[keen]

Hi folks.

I have just bought a passive probe and thought I would check it out for accuracy.

You will need at least 20 minutes to watch and digest this.... so grab a coffee!

Cheers, Keen

https://www.youtube.com/watch?v=Qtg6K27h_eI

https://www.youtube.com/watch?v=M6kL62v_ecc

[Hirudin]

The probe I use with a different machine measures the stylus tip diameter by probing against the inside of a ring gauge. Basically, if the ring gauge is exactly 199.998mm in diameter (for instance), and the machine only needs to move 197.004mm between probing in the X+ and X- directions, it will know that the stylus tip is 2.994mm in diameter in X. This particular routine only measures diameters, but you could even accurately measure the four most-used radiuses individually if you wanted to. This method also accounts for stylus flex, varied triggering pressures, and controller/motor reaction time as long as you preform the calibration at the same speed(s) and probe orientation that you will use while probing. Seems like a smart way to go and wouldn't be too hard to write up.

Of course, if you then rotated the probe the measurements may be off so you'd want to load the probe into the spindle as consistently as possible every time you used it.

[wtopace]

Hey Keen - I posted this on your Youtube video directly, but I figured it would be better to keep the discussion in this thread =)

It could be that the backlash in your machine is causing the 0.001" error between the Y-/Y+ that you are seeing.

Can you measure your Y-Axis backlash with a DTI and subtract that from the stacked error of 0.0011" and see what is left from the flex in the stem?

It would also be interesting to see your X-Axis backlash measured with a DTI as well since PathPilot shows it as exact from X-/X+.

[keen]

Hey Keen - I posted this on your Youtube video directly, but I figured it would be better to keep the discussion in this thread =)

It could be that the backlash in your machine is causing the 0.001" error between the Y-/Y+ that you are seeing.

Can you measure your Y-Axis backlash with a DTI and subtract that from the stacked error of 0.0011" and see what is left from the flex in the stem?

It would also be interesting to see your X-Axis backlash measured with a DTI as well since PathPilot shows it as exact from X-/X+.

Hi - Thanks for watching and digesting this. I had thought about your valid point of backlash. This machine has equal backlash after tweaking of nom 0.01 mm on the X and Y. Also the first video tests show the same error behaviour with a different test (single direction and rotating the probe) - the video is a bit long winded but if you persevere you will see the issue.

Cheers. Keen

[keen]

I am coming to the conclusion that the probe stylus is too long and thin (2mm stem) for such a stiff mechanical actuation.

I suspect the manufacturer copied a more expensive light actuation design but did not think about the flex problem fully.

I am planning to make another stylus that is shorter and out of 3mm dia carb rod with a 4mm ball . I think this will fix 75% of the problem.

keen

[keen]

The probe I use with a different machine measures the stylus tip diameter by probing against the inside of a ring gauge. Basically, if the ring gauge is exactly 199.998mm in diameter (for instance), and the machine only needs to move 197.004mm between probing in the X+ and X- directions, it will know that the stylus tip is 2.994mm in diameter in X. This particular routine only measures diameters, but you could even accurately measure the four most-used radiuses individually if you wanted to. This method also accounts for stylus flex, varied triggering pressures, and controller/motor reaction time as long as you preform the calibration at the same speed(s) and probe orientation that you will use while probing. Seems like a smart way to go and wouldn't be too hard to write up.

Of course, if you then rotated the probe the measurements may be off so you'd want to load the probe into the spindle as consistently as possible every time you used it.

Hi Herudin. The trouble is the flex or effective diameter is different between the X and Y directions and needs two different 'tool' diameter settings to account for this.

See my last comment - I think the probable stylus design fault needs to be addressed first to minimize the issue at its source.

Keen

[Hirudin]

Hi Herudin. The trouble is the flex or effective diameter is different between the X and Y directions and needs two different 'tool' diameter settings to account for this.

See my last comment - I think the probable stylus design fault needs to be addressed first to minimize the issue at its source.

