MechMate CNC Router Forum

MechMate CNC Router Forum (http://www.mechmate.com/forums/index.php)
-   70. Control Systems (http://www.mechmate.com/forums/forumdisplay.php?f=28)
-   -   Proximity Sensors | targets / detecting & wiring (http://www.mechmate.com/forums/showthread.php?t=4526)

Fox Sat 26 July 2014 14:58

Proximity Sensors | targets / detecting & wiring
 
I am about to install my proximity switches and I am wondering about some things:

1) Do those who run proximity sensors, use the lasercut eccentric target placed in the enlarged hole, according to the plans, and how do you stop it from turning ? Or did you just drill a hole ?

2) In the latter case how do you fine-tune the sensors to trigger exactly at the right time ? And in the first case: the eccentrics would move the hole also from left to right instead of only back and forth when turned, giving an off center target...doesn't this give unreliable triggering ?

3) How do you make sure the sensors on the X-axis trigger exactly simultaneous?

4) Did you wire each sensor individual or all together (see 5) ?

5) Does anyone use autosquaring on the X using their proximity sensors( my software provides for it and I think Mach3 also) but you need to run wire the sensors individually I imagine ?

pblackburn Sat 26 July 2014 23:39

I use individual axis proximity switches with my homing. X and B will home individually. Both will run until one made and the other will continue until it is made. I used separate setup rather than the eccentrics so I cannot comment.

Fox Sun 27 July 2014 02:41

What do you mean with separate setup ?

Tom Ayres Sun 27 July 2014 05:08

He has his X and slave triggered separate to allow auto squaring of the gantry. I have both set to trigger as a one. The accuracy (imho) only matters in Pete's case the rest are reference points for machine coordinates and of course limits. So concentrics are an easy way to allow adjustment for auto squaring. As far as attaching them thats up to you but look at some of the older setups where these thing were introduced. Pete give him the lowdown.

Fox Sun 27 July 2014 08:37

Hi Tom,
I was looking at the older machines already but have a hard time digging up pics/experiences on the eccentrics and proxies.

With teh stuff in front of me I was actually thinking it would make more sense if the hole for the prox in the stopblock would have been slotted, hence my questions. Still pondering on those.

And do you guys use only 1 or 2 target for the X proxy ?

Tom Ayres Sun 27 July 2014 09:46

I use 2 proxys but act as one, Pete I think uses 2 acting as 2. As far as the actual targets, I followed Gerald's original plans. I don't actually use concentrics just the target holes.

pblackburn Sun 27 July 2014 10:06

Most people set home and limit from the same switch for each relative axis. This is because of the limited number of inputs that are available with a standard setup and limited money. The switch for limit is normally wired in a fail-safe manner where the switch is always active and a drop of signal indicates a problem. The eccentric allows the hole to be moved or fine tuned for this trigger to occur. If you plan on correcting racking manually, you only need one switch. If you are going to have the software do it, you need two. One for axis X and one for axis B.

Fox Sun 27 July 2014 15:18

@ Tom, so how did you adjust those to sense simultaneous if you just have the hole ?
@ Pete, understood. I will use proxies as limit and home switches.

Remaining questions:

- The eccentrics would also move the hole from left to right instead of only back and forth when turned, giving an off center target...doesn't this give unreliable triggering ?

- How do you make sure the sensors on the X-axis trigger exactly simultaneous? Or is that not needed ?

- Do you guys use one, or two targets on the Z-axis ?

Tom Ayres Sun 27 July 2014 18:45

Fox, the hole (or lack of anything) is what triggers the switch, that is my 'machine coordinates' reference (doesn't really matter where that actually is as long as its the same every time)(truthfully I don't know which side triggers the switch first, personally, I don't think it matters, some find it more important). The two proxys are wired (can't remember if parallel, but I think it is, I'll have to look back in my records) to one input pin on the bob, that way if either side of the gantry is accidently lifted from the rail the switch is activated. It has been suggested by Steve at PMDX that the slave to X share the same pin as the Y car (if you have limited inputs) to make use of auto squaring and that mach already recognizes when the proxy is being used for squaring and/or limits. (never tried it though, but thats when concentrics are needed)

The concentrics were designed to allow something like 3mm adjustment, meaning the hole is offset only 1.5mm from center, not enough right-left movement to matter but still gives you that fine tuning if your proxys are located correctly.

