I “just” have a standard Formbot Voron, but I am quite happy with it. I had an “Omron” (maybe not the original) and it was pain in the a**. The change because of temp was specified with 20%, my printer was more a carousel doing the QGL as a printer, it drives me insane. So I changed to the klicky probe wich was much better but… see discussion here.
I would check, if this should not be +5. The belt starts “behind” the table… I will change it tomorrow…no today (shit it’s late)…
Looking at it again, I think it -5 is appropriate. When I looked at it after the original post, I forgot that negative values for y are outside the build envelope which is where I think the gantry corners should be.
A positive value will probably be problematic. As I said, it’s working well and I’m reluctant to change it unless there was a
OK, I checked the values for QGL, corrected them, set back the stop criterion to 0,005 and started QUAD_GANTRY_LEVEL.
Finally it failed due to the bad quality of my probe.
And I found a bug (!)
// stepper_z = 0.012786
// stepper_z1 = -0.013774
// stepper_z2 = 0.006494
// stepper_z3 = -0.005506
13:41:12 // Retries: 5/5 Probed points range: 0.011250 tolerance: 0.000000
I commented out the entry #retry_tolerance: 0.02 so klipper took 0,00 as tolerance. This should lead to at least a warning.
What type of probe are you using? I’m curious to make sure I avoid it.
Just as a comment, a tolerance of 0.005 is really tight (the basic Voron 2.4 uses 0.0075) and, when I was setting up the printer, I found that I couldn’t see any practical difference by going to 0.05, an order of magnitude higher.
I should also point out that the standard deviation of measurements for BL Touch, Omron TL-Q5M2 and Klicky are all more than 0.005 so you’re getting into the noise range of any sensor.
Actually, not specifying a tolerance should generate an ERROR, not a warning.
Klicky Probe: GitHub - jlas1/Klicky-Probe: Microswitch probe with magnetic attachement, primarily aimed at CoreXY 3d printers
But it’s not the design, I just used an end switch, that was available in my supply box.
for the 0,005mm sure! I was just joking with ReXT3D, because I use 0,02mm and he uses 0,005. See posting above! I am back at 0,02mm that works nice with my shitty probe. I changed my probing to 10x and average calculation. Seems to help, but I will replace my switch with a new one, because my impression at the beginning with klicky probe was, that it works very nice…
1 Like
FWIW, my experience with the OMRON TL-Q5MC2 has been rather good thus far:
The above not only includes the electro-magnetic probe repeatability but also the mechanical repeatability of my printer in the Z axis. The plot shows that the repeatability typically falls within approximately 0.001 mm band (+/- 0.0005 mm) with occasional outliers within 0.0015 mm band. It is certainly more than accurate enough to clearly show the initial mechanical drift (or settling) from ~1.4585 mm to ~1.4545 mm over the course of 500 probes…
This type of characterization is what I normally do to arrive at the some of the settings of my printers. The Voron now gives me practically perfect first layers, especially after I “cracked” the nozzle vs. Z switch repeatability “puzzle”: Z Homing Accuracy & Repeatability
EDIT 1: For completeness, I use 256 microstepping on the Z axis so the theoretical resolution is 0.00015625 mm per microstep.
EDIT 2: I decided to explore this a bit further and repeated the same experiment on my very heavily modified CR-10S Pro. This printer is equipped with genuine BL-Touch, Moons’ 0.9° steppers and LDO 4 mm pitch precision lead screws with radial type anti-backlash nuts. I do not have linear rails and still use the POM wheel trolleys, but the whole printer is mechanically adjusted to “perfection” within the limitations of its design and manufacturing quality. The following plot uses normalized results so that they can be easily compared - I normalized the data by plotting the magnitude of deviation from the average reading. The results are quite interesting and really speak for themselves.
It is worth mentioning that I can obtain slightly better repeatability from the BL-Touch & my CR-10S Pro if I reduce the probing speed from 4 mm/s to 2 mm/s, but the key word here is slightly.
Maybe my OMRON is a cheap copy. It was really driving me mad, because it takes 15 Minutes to become stable enough for the measurement. I dived into the specs and found:
Temperature
influence
±20% max. of sensing distance at 23°C in the temperature
range of −25 to 70°C
I read it that the probe is ±20% from -25 to 70 degr. and then I dumped it (because I saw exactly this behavior). Switching to klicky probe was a big step and felt much better than this… So I am confused again. First, I will try to replace the Klicky Probe by a new one, if that fails, I will cry some big tears and doom the day of my birth (twice!).