Keen

Yes, Tormach should employ the method that I described to measure the stylus tip. Two diameters would probably be sufficient (that's what my machine uses), but if you really wanted to be thorough "you could even accurately measure the four most-used radiuses individually". The single field to input a diameter is insufficient. It's not even worthwhile to display the tip diameter(s) anyway, as long as the machine measures it accurately during the calibration there's no reason the user would need easy access to that value. I forget that the Tormach version of Mach3 doesn't allow you to modify the screenset, so correcting their errors isn't possible even though it "wouldn't be too hard to write up".

Regarding the stylus flex, I'm 90% sure the Tormach probe has an M4 thread so you can simply pick up a stiffer stylus. M4 is a common stylus thread so there shouldn't be any shortage of choices.

[MichaelHenry]

Regarding the stylus flex, I'm 90% sure the Tormach probe has an M4 thread so you can simply pick up a stiffer stylus. M4 is a common stylus thread so there shouldn't be any shortage of choices.

Here are some other sources to try:

Styli That Work With All Probe Systems - Q-Mark Probe Styli
https://www.carbideprobes.com/
http://www.damencnc.com/en/tools/probing---toollength

[MichaelHenry]

Keen,

No offense, but did you axially calibrate the probe as described in the Tormach mill manual?

[keen]

Hi - In this video I show the issue more scientifically, suggest a reason for it, and propose a solution. Keen

https://www.youtube.com/watch?v=i-CnqYBIiaY

[keen]

Keen,

No offense, but did you axially calibrate the probe as described in the Tormach mill manual?

Hi Michael - Yes, don't worry I did all the study and prep before embarking on this.

See my new video above, which should explain the issue more clearly. Keen

[keen]

Yes, Tormach should employ the method that I described to measure the stylus tip. Two diameters would probably be sufficient (that's what my machine uses), but if you really wanted to be thorough "you could even accurately measure the four most-used radiuses individually". The single field to input a diameter is insufficient. It's not even worthwhile to display the tip diameter(s) anyway, as long as the machine measures it accurately during the calibration there's no reason the user would need easy access to that value. I forget that the Tormach version of Mach3 doesn't allow you to modify the screenset, so correcting their errors isn't possible even though it "wouldn't be too hard to write up".

Regarding the stylus flex, I'm 90% sure the Tormach probe has an M4 thread so you can simply pick up a stiffer stylus. M4 is a common stylus thread so there shouldn't be any shortage of choices.

Hi Hurudin. Sorry I have been slow to 100% follow you. You mean when calibrating, your machine allows for a different stylus diameter value to be set for the X and Y?

(This of course would take care of the tri swing arm issue if the Probe is installed in the same rotary position and the probe only used to probe as per the calibration directions).

Thanks for your last idea - see my latest video.

Keen



Keen

[keen]

Here are some other sources to try:

Styli That Work With All Probe Systems - Q-Mark Probe Styli
https://www.carbideprobes.com/
http://www.damencnc.com/en/tools/probing---toollength

Thanks so much for that Michael! The last link explains the issue I have found out the hard way! Talk about re inventing the wheel!

I quote: "However (the three point design) does have some problems, which become apparent when using it for long measuring jobs, or very precise jobs. Due to the way the 3 point mechanism is constructed, the stiffness is not the same under different approach angles. This can have influence on the results."

Classic! What a journey!

Cheers. Keen

[keen]

So - for those of you that missed the Michael Henry link:

damencnc.com - Probing & Toollength

- - - Updated - - -

Or more specifically:

http://www.damencnc.com/en/tools/probing---toollength/887

[bevinp]

Keen,
As you say, the design and consequent inaccuracies are due to price, a target probe price suitable for the level of machining that the Tormach is designed for. This is not a criticism of Tormach approach but a recognition that Tormach users would not spend 500 Euros for a probe, which is the price of the DamenCNC high end probe that does not suffer the 3 pivot problem.

