After reading some the community’s experience with mechanical end-stops, I decided to design my own hall-effect based sensor solution. I’m not saying I went about it the right way and there are probably better/cheaper/faster solutions, but it’s what I came up with. I should say, I was inspired by Kevin Patterson’s hall effect solution (love the red/green feedback LEDs.) https://www.youtube.com/watch?v=vqU_5wZAX3s
NOTE: This design has not been implemented yet. My boards from OshPark don’t come in for a few more days.
- 24V separate power supply for the hall effect circuit.
- Nominally high (active low) design.
- Opto-isolated sensors
- Higher sensor voltage (24V) in combination with some simple pull-ups an caps to help ride out any coupled transients.
I’ve built some acrylic holders for the sensor boards and a small enclosure for the main board. The concept is similar to Kevin’s in that I will have 6 sensors (2 for each axis) connected with inexpensive phone cord connected by RJ11’s to the main board. I’ll mount my main board to the side of my control box where the grblshield/arduino lives.
The PCBs are available to purchase or download the gerber files (PCB build files needed by a board house to make printed circuit boards) from Oshpark.com.
Main Board: https://oshpark.com/shared_projects/X1PVc1kz
Sensor Board: https://oshpark.com/shared_projects/B7TXAfxt
As I build and test this concept I’ll try to post my results. Hope this helps some of you.
Schematics and PCB Layout software --> OrCad
Housing design: ViaCad + MeshCam
Digikey_Order_Shapeoko_homing_circuit.pdf (23.7 KB) shapeoko_homing_circuit_main_board.pdf (22.3 KB) shapeoko_homing_circuit_sensor_board.pdf (12.8 KB)
Awesome! Love those custom acrylic holders.
Don’t have my PCBs from OshPark yet, but I have the components so I prototyped the circuit and have the following comments for anyone attempting to use this design.
- The Hall Effect sensor I chose is latching. This means once it trips, you either have to apply an opposite magnetic field to release it (flip the magnet) or just cycle power on the circuit. For me, this is no problem because I have each sensor broken out into individual RJ11 connectors back at the main board so I can either remove/insert the connector for that sensor or just cycle the main 24V power of all the sensors. If you don’t like latching sensors, there are alternatives with the same SOT23 footprint that will work for you.
- THESE PARTS ARE VERY STATIC SENSITIVE! I do a ton of prototyping and soldering and rarely run into parts this sensitive. I killed the first 3 this way without realizing it. I thought initially that Digikey had sent me the wrong parts or I had something shorted somewhere that shouldn’t be, but on try # 4 (grounding myself the whole time for this one) I got it to work. Just so you know…
Nice work! Why did you choose the latching sensors?
Love the box too.
I chose the latching sensors for a very logical reason… Actually, I didn’t even pay attention when I selected them. I was looking for a SOT23 package, open collector, >=24V supply capability but missed the latching feature… Now that I think about it, for end stops, the latching should be fine, but it may cause a problem for homing. Probably have to get different sensors. (Looking now, the Allegro Microsystems A1120ELHLT-T may be an okay non-latching alternative.)
Thanks for asking. I was just going to go forward with the latching version but forgot about homing. Saved me some hassle.
A better part then the Allegro for the application is TI’s non-latching version of the same sensor --> DRV5023AJQDBZR. Again, same footprint and use so should work with the design as-is.
$0.956 in quants of 10 or more from Digikey (PN: 296-38452-1-ND)
This might save you the hassle.
Zach, I’m disappointed in you man. What fun is it in taking the easy way out? I’m not 100% sold on the whole moving mechanical switch and possible misalignment over time “thing.” And there’s no LED lights on mechanical switches. LEDs make everything more fun (and possibly dork-like, but definitely more fun.)
Ok you got me there everything is better with LEDs. Also tonight I broke the limit switch on the Z by running gcode but forgetting to do G92 X0 Y0 Z0. It literally broke the clip off. So I learned my lesson only a few hours after that reckless post!
Everything is taking forever… I love getting PCBs really cheap, but I hate waiting 12 days for my boards.
…The footprints for the RJ11 connectors were backwards on the main board. Had to order some more. Maybe I’ll have this thing done by June at this rate.
For anyone planning on building a copy of this design, here is the link to the shared update of the Main board on OshPark:
Here are the PCBs from OshPark. Round 2 of the Main Board is on its way…
Looking good @DavidBurke !
Did you design the acrylic case too?
Thanks. Yes, I designed the sensor holders and main-board housing with ViaCAD. I used to sell acrylic fish traps on Ebay many many moons ago and had a bunch of acrylic scraps left over that I’ve been playing with on my Shapeoko. I learned the lesson that others before me have regarding the difference between extruded acrylic (the cheap stuff from Home Depot) and cast. The extruded stuff just melts onto the bit while the cast acrylic can be nicely machined. I use a little bit of dish soap on the bit and that seems to make the edges a bit nicer.
So are you going to stay with the mechanical switches or do you think you may try magnetic at some point? Send some of that 70’s weather up to Detroit!
I’ll try to add updates of the project to this forum but anyone who wants to keep up to date on it can visit the following location. I figured it was time to add a blog my website.