Usual preface: I’m with PreciseBits, so while I try to only post general information take everything I say with the understanding that I have a bias.
That very much looks like failure from too much torque. 100 ft-lb may be too much. I’ll explain but fair warning it gets complicated.
So to start with ER collets were originally invented by Rego-Fix. It’s actually what the “R” in ER stands for. Where pretty much all these torque numbers come from is their recommendation. However, if you look at their catalog that these come from these numbers are not listed as a “recommended” torque but MAX torque before damage. So right off the bat there a potential issue with materials and finish (hardness, matching taper, etc.) where these might not be correct for other manufactures. Also keep in mind that for most collet sizes max torque depends on the bore size and nut thread. Unlike other ER sizes it’s not split up very much for ER32 which might be a problem.
That out of the way there’s a few reasons that one would want to apply more torque to a collet. The biggest one is slip resistance. Basically how much force it takes to spin the tool inside the collet or pull the tool out of the collet. The second would be runout or how much the bit spins off the center axis of the spindle. There’s also a vibration dampening number but that goes even further into the weeds and I’m not going to address it other than you need a lot of cutting force (e.g. heavy metal cutting) for it to functionally matter.
For the most part unless you are aggressively milling metal with large tooling or trying to use a poor fit/worn collet you are never going to come close to needing the max slip resistance and therefore max wrench applied torque.
The runout number is more complicated. Basically there’s a minimum amount of torque you need to properly compress the collet and get good runout. However, in our testing this is not even close to max torque. In fact runout typically gets worse as you approach max torque. We don’t have enough data yet to publish recommended numbers. But here’s some of our initial data.
This is from our internal data and only on a single ER16 1/4" collet with a mini nut (max torque 24Nm/18ft-lb. It’s from 480 data points. The face measurement is at the face of the collet, RMT is 1" out. Again, this is not enough data to do more than show the pattern we’ve seen. We’ll probably put something out when we feel we have enough data. Units are in thou (0.001"/ 0.0254mm).
This pattern basically plays out in every collet size and bore we have tested. Meaning that at least for the data we have collected so far max torque will give you worse runout with greater slip resistance. However, unless you need the greater slip resistance (heavy metal milling, poor or worn collet) you should never use it. There’s much more advantage to the runout reduction than the slip resistance increase (plus not possibly damaging or wearing the parts faster).
I don’t have a lot of data yet on ER32. The initial set says torque for a 1/4" should be somewhere around 25-30 ft-lb for the best runout. But again I currently have limited data. What I’ve been using as a stop gap for now is the bore range to torque for a given collet. So for a ER32 1-3mm bore is listed as 18ft-lb and over 3mm is listed as 100 ft-lb. Max bore size is 22mm so 18ft-lb + ~4.3ft-lb extra per mm over 3 making 1/4" ~32ft-lb. That’s not a perfect number by any means but it seems to be working a lot better for our customers than going to MAX torque for the collet size.
To explain the above a bit. Looking at all the ER collet and bore sizes the smaller the bore the lower the max torque. There’s a number of reasons for that but I’ve been long winded enough. The calculation let’s you fractionally increment with less chance of higher runout or damage. But still getting more slip resistance per bore diameter.
One final thing. As a general guideline we don’t recommend using collets, toolholders, and nuts dry. You should never have a lubricant in the bore of the collet but all the tapered surfaces should have something like a dry lube or mono layer of a light oil (e.g. penetrating oil) on them. We have been involved in the steel rule die industry for decades and not doing this was the primary reason for spindle rebuilds and rapid collet, nut, and toolholder wear. It was so bad in that industry that they were having to rebuild their spindles every couple years. We eventually started making a lubricant for it and applying that to those parts basically eliminated rebuilds and tool holder sticking while increasing the life of the other parts massively. Probably didn’t help that the spindle distributor was telling everyone to clean all those with alcohol either…
Hope that’s useful. Let me know if there’s something I can help with.
