Fast Probing with QTPlasmac
11 Aug 2021 08:04 - 11 Aug 2021 08:06 #217482
by rodw
Fast Probing with QTPlasmac was created by rodw
A couple of guys were impressed with a short video of my Spaceship Plasma on a water table because how fast it was probing.
Another user shared a video where the torch did a little happy dance above the material after the retract.
This is using hypersensing on 2mm mild steel with a 5mm probe height
Cutting this part on a 1.2m x 1.2m sheet would have over 500 pierces and some of my other jobs on a 1.2m x 1.2m (4'x4') have over 1500 pierces.
So speeding up probing has a major impact on cycle times
So how does your probing compare?
Please feel free to share a video
Lets talk about how to improve probing speed.
Things I have found are
1. Keep any float switch hysteresis (between on and off) to be the shortest distance possible
2. Keep your probe height low
3. Set acceleration and velocities on your Z axis to be as fast as possible
4. Use float switch travel to absorb decelleration on contact.
5. Use fast relays if not using hypersensing. A 10msec on time can lead to a long time to back off the material.
6. Solid state relays with < 250 usec on time will speed things up and allow linuxcnc to respond in one servo cycle.
What are your ideas?
Another user shared a video where the torch did a little happy dance above the material after the retract.
This is using hypersensing on 2mm mild steel with a 5mm probe height
Cutting this part on a 1.2m x 1.2m sheet would have over 500 pierces and some of my other jobs on a 1.2m x 1.2m (4'x4') have over 1500 pierces.
So speeding up probing has a major impact on cycle times
So how does your probing compare?
Please feel free to share a video
Lets talk about how to improve probing speed.
Things I have found are
1. Keep any float switch hysteresis (between on and off) to be the shortest distance possible
2. Keep your probe height low
3. Set acceleration and velocities on your Z axis to be as fast as possible
4. Use float switch travel to absorb decelleration on contact.
5. Use fast relays if not using hypersensing. A 10msec on time can lead to a long time to back off the material.
6. Solid state relays with < 250 usec on time will speed things up and allow linuxcnc to respond in one servo cycle.
What are your ideas?
Last edit: 11 Aug 2021 08:06 by rodw.
Please Log in or Create an account to join the conversation.
11 Aug 2021 09:29 #217494
by Tesremos
Replied by Tesremos on topic Fast Probing with QTPlasmac
Ohe, Rodw is talking about me.
20mm probe height, 25mm safe height,
probe moves to probe height,
moves down for contact
makes contact
moves off slowly (this is what started the conversation) until contact is lost,
then advances towards material until contact is made again,
and rapid to pierce height.
i took a quick picture of a scope with probe vs z axis commanded position. shows the little dance.
20mm probe height, 25mm safe height,
probe moves to probe height,
moves down for contact
makes contact
moves off slowly (this is what started the conversation) until contact is lost,
then advances towards material until contact is made again,
and rapid to pierce height.
i took a quick picture of a scope with probe vs z axis commanded position. shows the little dance.
Please Log in or Create an account to join the conversation.
11 Aug 2021 10:29 #217503
by rodw
I Still don't understand why it bounces above the plate and mine does not. You said you are using Ohmic sensing with relay. My first thought was it was too much hysteresis in the float switch but you are not using one. I think the relay you shared had a 10 ms on time and 6 ms off or something like that. Plasmac by default has a 5 ms debounce time so that could add a bit more time.
But I'm wondering if you are contacting the material before you reach probe height? I think Plasmac backs off and reprobes in that instance
Any other ideas?
Replied by rodw on topic Fast Probing with QTPlasmac
Oops. But to be fair it was another guy who said my probing was fast!.Ohe, Rodw is talking about me.
I Still don't understand why it bounces above the plate and mine does not. You said you are using Ohmic sensing with relay. My first thought was it was too much hysteresis in the float switch but you are not using one. I think the relay you shared had a 10 ms on time and 6 ms off or something like that. Plasmac by default has a 5 ms debounce time so that could add a bit more time.
But I'm wondering if you are contacting the material before you reach probe height? I think Plasmac backs off and reprobes in that instance
Any other ideas?
Please Log in or Create an account to join the conversation.
11 Aug 2021 10:31 #217504
by Tesremos
Replied by Tesremos on topic Fast Probing with QTPlasmac
if that were the case, i see the error,
probe tripped before reaching probe high adjusting safe travel height.
probe tripped before reaching probe high adjusting safe travel height.
The following user(s) said Thank You: rodw
Please Log in or Create an account to join the conversation.
11 Aug 2021 12:04 - 14 Aug 2021 08:14 #217526
by phillc54
Replied by phillc54 on topic Fast Probing with QTPlasmac
It doesn't give an error for that, it will retract and then try to probe again.
Probe Height is the height above the minimum limit.
Safe Height is the height above the probed height of the material. Plus or minus any warping of the material during cutting if THC is active.
They have no relationship to each other.
Probe Height is the height above the minimum limit.
Safe Height is the height above the probed height of the material. Plus or minus any warping of the material during cutting if THC is active.
They have no relationship to each other.
Last edit: 14 Aug 2021 08:14 by phillc54.
The following user(s) said Thank You: CNCFred
Please Log in or Create an account to join the conversation.
11 Aug 2021 23:02 - 11 Aug 2021 23:05 #217576
by beefy
Replied by beefy on topic Fast Probing with QTPlasmac
Rod,
I use over-running mechanisms on my X and Y home switches. Nothing technical, just a bolt with a smooth shank, and spring over it to keep it pushed hard against a stop. When it hits the delicate home switch (basic plastic microswitch), so long as the deceleration is fast enough the over-running mechanism does not even operate. I basically put it there as a safeguard in case the axis did not stop for some reason. However if the machine was to run into a limit for some reason the axis would have gone past the home switch and the over-running mechanism comes into play.