Hope that helps and doesn't cause more confusion.

Fox Mon 28 July 2014 10:01

Ok. I think that helps :rolleyes:
I was just puzzled by the X-axis triggering, to be able to auto square I imagine fine tuning is vital for teh 2 sensors ? But I guess I just run them as is, and dive into that matter later to be able to move forward...

smreish Mon 28 July 2014 12:30

Mach has a homing routine that takes into account all axis sensors wired to one input.
It starts moving x, then y then z.
ANY sensor in the 3 or 4 that trigger while the active axis is moving is treated as "Homed" for the axis in motion.

If you trigger the z switch while x is moving. X will be homed at time of trigger on Z.

Makes sense?

pblackburn Mon 28 July 2014 15:13

Good post Sean, I forgot to mention that.

Fox Mon 28 July 2014 17:43

Hi sean I understand that behavior when using multiple sensors and one input..... It's logical behaviour...

But what I don't understand is how it does not matter the sensors are synched excactly to be able to perform an auto square when sensors are wired individually..... ( all my sensors are wired individually, I use edingcnc )
My understanding is that to be able to auto square it will 'home' x1 then x2.

So when Pete tells me he uses auto square ( Mach 3 I know, but software is irrelevant IMO ) I do not see how that would properly work when he did not make sure that his sensors trigger excactly at the moment they are ( actualy his gantry is ) in square position. Meaning FINE tuning of the target ( holes )... Unless he drilled them to a hundredth of a mm exact ?
But maybe I a missing something...???

pblackburn Mon 28 July 2014 18:03

I use a prox that is dedicated to its own pin assignment in Mach3. Each prox has its own port and pin assignment and none are shared. So if X triggers first B will continue to travel until the B home prox is met and Vise-Versa. I did not use the holes for the home. Home is usually a pulse location. So I wired it normally open with a raised slide that can be metered for fine positioning. The Limits on my mine is like an overtravel on a mill. Separate set of switches entirely.

Tom Ayres Mon 28 July 2014 19:11

Fox, let me throw in another thought, What about those who do not use proxys and use machanical stops to set the gantry square using the steppers to hold it, how automatic can that be? Most if not all gantries are not square so hitting the target simultaneously is nearly impossible anyway. I understand what you are saying if your assumptions are based on a perfectly squared gantry to start with.

pblackburn Mon 28 July 2014 22:54

Racking is generally not an issue. I have only seen it extreme once and the steppers growled the whole way until it was homed.

Tom Ayres Tue 29 July 2014 03:00

You got bigger problems if its growling :)

Richards Tue 29 July 2014 08:07

All the stepper motors that I've ever used "resonate" (growl) when they're run at a low speed where the "resonating" happens. The cure is to run them faster or slower. Usually the "resonating" speed is really slow, so going just a little faster cures the problem.

Some motors can loose steps when they run in their "resonating" range.

A problem with CNC machines is that 2D and 3D cuts can require a motor to run in its resonating range. I've never had a "square" motor loose steps when it resonates. Some of the older "round" motors lost steps.

danilom Tue 29 July 2014 13:55

With AM882 drives , the slower you go there is less vibration and noise, when zeroing I can't hear Z axis stepper spinning. I'll record it in next couple of days :)

pblackburn Tue 29 July 2014 14:24

I was being facetious. It was growling because it racked so bad. Not the steppers but the noise from the pinion and rack misalignment.

Fox Tue 29 July 2014 16:29

Tom you are misunderstanding the concept of auto squaring. In order to auto square, your targets need to be exactly at the square location, if your gantry becomes out of square, one target will trigger then the other, and it wil reset to those new values. Similar to driving your gantry into your mechanical limits, but then automatic.

pblackburn Tue 29 July 2014 19:41

I would say the locations need to be in a place that renders a square end result. They do not need to be at a square location on the rails but need to have the end result be the Y car is traveling square to X. Since some have to favor the Y rails a little to set square to X because of parallelism that sometimes occurs with the welding of the gantry. So the location would not be square across the machine in this instance.

