This must be backlash but how to fix?

[Shane Johnson]

Hi

I normally try to fix issues myself but I am really struggling with this one!

My Red & Black 80W Chinese has been working fine for a few years. Some weeks ago I remember the head got caught on a camera housing fixed to the inside front of the lid - I'm not sure if this is when the problem started or not but I mention it in case it might be relevant !

In the following picture the red circle indicates the start position. As shown, the circles are not completing when starting at the top or left positions.
There is no problem when the laser head moves from the bottom (front of the laser machine).

IMG_1168.JPG

Now, this looks like a backlash issue but I have spent hours and basically removed everything from the linear rails to the belts a number of times to try to fix it. I have tightened belts, loosened belts, tightened all of the grub screws on the pulleys, set backlash settings in Lightburn for the X and Y axis, installed RD Works to see if it made a difference - not once have I seen any change.

When powered, there is no discernible movement possible in either the X or Y axis.

I have watched the Lightblade Learning Lab episode on Understanding Backlash & Scanning Reverse Interval a number of times but it didn't help - adjusting the Y backlash setting fixed the issue at the top of the circle but resulting in a 'bump' at the bottom!

Any assistance would be very much appreciated.

[Pete Cyr]

I normally try to fix issues myself but I am really struggling with this one!

viewtopic.php?f=162&t=5826

review this post.

[Shane Johnson]

Hi Pete

I was hoping that you might reply

Yes, I should have mentioned that I have done the full calibration, but even if it was out, it should result in on oval shape rather than not joining ??

The circle is fine when the laser head moves from the bottom of the bed which would tend to suggest some sort of backlash, But I can't find anything that seems loose.

[Gene Uselman]

Take a look at this one too.
viewtopic.php?f=105&t=5733&p=29442&hili ... ash#p29442

[Shane Johnson]

Take a look at this one too.
viewtopic.php?f=105&t=5733&p=29442&hili ... ash#p29442

There's not much I haven't looked at on this forum

These backlash examples show a problem on opposite sides of the circles and setting the backlash compensation in the software fixes the problem. However, I only appear to have a problem at the start/finish point and so if I use a Y backlash compensation number it fixes the original problem but then creates a problem at the opposite side of the circle.

I will look at the reversal interval adjustment, which only effects scanning, once I get this problem sorted.

I do appreciate your assistance, it's just that there isn't much I haven't already tried

[Shane Johnson]

I didn't realise that I have been struggling with this problem for over 2 months!

But I am ...

I have just about pulled this machine apart on numerous occasions and the problem still persists!

I have checked the X and Y calibration - X was at 150.13 and Y at 150.09 on a 150mm cut - not perfect but probably within the margin of error for my electronic calipers.

I tried a new lens - no change.

The screws fixing the left hand Y belt to the gantry had stripped the thread in the gantry slightly - I thought that this might have effected things when the head moved down (towards the front of the machine). I replaced the screws this morning - no change!

A test today -

2021-09-05-0002.jpg

1 - the head moved down and to the right to get to the start point - the start/finish points didn't match. 2 - the head moved up and to the right to get to the start point - perfect result. 3 - the head moved down and to the left to get to the start point - the start/finish points didn't match. 4 - the head moved up and to the left to get to the start point - perfect result.

So, something is happening when the head needs to move down to get to the start point??

Could this be a stepper motor problem? Or a driver problem? I don't think it is a belt issue, or a backlash issue as it only seems to happen in one direction?

Any advice appreciated.

[Jeffrey Aley]

Does this problem occur at all speeds, or only when cutting "quickly"?
Can you identify any mechanical reason why the stepper motor might be losing steps? For example, missing teeth in the belt. I think you've already checked belt tension.

Are the circles on the blue paper cut clockwise or counter-clockwise? I expect the answer is counter-clockwise, because I'd expect the backlash to make the ending point of the circle lower than the starting point.

The start/end points of your circles are at 12 o'clock. If you changed the circles so the start/end points are at 6 o'clock, I'd bet you'd see the mismatch when the head approaches from the bottom.

[Shane Johnson]

Hi Jeffrey

Thanks for your comments.

Yes, happens at all speeds - I have slowed to 5mm/s and it still occurs.

I have closely inspected both of the two side Y belts and the X belt - nothing is obviously wrong. I have also swapped over the 2 side belts and there was no change. The fact that the problems occur on different positions on the bed would probably indicate that it is not a tooth issue - that would be my understanding anyway?

The picture above were all cut in a clockwise position.

I have done some more cuts with start points at 12, 3, 6 and 9 o'clock and in both a clockwise and anti clockwise direction -

The first row had a start point at 12 o'clock, second row 3 o'clock etc. The first column had the origin at top left, second column top right etc.

Clockwise

Clockwise.jpg

Anti Clockwise

Anti-Clockwise.jpg

There is some consistency of sorts - with start point at 12, the circles where the laser head had to move up were good and where the laser head had to move down were bad. But, with a start point of 6 the opposite occurred?? With a start point of 3 the circles where the laser head had to move right were good and where the laser head had to move left were bad. But, with a start point of 9 the opposite occurred??

To be honest, I'm not sure what to make of this...

Happy to do any further tests suggested - any suggested fixes welcome

[Jeffrey Aley]

Thanks for running the experiment. I'm still digesting all the data, but I realized that I should ask you a very simple question:

Have you turned on "backlash reapy optim"? It's on the Output tab at the right side of RDWorks.

[Shane Johnson]

The problem occurs using both Lightburn and RDWorks.

Using RDWorks there is no difference if I select 'backlash reapy optimize'. There is also no change if I select 'Hide backlash' which I understand is the Lightburn equivalent.