I am not criticising Tormach. I applaud Tormach (Greg Jackson, a sad loss for all of us) for the effort and achievement to produce a low price product and then support it with exceptional service that is, in the words of the military, "above and beyond".

Anyway, you too should be congratulated for your "journey" that has alerted Tormach users of the problems and shown how to minimise them if they require. Well done.

Bevin,
PS. And you have three videos of that journey, although I don't think it would make a good movie such as "Walk in the Woods" or "Everest".

[Hirudin]

...
You mean when calibrating, your machine allows for a different stylus diameter value to be set for the X and Y?
...

Exactly. It's a good thing too because the ruby ball that came on my first stylus wasn't even spherical. Thankfully, the major and minor axes of the oval-shaped "ball" roughly lined up with the X and Y axes of the machine. Unfortunately, I don't remember exactly how much variance there was.

If the machine hadn't measured the ball itself I would have assumed the actual diameter was the "6mm" listed in the specifications. Instead, if I remember correctly, at least one of the measured diameters was something like 5.89mm.

[Hirudin]

Considering the "Y" shape of the three prongs inside the probe, I would guess there might actually be three different trigger pressures. With the "Y" shape oriented the same way as the letter, I would assume X+ and X- would be about the same, but I'd guess Y+ would be different than Y-. Maybe employing three (or four) effective radiuses really would be the best way to go. At some point I would like to rewrite all the routines that came with my machine to use radiuses instead of diameters. It is probably worth considering for anyone trying to improve probing accuracy with any machine.

[popspipes]

I asked an old friend how he located centers on a mill, he said he always used a dial test indicator and a mirror, in his opinion this was the most accurate.

He is 88 years old and retired from Continental Can and has been a machinist since his US Navy days so I tend to listen to what he has to say............

This is definitely not as easy as the probes, but as old and klutsy as I am, its much cheaper..................................

I like your shield on the rear of the table, mine is made of rubber gasket material with a pvc top rail it is serviceable but doesnt look anywhere near as nice as yours!

[TurboStep]

I have just bought a passive probe and thought I would check it out for accuracy.

Hi Keen, thanks for publishing your measurements. These are exactly the values I'm interested in!
Although the stylus is long and thin, according to Tormach they appear to be made of ceramic. I'm not sure that ceramic would bend that easily (I could be very wrong). The central hub is apparently made of plastic and I was expecting this to be the weakest link - in particular where the stylus screws into the bottom because the thread is relatively short. I would be tempted to insert a large m5 washer or spacer between the tip and the plastic hub to spread the pressure better over the lower surface and improve the stability.
I've also been considering (if I ever get around to buying one) the possibility of cutting the plastic hub down to just about 1mm above and below the 3 pins. Then add a metal plate above, and a metal boss below to stiffen the construction. At the same time I had played with the idea of using a Haimer tip instead. Not only would they be slightly cheaper but I would only need to keep a reserve of the one type.
Step

[keen]

Keen,
As you say, the design and consequent inaccuracies are due to price, a target probe price suitable for the level of machining that the Tormach is designed for. This is not a criticism of Tormach approach but a recognition that Tormach users would not spend 500 Euros for a probe, which is the price of the DamenCNC high end probe that does not suffer the 3 pivot problem.

I am not criticising Tormach. I applaud Tormach (Greg Jackson, a sad loss for all of us) for the effort and achievement to produce a low price product and then support it with exceptional service that is, in the words of the military, "above and beyond".



Anyway, you too should be congratulated for your "journey" that has alerted Tormach users of the problems and shown how to minimise them if they require. Well done.

Bevin,
PS. And you have three videos of that journey, although I don't think it would make a good movie such as "Walk in the Woods" or "Everest".

Hi Bevin - Yes 'a walk in the woods'...just saw the movie too.

I am sure a shorter 3mm stiffer stylus will make the Tormach Passive Probe accurate enough for most jobs - perhaps reducing the X to Y difference from 0.03 mm down to around 0.01 mm.

Yes I applaud Tormach and Greg Jackson also - they have brought commercial CNC to the masses.

Keen