Fox Wed 30 July 2014 00:29

Hi Pete, you're totally right about that of course.

rischoof Wed 30 July 2014 10:59

Proxy
 
FOX,
Good to hear you make progress,
I have separate home and limit switch. Limit switch get activiteiten with hole. This is fixend and can't be moved, (like the originaliteit plans)
Homing is done on a nut which can be slided to adjust. At only one Side
I made a square plate from 3 mm aluminium 30x30 mm. It has two sleeves 15x3 mm to adjust on the rail. IT has a m5 hole in the middel. There is a bolt mounted. Thread Side is 3mm long, so it fits in the alu plate. Head Side is turned off. To 3mm.
For my z I use a slidable home it can be slide 200mm. I used a curtain rail to mount the slidable home. So sometimes I put it lower so I don't have to wait till it's completely up.
When my x and y Homing point are fixend. I drillend Some 6 mm holes at x-3 with a pitch from 200 mm
Also at y in drillend holes at y-3 pitch 200.
In those holes I put dowell pins from 6 mm.
Those dowelpins I use as stop ( Dutch:aanslag) I remove the dowell pins when the material is clamped.
I have seen Can's machine. He is using 2 hard endstops. Put the x gantry to those two endstops by hand and activates the drivers. It works quicker than automatic Homing and it's accurate. And he saves the money for the proxys.

Fox Wed 30 July 2014 13:09

Hey Rik !

All good info, thnx.
I already got my proxies and hard end stops mounted. Still undecided on how I will make my targets however, wiring the beast as we speak ( pulled 150 meters of cable trough chain today ) and still I got plenty to do before I have to.

I see in the plans (downloaded way back) a hole of 20 mm with an ex center which has a 10 mm hole. But the sensor mentioned everywhere is m12 ? Did Gerald account for an 8 mm sensor ? As i can't believe the 10 mm hole/target will work with an m12 sensor ? Plot thickens...

ps. good to see you are still alive ...How about your machine ?

Gerald D Wed 30 July 2014 13:20

The sense "window" is cone shaped, therefore a 10mm hole does work with a 12mm sensor (depending on distance).

https://pc-control.co.uk/images/proxy.jpg

I did test everything before it was put on the drawings - including the eccentric 10mm hole.

Tom Ayres Wed 30 July 2014 14:39

Fox I do understand that it must be perfect but the gantry does not have to be hence the need for auto squaring

Fox Wed 30 July 2014 17:16

Hi Gerald !

I was only wondering.... cause you did not put sensors on any of the machines youve build, I've read that somewhere.....
I should have known :D

Well I guess I'll stick to the plans...

My compliments to your thorough design, it's been a pleasure during the build to find out how much you have thought things trough ( and sometimes not so... When I deviate from your plans and find myself in trouble soon thereafter...);)

How's your son and his mm fleet doing these days !?

@ Tom, sure thing.

Gerald D Thu 31 July 2014 10:00

Son's fleet is working hard - already wore out a set of cables. Not killed a single motor yet. Still no proximity sensors. It wasn't easy to decide on a 10mm hole, it could have been 11mm. But somebody can always make a small hole bigger if needs be. Or, even file it square . . . . (never be scared of a file!)

Fox Thu 31 July 2014 14:06

Good to hear, let's hope I can follow in those footsteps and build a business out of the machine as well, hope to have it running within a month.

Amen. Manual files are great, first thing I teach my employees.
Nowadays everybody seems to think it's a bad thing when you sweat a little with them powertools lying around in the shop. I use my assortment of files coarse (standard) to very fine ( diamond lock files) almost daily, gives so much more control and often not even slower then a dremel or powerfile.


All times are GMT -6. The time now is 07:46.

Powered by vBulletin® Version 3.8.3
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.