Different steps/mm between X and Y

I'm running a Prusa Mendel variant, with T5 belts.

To get my machine dimensionally calibrated, I'm runnning different steps/mm on my Y and X axises.

65 steps/mm on X, 62 steps/mm on Y. Same belts, same gears, same motors.

Is anyone running with a steps/mm difference like this?

When I first set up my machine, I calibrated with a dial indicator, and steps were something like 64.xxx for X and Y.

Edited 2 time(s). Last edit at 11/20/2012 09:30PM by Dirty Steve.

That is definitely not right, you might have some sort of belt slipping/motor skipping. You should really calculate steps/mm instead of measuring it - you can do this by working out how many steps/revolution your motors do (making sure to account for microsteps), then how many teeth there are on your pulley, then what the belt pitch is. There's a calculator somewhere on the wiki that'll do this for you. If you have identical pulleys/motors/step settings, you should get the same steps/mm. The fact that yours are pretty close suggests that there is some small factor like slip at play.

That's the thing, I know it's not right.

No loose belt, no loose grub screws,not missing a step in a 20hr print, low backlash at .03 on X, and .09 on Y.

As stated, when I first set up my machine with a dial indicator, X and Y steps/mm are equal, and I can draw with a pen mounted to correct physical dimensions. I can set my machine right back to this step value, and get correct dimensions in a pen plot.

But when I print with ABS, to get correct dimensions in the cooled part, I have to run this offset between X and Y.

I mean, I am getting dimensionally correct prints, I just know the math isn't right.

Going to switch to GT2 belts and pulleys. It looks like I am getting some wear in my aluminum T5 gears. Wasn't particularly impressed with the quality of my gears when I recieved them.

Quote

You should really calculate steps/mm instead of measuring it

IMHO, calculation is good for starters, but as you see, there are things which can't be calculated. If theory and real world behaviour disagree, real world always wins

For example, one thing not precisely known is the diameter of the pulley. Sure, it's made to match the pitch of the belt, but at which radius precisely? Can't be said until you actually measure the resulting belt movement.

Quote

No loose belt, no loose grub screws,not missing a step in a 20hr print, low backlash at .03 on X, and .09 on Y.

If the machine runs that well, why change it?

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Quote

For example, one thing not precisely known is the diameter of the pulley. Sure, it's made to match the pitch of the belt, but at which radius precisely? Can't be said until you actually measure the resulting belt movement.

As I have said many times before, the precise diameter of the pulley does not affect the distance moved, only the number of teeth and the precise pitch of the straight part of the belt. What happens while it goes round the curve of the pulley is irrelevant unless it is so far mismatched that it skips or has backlash.

Simple thought experiment: consider what happens when the pulley rotates exactly 360 degrees. The straight part of the belt must move exactly that number teeth to still be exactly lined up with the tooth at the top of the pulley.

Edited 1 time(s). Last edit at 11/21/2012 11:56AM by nophead.

---

[www.hydraraptor.blogspot.com]

Just wanted to know if anyone else has a step difference like this for X and Y.

You're squirting plastic out of a hole, you have no idea is it's actually round, or if the path is straight, if changing the steps makes the part come out better I'd go with it.
Having said that adjusting by>5% seems like a lot.
Generally though I'd trust measurement in cases like this, but I understand the need to justify it.

As to does anyone else have different steps in X and Y? I don't but I do have very slightly different errors in X and Y, though nothing like the 5%+ you are correcting for.

Edited 1 time(s). Last edit at 11/21/2012 12:35PM by Polygonhell.

Thinking my belts are garbage, cheap white ones off ebay, steal reinforced.

There looks like there is alot of slop in the gear/belt tooth interface.

Wouldn't diameter also affect travel?

If you have the same number of teeth on two gears, but one with narrower teeth and less diameter that still engages the belt would travel less distance in one revolution due to diameter difference reguardless of tooth count.

Edited 2 time(s). Last edit at 11/21/2012 05:17PM by Dirty Steve.

I don't know what electronics you are using, But if you have separate motor driver boards like on Ramps, it might be worth switching the two over to see if the problem moves axes. Some of the cheap pololu type driver boards only do 8x microstepping even if they are set to 16x. This should in theory make the one axes twice the size of the other, but it depends on what scaling you have set in the firmware.

Tension of belts increase his length. In your way X tension is bigger than Y. Try set same tension X and Y and your steps_per_unit be same too.

@ radus

Belt tension does not affect steps, only backlash/slop. The belts could be a mile long, the diameter of the drive gear is a constant. With equal stepps I get mechnical dimensional accuracy, but not when printing.

@martinprice2004

No firmware here, running Mach3 and a TB6560 driver board, and it's an awsome setup because I can run as a CNC mill, an engraver, a vinyl cutter, a pen plotter, and if I had one, even a laser cutter. Same microsteps, values are not that far off, no scaling.

Designing through X and Y means it is 2 dimensional.
Yes you can edit both X and Y as it is not difficult whatsoever.

@ ben11c

?????

Quote

Belt tension does not affect steps, only backlash/slop. The belts could be a mile long, the diameter of the drive gear is a constant. With equal stepps I get mechnical dimensional accuracy, but not when printing.

Belt tension does does affect how far it moves, the precise diameter of the drive gear doesn't. Under tension the belt stretches and its pitch increases slightly, so when the pulley rotates a given number of teeth, the belt moves a little further.

---

[www.hydraraptor.blogspot.com]

Can you show me some math on that? By that figuring, a 5 mm diameter gear, if the belt is tight enough, in one revolution is going to move as far as a 10 mm diameter gear with a loose belt. That doesn't figure. My belts are equally tensioned anyway.

So what about a toothless belt on a grooved pully? The diameter of the pully directly determines the distance of travel in one revolution.

Quote

Can you show me some math on that? By that figuring, a 5 mm diameter gear, if the belt is tight enough, in one revolution is going to move as far as a 10 mm diameter gear with a loose belt. That doesn't figure. My belts are equally tensioned anyway.

The maths is dead simple. If the pulley revolves exactly one revolution the belt moves exactly its pitch times the number of teeth. A 10mm pulley will have more teeth than a 5mm pulley. If the pitch of the pulley is too far out the belt will slip. If not it deforms while it goes round the pulley but that bit doesn't affect the distance to the carriage. That is determined by the straight part of the belt.

Quote

So what about a toothless belt on a grooved pully? The diameter of the pully directly determines the distance of travel in one revolution.

Yes that is why timing belts have teeth and reinforcing cords to ensure they stretch as little as possible, but when you adjust a belt tensioner with the motor locked you will see the carriage move. You can easily stretch it a mm or two over its length, so that will change the distance the carriage moves by roughly its travel times that divided by the total belt length.

---

[www.hydraraptor.blogspot.com]

I was wondering if you figured this out. I have this problem as well, Y is calibrated and X moves exactly 104 parts for every 100.

Never mind. It's because of a damn crack in the pulley.