Is cutting the arm off the Z-axis switch a good idea?


We all know (or should!) that the repeatability of common spring-arm limit switches is crappy at best. I'm looking to build & install one of the precision height adjusters for the Z-axis limit switch, and noticed a post on some forum suggesting removing the arm and triggering the switch button directly (e.g. with a screw end).

Has anyone tried this, and if so has the repeatability of Z-homing improved any?


Sorry -- this is a stock Prusa i3, which depends on physical contact between the vertically-moving subassembly and a microswitch mounted on the frame.

Carl Witthoft

Posted 2016-11-08T19:22:30.303

Reputation: 2 918

Related question here

– Sean Houlihane – 2016-11-24T17:29:45.103



While I haven't seen this on a Z axis for a printer personally, there is no reason it wouldn't work, and would improve your repeatability in theory. Removing the arm on the switch is taking away the lever. Going back to simple machine mechanics, the lever gives you a larger range of motion in which the button could be triggered, with the tradeoff that you get a larger target to hit. How much of an improvement depends on the exact switch, where the button is compared to the axis of the switch, and how long the switch is.

My Shapeoko 3 CNC router has a switch for all three end stops that do not have arms, and my Original Prusa i3 mk2 has switches without lever arms on X and Y axis endstops, so there is is no reason it won't work for your i3 Z axis, you just need to make sure you can accurately hit the small button on the switch.


Posted 2016-11-08T19:22:30.303

Reputation: 417

Exactly. You remove the lever and the lever potentially bending/wearing on the plastic hinge over time. – Fred Hamilton – 2016-11-12T01:23:44.547

My only thoughts are to the delay between when the switch is pressed and the printer stopping. Also the risk of the mechanics getting out of place and the head crashing. – StarWind0 – 2016-11-13T05:09:03.420

1Those are both separate issues outside the scope of the question. The question did ask about the mechanics of securing the items around the switch, just if it can be done, would it help improve the printer accuracy. – Jexoteric – 2016-11-15T00:23:07.333


No. Buy a better switch if it's an issue (see below).

You would need to have some very tight tolerances to hit that micro button with whatever your arm is. If you had a machine with good tolerances you would not be considering this modification. That alone is why I would say this is not the greatest idea.

Following it might work if your Z is connected to the hot end and smashing into the bed. But I suspect you will still have a myriad of issues, such as the switch getting out of position enough to cause the head to crash into the machine. The real question now is how many rotations of the Z axis could happen if the printer is moving at maximum speed and the button is pressed? That metal arm is your grace period. Now your printer is potentially smashing into the switch.

Lastly, just get a switch with a more solid and less springy metal tab.

The real question is whether there is actually a variance caused by the metal arm? I would suspect that it hits the switch very precisely, consistently and within an acceptable tolerance. Removing the arm will buy you little. Replacing it with a stiffer-arm switch might serve you better.


Posted 2016-11-08T19:22:30.303

Reputation: 2 913

I'm not convinced. Take a look at some of the advanced Z-stop designs at Thingiverse. A solidly-mounted lead screw is not going to move in the X-Y plane and thus will always be lined up with the button. Similarly, these designs lock down the microswitch position far better than the stock Prusa mount. – Carl Witthoft – 2016-11-10T12:13:33.990

Well, you asked. Sorry I didn't say what you wanted to hear. That said go for it and write a blog post for the community :-) You do realize I have no idea what your setup looks like? If a Mendel max went with this it would be a disaster. – StarWind0 – 2016-11-10T20:06:13.263

I tend to be biased towards quantitative answers, and yours, while a fair opinion, didn't have any tolerance data or personal experience. Sorry if I sounded dismissive. – Carl Witthoft – 2016-11-11T12:19:03.643

1@StarWind Your answer is a very resounding "no", while I'm sure it isn't quite that black and white. Are there maybe some situations in which removing the arm does make sense? As it stands, your answer reads like a rant on OP's printer. – Tom van der Zanden – 2016-11-11T12:21:43.747

I would note that you explicitly asked for an opinion answer when you have a title asking "Is x a good idea?" That does not lend itself to a quantitive answer. – Jexoteric – 2016-11-12T00:22:24.907

@TomvanderZanden It is hard when there is no information of his printer. I stand by my answer and question your helpfulness in your reply.. It seems Jexoteric is more on the ball. – StarWind0 – 2016-11-13T04:51:03.243

1@CarlWitthoft I do want to address, if you want "quantitative answers" then you have to provide better questions. "Is this a good idea" is not a question that is formatted for this type of answer. Now Something specific such as "What are the down sides of removing a level" or better "What are ways I can increase precision in my Z axis" will yield what you were really wanting. – StarWind0 – 2016-11-13T04:59:15.650

@TomvanderZanden With the information provided I do feel No is a valid answer. Following you are correct I do not address the pros in this case. I feel the risks outweigh the benefits. I do not believe this is a rant but I will keep it in mind and simply flag vague questions in the future. – StarWind0 – 2016-11-13T05:06:22.173


The answer to this question is, no, taking the lever off will not do anything improve repeatability, but it will improve accuracy of the Z Axis. I have tried both ways, actually 4 with a Duet Wifi. 2x 8mm Leadscrews, 0.9 Stepper, Calibrated at 800 Steps per MM.

  1. Full lever size: After actuation I have to baby step up my Z Axis 13 x 0.05mm to let go, and then 13 of the same baby steps back down to actuate it.

  2. Short lever: After actuation I have to baby step 0.05mm 10 times, 10 more times to let go.

  3. Small enstop button only: 3 x 0.05mm baby steps is all that is needed.

  4. Larer button from roller lever endstops: 10 x 0.05 baby steps needed both ways.

The large endstop button only surprised me, thought it would be more reliable.

It is a bit more difficult to make the small button approach work, and it will increase the likelihood for failure and machine damage though.

Aaron Stoker

Posted 2016-11-08T19:22:30.303

Reputation: 11