Setup for squaring twin X-axis gantry

[routercnc]

No, this isn't the usual question about Mach3 settings for slaved twin X-axis homing.

I'm upgrading my cnc machine to twin X-axis ballscrews on the gantry. My new home switches have arrived, and reading the Mach3 manual and around the web shows me that I can independantly home each ballscrew to straighten the gantry. All fine.

But what I had been expecting was that you could tell Mach3 how much to back off the switch on the slave side, to tune the final position and make it perfectly square. Otherwise it would seem that you need to design some kind of micro-adjustment on the slave microswitch to fine tune the position.

Can anyone tell me which of these options is required - I was expecting to get the switch pretty close mechanically and then fine tune the amount of back-off in software to get it dead on.

Thanks

[JAZZCNC]

Can anyone tell me which of these options is required - I was expecting to get the switch pretty close mechanically and then fine tune the amount of back-off in software to get it dead on.

Take the time to Get it dead on to start with then you won't have to is the best option.!! . .. . .Then if any adjustments are needed at all your into fractions of millimeters and so just tweaking the switch is plenty.

Proximity switches are great for this has you have adjustment via the threads on the switch and altering the sensing distance if needed has well as any lateral adjustment you build into the switch.

[EddyCurrent]

I can't get my head round the fact that accuracy of the machine is down to the position and sensitivity of proximity sensors or switches.
I'm talking about repeatability I suppose, it's not so bad when there is one switch per axis but when there is a separate one for slaved motors it makes even less sense.
This is one area where absolute encoders are far superior and there are programmable limit switches based upon encoders but quite expensive.

[JAZZCNC]

I can't get my head round the fact that accuracy of the machine is down to the position and sensitivity of proximity sensors or switches.
I'm talking about repeatability I suppose, it's not so bad when there is one switch per axis but when there is a separate one for slaved motors it makes even less sense.
This is one area where absolute encoders are far superior and there are programmable limit switches based upon encoders but quite expensive.

Yep I agree 100% and exactly why I prefer screws connected Via belts has it's simple and accurate.! ...............Again thou it's down to usage and for machines which don't require higher accuracy or repeatabilty like routers just cutting wood, plastics etc then slaved motors are perfectly fine provided they are run within safe limits.

Edit: Daft as this may sound then if you want better accuracy from slaved motors than using switches just use Hard stops and slowly drive upto them before setting Zero.!

[Neale]

If you are fitting home switches, then in general terms what's the best kind? I'm looking for repeatability here, I think, and the choice seems to be some kind of mechanical switch or a proximity switch. I'm happy that I could interface either kind with whatever electronics/breakout board I eventually select. I have some slight concerns because my 3D printer uses simple little microswitches. In principle these should work OK but I've had repeatability problems with the Z axis switch as you are looking for better than 0.05mm or so and the original switch was not capable of achieving that. In the end I swapped switches around until I found one that gave better repeatability but as a result, I'd like to make sure that the new router, when it gets built, doesn't suffer this way. Could be a critical issue if you are using dual X motors when it comes to squaring/resetting, of course.

[JAZZCNC]

If you are fitting home switches, then in general terms what's the best kind? I'm looking for repeatability here, I think, and the choice seems to be some kind of mechanical switch or a proximity switch.

Question is how deep are your pockets.? High repeatabilty requires high quality switches and they cost plenty. I find for general DIY then your average chinese Proximity switch are good enough with much better repeatabilty than cheap MS and even better quality MS like Honeywell's.

Mechanical MS are ok for limits, thou I tend to use proximity for both has they are cheap enough and I just use the same 24V supply.

[EddyCurrent]

Repeatability tests on YouTube, there's plenty more too.

G0704 Home/limit proximity switch repeatability test - YouTube (http://www.youtube.com/watch?v=IGO47l62T2o)
Homemade DIY CNC Series - Home Switch Repeatability - Neo7CNC.com - YouTube (http://www.youtube.com/watch?v=i3nmp9AXccQ)
Mechanical switch repeat ability - YouTube (http://www.youtube.com/watch?v=5l7UMoEBP6g)

These seem to be a favourite and I think it's the ones in that first video.
http://www.ebay.co.uk/sch/i.html?_odkw=sn04&_osacat=0&_from=R40&_trksid=p2045573.m570.l1313.TR0.TRC0.H0.Xsn04+swit ch&_nkw=sn04+switch&_sacat=0

I got some of these for my limit switches, but will probably use proximity type for home switches. (postage a bit steep if you just want one but if buying more it's a good deal, or get from China if you want better)
http://www.ebay.co.uk/itm/Industrial-Limit-Switch-TZ-Series-/270745711183?pt=UK_BOI_Electrical_Components_Suppl ies_ET&var=&hash=item3f09b3b24f

[Ger21]

You can make your own very accurate switches for very little money. These are what I use.
Electronic home switches made easy! (http://www.cnczone.com/forums/open-source-cnc-machine-designs/101878-electronic-home-switches-made-easy.html)

[JAZZCNC]

You can make your own very accurate switches for very little money. These are what I use.
Electronic home switches made easy! (http://www.cnczone.com/forums/open-source-cnc-machine-designs/101878-electronic-home-switches-made-easy.html)

I remember this Gerry and Romani did a great write up thou he did go a bit deep with homing to Magnetic poles of Steppers.. .Lol

There are 2 things I don't like about these switches. Magnetic pick up and 5V.

