[Mooser]

Just wanted to get some feel for what you are getting.
When restarting the machine (next day or whatever) and re-homing (ref-all)
How far off are you finding your G54 co-ords to your fixtures, vice jaw or whatever?
I've been moving the axis about 3 inches from home before ref'ing them and when re-clocking I'm usually about .002-.003 on the X, closer on the Y, maybe .0005-.001
Anyone getting better? worse? tips?
Mooser

[philbur]

Mine (on average) are about the same, maybe slightly better.

Phil

Just wanted to get some feel for what you are getting.
When restarting the machine (next day or whatever) and re-homing (ref-all)
How far off are you finding your G54 co-ords to your fixtures, vice jaw or whatever?
I've been moving the axis about 3 inches from home before ref'ing them and when re-clocking I'm usually about .002-.003 on the X, closer on the Y, maybe .0005-.001
Anyone getting better? worse? tips?
Mooser

[philbur]

"Home" of course has not necessarily been designed with repeatability in mind. It is also clear that there is lots of scope to improve repeatability if the need is great enough.

Tormach (Greg Jackson) commented on "Home" repeatability somewhere/sometime. Possibly it was on the Yahoo group, or maybe here.

Phil

Just wanted to get some feel for what you are getting.
When restarting the machine (next day or whatever) and re-homing (ref-all)
How far off are you finding your G54 co-ords to your fixtures, vice jaw or whatever?
I've been moving the axis about 3 inches from home before ref'ing them and when re-clocking I'm usually about .002-.003 on the X, closer on the Y, maybe .0005-.001
Anyone getting better? worse? tips?
Mooser

[keithorr]

It is determined by the action of the limit switch. My X and Y got gunked up at different times and the switch action was really slow. A few times the Y axis crashed at the physical limit and other times ate up a couple of inches travel before releasing. Replacing with optical switches would increase accuracy but you would still want to re-zero your machine relative to your main fixture as a daily habit.

[Freddy Bastard]

Mine is about the same. Sometime I get my limit switch in X don't work very well, The machine will reach the limit switch but come back 1 or 2 inch. When it do this, I re-home... If we could get a better reference return it would of great help for me.

[Mooser]

It would be nice to turn on the machine without having to re-clock the fixture each time.

I'm trying to think of how the nc's at work run, Their home switches must be more accurate somehow. I've got to look at the mazak and see how it works, I'm pretty sure they are mechanical.

I've actually got a set of Fagor glass scales over one the bridgeport that I've been looking at. Seems to me that Mach3 can accept a feedback signal.

Then again, my touch probe is coming along nicely that should make re-clocking quicker.

Gives a person something to think about during longer cycle times
Mooser

[rzrostlik]

For what it's worth, I've been lurking on this site for a while considering a Tormach in my very near future.

It sounds like the Tormach uses a simple switch to determine home position. I know on some of the other machines I've worked on, they use a two-step process to determine home position. A limit switch is used to trigger a reversal in the direction of the ballscrew and then the controller looks for the next index pulse from the shaft encoder. These machines typically use a servo motor rather than steppers, but the concept would work for steppers so long as there is a signal equivalent (i.e. one pulse per ballscrew revolution) to an index pulse.

Obviously the controller must be programmed to reverse upon hitting the limit switch, but in theory, it's pretty simple and very repeatable.

Rick

[Don Clement]

I checked the repeatability of the X Y axis on the ref all command and both were about 0.001". When checking Home position for repeatability is the Ref All command done at the same temperature? Are the limit switches and limit stops clean and free of chips? I know in the mornings this time of year the temperature in my shop can be 40* F and by afternoon be 65*F. Remember expansion is in/in or cm/cm so the longer X axis will change more than the Y axis for the same temperature difference.

[philbur]

But if the limit switch triggers at + or minus 0.001" and then the controller reverses 1 step or 10 steps isn't the positional repeatability is still + or minus 0.001".

Phil

For what it's worth, I've been lurking on this site for a while considering a Tormach in my very near future.

It sounds like the Tormach uses a simple switch to determine home position. I know on some of the other machines I've worked on, they use a two-step process to determine home position. A limit switch is used to trigger a reversal in the direction of the ballscrew and then the controller looks for the next index pulse from the shaft encoder. These machines typically use a servo motor rather than steppers, but the concept would work for steppers so long as there is a signal equivalent (i.e. one pulse per ballscrew revolution) to an index pulse.

Obviously the controller must be programmed to reverse upon hitting the limit switch, but in theory, it's pretty simple and very repeatable.

Rick

[Andre' B]

It would be nice to turn on the machine without having to re-clock the fixture each time.

I'm trying to think of how the nc's at work run, Their home switches must be more accurate somehow. I've got to look at the mazak and see how it works, I'm pretty sure they are mechanical.

