ER Collet runout investigation Cheap Vs Brand

Whilst developing my CNC router for my Honours Degree, I had numerous issues with tooling and run out.

At the time, I found that the 1/8 collet I was using for my V tools was not concentric though the centre.

Since then, I have acquired several small diameter collets and despite still encountering run out have never really investigated why. So, this morning I went out onto my lathe and chucked up an ER11 8 x 100 chuck and dialed it up on the shaft and around chuck head so that there was a measurable 0.01mm of runout on both about an inch away from the lathe chuck.
My understanding that if there is 0.01mm of runout an inch away, that at 2 inches there would be 0.02mm.

For the benchmarks in this case, that would mean anything more than 0.02mm of runout 2 inches away at the tool would be tool runout and not chuck runout.

Enter the contestants for this benchmark, 3 ER11 collets used for 1/8 bit tool holding.

1: A 3.5 - 3mm collet purchased (I think) from Chronos tools in the UK
2: A 1/8 collet acquired with a dying spindle included with a second machine
3: A 1/8 collet purchased from china about a month ago from merlin tools in a blue/yellow box

My main concern is repeat-ability i.e. repeating the same run out after multiple tool changes so during each test of the collets I released the nut, twisted them about and simulated tool changes allowing for re-positioning of the collets.

First up, the most recent collet #3 was placed into the chuck. The quality itself is terrible, despite being in a sealed wrapper, there is still swarf between the slots. When cleaned and chucked up in the holder, I placed a carbide cutter in and measured the runout. This came in at an astonishing 0.08mm, minus the 0.02mm from the holder and thats a 0.06mm of runout.

Secondly, #2 collet. This was included with a spindle which was included with a cheap PVC Zen toolworks lookalike. Quality wise, it looked ok. It was by no far stretch as good quality as some of the other collets. Chucked up in the holder resulted in a reasonable 0.04 - 0.05mm of runout meaning 0.02 - 0.03mm of runout minus the 0.02 from the holder. That's still a thousandth of an inch of runout and for most engraving is simply unacceptable.

Lastly #1 Collet. I believe this was purchased from Chronos tools UK whilst waiting for the Original 1/8 collet to arrive from china (which was swiftly binned) Although it is 3.5-3mm size, it allows for 0.5mm of compression and under compression, accuracy of collets decrease so I was interested to see how this turned out. Chucking the tool up into the ER holder I measured between 0.025 and 0.03mm of runout. That is only 0.01mm of runout on top of what's already there and somewhat impressive.
Unfortunately, the issue with this is that the repeatability is a little off. Sometimes it can be a little higher making it difficult to gauge.

The conclusion of this little investigation found really something that anyone should really know anyway in that buying cheap may not always be the best solution.

I hope that someone may perhaps find this useful :)