[keen]

Hi Folks

Recent threads on gib adjustment etc make me think a thread on Tormach accuracy and how to measure it may be useful.

I start it off with a couple of basics most people agree on. Check the backlash with a dial indicator measuring traverse with fine steps - eg X + traverse. Now reverse the traverse direction and note the DRO movement before the indicator shows actual movement.

I gather from my trials and posts over the years that a lost motion of 0.0005" is very good, 0.001" is common, and 0.002" shows there is a problem.

But a even better way to check real world errors is to test machine a OD diameter test piece and measure it with a micrometer. (A 'vernier' caliper is too alignment and pressure dependant) But note, even this will not always measure symmetrical lobing errors. Better still to rotate it and indicate it eg on a V block or in the lathe to measure total out of round. This measures a real world combination of machine errors including backlash, flex, racking, oil film under vibration etc..

Here is this test done on my 2007 Tormach 1100 in 2008. showing a total out of round error of 0.025mm or 0.001"

Cheers Keen

[keen]

By comparison a high quality VMC would have backlash of around 0.0001" and machine a test piece like this to within 0.0002" of round.

But the machine will cost you four or more times as much !

Keen

[Superman]

I start it off with a couple of basics most people agree on. Check the backlash with a dial indicator measuring traverse with fine steps - eg X + traverse. Now reverse the traverse direction and note the DRO movement before the indicator shows actual movement.

I was in a different view to Backlash measurement, but similar operation

using a 0.002mm dial indicator, winding in 1 direction only, set DRO and indicator to zero,
- now wind axis past the dial zero point, then move axis to bring dial to zero (opposite direction),....... backlash is the DRO reading
( generally, the control would move the leadscrew by the backlash setting ( theoretically, it should involve no movement of the axis )
( - if the dial indicator needle immediately moves if you reverse winding direction, you have too much backlash comp in the control )

Not often you see a kiwi work in inches :stickpoke

[TurboStep]

Hello Keen

...Recent threads on gib adjustment etc make me think a thread on Tormach accuracy and how to measure it may be useful....

In the thread I assume you're referring to I asked the question:

...How EXACTLY are you/they defining accuracy?...

I have to ask the same question here! Your results also show errors due to causes other than backlash/lost motion. An expensive VMC may give you improved roundness but the resulting diameter won't necessarily be "accurate".
Step

[Steve Seebold]

My 1100 is 5 years old, and I can still do +/-.001 work on it all day long.

[keen]

Hello Keen


In the thread I assume you're referring to I asked the question:


I have to ask the same question here! Your results also show errors due to causes other than backlash/lost motion. An expensive VMC may give you improved roundness but the resulting diameter won't necessarily be "accurate".
Step

Yes good point Step - my post was quite general.

We ideally should agree on a desired end result before we can measure how close a machine gets to it.

OK how about:.... How repeatedly ROUND can an Outside Diameter be machined via circular interpolation with an end mill on a given machine. This is a practical real world test that gives useful results. EG... If the OD is 0.001" out of round it is not ideally suitable for a precision bearing ID press fit.

If it is 'repeatable' then the actual diameter is less critical as it can be accurately established via iteration.

Keen

[MichaelHenry]

Maybe a poll would be a good idea here? Get interested parties to run the test and report their results. Tormach had an early video back around 2008 that showed error in (I think) interpolating a round hole with an end mill on their 1100 model.

[kstrauss]

Maybe a poll would be a good idea here? Get interested parties to run the test and report their results. Tormach had an early video back around 2008 that showed error in (I think) interpolating a round hole with an end mill on their 1100 model.

Perhaps post the standard code and have the poll for 1100/770/440 and any others that are willing to try it? Roundness should be measured by sweeping the periphery of the hole with a DTI in the spindle.

[keen]

That would be very interesting. I hope it catches on - It would give us a really good benchmark comparison.

Yes I think sweeping with a DTI in the spindle wold be fine.

My old code was for Mach3 and so probably better to get one of you to post the code to be sure it suits PP.

Cheers Keen

[popspipes]

That would be very interesting. I hope it catches on - It would give us a really good benchmark comparison.

Yes I think sweeping with a DTI in the spindle wold be fine.

My old code was for Mach3 and so probably better to get one of you to post the code to be sure it suits PP.

Cheers Keen

I will have to try this, I made a bearing plate awhile back and I had to bore it oversize a thou or two as it had tight spots in the bore, I wanted a light press fit, it wasnt a critical part so I bored it a bit over and used loctite to hold it in.

[Steve Seebold]

I can get small OD's and ID's so close to being round that you can't measure any egg shape at all. I just finished a 150 piece job that had an OD of 1.3750 +/-.00025.

It was actually quite easy. I rough machined it leaving .010 for finish stock, then I put a boring bar in backwards and ran the spindle backwards and everything worked out perfectly.

[keen]

I can get small OD's and ID's so close to being round that you can't measure any egg shape at all. I just finished a 150 piece job that had an OD of 1.3750 +/-.00025.

It was actually quite easy. I rough machined it leaving .010 for finish stock, then I put a boring bar in backwards and ran the spindle backwards and everything worked out perfectly.

Hi Steve - yes single point sweep boring overcomes the issues we are talking about. I agree that type of approach needs to be applied to use a Tormach for high precision work.

But the reason for the test is to expose the machines limitations so that we can make comparisons. Some types of work are just too precision for a Tormach ...and it is useful to know where the boundaries are.

Keen

[CadRhino]

Keen, your videos are terrific. Could you do one on checking a tormach's accuracy. I don't know that I have been doing it correctly or using my instruments in the right way.
Much appreciated.
Nathan

Hi Steve - yes single point sweep boring overcomes the issues we are talking about. I agree that type of approach needs to be applied to use a Tormach for high precision work.

But the reason for the test is to expose the machines limitations so that we can make comparisons. Some types of work are just too precision for a Tormach ...and it is useful to know where the boundaries are.

Keen

[mountaindew]

Keen +++1 on the last request.
I have trouble with this also and not much confidence in my measuring tools at times also. Will fix that soon with much better quality indicator.
Thanks in advance

[keen]

Thanks guys. - Currently I need to put any spare time into the new crash resistant probe development project. (See thread).

But that would be an interesting video subject - it would need to be carefully thought through ...or it would be challenged on every detail!

Keen

[MichaelHenry]

+1 to the Tormach accuracy check video.

I'm going to be replacing my master collet and drawbar soon and will run a version of your code before and after.

[mountaindew]

Thanks guys. - Currently I need to put any spare time into the new crash resistant probe development project. (See thread).

But that would be an interesting video subject - it would need to be carefully thought through ...or it would be challenged on every detail!

Keen

Understood!
Always looking for ways to get better results and improve performance. Using info here and from tormach has helped so much and would like to see more!
MD