Secondary Dive height for probing

Hernicz

@sinned6915 I do G32 with 0.001 precision with really good repeatability. Probe also set up to 0.001.

Of course if you probe with different speed or dive height the measurement will be off, but RepRap only renders the probing successful if you have the same values on two consecutive probing moves.

The purpose of this request is to lower the dive height on secondary/finishing probing moves thus speeding up the process.

sinned6915

@Hernicz said in Secondary Dive height for probing:

@sinned6915 I do G32 with 0.001 precision with really good repeatability. Probe also set up to 0.001.

Of course if you probe with different speed or dive height the measurement will be off, but RepRap only renders the probing successful if you have the same values on two consecutive probing moves.

The purpose of this request is to lower the dive height on secondary/finishing probing moves thus speeding up the process

I disagree. That thinking is false economy.

By not reversing the Z motion system past the amount of inherent Z backlash, you are not measuring anything. All you are doing is 'measuring' the turn of a stepper motor withing the backlash amount.

Consider that your claimed 'precision' is 1/3 that of the thickness of the ink drawn with a Sharpie marker or 1/5 the thickness of a piece of lint. Measure those and then re-state your claim.

Hernicz

@sinned6915 The BL-Touch has more than 1mm unnecessary travel, many of us has anti backlash nuts.

I have a Retraction Z Hop of 0.05 mm without problems. I should see artifacts if I would have backlash, so I assume a 0.2-0.3 mm secondary dive height shouldn't cause any problems.

Also I can test it with a variable height probe repeatability test script.

I still think we would definitely benefit from this feature.

dc42

@Hernicz I considered this and decided it was a good idea, so I've implemented it ready for release 3.5.0-rc.1.

I tested it by probing 135 points on a E3D tool changer (probing speed 600mm/min) with the M558 A2 S0 parameters used to force two probes per point. With both dive heights set to 3mm the time taken was 3min 20sec. With the dive heights set to 3mm then 0.5mm this was reduced to 2min 27sec.

sinned6915

@dc42 said in Secondary Dive height for probing:

@Hernicz I considered this and decided it was a good idea, so I've implemented it ready for release 3.5.0-rc.1.

I tested it by probing 135 points on a E3D tool changer (probing speed 600mm/min) with the M558 A2 S0 parameters used to force two probes per point. With both dive heights set to 3mm the time taken was 3min 20sec. With the dive heights set to 3mm then 0.5mm this was reduced to 2min 27sec.

Can you post up a summary of the trigger values please?
Change it to min dive height of at least a 1 full mechanical step of the Z motion system (ie, > 2mm for a 2mm lead screw) and compare those numbers.
What probe are you using? If you want speed, Euclid with D2F-5 switch, we've gone as fast as 500mm/sec travel speed with repeatable and predicable results. OEM says not faster than 30 operations per minute for electical component life.

Hernicz

@dc42 Thank you very much. I'll do some repeatability and other tests.

Hernicz

@sinned6915

I'll do some tests.
2 mm is far more than 1 full step. Also full step depends on motor angle as well. (1 full step with a T8 is 0.04 mm with a 1.8° Stepper, half of that with a 0.9°, a full rotation of the shaft is 8mm travel, so 2mm is 50 or 100 full steps depending on motor angle) We need to keep in mind that the probe trigger point will be most likely in a microstep, so in my opinion full step probing isn't necessary as we dont even know if the motor is in a full step or not if microstepping is enabled we can only pass full steps but we dont know if we actually stopped at one.
Microstepping is pretty good on trinamic steppers so I dont really see the point.
Euclid is cool, but we use different probes and different machines. This feature is for touch probes.

sinned6915

@Hernicz you are still not understanding that if you dont reverse the Z axis far enough to clear the backlash/friction of the mechanical system this is not yeilding any real results. All you are measuring the relative slop in the system.

please reread my comment- i said one full MECHANICAL step of the system.

If you are trying to claim any accuracy by 'preloading' the touch probe so it does not have to travel as much, then what happens when you run over a crumb or scar in the bed?

dc42

@sinned6915 said in Secondary Dive height for probing:

you are still not understanding that if you dont reverse the Z axis far enough to clear the backlash/friction of the mechanical system this is not yeilding any real results. All you are measuring the relative slop in the system.

If you don't reverse the Z axis far enough to clear the mechanical backlash of the axis, then if the probe itself has any backlash at all (e.g. a switch or a BLTouch) then it will remain triggered, and probing will fail.

On my tool changer with the standard switch-type probe (which therefore has some backlash), I can set the second dive height as low as 0.05mm (but not as low as 0.03mm) and get successful probing. So the leadscrew backlash plus switch backlash must be less than 0.05mm.

Hernicz

@sinned6915 This is why I requested this feature so the first dive height is high enough that the probe doesn't touch the bed in X-Y movements, but after the first probing move it dives from lower height thus speeding up the sequence.

Whithout this feature the only option to speed probing is to slide the pin of the probe on the heatbed. It works, but definitely not a solution in the long run.

Low dive height fast probing @ 1mm/s, sliding the pin

Also as @dc42 said, if the Z axis doesn't move due to backlash the probe the sequence will throw a "Probe triggered before probing move" error.

Hernicz

@dc42 I've made some tests according to what @sinned6915 said and I have to say he's saying something. There's a minimum movement distance to actually get the commanded distance.

I've attached a dial indicator to measure Z axis movement then moved the Z axis up-down-up-down.

If I move the Z:

0.01 : I get no movement
0.02 : dial moves less than 0.01 both directions
But here comes the interesting part, if I move Z:
0.05 : I get 0.03, like 0.2 is deducted from it.
0.07 : I get 0.05
0.09 : I get 0.08 only 0.1 gets deducted

And it seems that from 0.125 I actually get the right movement distance that had been commanded.

Considering this there are two options

I probe with 0.05 dive height hoping that the backlash is equal among all the probing moves (I assume it is because I don't se issues on my prints with 0.05 Z-Hop, but testing it now with 0.15) or
I set the dive height to 0.125-0.15 which is still pretty low and gives true movement distance.

I would chose the latter.

Hernicz

@dc42 I tested Secondary Dive Height as well and it's wonderful, but it disables Secondary Dive Speed.