Hall effect sensors for end stops. Any tips?

[Gr8Scott]

I ordered some Hall Effect sensors from RepRapDiscount. Got them in the mail today (ordered them on Tuesday). Any suggestions on how to get these babies working their best on the Rostock Max?

[626Pilot]

I'm using the Hall-O board from mauk.cc to implement a Z-probe, which you can see here. There are some things I include in the instructions that might make your life easier, such as how to adjust the trim pot without destroying it and what kind of pin connector to use. I don't know if RRD's Hall sensors have the same pinout, but if they do, you want to connect STOP to the third pin of x/y/z_max. (If I'm wrong you might wind up letting the magic smoke out, so ask someone at RRD how to wire these up to a RAMBo.)

There is a guy who clamped magnets to his drive belts and mounted the sensors at the bottom instead of the top. Pretty smart. If you do that, make sure you clamp the magnets to the side that isn't going to make them go around pulleys when the thing is running.

Was there anything that made you want to do this instead of the regular endstops?

[Gr8Scott]

I feel that I can get more precision and repeatability from the hall effect sensors rather than the mechanical switches. I can't seem to get a great deal of repeatability out of my printer for some reason. Seems like my home settings wander slightly with each print. I want to be able to print .1mm layer heights even on the first layer without crashing the head. Maybe that is foolish or unrealistic, but that's my goal.

I have a MIC-6 aluminum plate mounted over the Onyx bed. It is perfectly flat even at very high bed temps. I calibrate z-height with the effector close to each axis and in the center and they all match perfectly every time. Then I start the print and the head eventually crashes into the plate when the head gets between the middle of the bed and the z axis. After I stop the print, I check the calibration at each axis and they are perfect again.

I thought about putting a magnet on the outside of the belt and mounting the end stop at the base like what you mentioned, but I am unsure of how to reliably mount the magnet to the belt. It would definitely be the easiest mount for me to make, but I don't know how to keep the little round magnets from falling off the belt. How did that guy fasten the magnet to the belt?

[626Pilot]

That's not caused by the endstop switches, unless they are so loose that they can move around. Replacing them with Hall sensors won't fix it. Do you have a piece of glass on top of your aluminum surface and is the whole thing clamped together somehow? Without that it's going to move around too much from thermal expansion. You should also make sure you're doing the 4-point calibration when the bed is already up to temperature, and that your PRINTER_RADIUS is set up right.

[lordbinky]

I'm playing with printing .05mm layers and my leveling is staying consistent. I was down to .02mm layers on a .4mm nozzle but the trace width was ~.6mm so anything that came to a point vertically came out crappy once it was smaller than .6mm.

Here are some things I did that I think helped get me there.

I got a 6" master precision square from Mcmaster-carr ($14 + S&H). Used this to get my tower alignment better, you have to check 3 out of 4 of the side for each tower.

I gave up on my magnetic arms and went to the tricklaser arms. I only have one printer so it being down while I work on improved arms was not fun. Completely my fault it was an issue to start with. If you have the old-style stock arms be OCD while you check for delta arm blues one more time.

I moved the microswitches so the end stop screw no longer hits the round part of the lever but the flat part. This did increase repeatability.

I cut out a part of the snowflake so the Onyx wiring has room. This increased my level consistency between bed warm ups.

I made sure the cheapskates were set per the sticky thread with gene's video demonstration.

hope some of this helps.

[Eaglezsoar]

I'm playing with printing .05mm layers and my leveling is staying consistent. I was down to .02mm layers on a .4mm nozzle but the trace width was ~.6mm so anything that came to a point vertically came out crappy once it was smaller than .6mm.

Here are some things I did that I think helped get me there.

I got a 6" master precision square from Mcmaster-carr ($14 + S&H). Used this to get my tower alignment better, you have to check 3 out of 4 of the side for each tower.

I gave up on my magnetic arms and went to the tricklaser arms. I only have one printer so it being down while I work on improved arms was not fun. Completely my fault it was an issue to start with. If you have the old-style stock arms be OCD while you check for delta arm blues one more time.

I moved the microswitches so the end stop screw no longer hits the round part of the lever but the flat part. This did increase repeatability.

I cut out a part of the snowflake so the Onyx wiring has room. This increased my level consistency between bed warm ups.