Magnetic is Ok for wood machine etc but not milling machines or near metal for obvious reasons.
Don't like 5v has it's susceptible to picking up electrical noise.

Inductive proximity switches are cheap these days and work just has good I've found and can be run thru the same 24V control system used for Limits etc.

[EddyCurrent]

You also have to consider switch hysteresis because that determines how much Mach3 backs off once activated.
These switches for example have to be backed off 5mm before they reset.
TZ-8104 Rotary Roller Lever Arm Enclosed Limit Switch | eBay (http://www.ebay.co.uk/itm/TZ-8104-Rotary-Roller-Lever-Arm-Enclosed-Limit-Switch-/290923040376?pt=UK_BOI_Electrical_Components_Suppl ies_ET&hash=item43bc5d4e78)

[EddyCurrent]

You can make your own very accurate switches for very little money. These are what I use.

Yes I was thinking about using a hall effect sensor, seeing as I cut wood they would work for me, thanks for the info.

[JAZZCNC]

Yes I was thinking about using a hall effect sensor, seeing as I cut wood they would work for me, thanks for the info.

To be honest it's not worth doing unless you buy the sensor and Led resistors etc cheap has you can buy inductive prox switches for approx £2.80 from Ebay.

I bought these to try and they are fine giving repeatabilty much better than mechanical switches and simple and easy to adjust. They also allow higher voltage so can be used with 24V and kept in line with rest of system. They are more than good enough for a Wood router without the hassle of soldering and making holders etc.
UK 10pcs Inductive Proximity Sensor Detection Switch NPN DC6-36V LJ12A3-4-Z/BX | eBay (http://www.ebay.co.uk/itm/UK-10pcs-Inductive-Proximity-Sensor-Detection-Switch-NPN-DC6-36V-LJ12A3-4-Z-BX-/400591616751?pt=LH_DefaultDomain_3&hash=item5d451efaef)

[Neale]

Do these need a magnet to trigger them, something magnetic, or what? Looks like a cost-effective device, in any case.

[JAZZCNC]

Do these need a magnet to trigger them, something magnetic, or what? Looks like a cost-effective device, in any case.

If your talking about what I posted then No they just need a Steel target to trigger them which the sensor detects as it passes the face. Adjusting the distance from target to sensor face as the affect of adjusting sensitivity.

[EddyCurrent]

I've been thinking about this gantry squaring issue when two motors are used for the X axis.
If the gantry design consists of two tall end plates with unsupported rails between or maybe aluminium profile and supported rail, I can understand there might be a fair amount of flex between the two ends and in this case it might make sense to have two home switches.
In my case where the gantry is like this http://www.mycncuk.com/forums/gantry-router-build-logs/6565-ready-steady-eddy-18.html#post55076
it's rock solid and virtually no way it can go out of square as it's like one big lump of metal. With this in mind I'm proposing to use just one home switch for the X axis because I think the steppers will stall if they try to home independently and the AM882's will then trip out.
Any thoughts ?

[Ger21]

Clamp one side so that it can't move, then push very hard on the other side. I wouldn't be surprised if it moves at least 1 mm.

Another thing to consider, is that when you apply power to the steppers, that they'll usually "jump" to the nearest full step. If one moves one way, and the other moves the other way, then they'll be twisting the gantry out of aquare, although the amount would be very, very small.

If it's as rigid as you say, then one switch would be fine. But you need to make sure that the gantry is perfectly perpendicular to the axis is moving along.
If you go with two switches, you need to position them so that they get the gantry perfectly square. Both methods require very precise measuring and fabricating.

[JAZZCNC]

I wouldn't be surprised if it moves at least 1 mm.

Wow Gerry your getting all Metricated now, Watch it you'll be speaking with Yorkshire Mans accent before you know it. . Lol

Eddy every thing Gerry says but I'll also point out the affect of inertia pushing the gantry can lead to error that without home switches on each axis you'll never know about untill the error as accumulated enough to rack the gantry and stall ONE motor. The Mess that can come from this is sickening if it happens at high speed which it often does because that's when torque is lowest.!!

Another way and one I know Gerry used or maybe still does is to bump upto a hard stop on each axis which squares gantry and use this as machine Home position. Then just occesionally bump upto stops to realign.!!

Another way is to permantently fit cheap dial indicator at each side and use these to check if it's losing position.

[EddyCurrent]

The only time you are going to know it's moved out of square, unless it stalls, is when the gantry is back at the home switches. So with this in mind I'm going with one switch and will create a datum point each side to enable checking with a DTI as suggested. If upon the first attempt at homing the machine I can see that Mach3 moves it only a very small amount I may then consider using two switches but the last thing I want is for the homing procedure to gradually 'loosen' the gantry structure.

[Ger21]

The only time you are going to know it's moved out of square, unless it stalls, is when the gantry is back at the home switches.

Or when your parts don't fit together.

Yes, I used to just bump into the stops before I had home switches. But it's much easier to just push a button and let the machine take care of it.

With 25 years of using 32mm cabinet hardware, and damn Italian machines, knowing a little metric comes in handy. :friendly_wink:

[EddyCurrent]

Or when your parts don't fit together.

True, but that's after it's already been squared by whatever method was chosen, unless it's set to home every now and again during the job.

[Clive S]

Wow Gerry your getting all Metricated now, Watch it you'll be speaking with Yorkshire Mans accent before you know it And I thought it was the official system over there, but I have had a few glasses of vino http://www.nist.gov/pml/wmd/metric/upload/1136a.pdf:very_drunk: ..Clive