Bed Leveling for Beginners (Page 1 of 5)

Nice post Carl !
I'd just like to add that there is an Android app available on Amazon which will save you a LOT of time with the initial auto calibration. It calculates the amount you need to turn each screw to get a pass with your chosen offset.

Thanks! I was just going to search for this!

many thanks Carl, This was a big help.

Thanks Carl! 

My new AIO machine arrived with two of the bed-leveling thumbwheels rattling around loose inside the case and I spent a couple of hours trying to use the printer's built-in calibration without ever achieving a pass.  I then found your suggestions and within five minutes I had a system which prints perfectly.

I have a da vinci duo, and learned the hard way how to calibrate.  I had the same issues - the literature that came with the machine warned me not to calibrate (since it was calibrated at the factory).  Very soon, the extruder(s) were hitting the model and the the nuts started falling off, so I had to learn how to calibrate.

This is a very long post, but I wrote it to help others understand that if you create a repeatable process for yourself,  calibration is a lot less of a hassle, even if what you are doing seems random and or backwards at first.

I created an excel file to remind myself that the three numbers displayed are in order:

Back (right), Right (front) Left (front)

and refer to the corners that the autocal measures in this weird Back, Right, Left sequence.

There are three support nuts - NOT where the corners are,  but in a triangle.  The 'back' nut is really in the middle of the plate! I taught myself to think about the three nuts as:

Back, Right, Left

And associated them in my mind with the three values.

The Back nut is kind of on the diagonal between the left front corner and the back right corner.  So . . . I'll return to that in a moment . . .

I realized the easiest way to think about turning the nuts was that if you put your thumb to the left of a front nut and rotate it to the right until your thumb hits the edge of the plastic surrounding the nut . . . that's about a 1/4 turn.  I think of it as rolling to the right.  I have to 'roll' the left and right nuts, but I can grab and turn the middle nut 'the back nut'...

When I turn the 'back' nut, I position my thumb and forefinger at 45 degrees on the nut (they are now on a line from left to right), and twist it through a 'roll' that puts my fingers on a new line, 90 degrees from the first (they are now on a 'front back line').  This means that I can roll the back nut through a quarter turn, just like the front left and right nut.  It really helped me to turn (roll) all three nuts the same amount.

ROLLING TO THE RIGHT RAISES

I then, through a lot of repeated effort, understood that everything is backwards.

If a number is higher than the target (for the duo the target is 100), then you have to raise the nut associated with that number to lower the number.

To raise the nut, and lower the number you roll the nut to the right.

To lower the nut and raise the number, you roll the nut to the left.

I visualize the plate.  If the number is 90, 100, 90, then the left corner and back corner are at the same height, and are low, meaning that those two corners are closer to the extruder - they are high.  If you can get two numbers close to the same value, you can then concentrate on the other number.

Having a good imagination is crucial to not going crazy over the headache of calibration.

so . . . back to the fact that the 'back nut' is not on the back of the plate.  It's really in the middle of the plate.  It's only 'back' in the sense that it's away from the 'front' of the plate.  - this is where some imagination helps - Picture the plate as having a hinge when you calibrate.  That hinge can be along the front of the plate, or it can be along the diagonal between the left (front) and right (rear) corners. 

It can also be along the diagonal between the right (front) and the left (rear). but most of the time we can ignore that.  All we really need to do is to pair two of the numbers and adjust to bring those two numbers closer together.

Often I'll get a result of 139, 90, 80, meaning the back is low and the front is leaning to the left.  Since 90 and 80 are close, I'll ignore them and work on getting the '139' closer to the other two.  If the right and left numbers are close, then the hinge is along the front edge.  Raising or lowering the back nut will open or close the 'hinge' and affect the third number, while having a minor effect on the other two. Remembering that a larger number means that part of the plate is lower than it should be, I roll the back nut to the right, which will raise the back towards the extruder and decrease the number.  It would have been so much more intuitive for me (on the West Coast of the US) if the numbers followed the direction of the nuts.  The engineers in Taiwan probably find it intuitive just the way it is . . . but whatever.  Culture is different depending on where you are in the world - you just have to live with it.

I used the spreadsheet to keep track of my moves and keep track of the result of the move.

Eventually I put a label on the front of the printer to remind me.

Actually I put two labels on the front with arrows, but I think I'll replace them with the simple statement:

ROLL TO THE RIGHT RAISES

That phrase helps me visualize the bed of the printer.  If I need to raise the back, or a corner, I roll the nut to the right.  If I need to go the other way . . .

There is probably a Chinese phrase that makes perfect sense and associates the movement of a nut to the change in the number.  If I understood Chinese it wouldn't seem backwards to me

So instead of getting hung up with the numbers being backwards to the direction of movement of the nuts, I concentrate on translating what they mean.

Back, Right, Left
156     35     90 

means the right corner is too high (relative to the left), and the back of the plate is too low.  The left is sort of close to the target, and I can ignore it for now.  So I'll roll the back nut to the right to raise the back and roll the right nut to the left to lower it.

