Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: thespindoctor on September 22, 2011, 09:09:32 AM

Title: limit switch accuracy
Post by: thespindoctor on September 22, 2011, 09:09:32 AM
What type of limit switch is the most accurate?

Thanks
Title: Re: limit switch accuracy
Post by: Tweakie.CNC on September 22, 2011, 11:59:47 AM
Hi Doc,

Good question - bit like asking "how long is a piece of string ?"   ;D

I suppose it all depends on how much you want to pay and the level of accuracy that you need.

I use ordinary Honeywell Microswitches but Baumer Electric produce a superior range and their mechanical push switches, in the My-Com range, have a repeat accuracy of better than 0.001 mm (constructed a bit like a touch probe).

Tweakie.
Title: Re: limit switch accuracy
Post by: thespindoctor on September 22, 2011, 12:14:44 PM
Thanks,  I think I need to modify the way I am doing the probing to this switch to
get repeatability so I will try that first then go to better switches and see if I see
a difference.  I have been using the switch to check the size of my grinding wheel but
only moving slowly in Y to the switch to trip it and noting the position but I need
to trip the swith and move past then reverse and note the closure point.  It seems
like moving toward the switch is better for this application though.  Otherwise there
may be some other mechanical issues in my system because I am not getting
expected consistency.
Title: Re: limit switch accuracy
Post by: Tweakie.CNC on September 22, 2011, 12:31:32 PM
The touch probe macro sounds like it should meed your needs then.

Tweakie.
Title: Re: limit switch accuracy
Post by: thespindoctor on September 22, 2011, 07:33:21 PM
I just looked at the speed of homing and I had it at 100% so I cut
it down to 5 % and now I am getting much more reproducible
results so it does not look like the switch was the weak link!
Title: Re: limit switch accuracy
Post by: Tweakie.CNC on September 23, 2011, 02:22:30 AM
You are nearly there then  ;D

Tweakie.
Title: Re: limit switch accuracy
Post by: BR549 on September 23, 2011, 08:46:24 AM
Might want to consider the G31 probing routine to do the Wheel diam function. The G31 sets the trip point on the way INTO the shape and is deadly accurate. and repeatable.

Just a thought, (;-) TP
Title: Re: limit switch accuracy
Post by: Jeff_Birt on September 23, 2011, 09:58:13 AM
I always like questions like this as the accuracy of the system in general is probably more important to know.  

Quote
have a repeat accuracy of better than 0.001 mm

Is that really the manufactures specification?  It is quite deceptive of them. Under what conditions would such a rating be true. That is about 4 hundred-thousandths of an inch. If you breath on it your likely to heat it up enough to change the reading not to mention how the reading will change as the machine itself heats and cools. A speck of dust in around 0.00002" so if the sensor or actuator is not perfectly clean then your reading will be an order of magnitude off (gasp only good to 4 ten-thousandths of an inch).

A typical mechanical switch will vary a bit in the exact position it opens and close due to several factors. The spring loaded parts inside will react slightly differently with temperature changes. The contacts also don't instantly change from open to closed (or closed to open) they 'bounce' a little so which of these bounces is Mach likely to see?

Think of the switch contacts as a resistor, when the contacts are closed you have close to zero ohms of resistance, when open the contacts have nearly infinite resistance. As the contacts start to open the resistance starts to rise every so slightly. Witch a good contact design this resistance change is fairly accurate and expensive contact probes look for this change is in resistance which provides a more repeatable indication that the switch is opening.

Back to the job that the OP is trying to do. You may find it is more accurate to move until you contact the switch and then back off of it really slowly.
Title: Re: limit switch accuracy
Post by: Tweakie.CNC on September 23, 2011, 10:57:50 AM
Quote
Is that really the manufactures specification?  It is quite deceptive of them. Under what conditions would such a rating be true. That is about 4 hundred-thousandths of an inch. If you breath on it your likely to heat it up enough to change the reading not to mention how the reading will change as the machine itself heats and cools. A speck of dust in around 0.00002" so if the sensor or actuator is not perfectly clean then your reading will be an order of magnitude off (gasp only good to 4 ten-thousandths of an inch).

A typical mechanical switch will vary a bit in the exact position it opens and close due to several factors. The spring loaded parts inside will react slightly differently with temperature changes. The contacts also don't instantly change from open to closed (or closed to open) they 'bounce' a little so which of these bounces is Mach likely to see?

Think of the switch contacts as a resistor, when the contacts are closed you have close to zero ohms of resistance, when open the contacts have nearly infinite resistance. As the contacts start to open the resistance starts to rise every so slightly. Witch a good contact design this resistance change is fairly accurate and expensive contact probes look for this change is in resistance which provides a more repeatable indication that the switch is opening.

It sure is Jeff  ;)

Although their switches are used with CNC many are used on spacecraft as well as military equipment.

Attached, typical construction of "My-Com" series.

