Performance

Wednesday, 07 June 2023

Home
Up
Performance
Firmware

 

Wireless probe performance

I'm frequently asked about the relative performance of the various probes on the market and the impact that the wireless probe has on that performance.  I added this page to try to explain what the limiting factors are give you an idea of what you can expect.  Below you'll find downloads of the test results and the Fusion model I used to generate g-code that produced the reports.  You'll need my post-processor to generate the g-code for yourself. The macros required to run it are already part of PathPilot.

If you want to know more about how the pdf reports were created, see this page.

This test was performed using a 2020 model Tormach passive probe, converted to wireless operation with one of my wireless kits.  The stylus I was using was from Renishaw.  It's 50mm long (carbide) with a 3mm ruby tip.  At that size, it's likely to be one of the bendiest stylii you'd use, so the results presented here are at the poorest end of what you can reasonably expect.

The test

As a test, I chose to use a 1-2-3 block and used the inspection report generation feature of my post-processor to create pdf files with reports of the results.  The 1-2-3 block was mounted on the moving section of my vise, retained by 4 small magnets.  The 1-2-3 block was pressed against the rear of the moving jaw, so the alignment with the mill's axes was only approximate.  Also, since I don't have access to any equipment capable of measuring the block's dimensions accurately, I can't assess the absolute accuracy of the measured dimensions.  In any case, that would only indicate the accuracy of the machine's ball screws rather than anything about the probe.

The test then consisted of the following steps.

  1. The effective diameter of the probe tip was determined using a 3/4" gauge ring and the tip calibration routine in PathPilot.  The effective diameter was measured at 2.839mm.  The difference between that measurement and the nominal 3mm is a mixture of the mill's lost motion and flex of the stylus and probe mechanics.  As noted above, this probe's flex is likely to be comparatively large and a shorter larger diameter stylus would  likely produce results closer to the nominal size.
  2. The mill's coordinate system was aligned to the centre of the 1-2-3 block using Fusion's probing routines to probe both sides in X and both sides in Y and take the centre.
  3. The program then made 3 separate measurements of X size and centre and similarly 3 measurements of Y size and centre, distributed across the length and bredth of the 1-2-3 block.
  4. The measurements in 3 above were repeated 3 times and tabulated as if 3 separate 1-2-3 blocks were being tested.  In fact it's the same 1-2-3 block being measured 3 times. The three sets of results should be identical and any variation indicates the repeatability of the probe system.
  5. The probe was rotated 90 degrees in the spindle and steps 2, 3 and 4 were repeated, creating a second report of 12 measurements repeated 3 times.

Downloads

You can download the following documents

  1. The Fusion model that generated the nc code and report results.
  2. Report with the Tormach probe's cable gland pointing in the -ve X direction.
  3. Report with the probe's cable gland pointing in the -ve Y direction.

Results

Things you should know and things you should be able to see from the reports.

  1. The smallest movement you can ask a 440 to make is 0.0025mm.  That's just a fraction under 0.0001".  So the reports can't show anything smaller than that.
  2. The lost motion in a decently adjusted Tormach mill is in the region of 0.001", so very roughly 10x the smallest move you can ask it to make.
  3. The 3 tables in each of the reports should be identical.  If you compare the tables in either one of the reports, the variation in any single measurement is at worst 0.0003" or around 1/3 of the mill's lost motion!  Given that that variation is the sum of the mill's repeatability, the probe's repeatability and any variation caused by the wireless system's latency it's hard to conclude that any of those components could be usefully improved.  The repeatability results are better than you could reasonably expect.
  4. If you compare the 2 reports, you can see that as you go from one probe orientation to the other, the x measurements shrink by about 0.001" and the y measurements grow by the same amount.  This is caused by the flexibility of the stylus and the fact that the force required to activate the probe varies depending on the direction you approach it.  That's a function of the 3 point suspension used in the probe and the strength of the spring that holds the probe in the central position.

Conclusions

The conclusion is that the biggest error in the whole system is caused by the amount of stylus flex you see varying because of the activation force of the probe being different at different probing directions.  None of that error is caused either by the mill or the wireless system and it's inherent in the 3 point suspension used in probes.

The things that would reduce that error are

  1. Use a shorter probe tip or one with a larger diameter shank so there's less deflection to vary.
  2. Use a probe with a weaker spring, again reducing the stylus deflection.

The downsides would be

  1. Fatter, shorter probe tips are limited in where they can reach, so less convenient to use.
  2. If you weaken the probe's spring too much, then the probe will start to lose reliability as the contacts won't always be re-established when the probe moves away from the contact point.

Home | Performance | Firmware

 

This site was last updated 06/05/21