Precision of end-stops?

I note that the normal Reprap machines have end-stops to detect the extreme positions of the carriages.

Am I correct in thinking that before starting a printing session the carriages are moved to the "zero" end-stops so that the firmware has a starting point for its position calculations?

I also note that the end-stops frequently use standard micro-switches. I have no experience of using micro-switches for that purpose and I wonder what degree of precision they give? For example will they routinely identify the position of the carriage to within a single motor step? If not, what level of precision do they give, and what level of precision is needed?

Thanks

...R

I can't answer the precision they give, but I can answer the precision required. The endstops are generally only used at the start of a print, so that the printer knows where the zero point is. As long as the stepper motors don't skip steps, the position will always be known from then on.

For X and Y, precision endstops are not needed. As long as you are "close" (within a few mm), that's good enough. The only reason to have better precision there is if a large print that takes up most of the printable area is desired, so that the print is kept within the printable area.

For Z, much higher precision is needed. If you aren't printing lower than 0.2mm layer heights, then 0.05mm precision from the Z endstop is good enough. For printing with lower initial layer height, better precision would be required.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

I've struggled with my Z end stop for a while, and then finally have given up...

The micro switch I used for Z end stop was initially "tied" to the vertical frame (I have Prusa i3) using a zip-tie. I later bolted it firmly to the frame using a couple of small screws and nuts. However, it didn't help much.

I used another screw installed on the X idler, which is what pushes the switch - I can easily and precisely how much it sticks out by screwing/unscrewing it. I had to readjust it every other day. I am not sure why. I thought maybe the vibration caused the screw to move a little bit, so I used a nylon nut, but it didn't help either.

Long story short, since I had to readjust the end stop, I just decided to not to use it. I removed the G code which "homes" X, Y, and Z from my slicer software, and manually register Z zero (G92 Z0) after adjusting the hot end position manually.

It's basically the same thing I had to do when I had to re-adjust the Z end stop. Now that I don't use Z end stop, I don't bother with Z end stop screws, and it's slightly less work.

I think it's more the implementation of the endstop than it is with the precision of the switch. Using a snap action hinge lever is going to be constrained by the quality of the lever as the first point of contact. I've got some Omron DF2-L-A switches in a prototype for some parts I'm working on. They're quite nice compared to the off brand copies from Shenzhen but are about 10 times the price(20 plus times at retail in the US).

I've had Z endstop issues similar to Yamster. At one point it was so bad the stop didn't trigger and from the other room I heard the grinding noise of the X axis digging into the bed and bending the Z lead screws. I think if a good switch is used, with a proper mount many of the issue will be mitigated. The stock i3 Z switch mounting parts are minimal and not best of breed for that application and there are more than a few that have other solutions and some that are yet working on others.

I've also got some prototype boards coming in to see how well a Hall effect switch works and if its more repeatable than a mechanical swtich with the issue of the mounting taken into account. I see some of my issues are due to the X axis physically landing on the switch. I've taken now to homing then raising the head sothe axis isn't resting on the end stop while the hot end is heating.though my goal is to see, at least on the Z switch if the Hall works better.

Edited 1 time(s). Last edit at 09/11/2013 10:45PM by vegasloki.

make sure that if using a lever switch on the z axis that the spring lever is removed.

also make sure that anything holding the z axis switch does not bend, so make sure it is thick and large. also lower firmware home rate. make it slow.

jamesdanielv Wrote:
-------------------------------------------------------
> make sure that if using a lever switch on the z
> axis that the spring lever is removed.
>
> also make sure that anything holding the z axis
> switch does not bend, so make sure it is thick and
> large. also lower firmware home rate. make it
> slow.

Removing the lever is a good idea. Hadn't thought of that. Takes the variable of the lever out of the equation. The bending of the mounts I see as more of an issue with some of the designs that attach directly to the Z smooth rod.

Thanks all for the information.

I wasn't actually thinking of the Z axis when I wrote my question but the information about it is obviously essential. I reckon I will start with manual positioning for the Z axis zero point.

...R

i have omron switches from ebay, but not micro, little larger, precision is better then 0.005mm(5 microns.)

I would appreciate it if you could post a link to the exact switch you are using.

...R

woo Wrote:
-------------------------------------------------------
> i have omron switches from ebay, but not micro,
> little larger, precision is better then 0.005mm(5
> microns.)

If precision is needed then a good and cheap solution would be an optical sensor similar to this one:
[www.ebay.co.uk]

If you have a scrapped inkjet printer lying around you will almost certainly find at least one of them inside it for zeroing the print head carriage. Precision is usually very good, typically a few microns. It will however need at least a simple electronic circuit to make it compatible with the end stop input of the 3d printer. This can be as simple as three resistors and a transistor.

@lister6520, thanks for that. I had no idea what their precision was. They are much cheaper at Farnell.

...R

... for more precise calibrating/referencing than needed with RepRaps I'm sometimes using inductive sensors from Pepperl&Fuchs which have a hysteresis of around 50 microns but are precise to 5 microns (axial types) to 2 microns (planar types) ... and for better resolution Omron-sensors too with analogue drivers, that are precise to 0.6 microns (600 nanometers )

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I think you are making mountains of mole hills. The precision of a micro snap switch is more than precise enough for what we are using them for. Heck i have a shop full of CNC machines that use micro switches to home the axes, if its precise enough for a $120,000 CNC its precise enough for a $500 reprap. If you are having problems with homing precise evey time you probably need to slow down the homing feedrate or you have a mechanical issue. Unless something mechanical has changed there is no reason for the machine to not home out almost exactly in the same position every time.

