No physical endstops with Archim 2?

I am venturing from a 3D printer to a MPCNC machine to work with my woodworking hobby.

While doing my research (I believe) I’ve decided on the Archim board but, as Google will lead one down the path of too much information, I ran across this article concerning the drivers (Trinamic TMC2130) on the Archim 2 board at toms3d which has a tmc2130-guide:

This article states that with these drivers, the need for physical endstops on X and Y is moot.

Quoting -

“the 2130 adds a communication interface that hooks up to your controller. This allows the stepper driver to report things like whether it thinks the motor its driving is about to lose a step and it lets the controller set motor current and many details about how it drives the motor. And this means a few things:
Your printer’s firmware can auto-tune how much current your motors need to run as cool and quietly as possible without losing steps, so no more fiddling with those tiny potentiometers.
It means you don’t need endstops for the X and Y axes as you can just ram an axis into the end and the driver will sense that resistance once it touches the end, and that totally works.”

I did read the archim-or-Rambo thread which discussed board options.

With the discontinued sale of the V2 board here, not sure how I want to proceed.

I found the Archim 2 on Amazon but the funny thing is, they sell it with the physical endstops for ~$170

Any advice would be greatly appreciated,

–Loose

1 Like

Welcome @Loose

One has to make the distinction between endstops for keeping the axes from crashing into things and endstops that are for an MPCNC which are only for autosquaring and don’t function as normal endstops.

The trinamics do work great for keeping the crashes from breaking stuff. I have the Archim 2.0. But the auto square function is helpful for using the physical endstops. I like my Archim. one of the reasons I went with them was for sound abatement and crash minimizing, in addition to having the wiring for the drivers very easy to plug into and not worry about configuring voltage, etc.

1 Like

Appreciate the welcome @scrounge79

Your answer just about seals it for me. Just so that I have this clear though, the endstop ability of the Trinamic drivers, while prevent breakage due to over travel, can not be used for the auto-squaring function; is that a functional limitation or a firmware/programming limitation? <— Hopefully not to dumb a question as my only experience with this is manipulating gcode for the last few months and a version of Arduino as I setup my aquarium controller (Reef Angel).

1 Like

I believe Prusa uses these drivers to avoid using physical end stops, but then their firmware enables this. Not at all knowledgeable about how this works. This might get us both a little more knowledgeable about them.

https://www.reddit.com/r/3Dprinting/comments/753veo/how_x_y_axes_know_the_limit_without_endstops/

1 Like

After I read through the first article I thought, ‘I’m not sure I want the motors to skip steps in order to square it,’ but then in the reddit thread you linked, screwyluie states “… having the motors skip steps doesn’t hurt them.” I did learn something there as I’ve had my glass plate on my Ender 3 shift and stop the Y travel by running into the verticals (until the bed clips gave up and became mini-missiles :slight_smile:)

Ordered my Archim v2 yesterday and continued printing MPCNC-C parts. I have a lot of time to study up as I finish printing; for some reason, my parts are taking ~2x longer to print than the estimates using CURA 4.4. That coupled with my pandemic work schedule, I expect I’ll have a bit of experimenting time too (that reminds me, I need to order that emergency stop switch).

1 Like

Gyroid fill, by chance? All the direction changes mean a lot of accelerations, which cura doesn’t handle well. I had that issue, anyway. It’s much closer with straight - line infill like cubic or something.

@turbomacncheese I spent about 20 minutes trying ALL of the infill options in CURA 4.4 and then tried again in 4.5 and each one sliced at more than double the estimated print times. For example the XYZ_C_Burly was estimated to be 5 hours and ~71.5g, mine sliced at 10 hours 58 minutes and 90g. I think I settled on either zig-zag or grid, bit the bullet, and hit print; took 11 hours, 2 min to print (each). I’ve tweaked the profile multiple times, to include using the CURA defaults and using ones from Filament Friday:

The parts are printing nicely so, for now, it’s not a big deal; just an investment in time.

I’m fairly positive it’s my lack of profile adjustment knowledge that’s slowing it down. Just when the stock profiles are do the same thing, I don’t know what to adjust. :thinking:

Ohhhhh, compared to the V1 estimate? That could be speed, nozzle size, layer height, even. I thought you meant actual print time compared to Cura estimated print time. My bad.

1 Like

:grin: Thanks for that list, I now suspect nozzle size since that’s about the only thing I didn’t change; I’m using a 0.4mm. I just might (might) have to change that for the Rollers since there are four total with an estimated 5:35 each. I figure that’d be ~44 hours of print time for my setup. From the pics it looks like I ‘could’ get four on the build plate at once but will probably do singles or 2 and 2 max; 26 hours is my longest print so far.

Pretty sure those estimates are for a .4mm nozzle as well.

1 Like

Well, then I guess I’m back to:

Forum etiquette question, if I want to ask for assistance with my Cura profile (or could it be my Octoprint profile/setup), would I continue here or start a new thread?

Going to read through his thread as it seems to have a similar situation:

I think Ryan uses something like .32mm layer height. Cura probably defaults to .2mm. I think the rule of thumb is that max layer height can equal about 80% of nozzle diameter. Changing that could make a drastic difference in print time.

We pretty much don’t care, as long as the thread is yours. It might help to make a separate post, but no one complains if it is your post.

2 Likes

Just to finish this time estimate portion off, I did a few more tests and it did come down to layer height. While some of the larger parts were longer than the V1 estimate, using my profile and the .32mm layer height, the smaller ones were spot on if not less than the estimate. Happy to say that I’m not as messed up with this as I thought. :sweat_smile:

2 Likes