Probing question again...

[A_Camera]

In fact, it must not be a fixed plate at all

There are thousands of people using fixed plates for probing, with no issues at all.

Solid, fixed plates can only be probed slowly. Probing at high speed will smash into the surface, damaging the surface or the tool. This is what happens when you probe at 4000mm/min:



Even if you would increase acceleration a lot more, you would still damage something because the overrun needed for deceleration would still be noticeable. In fact, I found that even 400mm/min gives inaccurate results with rigid probing if done in one single pass, which means that the overrun is too much even at 400mm/min.

[Battwell]

I'm testing probing repeatedly with the same ball nose tool in exact same position on the very accurate tool probe I have.
Not removing tool or rotating spindle.

I'm still finding I need to double tap to get repeatable results within 1/100 mm (0.02 is my average deviation. ) I'm using 32bit encoder servos which are fast enough to go from 4200 rpm clockwise to 4200 rpm counterclockwise in 1 second off load .

Obviously with my realistic tuning lifting my z axis acceleration is only set at 1000.

I only get an over run of approx 1mm after my first fast probe at 700mm/ minute
Then I use a 2mm retract and slow probe at 50mm/ minute

At least probing isn't broken like versions 22 to 26 which cost me a lot of money- having the probe I've had developed redesigned and manufactured to include an over travel limit switch circuit too . Those versions drove me nutz missing probes!

[A_Camera]

I only get an over run of approx 1mm after my first fast probe at 700mm/ minute
Then I use a 2mm retract and slow probe at 50mm/ minute

1mm overtravel at 700mm/min is a lot. What's the acceleration of your machine?

[Battwell]

i should have said maximum 0.2mm (and this was showing up at high speed
using the old method taking reading after axis had stopped 0.01 average - usually 0.005mm (recorded)
ive left my tool probing as is at the moment- as it works fine for me.
i will update to the newer method using var # if balacz changes it to be like homing where the var is captured on the probe backoff.
as my machine is atc i only need to run the probe routine for the first run of the program to ensure all tools used are entered and measured correctly in the tool table.

[Battwell]

not fully as im developing it as a commercial product.
but inside its the same theory as renishaw- 6 ball contact around a ring. - requires just a tiny movement to break the circuit

[Battwell]

The current is controlled by the opto isolator.
Mine is on a 24v circuit
The higher the voltage Lesser current and lesser chance of noise interference

[A_Camera]

Never heard of "direct with full current signals or though an interface box measuring resistance"... What kind of apparatus is that?

[A_Camera]

The current is controlled by the opto isolator.
Mine is on a 24v circuit
The higher the voltage Lesser current and lesser chance of noise interference

Unless I misunderstood you, that's incorrect. If it is an opto isolator than the higher the voltage the higher the current through the LED. Normally a resistor is limiting the current to the LED and if you dimension the resistor for 12V and 10mA than you will get 20mA if you use the same resistor and increase to 24V. Optimally, LEDs are driven by constant current, so at 24V the resistor should be doubled. Ohms law.

[Battwell]

Of course
But mine have correct resistance to produce correct current set for 24v
Maybe I should have said limited by the opto isolator circuit

Better noise immunity than a 5v signal
So the current flowing through my probe is only what is required to light the opto emitter

[Derek]

Not sure how far into the design stage you are but I have 2 probes. Both in the $250.00-$300.00 range. One is the Tormach and one is from IMSRV. The one from IMSRV has 4 set screws on the side of the probe to set the tip runout. I have no idea if this is more accurate or less accurate than the Tormach style but it is a whole lot more convenient. The Tormach uses a gland nut to hold the cable which is light years ahead of the rubber grommet that IMSRV uses to NOT hold the cable. I've had 2 of these probes and the wire pulling out of the grommet has been an ongoing problem. The IMSRV uses a V block arrangement for the contact and the Tormach uses ball and pin. When the Tormach completes the probe the Probe LED in UCCNC is lit. When the IMSRV complete the LED is unlit so there must be some spring back I guess..


Just throwing that out there since I personally think there is real room for improvement in the sub $500.00 range of probes. It's like you jump from $250.00 to $1500.00 with not much in between.
Derek