Chinese engraver cutting incorrectly

[Keith Winter]

Hi friends,

My 2nd Chinese engraver wasn't making perfect squares so I went into rdworks settings as the vendor Thunder Laser told me too and adjusted the step length. Good news is the 100mm x 100mm square is now perfect. Bad news is now it's hitting the right side of the bed when I go around it or try to run a job that is the full length of the bed. Apparently by making the step longer to fix the mis-shaped squares it now is stepping too far on the X axis and hitting the right limit switch. I've run this file maybe 100 times before so I know it's a good file.

Things I've tried
1) Redid the stepper test after reseting from a full power down, still cuts 100mmx100mm square perfect
2) Redownloaded the file, still hits the limit
3) Reset origin to the top left multiple times
4) Retightend the x axis belts, no change

Ideas on how to fix this?

[Scott Marquez]

Can you adjust the limit switch for more clearance or will you physically hit something?
If you ran this file in the past, with the machine cutting it under sized, can you shrink your part size slightly to fit within your working area?
Scott

[Keith Winter]

Thanks for the reply.

It's not related to the limit switch placement. The limit switch is in the correct place , it matches my other laser and I have run that file mentioned above over a hundred times on both lasers.

[Matt McCoy]

If you have adjusted the steps, you might now be outside of the cutting envelope. The limit switch can be moved or the nesting reconfigured to get you within the envelope again.

[Jeff Body]

Hit the reset button on the control panel.
You need to get the machine to relearn the limits with it's new step lengths.

[Jerome Stanek]

Go into the vendor settings and adjust the table size

[Keith Winter]

Hit the reset button on the control panel.
You need to get the machine to relearn the limits with it's new step lengths.

Hi Jeff,

I've turned the laser on and off a few times, are you speaking of a software setting or is turning it completely on and off the same thing?

[Keith Winter]

If you have adjusted the steps, you might now be outside of the cutting envelope. The limit switch can be moved or the nesting reconfigured to get you within the envelope again.

Hi Matt. What is nesting reconfiguring?

[Kev Williams]

How big is your machine? for fun I'll work with 900x1200...

How far out were the 100mm squares? If they were 99.5mm, then to compensate across a 1200mm table you've lost roughly 6mm (1/4") of space. That's plenty too much if you've been running right at the limits.

If the machine will physically go far enough with no interference, I'd just move the right limit switch- or REmove it. FWIW, I've never seen a right-side limit switch anyway.

Or, can you physically move your LEFT limit switch so your zero will be a bit farther left?

My Triumph can't utilize the entire 1300mm length because the left limit switch is mounted so that 'zero' is a bit too far right. The machine will physically still go 1300mm, but the head runs out of bearing rails! I've been going to move my switch but I never have, simply because I've never needed the full 51".

nesting reconfiguring--just means repositioning whatever you're cutting to fit better within your boundaries (if possible)

[Keith Winter]

Hi Kev,

Thank you for the explanation. My bed is 1300x900 They were quite far off. Like 94mm so that means I've lost 6mm every 100mm so 76mm (3") of the usable table space to compensate for the servo being inaccurate? Basically they sold me a defective servo, is that how I should interpret being off by that much? Or what does that mean?

[Dave Sheldrake]

Step length or current may be wrong Keith, step length controls the distance per pulse but if the current is off then it may still go too far.
Zero is usually X 10 y 10 so you lose that right off the bat and add in a bad current setting on the step driver and you could lose the rest easy enough

[Keith Winter]

Step length or current may be wrong Keith, step length controls the distance per pulse but if the current is off then it may still go too far.
Zero is usually X 10 y 10 so you lose that right off the bat and add in a bad current setting on the step driver and you could lose the rest easy enough

Thanks Dave, how do you setup the current? I've been having a poor engraving problem with this laser as well which is making me wonder if the servo missing steps. Causing this problem and the poor engraving issue.

[Dave Sheldrake]

Usually a set of 6 - 8 dip switches on the side of the stepper driver. I can't recall what they are supposed to be set to (what positions) but it could well be some of the problem

[Kev Williams]

You were off nearly a full 1/16" inch for every inch? Holy crap-- maybe you have the wrong stepper, or maybe the wrong belt cog...

I just bought a stepper I didn't need for the Triumph, old one is right here, doing a little checking...