I found it really helps to have something else to do while calibrating.  I'll start an autocal, go in another room and focus on some small task, then return to the machine when it gets quiet.  Do another round and do some other small task.

This means I get about 15 minutes of work done on something during the 30 minutes I'm calibrating the printer.  Watching TV helps, or watching YouTube.  I can go back to the printer during the commercials, or between videos.
 
But wait!  Why does the rolling the nut in one direction sometimes appear to have a random result?

There is a spring between the bed and the nuts.  Sometimes when you turn the nut, you turn the spring too.  This creates a movement of the spring end on the surface of the nut or the underside of the bed.  When this happens, you get a result that is totally unexpected.  1 quarter turn typically changes the number by X . . . except when something moves that you didn't want to move (like the spring).  Instead of just compressing it, you turned it and it resulted in an unexpected change in the numbers.  I may introduce a couple of smooth washers into the stack between the nut and spring and the spring and the bed and see if that helps make the movement of the nut more regular.

If it works, I'll post it.

All of us that have had the machine for a while and gave up on the auto function all understand what the numbers and what they mean. The issue is though that the range that the machine allows for pass and fail is wrong for correct bed height.

The first issue is that the cal is ran without knowing what layer height you plan to use. Proper bed height should be half or even a third of your layer height. So that is the first flaw. The next issue is the range. If you go to small on the range it will fail and that small range is still too high. Of course if your range goes too high it will fail there as well. The best bed height should be about .1 to .15 for this printer. That is impossible to achieve through the auto cal because it will fail being that low. So that is the next flaw with this feature.

So those two reasons are why I posted this guide months ago. I finally convinced a mod to make it a sticky and that is why you just saw it only after making your long essay.

You really don't need to use auto cal again once you get a pass. Just calibrate manually straight away as per Carls' post and you're good to go.
I haven't had to level the bed for weeks and first layer is still as good as the last time I levelled.

Carl mentioned:

As a someone who hasn't had my machine for a while, I'm obviously a beginner with this machine.  I'm used to much higher quality machines which are a lot more fire and forget (but are insanely expensive for what they really are). I wrote my essay to address how I gained an understanding of what the numbers mean.  I did that to benefit others who have just found this forum, and might want to stick with the stock process.

I really appreciate your guidance in the two issues you talk about.  Perhaps you can help me understand what the value in parenthesis after a successful 'Cal' represents.  Does that represent the gap between the extruder head and the bed that the software will use during printing?

Thanks for this! 

I am new to 3D printers and the calibration on this printer is finicky.  I tried for 1 1/2 hours last night before giving up.  Then read up on it today and found the online calculators, which where helpful.  When I left off last night I had readings of 144, 186, 239.  It would seem like I was getting random readings at times.  I took the suggestion of some others and cleaned the contact points on the plate with a little acetone as well as the sensor on the extruder.  I did this first tonight and on the first calibration run I came in at 233, 230, 245 without changing anything.

I tried a test print and the key chain thing printed ok but I could tell the head was still too close on the left side.  I then tried a larger print (Coat Hanger for a door) and it would not print correctly.  I remembered reading your advice and did it exactly as indicated and when I was done it printed nice on the first try.

I find that the key to not pulling out your hair is to:

• Learn what is really going on

• Multitask while using the stock calibration utility

What is really going on

The different models have different arrangements, but I think the following can be said for all Da Vinci systems.  The print head has a sensor on it.  The Calibration Utility verifies a bunch of stuff, then raises the print head to touch the sensor at three places.  If we've passed high school geometry, then we all know that a plane can be defined by three points in space.  That's what the utility is doing.  It's measuring three points and reporting to the user those three points.  The utility will 'pass' if those three points are within a range.  The Utility then performs some sort of calculation and uses the resu
lt in some fashion.  Only the Da Vinci engineers know for such what that calculation is and how the result is used.

It's not an automatic calibration, the user has to make adjustments.  How you make those adjustments is really weird. That is annoying as hell, but whatever, it is what it is.  You make the adjustments by turning nuts that are on threaded studs at three locations under the print bed.  That the nuts and studs are not in the same location as the points the sensor touches is . . . NUTS!  It make the process really difficult, and it's finicky as well.  The nuts are on a stud and there are springs between the nuts and the bed itself .  .  . which is great, since that protects the print bed from really bad crashes against the print head.  But, the design of the nuts, studs and springs are . . . cheap, just like the printer . . . which is why we bought it . . . so there.

ah, can you post a link that works?  I searched the XYZ tutorials site and did find a 'heated bed calibration' video, but I'm limited to the English language, and the only one I could find was in some other language.

Unfortunately I'm limited to one language and 'Ces trois valeures correspondant au niveau des trios points de mesure" is close . . . but

I think it confirms my thought that "the three values correspond to the three points measured"

I hope I'm not lost in translation