[strohkirchw]

Decided to add a treadmill motor and a prox sensor to my mill recently and am having some issues with the prox sensor. It's a NPN NO inductive type sensor and a g540 setup.

I wired up initially like the first picture and used a 18v power supply I had laying around. It would flash between 91xx rpm and 27xxx rpm in mach3 and then stay at 71xx rpm if I triggered the sensor. I checked the voltage coming out the NO connector and +v. With nothing triggering the sensor it read 21.9v. When something triggered the sensor it would read 22v.


I didn't think that was right so I wired it up the like the second picture even though I didn't see the difference. When I did this and triggered the sensor I heard a loud thunk from the sensor and it doesn't trigger any more(assuming it's broke now). I pulled reference for the wiring from this thread. http://www.cnczone.com/forums/gecko_...ex_sensor.html

Any ideas what I'm doing wrong?

[Al_The_Man]

What are the details of the Prox?
The fact that the inputs apparently already have a source, you will need to use open collector output for the prox type.
I am assuming the 48v is providing the input source on the G540, unfortunately they do not show the nature of the input configuration.
Al.

[SCzEngrgGroup]

If that output is open collector, as I suspect it is, then you need a pullup resistor from the output pin to +5V. 220 or 330 ohms should be good.

Regards,
Ray L.

[wildwestpat]

What is the sensor? Maker and type number as your description as NPN NO/NC only gives us the information on the output transistor (npn type) and the normal at rest state with the option of closed or open connections. The mechanical construction does not define the sensing technology. You may find that a pull up resistor is required if the sample you have is open collector. It is also possible that you have an proximity detector rather than a pulse counter. The proximity detectors have a limited operating speed. The fact that it has made a noise suggests it has become toast.

The links in the thread from which you have lifted the circuit do not work for me.

Regards - Pat

[strohkirchw]

Here is the sensor I bought. How do I tell if it's OC type? I also just realized it says switch :idea:

LJ8A3-2-Z/BX NPN NO DC 6 to 36V Proximity Switch Sensor | eBay


Can anyone point me to a sensor that will work in my application with similar mounting style?

[Al_The_Man]

If that output is open collector, as I suspect it is, then you need a pullup resistor from the output pin to +5V. 220 or 330 ohms should be good.

Regards,
Ray L.

The only thing wrong I can see is it will conflict with the Input source voltage?
The Gecko input itself should have some kind of source resistance, the fact that it will be operated by contact closure.
This is one problem I have with alot of boards sold out there when they do not indicate the exact nature and details of the I/O
The google search for the details of the prox did not show if it is open collector, AFAICT.
Al.

[wildwestpat]

OK so you have a proximity detector that can operate up to a maximum of 150 Hz ie with a single tooth sensor it will detect upto a maximum of 9000 rpm under ideal conditions. I say ideal as these detectors will have a zone where they have hysteresis. If you are pointing it at a gear wheel or multi slot vane than this lowers the maximum RPM in direct ratio to the number of teeth / vanes assuming you can find a mounting position that will permit each tooth to be detected individually.

I suggest that it might work best with a single vane of 180 degrees but there is no information on the ebay site about the zone size for detecting and release.

If it is spindle speed sensor you are looking for then look at the X2 sensors on the CNC threads. From memory there is one using an optical sensor and another using a hall effect chip. Yes there is a bit of electronics involved but not much.

You will need that 270 ohm pull up resistor and you should chack that the sensor when switched by a vane gives a clean zero which will be less than one volt and when switched off goes to the power supply voltage. I think you said you had a volt meter so you can do this test.

If this works try placing it in position and rotating the shaft by hand. If you get a clean on off for each tooth then ramp up under power. I would expect the sensor to stop well before you reach that 9,000 rpm due to the dead zone in the detector range.

Regards - Pat

[Al_The_Man]

You will need that 270 ohm pull up resistor

Surely this will depend on what the Gecko internal source pull up is, is it known?
Al.

[strohkirchw]

Hmm, looks like I'll need another sensor then. This is going on a Taig and with my pulley setup I'll be looking at just under 11,000 rpm. I also planned on epoxying something small to the pulley to pull the signal.

[strohkirchw]

I went ahead and grabbed the iMach spindle tach. The usb connection is going to save me a bunch of time now that I don't have to pull apart my enclosure, cut and mount the connectors for 12v/prox sensor.

[wildwestpat]

Search on ebay for 'magnetic hall proximity' and you will turn up a thread mounted hall effect detector. This would require a small permanent magnet to be fitted to the spindle gear. Please check with the seller what the cut off frequency is.