It's a 1.8° motor, so 200 steps per revolution--

My cog has 24 teeth, and my trusty (ahem) HF caliper showed .118" from tooth to tooth, which is exactly 3mm movement per tooth...

This = 72mm per revolution...

Maybe not ironically, the 3mm distance per tooth is exactly half your 6mm shortage per 100mm....

[Keith Winter]

Thanks guys. Kev I'm ball parking but yes it was around 94mm on the x and about 106 on the y, quite a rectangle! Want to hear the craziest part? The whole reason I went down this path is because we were having blurry engraving issues, but the cutting seemed ok until I had to cut and precise size for jig I was making. Essentially a 610mm high by 406mm wide rectangle for a jig. It was driving me mad because it was always off, then I cut it on my other Chinese and it was very close to perfect. Emailed the manufacturer and they said to the do 100mm x 100mm square and adjust those setting in the settings. So I was very off on the 100m x 100mm square but I was off by only 1mm on the height and 2mm on the width of the larger 600x400 piece. Now tell me how that's possible that I'm actually off more on the smaller 100mm x 100mm piece?

[Dave Sheldrake]

So I was very off on the 100m x 100mm square but I was off by only 1mm on the height and 2mm on the width of the larger 600x400 piece. Now tell me how that's possible that I'm actually off more on the smaller 100mm x 100mm piece?

That does indeed suggest incorrect dip switch settings on the driver card.If it was step count (distance) it would be incremental with distance

[Rich Harman]

It is easy to tell if you are missing steps. If you can repeat the job and the laser follows the same path, then you aren't missing steps.

What follows is strictly my opinion, which is backed up by facts and personal experience.

The best way to set step length is through math. Forget about that software "feature" where you cut and measure, then input new step lengths. That will only work on that one particular size of part. Do that on a small part and then when you go to run a big one, it will be way off.

Use math and all your cuts, no matter the size, will be as close as possible given the limits of the machine. The western machines have an advantage here because they use encoder strips. The small inaccuracies are accounted for. Chinese lasers have no such feedback, they just do the best they can.

You need to know the number of steps per revolution of your stepper. This can vary depending on how your driver is set. Then you need to know how far the laser moves in one revolution. You figure this out by knowing the pitch of the belts and the number of teeth on the pulleys. With this information you can calculate the exact step length and enter that into the vendor settings.

When I have parts that need to be exactly the same size, I cut them all at the same spot on the table. If I were to cut them side by side then there would be some variation in size, a couple thousands, which is unavoidable on this type of machine. There is no step length that will fix that. It is the nature of the machine.

[Kev Williams]

I've attached a couple of test scales (Corel 10), the upper is a 1/4" graduation scale that will span your table, less 1" on the left and around an inch on the right. The numbers can raster, the scale is to vector. The scale is in 3" sections, so the bottom line can be removed or not engraved. My idea is to have it vector engrave the scale all the way across and set an actual scale next to it to see where you variances, if any, are occurring, and how much...

The bottom scale is to be rastered across, they are to test your X axis backlash and left-right/right-left alignment. Do a few tests with different lengths in different areas. And test using different speeds. The lines are .003" thick, so the laser should just barely fire them. Raster these with a lot of gap, like .2 or more, so each direction's firing will stand out from each other. Ideally, you want perfectly vertical lines, but since you're having raster issues, I'm betting you wont at first! But you should have a software adjustment for this. Mine is called "carving backlash", in the "Crafts Parameters" section. NOTE this is a screenshot of the default page of the software, I have LOTS of settings-- and the funny thing is, to get mine to work I have to enter negative numbers...
338883
Anyway, I've painstakingly tested mine using a scale like this from every speed in 50mm/sec increments from 100 to 800mm/second to get the values dialed in. And as big as this machine is, it engraves fantastic. I'll put it's engraving alongside any other machine's. :)

If you don't have lots of scrap to test on, just use paper... :)


And Rich-- many a western machine uses stepper motors, like my LS900. Only machine I have with an encoder came from Taiwan! :D

[Keith Winter]

Thanks kev!

Got that in AI or EPS :confused:

Question I have a couple of these engraving bad and one engraving good, all on the same settings right now. When I apply the settings to try and fix one will that mess up the others? Or does every servo need slightly different settings even if they are essentially the same lasers as the others in every way?