How much did you paid for you probe? I see prices of around 70€ and around 15€ for the same “Name”…
One more question: What is you explanation for that curve? Is it going down because of temperature?
I assume, you have the standard surface for a Voron: Aluminium table + magnetic tape + spring steel PEI surface?
Taken at its face value this means that the probe accuracy is referenced at 23°C and can change quite a lot as a function of the complete operating temperature range. So in practice one could expect up to 20% error in the 5 mm trigger range when the Voron chamber is heated from room temperature to print temperature. This is 1 mm of potential error and clearly not acceptable for Z homing.
However, my data shows that once the probe is at a stable temperature, it is extremely repeatable. Therefore using it for QGL or bed mesh mapping will result in extremely low errors, on the order of 0.001 mm. I am certain that this was the intent of the Voron team when they chose a mechanical switch for Z stop calibration and inductive probe for QGL. Prusa does it differently with the current iterations of their Pinda probe, based on what I know - they attempt to use a single probe for both functions by thermally correcting the the absolute error of the inductive probe. Of course they do not have the additional challenge of using a heated chamber.
The current price is $61.51 at Misumi, which is where I got it. If you want to make sure that you get the original, you have to stick to reputable suppliers. But if you use good quality switch in your Klicky, you should be getting very good accuracy…
Each solution has some compromises and some benefits. Ultimately one has to be prepared to make adjustments whenever either the bed surface is changed (spec configuration like mine), the nozzle is changed (Klicky configuration), print temperature is changed significantly (Klicky configuration), etc…
I am not sure to be honest, but you have to look at the magnitude - it is a “drift” of 0.005 mm. This could be as simple as the stepper motors heating the belts and changing their tension. It has been repeatable though in my testing. If I perform two batches of 500 probes back-to-back, the second run has a substantially reduced amount of “drift”.
If I get bored I might repeat the same test with the chamber at stable printing temperature, where I expect the repeatability will be degraded due to thermal effects.
EDIT: Yes, I have a pretty standard print bed setup but I use Mandala Rose Ultraflat 6 mm bed with Subtle Designs Staymag magnet (now unavailable). The Energetic spring steel PEI surface that’s currently on top has larger flatness deviations than the bed & magnet combined - I can easily see the steel sheet “waviness” due to the rolling process when I visualize the bed topology…
How did you receive your Omron sensor? Here’s what I get the genuine article in:
I’ve used a number of them and I have to say they’re rock solid from 20C to 70C ambient temperature with no plastic softening (which seems to be a problem with knock offs). I made up a plastic gauge with different height pads to test its operation so I feel pretty comfortable saying the product is very stable over 3D printer enclosure operating temperatures.
The best price you’ll see for the genuine product is around $20 USD at AliExpress. I was shocked to see that it’s $92.90 USD at Digi-Key. I wouldn’t trust Amazon sellers, but I did see one in the packaging for $19.99.
Maybe somebody is counterfeiting the packaging but, if you look around, you will see that vendors make a big deal about showing the packaging in their product image to show that they are genuine and you can be pretty sure of their performance.
This is the beauty, I trowed out from my system. Seems to be ok so far… Formbot should not cheat knowing that they cheat… So I will not buy a new one and give this one a try…
And you used this with what voltage?
Yep. But I will check, if I will used it again. I used the 12V output as far as I remember…
1 Like
I replaced my old microswitch from the Klicky Probe and Voila! convergence is more than perfect. So finally it was a bad (damaged) probe.
OK, Thanks all for the contribution. I’d like to continue the conversation about QGL and Z-Tilt in the Feature Section, if there is interest in. What I want to do is to change back from “NG Klicky Probe” to normal “Klicky Probe”, if I got some spare time (never).
EDIT: This is the repeatability of the new micro switch (again: 0,10€/part). Standard deviation is 0,004mm which should be good enough for my case (Hey REXT3D: Should I try the 0,005 again? 
)
09:20:25 // probe at 300.000,275.000 is z=7.464237
09:20:27 // probe at 300.000,275.000 is z=7.467987
09:20:28 // probe at 300.000,275.000 is z=7.467987
09:20:29 // probe at 300.000,275.000 is z=7.471737
09:20:30 // probe at 300.000,275.000 is z=7.475487
09:20:31 // probe at 300.000,275.000 is z=7.474237
09:20:32 // probe at 300.000,275.000 is z=7.474237
09:20:34 // probe at 300.000,275.000 is z=7.476737
4 Likes