is from flipping and probing for the correct origin.
Here it is with v2 touch probe running 24.1.1
I am happy with thw results. Even the chamfer that was placed on the hole from the previous setup is centered.
That seemed to do the trick, and all of a sudden im back to feeling good about this machine.
Glad you are happy, my friend.
Something else I need clarification on. I just discovered that the touch probes concentricity is reliant on the collets ability to center itself.
I have the langmuire provided collets and just noticed that every time I have to load the touch probe, I also have to re-adjust the runout.
Before I go down this rabbit hole of searching for the perfect 3/8" collet as well as possibly swapping out the set screws with tiny thumb screws. Is there a method for ensuring precise loading every time?
What are your guysâ practices?
I have never experienced that issue before. I would suggest checking the spindle bearings or the spindle itself. Collets, 3/8" in size, with a grip range of 0.359" to 0.375".
McMaster Carr offers high-quality collets. A 3/8 collet costs $80 or more. I have several of them. However, I still believe there might be another underlying issue causing this.
Quality collets are an essential part of achieving parts in tolerance. Now, I donât make parts for SpaceX, either. Parts that are .0005 out are just fine for me. Cheap collets are like putting Walmart tires on a McClaren.
Can you run me through this check? Or provide a link on how to inspect it?
I guess I shouldâve quantified my run out issues.
Everytime I remove and install the touch probe, there is up to 0.005" run out (this last time it was 0.003").
Short of inspecting the spindle bearing. Iâm going to assume its the collet and purchase the one you linked.
Whatâs the consensus on the collets langmuire provides? Are they subpar, above average, etc?
Thanks for your help.
The collets provided by Langmuir are ( I donât remember the price) $50 for 10. The one you are buying is $80 For one. They come with certificates also.
I would find a ground 3/8 steel bar or something like that, and the dial indicates the spindle runout.
I donât have a procedure for this because I have never done it. Langmuir should be able to provide the details. Maybe some of our MR1 rocket scientists will chime in. Several guys have tackled this issue before. The bearings are supposed to be of a super precision type, but Iâll bet they are not. My machine is one of the first two that have been produced. They ensured that things were correct on that first batch after the spindle assembly delay on the initial batch. The first machines had bearing problems that were corrected with a new style bearing set Before the initial release.
I donât believe it is the probe unless you just got a bad one. I keep 3 of them just in case I self-destruct one.
Treat it like a lathe. Remove the collet and mount a test indicator upside down and reference on the taper that the collet seats against in the spindle. Check for slop first in each direction (X,Y,Z). That will give you an idea of bearing condition. Record things like this for future troubleshooting.
Ive never tried this, but if you can set manual spindle control to 10 rpm, you will be able to get a good idea of actual spindle runout. If the control doesnt go that low, gently spin it by hand to get an average. I would suggest gently cleaning that bore every now and then with something no more aggressive than used green or gray scotchbrite that youve cleaned up with brake parts cleaner and coompressed air. Then gently wipe out the taper with a lint free rag. Collet tapers get nasty quick and no one cleans them. If you havent cleaned the taper in say, 6 months or so, there is at least .0001-.0002" of runout to be gained back.
If you have very little runout, your problem is entirely in the cheap collets from LS. Full disclosure, i use those same collets but i never bothered to check their runout. I dont use small endmills often though.
Is this your first time using ER collets by any chance? Are you loading the collet into the nut before putting it into the spindle? If not there will be uneven loading and the collet wonât be true.
The Langmuir collets are inexpensive but have had reasonable runout for me. Inexpensive usually means less consistent, so my small sample size might not be representative.
With the V2 probe and concentricity of the tip set in a lathe, Iâve not once seen any reason to adjust while in the ER collet. Sure, there is probably some runout but flipping parts for me has been dead on. I didnât have a whole lot of issue with the V1 probe though either.
I went througj and checked the runout in the spindle. Surprisingly, it is roughly (guessing based on distance from graduation lines) 0.0002"-0.0003".
As for checking the bearing tolerances in the xy axisâ was a taller order because it was hard to descern between gantry flex and bearing slop. I can get 0.002" movement but I can also recreate that by pushing and pulling the at the top of the spindle. I probably shouldâve anchored the magnetic base to the spindle housing, but quite honestly, these results seem to point the finger in the other direction.
I agree. Would you like me to send you one of my extra probes to see if it could be narrowed down to it?
I honestly think the issue is something fundamental in how Iâm locking the nut. Although I do snap the collet into the nut before sliding in the touch probe. What I discovered (and these are early stages of my hypothesis) is that these langmuire provided collets have a tendency to slide around in the nut and prevent the collet from always seating properly.
While waiting for $105 3/8" collet from McMaster and Carr, I focused on ensuring that the collet was flush with the nut, and I was able to load and unload the probe with repeatable accuracy (+/- 0.0004").
The new collet arrived yesterday, and one thing that I noticed is that this collet actually distinctively snaps into the nut opposed to the langmuire provided 3/8" collet that kind of loosely clips in.
I have only loaded the probe in once with the new collet. In habit form, I got out the dial indicator thinking that I was going to have to zero it again. Surprisingly, it was within my tolerances.
Time will only tell whether or not, I have this licked. But for the foreseeable future, I will be verifying runout before probing.
Thats amazing and very generous of you.
I hope not to take you up on this offer. But if things still continue to go down this path, I may reach out.
One other thing that I didnât mention was that I also pulled the probe apart and put the shaft/flange on on my lathe to test run out. To my surprise, the needle barely moved. Iâd say 0.0001" runout on concentricity as well as placing the indicator on the face to ensure that the shaft wasnât bent.
Iâm hoping that I narrowed it down to your initial thought of the collet and nut combination.
After all the Langmuir probe is a $100 item with great tolerance. A $ 1000 Reinshaw itâs not.
I stopped using the touch probe. Albeit it is accurate, what I found is that it wasnt consistently accurate. I have since shelfed the probe and just use a jiggler. No issues since doing that. The extra 2-3 minutes of time to set a WCS easily outweighs the issue with tossing the dice if the probe was going to be accurate or not.
My probe is fairly accurate, but if I am needing to be precise on an offset I manually indicate or touch off of the part. Iâve had a couple flip-over operations that ended up having some alignment issues, and it drove me back to the manual methods for accuracy.
When it comes to machining from rough stock or billet material I will just use the probe then.l because the 0.005" or so variance isnât enough to matter in most cases.
Is there anyone on here that has noticed a dead spot in their probe? If the spindle is in the right spot the probe will move an extra .050" or so before its triggered. If you are not watching you will miss it big time. Move the spindle and it will trigger. I have had this thing apart a half a dozen times with either a bad board or updating or something and I am over it.
Does anyone have a good lead on a probe that works?