On most machines they use a combination of a home switch and the Z output on the encoders, this is a single index pulse output that happens once or on some machines a few times per revolution of the ball screw.

They look for the home switch to close and then start looking for the index pulse from the encoder. By doing this repeatability of the home position about the same as the resolution of the encoder.

[rzrostlik]

But if the limit switch triggers at + or minus 0.001" and then the controller reverses 1 step or 10 steps isn't the positional repeatability is still + or minus 0.001".

Phil

Phil,

I believe that using a two-step homing process will improve the situation if the positional repeatability is due to the switch (and not backlash in the ball nut).

In the two-step process, we are actually using the index pulse of a shaft encoder to represent 'home', rather than the position of table is at when the limit switch changes state.

The assumption here is that the index pulse on an encoder is much more repeatable than a mechanical switch changing state. This process requires that the switch is repeatable within one full revolution of the shaft encoder. This is usually not a problem. So, we use the mechanical limit switch to trigger an event that reverses the stepper motor and start looking for the very next encoder index pulse. Once the index pulse is found, we call this 'home'. Everything is referenced from this location, not when the limit switch changes state.

The beauty of this approach is that you can have a fairly sloppy limit switch and it should not impact home reference repeatability.

Rick

[philbur]

Ya sorry, I missed the significance of the "index pulse" as opposed to a "step". I guess that obtaining the pulse on a stepper setup is a whole different ball game though.

Phil

Phil,

I believe that using a two-step homing process will improve the situation if the positional repeatability is due to the switch (and not backlash in the ball nut).

In the two-step process, we are actually using the index pulse of a shaft encoder to represent 'home', rather than the position of table is at when the limit switch changes state.

The assumption here is that the index pulse on an encoder is much more repeatable than a mechanical switch changing state. This process requires that the switch is repeatable within one full revolution of the shaft encoder. This is usually not a problem. So, we use the mechanical limit switch to trigger an event that reverses the stepper motor and start looking for the very next encoder index pulse. Once the index pulse is found, we call this 'home'. Everything is referenced from this location, not when the limit switch changes state.

The beauty of this approach is that you can have a fairly sloppy limit switch and it should not impact home reference repeatability.

Rick

[keen]

But the Tormach is open loop - without shaft encoders.

[philbur]

I'm not sure anybody was proposing that it can be done on a stock Tormach, it was more just a technical discussion on what might be required to achieve more precise homing positions than can be obtained for a simple switch.

Phil

But the Tormach is open loop - without shaft encoders.

[keen]

Yes yes .......but this - and me forgetting to home at start up today has got me thinking. Correct me if you think I am wrong - but on the Tormach the Homing process is only useful for saving time resetting the work axis positions after a start up (and then only for non high precision work). The overun prevention is on automatically and as PCNC3 does not have 'soft limits' configured there is not this other layer of protection). The act of homing soley sets the machine co-ords to zero to allow re establisment of work position to the accuracy (of - as above threads). So actually referencing can be done after the work zero is set (this is useful if you are doing a quick job that has to be extended, of if you forget to reference before setting work zero - as I needed to do today)

Keen

[mc-motorsports]

IMO, the home switch's should be an accurate way to reset the work coorinates in general. If your using a plunger type mechanical switch, good luck, if it has a roller top, you need more than luck. The best way is supposed to be proximity switch's, the cheapest way is those crappy looking little micro switch's, they are actually pretty repeatable, and cheap, but they won't handle a bit of dirt or swarf. I quit using the home switch's (plunger type), mine are even less reliable than yours, lucky if it stays within .005"

[highspeedmazak]

Cant you just edge find each part after you rehome the machine?

[keen]

You mean accurate homing is not an issue if the work is re zero'd after start up each time? If so - sure. But in some situations this can be akward to do.On the Tormach the homing is only there to to save this step. Keen

[mc-motorsports]

You mean accurate homing is not an issue if the work is re zero'd after start up each time? If so - sure. But in some situations this can be akward to do.On the Tormach the homing is only there to to save this step. Keen

That's exactly it, but wouldn't it be nice if you could just re-home the machine and your right back to where your supposed to be? Within .0005"? Some of the fixtures that I run have a hole in the fixture that you indicate for work-piece zero. Problem is when you bolt down the work-piece, that hole is covered up, my mistake, have learned better over the years, so if anything happens, I have to unbolt the work-piece and re-indicate or tram the locating hole. Would be nice to just re-home the machine and know I'm right back where I need to be... agreed?

[highspeedmazak]

Ok how about at the end of your program just having a last point of return to a point like lets say we end at x=15 and y=1 and z=10 then when you start another part you could just go from that point and only go home at the end of the day. Would that work?