On my Tormach I use TTS 6 and 10mm sidelock holders for my two standard size shank cutters that do the majority of my basic machining. I also bought 3 TTS ER20 collet holders. I bought my ER20 collets from a commercial supplier, Carillion Industrial Services, that at the time were fairly local. I doubled up on 3 and 4mm collets. If I needed more ER collets I’d buy from Cutwel; ARC don’t seem to give a runout figure.

I learnt early on that on the Tormach it is better to use a small(ish) cutter at high speed and feedrates than a larger cutter. I’ve never felt the need to use facemills on the Tormach, only used them on the Bridgeport and horizontal mill. I use the Tormach TTS system on the Bridgeport, only changing for special tooling such as large endmills (>16mm), face mills, a Clarkson style threaded endmill chuck, slitting saw arbors and boring heads.

When I first bought the Tormach I also bought some imperial sidelock holders, but have never used them. All my CAM programming is done in metric, with metric cutters, irrespective of whether the part is designed in imperial or metric.

Andrew

On 12 July 2024 at 04:42 Ketan Swali Said:

ARC does give runout figures for ER collets

For the sake of discussion, education and insight into the world of collets, how is the given figure determined by any collet supplier? Would the supplier simply report or pass on a number provided to it by the collet manufacturer?

Do suppliers have a quality regime where they pick samples from the collets supplied to them by a manufacturer and send the samples for independent verification of runout (like a UKAS lab. for more mundane items such as soil and concrete)?

It would be interesting to find out what standard tests are used to arrive at this number (is there an ISO collet runout Standard?) and who (if anyone) in the UK has facilities to carry out these tests.

I guess you need some fairly specialised kit to carry out this test. I wonder what extra or different manufacturing steps the ‘better’ collets go through to attain the lower runout.

On 12 July 2024 at 06:11 DC31k Said:

On 12 July 2024 at 04:42 Ketan Swali Said:

ARC does give runout figures for ER collets

For the sake of discussion, education and insight into the world of collets, how is the given figure determined by any collet supplier?

We don’t know!

Would the supplier simply report or pass on a number provided to it by the collet manufacturer?

Always exceptions,  but it’s unusual for wholesalers or retailers to retest manufactured products. They pass the spec on.

Do suppliers have a quality regime where they pick samples from the collets supplied to them by a manufacturer and send the samples for independent verification of runout (like a UKAS lab. for more mundane items such as soil and concrete)?

Maybe.  Quality processes are a major subject in their own right, and there have been massive improvements since the 19th century.   In the early years of WW1, there was a major problem with British shells.   Production was low, costs high, and quality poor.   This was partly due to a quality system in which an inspector would sample one of a batch, and reject the whole batch if it were out of spec.   As applied to shells in 1914-1915, a bad system, very wasteful, and it allowed lots of duds to reach the front-line.   Today, the best manufacturers apply a quality system called ‘First Time Right’.  FTR rejects the idea of inspectors and statistical sampling in favour of building quality into all stages of the production process.   If, later on, a part is found to be out-of-spec, a team of engineers descend and fix whatever was causing the quality problem.   Could be due to a gap in a machine’s maintenance routine, different machine needed suboptimal cutter or material, operator deficiency etc.   FTR works by tackling root causes, not individual malfunctions, and it is the deadly enemy of bodgers!    Whether or not a collet maker applies FTR or not is down to them.   It’s expensive to set up, so a smaller firm might compromise, in the worst case making collets with a clapped out machine driven by a child, who just adds a sticky QA label before packing them.    

It would be interesting to find out what standard tests are used to arrive at this number (is there an ISO collet runout Standard?) and who (if anyone) in the UK has facilities to carry out these tests.

I doubt there’s a run-out standard for collets.   The manufacturer says what it is, and the buyer selects accordingly.  I guess any calibration lab in the UK could measure run-out.  

I guess you need some fairly specialised kit to carry out this test. I wonder what extra or different manufacturing steps the ‘better’ collets go through to attain the lower runout.

Done by working to closer tolerances than doing anything extra.   Of course achieving higher tolerances may require  different production techniques.

The most expensive ER collets are rated to run at high speed, 30000rpm or more.   These have to be balanced as well as low-runout.  Don’t see much value in buying these for my 2800rpm max hobby mill!    And I’m very suspicious of the notion that buying better tools will improve a basic machine.   My WM18 is good for about a thou, and spending extra on posh collets won’t improve it much.   If I want it to work in the tenths range, I’d have to work through the entire machine, applying a long list of tweaks and upgrades.

Dave

Hi DC31k,

I can only speak for the procurement process we follow for ER collets.

Around 18 years ago, the late JS encouraged me to look at ER collets as a standardisation for the hobby market, which before then was used to using a Chinese standard collet which is a longer collet which looks similar to ER.

A factory was found with whom I was happy. In addition to ER collets, we purchased other items and developed some items with them which JS created, with the help of a specific engineer in that factory. Over the years we worked closely with a set of people in the factory. They have a set of procedures in place to check each product on specific product related testing equipment.

About nine years ago, the specific engineer with whom we worked, left and started his own factory which specialises in ER collets. Two years after he started his factory, we moved our procurement of ER collets to him. We continue to buy a spread of items from the original factory, along with ER collets specifically manufactured by him. I have visited him on several occasions and seen the checking process.

Checking is done manually by an experienced checker who checks each collet for runout on a specific test rig. the collets are then separated into small bins by runout range. There is a market for every runout spec. Price we pay for 0.008mm is 10 ~ 15% higher than 0.015mm.

We are regularly offered prices cheaper than what we pay our manufacturer supplier, however I prefer to buy from people I am happy to work with, and I have been working with this team for around 18 years now.

So, above is the process ARC has for quality control. Any problems are usually flagged up after receiving three consecutive complaints with which we agree, for the price range at which a product is offered.

If certification is required, I would suggest Regofix.

Checking for accuracy outside a test rig, and on a milling machine can achieve mixed results as there are too many factors involved, including the accuracy of the spindle, milling head, arbor, nut, cutter, column alignment, how hard or not the collet system is tightened. For these reasons, in the early days, we were less confident of specifying ER collet runouts, even we were originally offering 0.015mm runout. An observation on one mill was found to be different from two to three others.

However, over time we saw others offering Chinese origin ER collets specifying runouts as a marketing tool, in the hobby environment, and we are confident that the prices suggested were difficult to achieve for the runout specified. So we decided to specify the runout as specified by our manufacturer, a few years ago.

Manual machines are less accurate than CNC. They work at lower speeds than CNC too. Although there are specific CNC machine specks ER collets, different from what we sell, there are CNC machining workshops in U.K. and Europe, to whom ARC supplies ER collets for use on their CNC machines. The ER collets we sell are fine for use on Tormach machines.

If one was to use seriously big/multi axis CNC machines for something like aerospace component making, then a different type of ER collets should be considered.

Ketan at ARC