I made sure the cheapskates were set per the sticky thread with gene's video demonstration.

hope some of this helps.

Were you having problems with the magnetic arms?

[lordbinky]

I think my arms theirself were actually done well and my problem was in the platform or skate connections. I spent a correspondingly less amount of time on those parts. I was just getting tired of the time I was putting into it causing the printer to be down or in unavailable to do other things as well as throw off my bed leveling.

[Gr8Scott]

I have a thin layer of aluminum tape as a slight heat barrier to the Onyx, which is mounted directly to the base without the snowflake. The 5/16" thick MIC-6 aluminum plate is on top of the Onyx and the screws press the MIC-6 aluminum down onto the Onyx and the base.

Maybe it is my radius (199.5) or arm length Delta Diagonal Rod (269.0) I still have the stock arms from April on the machine, though that may change one day soon. I think I do a 4 point zero each time. I measure in the center and directly in front of each extrusion.

[626Pilot]

If you bring your aluminum plate up to operating temperature and you put a level on it, and spin that level 360 degrees to check flatness all around, does it stay flat the whole time? I have a feeling the center might be warping because the Onyx is hotter in the middle than at the periphery. Part of the reason to use glass is that it will expand, but not warp.

And, yes, check your radius and delta diagonal arm length. Radius first. With stock arms I wouldn't touch the diagonal arm length unless you have swapped the arms out for something else.

[lordbinky]

Changing DELTA_DIAGONAL_ROD is an easy way to fine tune your X/Y scale. Only do this after you're satisfied with your DELTA_RADIUS though, X/Y scaling depends on both those settings. It definitely shouldn't change much though.

[Gr8Scott]

If you bring your aluminum plate up to operating temperature and you put a level on it, and spin that level 360 degrees to check flatness all around, does it stay flat the whole time? I have a feeling the center might be warping because the Onyx is hotter in the middle than at the periphery. Part of the reason to use glass is that it will expand, but not warp.

And, yes, check your radius and delta diagonal arm length. Radius first. With stock arms I wouldn't touch the diagonal arm length unless you have swapped the arms out for something else.

I'll check the plate at operating temp to make sure it's flat. I think it should remain flat at operating temperature because it's cast and 5/16" thick. That is an assumption on my part and I really need to check it out though, so thanks for the suggestion. I'll check it within the next week or so.

[Gr8Scott]

Changing DELTA_DIAGONAL_ROD is an easy way to fine tune your X/Y scale. Only do this after you're satisfied with your DELTA_RADIUS though, X/Y scaling depends on both those settings. It definitely shouldn't change much though.

As soon as I get through upgrading the hardware (I'll hold off on the endstops for now), I'll give this a shot. Sounds like I might need to keep those hall effect sensors handy for an auto-level/tune feature later on.

[626Pilot]

Even if it's flat, you should still use glass. You are going to scratch the hell out of it if you don't.

[Gr8Scott]

Even if it's flat, you should still use glass. You are going to scratch the hell out of it if you don't.

I am OK with light scratching. Deep gouges from crashing is an entirely different matter LOL. I have another plate as a backup if this one gets too gnarly.

[626Pilot]

Yeah, but glass isn't going to scratch. Between head crashes and the need to pry pieces off the print surface from time to time, the best solution is to just solve that ahead of time. You can get a borosilicate glass plate from SeeMe or have one cut in town from oven glass. Bonus: Your pieces won't have a negative copy of a bunch of scratches!

[Flateric]

Hey 626Pilot, I am unclear on the exact way that the hall-o sensor is assembled. I can't quite picture it. I have the hall-o and I have you stl printed out but can't get the assembly of the magnet and rod portion right in my head.

Could you show a close pic of yours, or perhaps a diagram of some kind?

Many thanks.

[bvandiepenbos]

I use thumb screws so endstop switch lever has a nice flat area to contact.
these http://www.mcmaster.com/#92421a151/=sa1xxr

[626Pilot]

Hey 626Pilot, I am unclear on the exact way that the hall-o sensor is assembled. I can't quite picture it. I have the hall-o and I have you stl printed out but can't get the assembly of the magnet and rod portion right in my head.

http://www.thingiverse.com/thing:161753/#instructions

If you follow the instructions, it should go together fine. If anything is less than clear, let me know.