Home Switch Repeatability

TomS

Active User
H-M Supporter Gold Member
Joined
May 20, 2013
Messages
1,906
Just curious. What is a reasonable expectation for home switch repeatability? Checked my X axis with an indicator and it bounces around a bit (~ +/- .010"). I set the homing speed to 20% from 35% and no change in repeatability. I read on the internet that lever type switches are not repeatable but wanted to check with the resident experts.

Tom S.
 
To get better accuracy you want a slow make/ break rather than a snap action. +/- 0.010 is not too bad for a snap action. Lever switches work better with a cam actuator. You are best to use a two step process to home. Move to switch until it changes state, then very very slowly move off of the switch until it changes state again.

On my router I have slow make/break switches. It homes consistently at 0.000 measured with a dial indicator, but it comes off of the switches at 10 pulses/sec with a resolution of 25400 pulses/inch (about 0.023 IPM) in the second homing stage.

If you want real accuracy get these guys, 5 micron repeatability, they also make a 1 micron switch.
https://www.amazon.com/gp/product/B...rd_wg=JKhXL&psc=1&refRID=JJTSV8EG9C1V9BVP6QC3
 
Last edited:
The Metrol switch looks like an interesting alternative to the typical microswitch and ramp. It is still vulnerable to swarf though

I designed and installed an optical homing system on my Tormach mill and have .0001" repeatability. The OEM microswitch and ramp could vary by as much as .003" which is why I made the modification.
 
Just curious. What is a reasonable expectation for home switch repeatability? Checked my X axis with an indicator and it bounces around a bit (~ +/- .010"). I set the homing speed to 20% from 35% and no change in repeatability. I read on the internet that lever type switches are not repeatable but wanted to check with the resident experts.

Tom S.
Use an optical interrupter.
 
To get better accuracy you want a slow make/ break rather than a snap action. +/- 0.010 is not too bad for a snap action. Lever switches work better with a cam actuator. You are best to use a two step process to home. Move to switch until it changes state, then very very slowly move off of the switch until it changes state again.

On my router I have slow make/break switches. It homes consistently at 0.000 measured with a dial indicator, but it comes off of the switches at 10 pulses/sec with a resolution of 25400 pulses/inch (about 0.023 IPM) in the second homing stage.

If you want real accuracy get these guys, 5 micron repeatability, they also make a 1 micron switch.
https://www.amazon.com/gp/product/B...rd_wg=JKhXL&psc=1&refRID=JJTSV8EG9C1V9BVP6QC3

I get what you are saying. I'll slow the feed rate way down and see what I get. My setup uses slow make/break switches (pictures attached). I'd be happy with +/- .005".

Tom S.

X axis
20150323_103336.jpg

Y axis
20150319_135317.jpg

Z axis
20150323_103402.jpg
 
That looks like a good setup, just the way I would do it. :encourage:

But..........Now I get to climb up on my soapbox :grin:

I see no reason to have home switches on a mill at all. The only exception to this would be on the Z axis to insure a parking position for an automatic tool changer that requires the Z to be in a specific position.

IMHO the ''home position'' should be a variable ''parking position'', settable on the fly. All coordinates should be relative to the work 0,0,0 position.

To accommodate a power fail or other catastrophe that causes a zero loss, a datum point can be set up somewhere within the work envelope of the machine. When using a vice, I normally set the datum to the vice surfaces, and in most cases this is also the work zero. When work is bolted directly to the table, I know the location of the fixture or work edges or the location of some feature on the work, or in the rare case that it is not possible to do that, will set a datum by bolting a small piece of scrap to the table somewhere and just move over it and spot a small divot at a known location.

The idea of having two coordinate systems (machine and work) on a machine is totally crazy. When setting up a job, you don't set the job to the machine coordinates, you set the job on the table then set the machine to the work. Now for a manual tool change on a knee mill, you normally need to move the table (and the work) to a location that allows clearance to make the tool change. On your mill with a mile of Z travel all you really need to do is move Z up to a parking position that gives you enough clearance to do the tool change.

Once an operation or job is complete, then the axes move to the convenient parking location that allows access to the work for measuring, removal, etc. But this is still relative to the work coordinates.

Now on my router, this is a slightly different situation, since my work is normally aligned to the machine, the top, right corner of the table, I use that as the ''home position'', but that can still change if needed to accommodate the work. But this is still the work coordinates and only coincidently the machine coordinates.

OK, I'll climb down off of my soapbox now ;)
 
Last edited:
Tormach's Path Pilot requires homing the three axes before work. Personally, I like the ability to accurately and repeatably rehome the machine should an e-stop occur or should I have to shut the machine down as in overnight to continue the next day. There are a number of jobs that I have done where the workpiece reference disappears once machining begins. I did build an external reference into the table which would allow me to re-obtain an accurate reference but Mach 3 wouldn't reset my G55, G56, etc. work offsets to the updated reference coordinates. Path Pilot goes one step further in that if an external reference is used, it has soft stops which prevent the machine from traveling past that reference point.

When I installed the optical homing I left the OEM homing intact. This was done partially as a backup if the optical system failed but also because the homing switches also served as limit switches for the extremes of travel. Path Pilot has eliminated the need for second feature with its soft stops as it very neatly prohibits movement past the soft stop. Even with a G0 move at 135 ipm, it anticipates the stop and the software will stop exactly at the soft limit. Since switching to Path Pilot, my OEM limit switches never are activated.

Bottom line, yes I could function without the homing but it sure is convenient to have. If you are going to use homing though, IMO it should be as accurate as your work requires. Otherwise there is no use for it.
 
Other advantages of a Home position in Mach3 are:

1) If you setup the machine parameters Mach3 will display your work position on the table. Useful to verify the code is correct.
2) If you setup soft limits Mach3 will tell you if the tool path exceeds the travel limits.
3) If you use multiple work offsets for multiple fixtures/vises.
 
another nice feature of having home/limit switches is that if you keep your vice on your table you can use the surfaces on your vice at your work coordinates. Home your machine then indicate the edges you want to use as your reference and record the machine coordinates for that location. You can now just punch in those coordinates for your G54 when you restart your machine without the need to indicate anything. I actually use the left side of my vice as my G54, and the right side is my G55.

ZK9vQtt.png
 
Back
Top