An Analog Electronic Edge Finder

**rgsparber** · Jan 14, 2022, 10:45 AM

This Electronic Edge Finder detects touchdown using either a cutter or a rod mounted in the spindle of a mill or the chuck of a lathe. Unlike my last design, this one is all analog, no Arduino.

It is a simple circuit using seven resistors, two capacitors, one dual opamp, a diode, two LEDs, and two pushbuttons. Power is from a USB wall wart or battery pack. If you replace one of the LEDs with an opto, the circuit is compatible with CNC systems.

Click here if you are interested.

Your comments are welcome. All of us are smarter than any one of us.

Thanks,

Rick

**DIYer** · Jan 15, 2022, 01:56 AM

Thanks rgsparber! We've added your Edge Finder to our Measuring and Marking category,
as well as to your builder page: rgsparber's Homemade Tools. Your receipt:

Edge Finder by rgsparber

tags: LED, edge finder

**Claudio HG** · Jan 15, 2022, 09:15 AM

Elaborated circuit, the schematic to me is a tad confusing because the touchdown contact is drawed as if it would be a pushbutton so a something like an arrow and a dash would have been clearer in my opinion. Also, if I got it right, the R part appears to be connected at the node +Vx +Ix where actually it is at the node -Vx -Ix: this node is attached to the part and then the part get in contact with the tool+spindle where the +Vx+Ix lais, right?
Apart this minor point I see an other mechanical aspect: wasn't better to have some kind of a spring as a probe to prevent the risk of hitting the part with the tough metal of the tool installed on the spindle (as I've seen in the video)?
Kudos for your work, anyway. Thanks for sharing.
Claudio.

**rgsparber** · Jan 15, 2022, 01:53 PM

Claudio HG,

Your comments are immensely helpful to me. I have updated the article including adding you to the acknowledgment section.

Please let me know if I missed anything or created any new confusion.

A 1/8" diameter 1" long hardened steel dowel pin can flex a few thousandths of an inch yet spring back. This is more than adequate. During the touchdown process, I start about 0.01 inches from the reference surface and advance the probe 0.001 inch at a time for a rough touchdown. Then I back out about 0.002 inches and advance again 0.0001 inches at a time for the final touchdown.

Thanks for your help.

Rick

**Claudio HG** · Jan 17, 2022, 02:42 PM

Rick, you are too kind, my contribution was almost nothing and not deserving any citation.
About the flexibility of the dowel pin you are right, I was fooled seeing an end mill bit. Back to the circuit, I think the test button would fit its function better if it would be replaced with a switch (a toggle switch) so one of your hands is not busy to keep the button pressed while the other hand has to perform the test. Just use a double pole switch connecting the second pole to lit a further LED (maybe bright yellow?) to signal the test mode status.
Another point is the drawing where you represent the resistances. I refer to the second picture at page 3.
It is my understanding that the actual circuit is as in the following drawing. If I didn't mistake your circuit the representation I provided would be more accurate, even though somewhat more confusing for those who are not familiar with electronics/EE, but what you call R spindle definitely should be the equivalent R of the machine's body, spindle included. Did I mistake something?

Claudio.

(P.s. If you still want to put my name on the acknowledgments, please do so using the following: "Claudio H.G. of Accidental Science", thank you.)

Greetings from the Alps, Rick.

An Analog Electronic Edge Finder-rick.png

**nova_robotics** · Jan 17, 2022, 11:11 PM

Good stuff. Have you given any thoughts of powering it with a little isolated power supply and using an optoisolated output?

I was going to make a touch sensor circuit for my mill, but it's very dry in my shop and I get static shocks off of everything. I have to ground myself on metal surface probably every 15 seconds or so. It's just second nature to brush my hand against every metal tool as I'm walking around. I'm worried about zapping the GPIO pin on my DRO. So... I figured completely electrically isolating the input and output side of the touch circuit might be a good thing. You probably wouldn't even have to modify your circuit.

**rgsparber** · Jan 18, 2022, 07:27 AM

Originally Posted by Claudio HG

Rick, you are too kind, my contribution was almost nothing and not deserving any citation.
About the flexibility of the dowel pin you are right, I was fooled seeing an end mill bit. Back to the circuit, I think the test button would fit its function better if it would be replaced with a switch (a toggle switch) so one of your hands is not busy to keep the button pressed while the other hand has to perform the test. Just use a double pole switch connecting the second pole to lit a further LED (maybe bright yellow?) to signal the test mode status.
Another point is the drawing where you represent the resistances. I refer to the second picture at page 3.
It is my understanding that the actual circuit is as in the following drawing. If I didn't mistake your circuit the representation I provided would be more accurate, even though somewhat more confusing for those who are not familiar with electronics/EE, but what you call R spindle definitely should be the equivalent R of the machine's body, spindle included. Did I mistake something?

Claudio.

(P.s. If you still want to put my name on the acknowledgments, please do so using the following: "Claudio H.G. of Accidental Science", thank you.)

Greetings from the Alps, Rick.

Click image for larger version.

Name: Rick.png
Views: 156
Size: 62.7 KB
ID: 41567

Claudio,

You continue to provide extremely valuable insights.

Pushbutton versus toggle - I failed to explain my design logic here. I didn't want the user to forget and leave the circuit in test mode so went with a pushbutton. It is easy to hold down the button with one hand and use the other hand to touch a bar to the gap between cutter and machine body.

As you suggest, I have changed the schematic in the body of the article to not show Rspindle and Rpart. It was easier for me to use text rather than drop in the diagram but hopefully, it is clear enough. I do show these resistances in the detailed circuit description. Rspindle is the resistance measured between the spindle and the machine body. This is due to the electrical resistance of the two spindle bearings.

I have updated the article.

Thanks!

Rick

**rgsparber** · Jan 18, 2022, 07:32 AM

Originally Posted by nova_robotics

Good stuff. Have you given any thoughts of powering it with a little isolated power supply and using an optoisolated output?

I was going to make a touch sensor circuit for my mill, but it's very dry in my shop and I get static shocks off of everything. I have to ground myself on metal surface probably every 15 seconds or so. It's just second nature to brush my hand against every metal tool as I'm walking around. I'm worried about zapping the GPIO pin on my DRO. So... I figured completely electrically isolating the input and output side of the touch circuit might be a good thing. You probably wouldn't even have to modify your circuit.

You are right on both counts. Isolated power is necessary because we do not want any stray current entering the EEF. If connecting to a separate circuit, an opto is the way to go. You can see the isolated power supply on page 10 and mention of the opto on page 9.

I live in the Sonoran Desert so am no stranger to dry conditions. When I first moved here and set up my DRO, I blew it up with the first static discharge.

Rick

**nova_robotics** · Jan 18, 2022, 05:00 PM

Originally Posted by rgsparber

You are right on both counts. Isolated power is necessary because we do not want any stray current entering the EEF. If connecting to a separate circuit, an opto is the way to go. You can see the isolated power supply on page 10 and mention of the opto on page 9.

Whoops. Sorry about that. I read the first half dozen then scrolled down to the schematic.

**neilg4dbn** · Jan 22, 2022, 01:18 PM

I wonder if this fixes the usual problem with electronic edge finders: fine chips adhering to the part being measured (or the finder itself) and giving false indications. I haven't tried measuring the resistance of the dust or fine slivers, but it should certainly be considerably greater than the resistance through the machine column and vice, so this approach might be a perfect way to avoid those false readings. In fact if the slivers are very small, it might even fail to detect a contact where a sliver is trapped between part and tool. That would give you a positive indication that the part or tool need cleaning. Interesting project!