Some calibration questions

Now I’m starting to print some parts that need to be as accurate dimensionally as possible, I’ve decided to try calibrating the steps for each axis for the first time.

I’ve been taking measurements from each side of both the x and y axis’ and they don’t match. For instance I set the steps at one side of the printer moving the x or y in one direction and it’s perfect, I then go to the other side of the printer moving it the opposite way and it’s some way off. What I’ve done for now is set the steps at halfway between the calculation for either side and, although neither direction is moving to a hundredth of a milimetre accuracy, it’s at least moving the same distance each way, but a 100mm movement isn’t quite exactly 100mm.

Is this the correct way to do it?

The z axis I calibrated using a 10mm movement and my dial gauge, and that seems to be absolutely spot on.

Another question is, is there a way to raise/lower either end of the y axis? When fine tuning the height of each side of the x axis, ive found the the aluminium plate is actually sitting 0.5mm higher at the back end, which could go some way to explaining the 0.1mm dip I have right across the centre of the bed due to twisting. Or is this normal?

Thanks in advance.

The stock X and Y steps values in the Taz firmware should be very accurate, They are designed to allow for 1% plastic shrinkage on cooling. If your measurements are off, it may be to a loose belt. THe best way to ensure that X and Y are accurate and true is to print a hollow cylender. The outer diameter and the inner diameter of the cooled part should be the same in any direction and match your intended dimension, and the wall thickness should be uniform. If it is off in either X or Y direction with the values otherwise the same in firmware, it indicates one of the belts is loose enough that it has some slack in it allowing for excess travel. If the walls are too thick or thin it usually indicates overextrusion or underextrusion. if all of the dimensions across the board are too big or too small, then you want to look at the firmware values.

As far as the difference from one side of the bed or X axis to the other, it could be the part offset of the belt. The motor shaft doesn’t have a recieving bearing, so it can deflect a tiny amount. It’s not really enough to be noticible, but it might account for your difference.

The Y bed plane is adjustable by adding shims under the main frame rails and feet if you want. There is a small chance that the main frame to bed bolts are pulling the heat set inserts out of the blocks as well. I thought about makign a variant of the openbuilds Y axis with adjustable leveling feet, but i haven’t got around to that project yet.

The dip you are seeing in the bed isn’t the bed, its rod deflection due to gravity. a 0.05mm dip is normal, but you might just have a really heavy dial gauge. The only way to eliminate that dip are: 1. Hardened 10mm rods, 2, larger (12mm) rods, Taz 6 prototype style, or 3. a boxed rail (take a look at the openbuilds Rail Taz modifications in the development forum)

Thanks for the reply. I noticed stock x & y values were very near, certainly a lot more accurate than the stock Z and E values were, but doing the calculations brought them even closer. I’ll try put them back to standard tomorrow and have a look at the belts, I’m not sure how tight they should be but it was the first thing I checked before calibrating anything and they felt pretty tight.

For the Y axis, good points.

With regards the bed, it’s actually .1mm lower in the centre than at the corners, not higher. This dip varies across the centre of the bed from left to right but is at its worst at extreme left on the X axis, where it measures 16mm lower, with X at home and Y somewhere between 100mm and 150mm away from home. I dont think this dip is due to the X rods sagging, not only because the bed is getting further from the dial gauge, but also because when running the gauge across the top or bottom edge of the bed, the furthest it dips is 4mm.

When I get my Z to the height of a piece of paper in the centre of the bed, the nozzle actually hits the bed and pushes it down slightly in the corners.

The V slot rails are something I’m seriously considering, in fact im almost certain I will do them, but haven’t the time for a few weeks at best and need to get as close to total accuracy as I can ASAP. My dimensions are all currently up to .2mm off. I haven’t scaled the model down in Cura yet, I’m sure I’ll have to but I want to get everything as even as possible first.

I left my X and Y axis steps to the new ones I put in, after running a test print there was a notable improvement and saved me faffing around with the belts without really knowing what I’m doing. A temporary fix probably but it will do for now. I put my Z steps back to the original settings. As with the X and Y it showed different measurements in each direction, and the original settings were spot on for moving up, which I assumed is more important than down, as the only time it goes down is to start the print.

My aluminium bed plate is twisted. It sticks up about 1mm in the rear right corner. I’ve looked on Google for this and found a fair amount of people reporting this problem and things they might try but never seen anyone mention a working solution. Is there a way to sort this out, or is it a new plate job?