With an over-running mechanism you can basically go towards a home switch at rapid speed but back off at a much slower "back-off" speed for accuracy.
So for the Z my thoughts are you don't want to be running an expensive machine torch into the steel sheet at a fast speed. I've therefore been toying with the concept of a separate mechanism (like I say, no design, only concept for now) where somehow a touch off arm activates and puts a probe arm directly under the torch. If your Z axis can move at 20 m/min then you can probe at 20 m/min as the proposed mechanism would have an over-run feature, then once the switch is made it backs off at a much slower speed to remake the switch. Like you said the faster the acceleration and deceleration and the lower the switch hysteresis the less total over-run / distance for the slower backoff.
To take advantage of such a system I'd like to install maybe a servo motor and planetary gearbox on the Z to get fast rapid speed. I like the idea of high Safe-Z to jump over tip-ups while having plenty speed to make up the distance fast.
So such a system requires an activation output to bring the touch off mechanism into position, and perhaps a user adjustable delay before the Z probe motion commences. It also requires a user definable Z offset (probe mechanism to torch nozzle distance).
So although we have the addition of the over-running probe mechanism, as compensation we can lose the separate floating Z axis. With the torch being fixed solid to the Z axis and no float switch that acts as a limit during cutting, I'd certainly want a breakaway.
Maybe you think the idea is a load of rubbish, but you did ask for ideas LOL.
Keith
I use over-running mechanisms on my X and Y home switches. Nothing technical, just a bolt with a smooth shank, and spring over it to keep it pushed hard against a stop. When it hits the delicate home switch (basic plastic microswitch), so long as the deceleration is fast enough the over-running mechanism does not even operate. I basically put it there as a safeguard in case the axis did not stop for some reason. However if the machine was to run into a limit for some reason the axis would have gone past the home switch and the over-running mechanism comes into play.
With an over-running mechanism you can basically go towards a home switch at rapid speed but back off at a much slower "back-off" speed for accuracy.
So for the Z my thoughts are you don't want to be running an expensive machine torch into the steel sheet at a fast speed. I've therefore been toying with the concept of a separate mechanism (like I say, no design, only concept for now) where somehow a touch off arm activates and puts a probe arm directly under the torch. If your Z axis can move at 20 m/min then you can probe at 20 m/min as the proposed mechanism would have an over-run feature, then once the switch is made it backs off at a much slower speed to remake the switch. Like you said the faster the acceleration and deceleration and the lower the switch hysteresis the less total over-run / distance for the slower backoff.
To take advantage of such a system I'd like to install maybe a servo motor and planetary gearbox on the Z to get fast rapid speed. I like the idea of high Safe-Z to jump over tip-ups while having plenty speed to make up the distance fast.
So such a system requires an activation output to bring the touch off mechanism into position, and perhaps a user adjustable delay before the Z probe motion commences. It also requires a user definable Z offset (probe mechanism to torch nozzle distance).
So although we have the addition of the over-running probe mechanism, as compensation we can lose the separate floating Z axis. With the torch being fixed solid to the Z axis and no float switch that acts as a limit during cutting, I'd certainly want a breakaway.
Maybe you think the idea is a load of rubbish, but you did ask for ideas LOL.
Keith
Last edit: 11 Aug 2021 23:05 by beefy.
Please Log in or Create an account to join the conversation.
11 Aug 2021 23:26 #217578
by rodw
Replied by rodw on topic Fast Probing with QTPlasmac
I wonder if an air ram could fire the probe down and there was a dog in a "thread" that rotated a probe element so it was under the torch at the bottom of travel? What I thought of was a fixed pin in a slot running full length of a rod that transitioned to a 90 deg thread to rotate the probe under the torch . That would be quick!
Please Log in or Create an account to join the conversation.
12 Aug 2021 03:22 - 12 Aug 2021 03:31 #217588
by beefy
Replied by beefy on topic Fast Probing with QTPlasmac
You're on the same page as me Rod.
A cam style slot that dictates the travel path. The poultry processing machines I maintain have zillions of cams slots with rollers running in them, to create all sorts of motion at once.
Extremely simple concept but what a wow to watch.
For our purpose the slot can just run vertically down then bend down at an angle (not fully horizontal or the cam roller could lock with a vertical pull back) to bring the probe tip under the torch..
An issue I can forsee is the end of the cam slot would wear with repeatedly being bashed into, so a good idea would be a stop plate so the cam roller/slider never actually hits the end of the cam slot.
Finally managed to find a video of the machinery I'm talking about.
Keith
A cam style slot that dictates the travel path. The poultry processing machines I maintain have zillions of cams slots with rollers running in them, to create all sorts of motion at once.
Extremely simple concept but what a wow to watch.
For our purpose the slot can just run vertically down then bend down at an angle (not fully horizontal or the cam roller could lock with a vertical pull back) to bring the probe tip under the torch..
An issue I can forsee is the end of the cam slot would wear with repeatedly being bashed into, so a good idea would be a stop plate so the cam roller/slider never actually hits the end of the cam slot.
Finally managed to find a video of the machinery I'm talking about.
Keith
Last edit: 12 Aug 2021 03:31 by beefy.
The following user(s) said Thank You: rodw
Please Log in or Create an account to join the conversation.
Moderators: snowgoer540
Time to create page: 0.103 seconds