Y1&Y2 difference

Measure the heal and toes distances of your YZ plates. Then need to be equal.

@vicious1 they are 1472mm on both sides (outside to outside 12mm mdf) from the outside to the outside of the table (100% square) both heel and toe have same distance (on both sides).
That makes me to believe the gantry is square?

Does your rail bearings tap on the rail if you tap the toe?

At this point the only thing it can be is a loose grub screw. Your crown is not off a full step, and it is not repeating, grub screw.

1 Like

@vicious1 the bearings hit the rail perfect no clicking when I tap or push them.

Grub screws are 100% snug and don’t move ( put a pen stripe on the axle and belt wheel and the line up after every move. (Took them apart 3 times to check again as loose grub screws was my first mistake).

In the meanwhile I replaced the Y drivers with fresh ones and makes no difference.

If I measure the crown long “mouth” @ the bottom the offset of the start and stop is 0,4mm @72 steps/mm =28,8 steps out?

I don’t really know what to test further….

From the beginning:

Gantry is straight
Bearings are on the rail
The Ystops line up the gantry straight.
The wiring is correct (stil thin for 2A)
The scrub screws are tight and secured with loc tide
Firmware is latest stock LR dual stop 2209
Firmware tft is also stock
It has the error on both Y and X
No blockages or things that could block the bearings during travel
Tension on the belts is loose but tight enough to stop skipping
Current doesn’t seem to effect the error

In my opinion this leaves 4 things:

Firmware (but other people have it working so we could stripe that down)
SKR pro board
Belt
Motors (as it does the same error on both Y1&2 doesn’t seem to be logic).

Did I miss anything?

Do the crown several times. Several in the same place, is the error in the exact same place. Do some in a new area, is the error in the same place.

The issue you are having makes no sense at all.

Make sure you do all your testing with my gcode.

Current would equal large skipped steps that progressively get worse, either high or low. My gcode is pretty tolerant to current since it is so slow.

Belt tensions would equal 2mm skips.

Nothing we have experienced makes small errors other than grubs, or overstretched belts but those usually end up home in the same place.

The only other thing is physical like cords or hoses hanging and pulling on your build weird.

The shift in the drawing could be pen pressure that is why I am saying do several make sure the pen is ever so slightly touch , just enough to make a mark, even have some light spots.

2 Likes

Please upload several pictures of your entire build so we cal look for obvious issues.

1 Like

@vicious1 start where the machine has stopped or start from the same position each time?

Don’t do anything between running the gcode. Just run it. And then run it again. No homing or anything.

@vicious1 Ok will do
Bear with me for a few days as I took everything apart yesterday and started printing YZ plates to make sure everything is up to spec.
(And somehow 3D printing takes a bit of time :innocent:).

Will be back when al is mounted again.

Thanks so far!

@vicious1 @jeffeb3

I printed the YZ plates and all fits nice and snug.

After rebuild I did not set correct sizes and the offset is stil there.

Dropbox - IMG_3544.MOV - Simplify your life Video of offset.

Crown 3 times only thing i did in between prints is move Z down 5mm to prevent air prints.

After the prints the grub screws did their work

and nothing has moved.

Pics of the build:













Heel and toe are the same

The vacuum tube holder and cabling is temp as I still have to make the strut plates

This is super weird. It is perfectly off, that is not normal for something being wrong. Issues are usually random, and get worse…and not on both axis at the same time.

Draw a square on that paper and measure each side lets see if they are all the same.

As for the video, your Y belt should be in line with your Y driver, you have a pretty steep angle (not good but not the problem either).

Lets rule out that screen. Please use a USB cable and use repetier host to drive the machine and run any tests, do not touch that screen. There is absolutely no reason I can think of that you would drive out 600 and back and be off 4+ mm and not hear something going wrong. Those screens have built in routines and I did not make firmware for that so it could be funky.

1 Like

To me, that looks like it makes every move perfectly during the entire crown. And then it moves +5mm in every direction…

Something is doing that deliberately.

I have not seen this before. So I am really interested in knowing everything you’re doing, to try to figure out what could possibly be different.

@jeffeb3 @vicious1

The crown is done with repetier (using parallels on a Mac)

Done the crown twice.

The square is done with manual commands from repetier with the G0 command.
(G0 Y100 - G0 X400 - G0 Y0 - G0 X500) doen this 3 times and the offset is less than the crown.



The touchscreen connected or not makes no difference
So that leaves the only place for an error in the SKR pro board.
Te firmware in the board is v1cnc_skrpro_DualLR2209-2-2

Anything you want me to do in a specific order anything I need to record?

2Crowns.gcode (213.7 KB)

Run that file.

Run it from Repetier, then move the machine and run it again from your screen and memory card.

Let’s see which ones messes up that will tell us where the extra code is being added.

Run that anywhere other than your machine’s actual 0,0. As in move it away from your end stops a little bit on X and Y (10mm) is fine.

I have a few guesses as to what is going on. This will narrow it down.

The firmware on your SKR is ours not edited in anyway?

Thanks for your time @vicious1 !

It’s stock firmware from you today I loaded the latest version as I have seen the bed size was upgraded :joy::innocent:. No alterations whatsoever.

Only settings in the board are the mm per step set to x71,28 y70.40 z396.04

This is repetier :slight_smile:

This is the tft:


They look quite similar ?

(Really hoping you still have a hunch!)

No, nope, not at all.

That gcode was just one on top of the other. There is no way that they should not be on top of each other. There was no releases or anything funny.

There is no way it would mess up precisely between two lines of gcode and not anywhere else.

What power supply are you using?

Run a M500, then run M502. Reset everything just in case.

It is only moving in Y now…Run it in a different spot in your bed.

G01 X41.7403 Y79.1580 Z-1.0000 F720
G01 X41.9597 Y79.5635 Z-1.0000 F720
G00 X41.9597 Y79.5635 Z5.0000 F500
G00 X0.0000 Y0.0000 Z5.0000 F2100

;No. 1: Engraving 6
G00 X75.2682 Y15.5365 Z5.0000 F2100
G00 X75.2682 Y15.5365 Z0.5000 F500
G01 X75.2682 Y15.5365 Z-1.0000 F210 S20000

That is the transition from the first crown to the second. It does go back to 0,0,5, which is what was exciting the issue in the first couple of posts. As far as I can tell, that is the only time it goes to 0,0.

Why would the act of moving to 0,0 cause it to skip by several mms? When we figure this out, we are going to have a beer. :beers:.

1 Like

I was first thinking the G00 was just defined too high, but it is set several times in this file so it can’t be that.

Moving the drawing away from the actual 0,0 is how I was trying to rule out the G0 x0 y0 thing.

The only hope I am seeing is that it is sometimes X or y OR X and Y… Randomness somehow is comforting.

WHY does it not mess up during the drawing at all?