M2 Auto-Leveling Kit (Beta)

[PcS]

Ultimachine has a working optical sensor for the rambo. The inductive probe is 5 usd delivered to the house from ebay I can keep a spare.If you look at the actual repeatable tolerance of some of these sensors they are amazing. I am running it at 24v. I never switch print bed . One bed works for everything I print. I look forward to see how the optical works for you as I was looking into it myself. The beta popped up here and I couldn't resist. Will you be doing auto leveling with optical ? Or not ? Might be good to compare results.

[Rara]

Are there higher end sensors that wouldn't be questionable temperature-wise?

[innkeeper]

That being said I think another type of sensor, like the one you mentioned is worth looking into. Had not considered an optical solution, would it work for a matte grey aluminum plate, not the most reflective surface.

well you wouldn't be reflecting off the plate, that's the whole idea, you would be reflecting off the glass surface. in theory it should work unless the underlying material under the glass is highly reflective like a mirror or a white background, raw aluminum .. so ..grey anodized would be preferable then..and the color shouldn't matter for the surface your reflecting off of though... so in theory, white gray aluminum etc shoulnt matter..

[innkeeper]

Are there higher end sensors that wouldn't be questionable temperature-wise?

i've seen them, but they were expensive like 100+, id rather do what pcs says, keep a spare or two and just deal with it if it happens.

[Josh]

PcS - what optical solution from ultimachine are you talking about? https://ultimachine.com/content/optical ... -assembled ?

innkeeper - I do agree that it's better to sense the top surface of the build platform, rather than an inductive surface somewhere below it; the inductive sensor, however, is a good trade-off between ease of use (don't have to move anything, manually or with a servo or similar) and sensing quality (as others have attested to). What optical solution are you looking to test out? The VL6180 or similar? I haven't found an easy-to-implement optical solution with high enough resolution for this purpose, but would love to hear if you've had better luck there.

[innkeeper]

Ultimachine has a working optical sensor for the rambo. The inductive probe is 5 usd delivered to the house from ebay I can keep a spare.If you look at the actual repeatable tolerance of some of these sensors they are amazing. I am running it at 24v. I never switch print bed . One bed works for everything I print. I look forward to see how the optical works for you as I was looking into it myself. The beta popped up here and I couldn't resist. Will you be doing auto leveling with optical ? Or not ? Might be good to compare results.

i'll see if there is any info on how ultimachine implemented it.

yea, from what i read its better to run the inductive sensor at 24v then to try at 5v. but also what i read you have to level convert the output cause it will also be 24v then.

i don't think i would be looking into optical if i didn't want to change out beds... wonder how hot it gets being only 5mm or less above the surface...since heat rises...

For the price, the inductive sensors are very very attractive.

Who knows this may fail epiclly and end up using an inductive sensor - i;ll defiantly post any results once i get it going to compare. i will assume they both will have plus's and minus's
(edit:minor typo;s as usual)

[Tim]

yea, from what i read its better to run the inductive sensor at 24v then to try at 5v. but also what i read you have to level convert the output cause it will also be 24v then.

The devices are basically just transistor switches with a 10kohm pull-up or pull-down depending on the version. Switching on shorts the output to one side or other (again, depending on the sensor version, because they have both NPN and PNP types), so yes, it's going to want to pull the output up to 24V if you let it, and the controller board doesn't want to see more than 5V on the min/max endstop connectors. But it's almost trivially simple to put another resistor from the output pin to ground as a voltage divider. First make absolutely sure that the sensor's internal resistor really is 10k, but all of them that I have seen (and they all seem to come from the same manufacturer) have the same internal resistance. The voltage divider is then (24V - 5V) / 10k = 5V / R, which solves to R = 2.63k ohms. That or a little lower (2.5k to 2.6k ohms) should work well.

For the price, the inductive sensors are very very attractive.

This is actually something new. I looked up "proximity sensor" in the Digi-key catalog and most companies are selling sensors in the $25 to $50 range, mostly for the automotive industry. These new cheap sensors are made in China and seem to come from one company, likely in Shenzhen. Sometimes it's just a matter of realizing that the solution has been there all along, only in military/medical/automotive specification, which drives the prices way up, even when the devices are really simple, like these inductive proximity sensors. Downgrade to commercial specifications, and it becomes really cheap, as long as you can find enough buyers outside of the military, medical, and automotive industries. Get enough 3D printers out into the world, and there are your buyers. . .

I have always heard that optical sensors are not accurate enough for the bed leveling job, but proof-by-experimental evidence suffices.

[innkeeper]

innkeeper - I do agree that it's better to sense the top surface of the build platform, rather than an inductive surface somewhere below it; the inductive sensor, however, is a good trade-off between ease of use (don't have to move anything, manually or with a servo or similar) and sensing quality (as others have attested to). What optical solution are you looking to test out? The VL6180 or similar? I haven't found an easy-to-implement optical solution with high enough resolution for this purpose, but would love to hear if you've had better luck there.

Josh, I agree, and that is what makes the inductive sensor so popular.

I looked into using a vl6180 and similar devices and and using an avr. but.. the problem is, due to the nature of it, it bounces though the glass onto the opposite surface and subsurface since the beams are at a small angle of incidence you get double and triple reflections. not to mention any delays in the coding that might though you off. . just a mess..not going to work.

The whole concept of distance measuring is making it more complex then it needs to be.

The way to do it is to increase the angle of incidence for the IR, and also not to try measuring distance but sensing an exact distance threshold...use a narrow beam angle to eliminate off angle detection. I've not found package that will do it, which means going it with discreet components. I've also found an implementation of something similar to what i'm doing, so i have that coming in parallel to test along with what i'm making ... see how it goes. My tests so far are positive so far so the concept works..and in theory should work straight off the power from the end switch connector. so simple connectivity. whats behind it is still more complexes then i think it needs to be, so also working in simplifying it, so there would be less components.

in any event, i will post the results as soon as its in a usable state, for better or worse.

[innkeeper]

I have always heard that optical sensors are not accurate enough for the bed leveling job, but proof-by-experimental evidence suffices.

this is very true, but not because its optical, but due to the methods used... i roughly outlined why in my reply to josh.

[PcS]

Yes josh that one. I just remembered seeing it but did not really investigate it .