I'd agree GITRDUN but your six figure CNCs probably aren't using 13 cent switches from Huaqiangbei. Looking at the cheap switches I first used there is a mechanical difference and the lever is big part of that difference as is the actuator on the switch. A real Omron is a different story, very solid and consistent. What I'm doing (and I'm certainly not the first) is eliminating the need for physical contact though in the long run it's going to be much less expensive to design a more robust mount (or use one from Thingiverse) for a few cents in material than it is to build an endstop that may cost 7-8 bucks in parts.

I hope there are some affordable solution using laser measuring; let the printer figure out where the heated bed is by itself.

There could be other applications than just homing in with Z axis with this laser measure capability.

For example, we can print on uneven surface with this, since the printer will be able to figure out how much to lower/raise the extruder to add the next layer (of course, the slicing/printing software should become much more sophisticated and Z movement may have to become faster).

We would be able to print some part with one color, then change to a different color filament and continue printing, and on and on, all with a single extruder!

Also, if something goes wrong just before completing the print after hours of printing (which happens all the time!), we can fix the problem and continue on from where we left off as long as the incomplete print out isn't terribly messed up.

Well, even if it's somewhat messed up, as long as it's the top portion instead of the bottom or middle, we can cut off the messed up top portion and let the printer reprint the cut out section and the remaining part.

Sticking a new layer on top of cold plastic may present some challenge, especially with ABS, but this will be so cool.

Actually, aren't there some 3D printers with 3D scanning capability? Do they support this kind of operations?

I also think there is too much being made out of the absolute accuracy deal.

As mentioned it only really matters in the Z-axis for most printers.

There is one situation where it is important to have repeatable homing in the X and Y axis as well and that is on the printers that use line and capstans or smooth belts and flat pulleys and home on every layer to cancel the potential slip. In this case you want to be repeatable to the typical nozzle diameter at least.

A middle of the price range snap-action sub-miniature limit switch (Microswitch (R) is a trademark) without a lever should have adequate repeatability.

I have an expensive diamond drag CNC scribe marker and it uses a non contact proximity sensor that probably repeats at 0.02mm but could be as bad as 0.5mm if it got some gunk on the tab I expect. It also homes both X and Y with the same sensor as each axis has a tab that it moves to the sensor. I don't think the set point is known to closer than 1.0mm but the repeatability is better than I can detect with a magnifying glass.

I think what might be an option is to use a switch with a lever that has 1-2mm of flex. and then have a hard stop 1mm past the switch position. Then you run for 3mm after the switch triggers and let the stepper run against the stop slowly. The repeatability should be one full step of the motor as it will jump by that amount. If this is adequate then the switch error becomes irrelevant.

The slotted opto detectors are available in wide, narrow and lensed (as shown earlier) configurations that have more accuracy than needed, it takes a lot of ambient light to defeat them and even more dirt, a RepRap does not make sawdust so it should take a few years. They are also available (for quite a bit more money) with the internal buffer circuit in the same size housing that have just 3 wires and will plug straight into a controller. Or you can just make up some of the Gen7 end-stop boards for cheap.

---

Kalle
--
Lahti, Finland
The only stable form of government is Open Source Government. - Kalle Pihlajasaari 2013

Besides obvious quality differences the only real difference between these cheap micro switches and the micro switches we use on the CNC machines is the lever, CNC's usually have a dog that contacts the switches small bearing that pushes the trigger arm straight down into the switch, these cheap ones just have a thin metal lever. Removing the metal lever IMO would decrease the accuracy of the system, the unattached end of the level has to move much farther than the fixed end which is where the switch is. By removing it i would expect that you would have to drasticly slow down the homing speed to keep the accuracy.

I also have more trouble from the optical switches than the mechanical, optical switches wear out much more often than the mechanical. Its almost rare to have to change a mechanical micro switch, but we keep a backup supply of the optical stop's on the shelf for emergency repairs. The mechanical switches usually just need a good cleaning to keep them free.

Robin2 Wrote:
-------------------------------------------------------
> I would appreciate it if you could post a link to
> the exact switch you are using.
>
> ...R
>
> woo Wrote:
> --------------------------------------------------
> -----
> > i have omron switches from ebay, but not micro,
> > little larger, precision is better then
> 0.005mm(5
> > microns.)



i use those:
[www.ebay.com]

they are better then 0.005 mm, but it waz not worth to do preciser meassurement, because this value is OK .

Thanks @Woo

...R

Mcroswitches have been used for decades in the industry. All normal printers work with microswitches, even far less fancy than the ones that are beeing used for the 3D printers. They are good enough for the purpose, donm´t worry. Don´t just drive up costs without any benefit. If there are precision issues, you will find them elsehwere in the machine,

---

Blogs:
Meine 3D Druck Abenteuer
[3dptb.blogspot.de]
FLSUN Delta Drucker für Deutschland
[flsun-deutschland.blogspot.com]
Books on 3D patents:
[goo.gl] (english)
[www.amazon.de] (deutsch)