Tweakie.
Title: Re: limit switch accuracy
Post by: Wallerawang on September 23, 2011, 07:57:59 PM
I was reading this post with interest as I am looking for homing switches for a lathe project.  I found the Australian Distributor and asked for a quote on the MY-COM D250/80 which looked like a suitable switch to me for job I wanted it to do. The quote was $748.00 + $25.00 Delivery + another 10% Tax.  To me a little expensive so I will have to keep looking for an alternative. (this is more then I paid for a 1.2 kW servo motor and drive)
Steve
Title: Re: limit switch accuracy
Post by: Tweakie.CNC on September 24, 2011, 01:18:10 AM
 ;D ;D ;D

Tweakie.
Title: Re: limit switch accuracy
Post by: Hood on September 24, 2011, 02:43:34 AM
Optical switches are cheap and very accurate so if you can keep them dirt and coolant free they work very well. Housings can be made with a rod sticking out either side. The rod has a flag in the middle that interrupts light and is held in position with a spring either side. When the rod gets pushed the flag will move away from the switch and let the light through and your switch will send the input to Mach.
I got my original ones from I think Industrial Hobbies but the dont do them now since the original guy sold the business, shame really as they were reasonably priced.
 I did tests when I first got mine, I had a glass scale and external DRO on the mill and I sent the axis to a position and zeroed the external DRO, I then jogged away and then sent it back to the initial position and the DRO read zero. I did this about 100 times and only once or twice did the DRO not read zero and even then it was only 0.005mm. This was on a clapped out old Bridgeport that was my first conversion and I would guess the few times it showed as not zero was due more to the slop in the axis rather than the switch.

Hood
Title: Re: limit switch accuracy
Post by: thespindoctor on September 24, 2011, 07:27:49 AM
I guess that is where I am concerned.  It seems like the switches I am using are pretty good.
When I repeatedly probe my grinding wheel on the mechanical switch I am using, I get the same value
each time as long as I do a homing move first to machine zero.  But when I turn the machine off
and start up again, the value tends to be different.  I am wondering where the difference in coming
from so trying to investigate other sources of backlash.

Thanks
Keith
Title: Re: limit switch accuracy
Post by: Hood on September 24, 2011, 12:00:54 PM
When yo turn off the machine your axis will likely move a bit, for example with a stepper you may switch off on a microstep but then the motor would tend to move itself to the nearest full step so your axis will be different to where it was. You should always home your axis after starting Mach and after any E-Stop or drive fault etc.
Hood
Title: Re: limit switch accuracy
Post by: thespindoctor on September 24, 2011, 12:19:24 PM
Hood, that is what I have been doing but even after starting up the machine and rehoming the machine,
the next time I believe the probing of the wheel is different but I need to double check it to see if
it is repeatable.  I will check to see if a timing pulley might be slipping or ball screw nut not secure.

Thanks,
Keith
Title: Re: limit switch accuracy
Post by: derekbpcnc on September 25, 2011, 05:06:34 AM
RS sell the My-Com range.
This one with 1um accuracy is about £63.00
http://uk.rs-online.com/web/p/precision-position-switches/0341418/
 (http://uk.rs-online.com/web/p/precision-position-switches/0341418/)
It says the hysterisis is 2um..but I recon we can live with that amount of slop  ;D

ATB
Derek
Title: Re: limit switch accuracy
Post by: HimyKabibble on September 25, 2011, 09:06:39 AM
When yo turn off the machine your axis will likely move a bit, for example with a stepper you may switch off on a microstep but then the motor would tend to move itself to the nearest full step so your axis will be different to where it was. You should always home your axis after starting Mach and after any E-Stop or drive fault etc.
Hood

It's actually worse than that - Stepper drivers will typically start up in the same full step position each time, as they have no way of knowing what position the motor is in.  So, when re-powered, the motor could end up several full steps off its last position.

Regards,
Ray L.
Title: Re: limit switch accuracy
Post by: Jeff_Birt on September 25, 2011, 12:23:00 PM
Quote
This one with 1um accuracy is about £63.00

You have to be very careful when reading a data sheet, it does not say  '1um accuracy', it says 'switching point sensitivity' (which is what?).

I've attached the mechanical specs of the switch below. Its repeatability is shown as +- 0.001mm with hysteresis of <= 0.002mm. While these are still impressive sounding numbers the typical well adjusted machine might have backlash as low as of 0.01mm, add in less than perfectly clean activating surfaces, temperature changes in teh machine ,etc and the 'real world accuracy' of even a precise switch like this starts becoming less clear.

I'm not wanting to be a negitive nellie but rather want everyone to get the 'whole picture'. There are many factors involved and you can't just fixate on impressive sounding numbers from one sensor/switch and ignore the rest of the system.
Title: Re: limit switch accuracy
Post by: derekbpcnc on September 25, 2011, 12:52:25 PM
Quote
This one with 1um accuracy is about £63.00
I'm not wanting to be a negitive nellie but rather want everyone to get the 'whole picture'. There are many factors involved and you can't just fixate on impressive sounding numbers from one sensor/switch and ignore the rest of the system.

No one is fixated on the numbers or ignoring the rest of the system- rather the price to performance which appears to be rather good.
I don't expect a My-Com switch to eliminate all the inaccuracies from a machine / machining process, or anyone elses for that matter - that would be silly :-)...but it looks to be a good starting point for anyone looking for a accurate home switch.
Title: Re: limit switch accuracy
Post by: thespindoctor on September 25, 2011, 01:01:13 PM
I appreciate the discussion and the sources for more accurate switches.

Thanks!
Keith