I've attached a couple of test scales (Corel 10), the upper is a 1/4" graduation scale that will span your table, less 1" on the left and around an inch on the right. The numbers can raster, the scale is to vector. The scale is in 3" sections, so the bottom line can be removed or not engraved. My idea is to have it vector engrave the scale all the way across and set an actual scale next to it to see where you variances, if any, are occurring, and how much...

The bottom scale is to be rastered across, they are to test your X axis backlash and left-right/right-left alignment. Do a few tests with different lengths in different areas. And test using different speeds. The lines are .003" thick, so the laser should just barely fire them. Raster these with a lot of gap, like .2 or more, so each direction's firing will stand out from each other. Ideally, you want perfectly vertical lines, but since you're having raster issues, I'm betting you wont at first! But you should have a software adjustment for this. Mine is called "carving backlash", in the "Crafts Parameters" section. NOTE this is a screenshot of the default page of the software, I have LOTS of settings-- and the funny thing is, to get mine to work I have to enter negative numbers...
338883
Anyway, I've painstakingly tested mine using a scale like this from every speed in 50mm/sec increments from 100 to 800mm/second to get the values dialed in. And as big as this machine is, it engraves fantastic. I'll put it's engraving alongside any other machine's. :)

If you don't have lots of scrap to test on, just use paper... :)

[Rich Harman]

And Rich-- many a western machine uses stepper motors, like my LS900. Only machine I have with an encoder came from Taiwan! :D

Use of steppers does not mean no encoders.

[Kev Williams]

Best I can do is DXF, every other way I export it the file is too big to upload? And I had to export it from Casmate to even this file small enough- which is funny because it's just lines..

As for adjusting the steps of the motors themselves and what that may affect, I'll leave that question to Dave!

I will say that as of late, I've had a rash of machine malfunctions, most of which were due to wire connection problems; data transfer errors with the LS900 due to a dirty LPT cable switch, crazy spindle activity on a 3400 due to a loose wire in a plug connector, broken stepper motor wire on the Triumph, bad data transfer via USB/Serial adapter cable, bad USB cable on the IS400, cold motherboard on the Explorer which was brought to life by cleaning & reseating all plugs & connectors... Your intermittent bad-good sounds almost like the cutting problems I was having with the Triumph. The broken wire was okay until the gantry reached the point where it curled up in the drag chain and would lose connectivity...

Short version, you might just have an iffy connection somewhere...?

--and fine, steppers with no encoders then. :)

[Keith Winter]

Ok so..

1. I checked all the dip switches on the blue and black cards, they are all the same as the "good laser"
2. I changed the servo/stepper motor + the "HBS57 Hybrid Servo Driver" for the x axis (blue card in the photo)
3. I could not change out the black "3DM580 Microstep Drive" because I don't have one

No effect, to the cutting or engraving. Cuts a perfect square with the same settings as the other motor I had at 7.13 step X and 8.27 step Y. When I put it back to the default 7.87 step x and 7.87 step y it stops running into the edge but cuts a malformed rectangle instead of a square again in the 100mm x 100mm square test.

Any thoughts, what the issue is?

[Dave Sheldrake]

can you check the corner angle for me Keith with a square on a cut out item?

[Keith Winter]

can you check the corner angle for me Keith with a square on a cut out item?

Sure dave you want me to take a photo of a corner cut out of a piece using the default settings yes? Not the corrected settings I found doing the 100mm x 100m.

[Dave Sheldrake]

yup...default settings

[Keith Winter]

Hi Guys,

I wanted to let you know I resolved this. I'm not sure what I did but I somehow messed up the alignment, maybe transposing a number or something. When I started 100% back over from scratch alignment it seems to work now. Thank you all for your help.

[Steven Yates]

What happens if you try to cut a 1200x800 rectangle? Does it hit the limit switch? If not measure the cut piece. Is it 1200x800?

If it is then the cutting is fine but your limit switch needs to be moved.

If the cut piece does not measure as you would expect, try recalculating your step from the larger rectangle and then try measuring a small rectangle again to see how much that is off.

[Matt McCoy]

Hi Guys,

I wanted to let you know I resolved this. I'm not sure what I did but I somehow messed up the alignment, maybe transposing a number or something. When I started 100% back over from scratch alignment it seems to work now. Thank you all for your help.

Good news Keith. Just to clarify: It wasn't the alignment (mirrors), but the bed size in the driver which was putting your file outside of the working envelope -- correct?

[Keith Winter]

More like user error I think unfortunately ;) I must have messed up something when I did the first test which caused all of the tests after that to mess up