[Hirudin]

I stumbled on these two videos and I think I will follow his recommendations to try to reduce my runout.
TTS Runout - R8 Collet - YouTube
TTS Runout - Holder - YouTube

I also plan to take it a step further and add a chamfer to the bottom 3.5 mm of the collet. I'll do this because I noticed a little gouge on the inside of my TTS collet created by the relief edge near the bottom of the shank of TTS tool holders. I believe the gouge is increasing the runout of the tool holders.* A chamfer there will also make inserting tool holders easier. Lastly, since TTS tool holders have the relief on the bottom of the shank the collet isn't clamping in this area anyway.

* I actually just finished boring out the first 3 mm of the inside of my TTS collet. I tested the taper of a single TTS tool holder before and after the bore. Before it had a consistent TIR of 8 tenths and after it was 3 or less.

[Hirudin]

After that I am considering boring out deep inside the R8 collet.

It seems like the ground surface inside the collets might be too short. Or perhaps they lose tolerance that far into the collet. Basically, when I insert a TTS collet chuck into the R8 collet I can feel a difference in the amount of friction at a certain point. I figure the different feel is attributable to the diameter changing slightly. If I'm able to I will try to correct that diameter difference.

I'm not sure if I will actually go through with this or not. If I do I'll try to add an update to this thread with whether I think it did any good or not.

[neilw20]

Nice work. I bookmarked this for later use.

[Hirudin]

Here's the collet after I added the chamfer. This was a noisy/shaky process. :/ Perhaps an intrepid company could have a bevel like this added at the factory... hint hint. Heck, maybe some stooge at the patent office would grant a patent.

I used a boring bar sticking straight up from my vise.
Feedrate: 10 mm/min
Speed: 400 RPM
I cut 0.25 mm per pass, but I ended up taking multiple passes at each step. I basically repeated the same cut over and over until it made less noise. Maybe ~0.125 mm per pass would have been better. A different tool could probably do a better job as well. I had a boring bar... it seems to have worked.

At the end I stuck a wadded up napkin with sandpaper around it to smooth out the roughness. After removing it from the machine I also sanded all the edges in the three slits that run up the collet. I also sanded the bottom - just to make sure there weren't any burs or anything.

I put a 45º chamfer on this one but now that I've seen it I think a sharper angle would have been better. Maybe take only 0.33 mm off laterally* for each 1 mm axially*. Oh well, next time I guess.

* I hope I'm using those words right!

[Hirudin]

Welp, I've been inking, tightening, and grinding/honing on my collet for hours. I'm thinking three or four hours... I basically followed the instructions in the videos except I didn't wipe the blue ink off before filing (I figured being able to see the high spots so I wouldn't have to "remember" where they were would be helpful) and I didn't check the runout periodically.

I think neglecting to check the runout periodically was a big mistake.

I guess a recap of everything I did probably won't be very interesting... In the end I'm getting 4 - 5 tenths of runout directly under the TTS collet. Much worse than I would have hoped for!

The contact patches between the collet and the spindle taper do seem to have grown, so that's good I guess. :/

I'll probably grind away on it some more tomorrow... I might change-up my strategy too - make the machine do more of the work.

[SCzEngrgGroup]

It seems to me there are a whole host of things that can go wrong with that approach. I'm not convinced the guy in the video didn't just get lucky that his improved.

I would *expect* the primary contact points on the collet to be along the kerfs, and I'm not at all convinced it's necessarily a bad thing. It would also be interesting to see what happens if you go through all that, then DON'T put the tool back in the collet with the correct orientation - things may well be worse than before. In addition, the collet is a "wear item". Once it's time to replace it, you may be screwed....

As I will soon have an ATC, I would not do that to any of my tooling.

Regards,
Ray L.

[LeeWay]

Did he ever check the spindle taper itself? Some error could be right there and if it is, and you hone all your collets and find out later that it was a taper error, you're screwed too.

[Hirudin]

I think the guy in the video is on to something. Three tenths of runout that far from the spindle nose is amazing! Being that he installed and removed his collet a couple times in what appeared to be a single "take" and didn't seem surprised by the end result I'm guessing it wasn't luck.

My spindle had ~2.5 tenths of runout the only time I tested it; this was a few days ago. That was with the indicator right on the taper.

I'm assuming grinding on an R8 collet would affect its runout.
I'm assuming an R8 collet with the same amount of runout as the spindle (but oriented in the opposite direction) would nullify the spindle runout. Not perfectly, but it would help.

