Recently I purchased two half inch bits from bits and bits to use with the 1/2 ER20 collet from the set I bought from Avid. Despit ensuring that it is perfectly clean and well tightend, both brand new bits slip down while being used. Im afraid to use the bits and they were not cheap. Anyone else have issues with this collet?
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.
Easy things first.
Are you snapping the collet into the nut before inserting it into the spindle/tool holder?
Have you cleaned the tapers in the nut and spindle/toolholder?
Is there any oil in the bore of the collet?
Have you measured the shank of the tools that are slipping?
Are you inserting enough of the tool in the collet (at least enough to have the collet “filled”)?
Those aside, the next obvious issue is if you are getting enough torque on the nut. The spec for MAX torque on a ER20 1/2" bored collet with a standard/A/UM nut is 80Nm or 59ft-lb. While MAX torque is rarely required or even ideal, the torque applied to the collet is directly related to the axial and radial slip resistance (ability to hold the tool under load). 59ft-lb is a LOT to ask for on a short wrench.
Hope that’s useful. Let me know if there’s something I can help with.
In addition to those suggestions, did you clean the bit’s shaft too? I find they tend to have some shipping or manufacturing oil on them, and a bit of alcohol or acetone makes the squeaky-clean.
My rule is all new bits get a degreaser wipe-down.
Is the bit in any way going above the collet ? I was experiencing bits falling using a long 3/8" upcut bit. I was inserting the bit so far that it sat above the collet. Once I kept it below the top of the collet - worked fine.
-scott
Opinion mode activate! 
Bits should never be inserted so far that they “bottom out” in the holder. That would prevent the collet from tightening all the way. However, there’s usually plenty of room in there. As TDA noted, you should try to fill the collet - which means at least a bit of shank poking out. IMHO people who say “only grab at the front” are really just not putting enough torque on; these collets are designed to hold tools very rigidly, which means as much contact area as possible, but more area means more force required.
Generally speaking the more smooth shank of the tool you can get into the collet without bottoming out in the spindle the better. It doesn’t make any real difference in the slip resistance (as long as you are at least the length of the collet). However, the less stickout you have the more rigid the tool leading to less deflection and potential chatter.
Quick copy paste example from another post I did recently:
Using millalyzer, the rigidity on an 1/8" shank, 1/8" cutting diameter, 0.25" LOC tool. The force required to deflect (bend) the tool 0.001" goes like this for the stickout (amount of tool sticking out of the collet):
1.0" - 3lbf
0.8" - 5.7lbf
0.6" - 12.4lbf
0.4" - 31.54lbf
For a 1/4" shank, 1/4" diameter, 0.50" LOC:
2.0" - 6lbf
1.6" - 11.4lbs
1.2" - 24.8lbs
0.8" - 63.1lbs
I’ll throw in a 1/2" shank, 1/2" diameter, 1" LOC too:
4.0" - 12.0lbs
3.5" - 17.6lbs
3.0" - 27.2lbs
2.5" - 44.6lbs
2.0" - 77.9lbs
1.5" - 142.5lbs
There is a whole range of issues if you bottom out the back of the tool in the spindle. Mostly it will keep the collet from properly clamping leading to runout. It can also damage the tool, collet, nut, and/or spindle in extreme cases. Off the top of my head it could also force the tool to skew making less contact to different leaves of the collet. That would reduce slip resistance.
Here’s a link to a my post a while back that was similar to this problem - Bit being pulled out of collet during cut
The YouTube video that I linked in that previous post was very helpful to me.
Cheers,
-scott