For a simple optical detector for use with a vane (tab stuck to the spindle) look at Optek Slotted Optical Switch - OPB917 as this is a photo diode and detector transistor in a fork type mount. You would require two external resistors - one to limit the current in the diode (LED) arm and the other to act as the a pull up resistor if you purchase the OC version. The non OC version has an internal resistor.

For a simple DIY hall effect look at Hoss X2 pulley conversion.

Also there are some cheap optical hand held rpm meters on ebay found by searching for 'Non contact Digital Laser Tachometer Wheel RPM Meter' The price is sub $10 has any one purchased one and if so how does it perform?

Hope this helps - Regards - Pat

[Crevice Reamer]

Hall Effect vane sensors come fully assembled, are very cheap, and plug directly into the G540. I used this one from Digikey:
Digi-Key - 480-1993-ND (Manufacturer - SR17C-J6)

A tab on a simple metal disk is all the trigger it needs. This thread covers the entire installation and use:
http://www.cnczone.com/forums/vertic...step-6.html#61

CR.

[wildwestpat]

Hi strohkirchw

When you have used your USB -iMach spindle tach by Vista CNC would you be so kind as to post your experiences in this thread as the device might be the easy answer for ArtSoft Mach motion controller users.

Regards - Pat

[strohkirchw]

Ya, no problem
Should have it delivered and installed by this weekend.

[strohkirchw]

Ok got the sensor and have it installed.

As of right now it seems to be working decently well. I was getting super erratic readings with only 1 coat of flat black paint on the spindle. It didn't cover the aluminum of the pulley very well so I put another one on. It seemed to help quite a bit. Now if it's above 1500rpm it'll read the correct spindle rpm for about 4-5 seconds then flash a number 500-1000rpm higher for 1/2 second then repeat the cycle. Below 1500rpm it's not that accurate. Exampling being it should be reading 400rpm but it's reading 400-800 up and down.

I'm going to put another coat or 2 of black paint on the spindle so it hopefully cleans up the accuracy even more. It's also super easy to install and took me all of 15min to have it setup and a bracket made with it running. It also has a neat setup in the plugin that will perform an action if the spindle is x rpm less than the require amount. Even with the erratic reading the spindle read out is never less than the actual amount only more.

[wildwestpat]

Thanks for posting your test results. Might be worth trying a bit of shiny metallic foil or metallised plastic as a replacement for the white patch on the spindle. It is the contrast that is being detected so any glinting of the surface will read as extra pulses per revolution. I have found in the past with optical reflection detectors that they have a sweet spot where the beam of light is just touching the shaft at the same point the detector is focused so it may be you will get better more consistent detection at a different distance between the shaft and the detector's front face.

How quickly does the reading settle on start up and shut down of the spindle. What I am interested in is does the indicated speed appear to follow the shafts rotation or is there a delay. The delay would be due to the sampling period, if any, used to smooth the pulse to pulse variations.

Many thanks - Regards - Pat

[strohkirchw]

Thanks for posting your test results. Might be worth trying a bit of shiny metallic foil or metallised plastic as a replacement for the white patch on the spindle. It is the contrast that is being detected so any glinting of the surface will read as extra pulses per revolution. I have found in the past with optical reflection detectors that they have a sweet spot where the beam of light is just touching the shaft at the same point the detector is focused so it may be you will get better more consistent detection at a different distance between the shaft and the detector's front face.

Hmmmm...should I just have left the spindle without the white paint since it's bare aluminum with a turned finished? Seems like it would catch the light in the room much better than the white paint.

How quickly does the reading settle on start up and shut down of the spindle. What I am interested in is does the indicated speed appear to follow the shafts rotation or is there a delay. The delay would be due to the sampling period, if any, used to smooth the pulse to pulse variations.

There is a delay that goes up if as you increase the smoothing. You can choose 0-10 smoothing and with smoothing around 3 to 4 it takes around 4 seconds after the motor starts to get up to the actual rpm. With the same settings it takes one second or less to go to 0 after the motor stops spinning.

[wildwestpat]

Thanks for the information I will now put this on the mods to do list.

Regards - Pat

[strohkirchw]

Added another coat of white and black paint and it's much better now. When spinning at 2500rpm it would flash an odd reading once every 30-40 seconds. When spinning around 600rpm it would read correct for 2-3 seconds then incorrect for 2-3 seconds and repeat.

I'm going to try adjusting distance like you said to help low speed.