I'm curious to find out whether the direction of the spindle runout stays consistent or not. I definitely have my doubts. If I find out the runout orientation does not stay consistent I guess the goal of hand-honing the TTS collet should be to get its runout as close to 0 as possible - so it doesn't reduce OR amplify the spindle runout. Either way it seems logical that the average amount of runout can be reduced by honing the TTS collet. Since I have no plans to use any R8 tooling other than that single R8 collet I think the time taken to optimize it will be time well spent.

[Hirudin]

I also don't want to have to line up little marks every time I change tools. I would greatly prefer to be able to slap them in and hit "go". Unfortunately, in my limited experience, I have found that isn't possible. A TTS holder might have 3 tenths of runout oriented one way and 15 oriented another.

...
I'm not saying I can explain it, it's just what I've observed.

[Hirudin]

I gave it another shot today. I had forgotten that I had put in a new TTS collet yesterday so I inadvertently started on a new collet today. This time I had a singular goal: reduce runout. I didn't care about the contact patches or anything else. I still inked the collet taper to make sure I was scraping an area that was touching, but that's about it.

Anyway, here's the result from today's work... (about 2 hours)
00089 - YouTube

Roughly half a tenth right under the collet, I'd say that's pretty damn good! I have a longer 0.75" bar but I don't know how accurate it is. I'll try it tomorrow.

[Hirudin]

I got the TTS collet down to where it had almost no runout what-so-ever. It did not have any antiseize on it though. After adding antiseize the runout jumped up to ~3 tenths. The antiseize I'm using does seem to be fairly gritty. Perhaps the stuff I got from Tormach is finer. I dunno.

I've been wanting to do a comparison between my Tormach collet chucks and my YinSheng (AKA "eBay") collet chucks. I recorded myself measuring 10 Tormach holders, 10 YinSheng holders, and 4 Novakon holders. Here's the vid...
http://www.youtube.com/watch?v=eBpjOCYVaD4

Given the results here I would say deciding between Tormach's holders at ~$60 apiece and YinSheng's holders at ~$15 apiece is a tough decision. I would personally lean slightly to YinSheng. At $15 a pop they're still half the price even if you have to throw half of them away. And the ****tiest collet holders could be used as very nice drill holders. Also, for me, the shipping was faster than Tormach and actually a little cheaper, despite coming directly from China.

To the best of my knowledge there has been very little information released about the Novakon automatic tool changer being built by SCzEngrgGroup‎. If you're planning to buy one when they become available it might be prudent to ensure the collet chucks you buy will be compatible with it. I suspect that these particular YinSheng holders will not be compatible straight out of the box. It seems possible that they could be modified to work though. I imagine they would look something like the second tool loaded in the video demonstrating SCzEngrgGroup's PDB in this thread.

[Hirudin]

Since this is so closely related I'm going to copy and paste a message I posted in another CNC Zone threead: TTS Tool Holders...

I got my Darkon ER20 ATC Standard Profile Precision Chucks yesterday. It was a nice surprise to see an extra one included in the package! They look very nice! They were also basically clean inside their packages - I cleaned them with alcohol anyway though (I hate how Tormach's TTS collet holders are practically dripping oil). I just recorded a video of myself testing the runout of all 11 and I've uploaded it to YouTube.

Here is the measured runout of each of the collet holders that I tested - in the orientations in which I tested them*. The runout of my spindle and R8 collet measured at 3.0 tenths or better.

[values are to the closest half a "tenth" as measured]
25. (D) less than 4.5
26. (D) less than 7.5
27. (D) more than 9.5
28. (D) close to 2.5
29. (D) more than 3.5
30. (D) close to 5.5
31. (D) more than 5.5
32. (D) less than 12
33. (D) close to 15
34. (D) less than 4.0
35. (D) close to 6.5

* I've found that rotating TTS collet holders in the R8 collet/spindle will give you different runout measurements. I think these numbers could be better (and could also be worse) if I was to test multiple positions.

Runout test: 11 Darkon quick change ("TTS") collet chucks - YouTube

[Hirudin]

A few pics of the collets I've tested... If price isn't a factor I'd say the Tormach tool holders are the best.

Tormach
+ high quality
+ low runout
- high price

YinSheng
+ low price
= decent runout
- not as robust (maybe not an actual issue)

Novakon (note: low sample size)
+ solid shanks
- quality issues
- high runout

Darkon
+ nice quality
= decent runout
- [nothing worth mentioning]