What switches will you be using. Unless the are proximity sensors that require a 3rd wire, wire the ground from the switches in the place where the com is in my drawing. All mechanical switches do is connect or disconnect  2 parts of a circuit. 

Those: https://www.aliexpress.com/item/32951447527.html?spm=a2g0s.9042311.0.0.de1f4c4degWLQ6

I updated my diagram to reflect your comments, including choosing a 4 Channel 12v-3.3v to reuse the board's power supply (12V 5A 60W). Or should I keep it separate?

Hopefully, this is better now --> GRBL_ESP32 for Beginners

Hi. Just came across this project and I am planning to build. I have read thought this thread and have a question. Has anyone looked at using Brian Dring's "6 pack CNC Controller" here https://github.com/bdring/Grbl_Esp32

Brian recommends using the Trinamic TCP5160 drivers.  Can these be used instead of the external drivers for the NEMA-23 steppers?

@turlette if you are using Normally open limit switches your looking good. If you are going to use Normally closed, the y limit switches need to be wired in series, you have them in parallel. FYI those are the same limit switches I'm using. The 3 terminals you will see are NC, NO, COM. 

@sslupsky, "Bart Drings" 6 axis is great if you have a need for the extra axis outputs. If you are building Topsies cnc it will be overkill. I am running one of Barts 4 axis SPI Grbl esp32 boards, with Trinamic 5160 drivers for 2.4 amps, they definitely get warm. I have a 120mm fan cooling the contents of my control box. It all depends on the the amount of current you want to push through your steppers. I think you would be fine with the TMC 5160 drivers on the NEMA 23 steppers. 

The heatsink adhesive gets soft and loose when they get warm. I'll be using thermal epoxy to make sure those heatsinks don't loosen up

Thank you for the feedback Peter.  I will plan to use heat sink compound / epoxy on the heatsinks.

I understand the TMC5160 step sticks are rated for 3A rms and 4.2A peak.  Actually, the FET's are rated to 20A but because of the small size of the step sticks they are thermally limited to 3A/4.2A.  The NEMA 23 steppers referenced for the TopsCNC build are from Steppers Online which indicates the "current per phase = 4.2A".  Do you know if the stepper current rating is an rms or a peak rating?

I went through and reread this entire thread and saw your message about using the DM542T.  Is there a benefit to using the DM542T versus the Trinamic drivers?  It appears to me the Trinamic drivers get a lot of positive reviews online and they are very quiet.  Looking at the DM542T they appear to be in an enormous enclosure by comparison.  So, I am not quite sure how to compare these.

One more question regarding the discussion of the opto isolators and limit switches.  The Trinamic drivers have a "StallGuard" feature.  Is this intended to take the place of the limit switches?  I have a Prusa i3 MK3 and it appears to use StallGuard instead of limit switches.  If using the Trinamic drivers, is there any reason to use the limit switches?

Peter, thank you for the link to the video about the whip.  That explained a lot.  One of the comments indicates the whip is due to the sag of the lead screw under its own weight and suggested a larger diameter is really what is required.  To you know how to calculate what the correct diameter should be?  I have seen other comments in the forum referring to T10 and T12 lead screw diameters and I think some of the parts have been modified for the larger diameter.  Your thoughts?

@sslupsky, regarding the T8 screws: I went through the same thought process than you. I even started redesigning to accept 12 before I saw that post on the forum. Once I knew what I needed, I could calculate the cost: Roughly 80$ additional for a 1500x1500mm.

The "problem" is: The ultimate choice would be Ball screws, not lead screws, which for my config would roughly be 100$ more than the T12. Therefore, I decided to test it out with the T8, apply the whip fix described on other pages of this forum  (apply tension), and save up for ball screws.

SSlupsky I would probably opt for the external drivers for the NEMA 23's, It will give you more headroom. The higher ratings on the TMC5160 are also with active cooling. As for the audible volume of the drivers, yes the 5160's are virtually silent, but in my workspace, the router, the shop vac and ambient is wa worse than listening to a slight driver whine. ( Different than my 3d Printer which is "Near Silent" TMC 2209 Drivers and silent fans all around.) It needs to be it sits on my desk. If I sat at my desk with the printer going after a few hours my ears would be ringing.

I do not know the proper calculation for determining screw size, for me cost was the determining factor. I mean  any lead, ball screw, etc will sag somewhat it also depends on the speeds at which you need the axis to move. What I I have found, is that the T8_8 under tension @1500mm length can run at 6000mm/m. That far exceeds the cutting speeds that the Z carraige can handle in its current state. 

In my quest to eliminate the whip, I also found a video of a 1" ballscrew on an industrial machine that needed to be supported half way so that I doesn't sag and induce whip.  The conclusion is what is the end goal. I would say anything longer that the 1500mm I would increase screw size "or" if you are going NEMA 23 the pats and plans are already designed for larger screws.

I probably would have gone NEMA 23 but I had already printed most of the large parts that I would have had to reprint, I also had my t8 screws on order as well as all of the steppers. Its only after I had everything ordered did I see the NEMA 23 Version.

Regarding the Opto Isolators, yes using the Stallguard Function would eliminate the need for the Limit Switches. With so many other variables, to get worked out, I am currently running on physical limit switches. There is little support for the tuning of the stallguard via SPI , when I get the rest of the machine swuared away, I may revisit the idea of testing and tuning the stallguard.