PDA

View Full Version : Backlash Problems?



Mason Atom
09-24-2012, 9:55 AM
Hello everyone, first time poster here. I'm a new laser engraver owner, but have used Epilogs fairly frequently over the past 3 years so I do have some experience with these types of machines.

The machine I recently bought is a Chinese 50W Artsign JSM-3060N from MBKP International. The engraver came with LaserWork 5.0.21, which I am using as a plugin inside CorelDraw X4.

I have the instrument up and running and for the most part it is working well... but I'm having an issue with bi-directional engraving. Whenever I engrave using "X-swing" or "Y-swing", I get some shudder on the image, such that it looks like there is a shadow on the outer edge of the image in whatever axis I've done the engraving in. I've measured this to be about a 0.45 to 0.55 mm offset.

When I use "X-unilateralism" or "Y-unilateralism", images come out great with no shadowing. The downside though is that I can only engrave at half the speed of bi-directional (swing) engraving, and obviously I'd like to get the engraving speed up because I have a few big jobs lined up.

Here is a picture showing what I'm talking about, engraved in reverse black/matte clear IPI plastic. On the left is a 'unilateralism' cut that looks very good, on the right is an unacceptable shadowed 'swing' cut. Both were cut at 350 mm/s.

241697

I've independently checked both the X- and Y- axes and the shadowing is not axis dependent; I measure about 0.45 to 0.55 mm in whichever axis was chosen for the job. Both axes engrave great with no shadowing if I use Unilateralism.

My initial thought was to add backlash compensation within the software. I have tried about 10 different settings, writing the config file to the laser, but it does not appear to make any difference whatsoever - same offset every time. I know that the config files are getting written to the laser because I modified some of the acceleration settings and observed that they were in fact changed the next time I ran a file. But the backlash compensation doesn't seem to have any impact regardless of what the settings are.

I have checked all of the mirrors to make sure that they are tight in their mounts and not moving. I did a re-align of the beam the other day and to the best of my ability it appears to be fairly well aligned. I literally took the beam completely out of focus so that the cutter wouldn't cut, then re-aligned everything and re-measured the backlash, but it is still exactly the same (0.45 to 0.55 mm). I am planning to have one of my friends who has a PhD in optics re-check the alignment for me to verify it is in good shape.

I guess the next thing to do would be to adjust/tighten the belts? Failing that, I'm pretty much at a loss for what could be wrong. Is there something that I have to do in the software to enable the backlash compensation?

Any suggestions/advice would be great.

Thanks,
Mason

Mike Null
09-24-2012, 10:01 AM
Mason

welcome to SMC.

Are you absolutely certain that the material is secure on the table? Vibration of the machine or table could cause the material to move.

Mason Atom
09-24-2012, 10:04 AM
Mason

welcome to SMC.

Are you absolutely certain that the material is secure on the table? Vibration of the machine or table could cause the material to move.

Well, I wouldn't say that I'm absolutely sure the engraving material is not moving... but why would it be in an issue only with bi-directional cutting? The laser head is moving the same speed regardless. Wouldn't I also see effects of material movement when using uni-directional cutting?

Glen Monaghan
09-24-2012, 11:01 AM
It looks to me like the problem here is primarily on the right side but I can't tell whether the "fault" is on the end of right-going passes or the start of left-going passes. There is also a glitch in the fine details of that little "tail" on the left side of the bottom stem on the "4", just below the corner. Might this be a case of needing the "tickle" voltage adjusted?

-Glen

Mason Atom
09-24-2012, 11:12 AM
It looks to me like the problem here is primarily on the right side but I can't tell whether the "fault" is on the end of right-going passes or the start of left-going passes. There is also a glitch in the fine details of that little "tail" on the left side of the bottom stem on the "4", just below the corner. Might this be a case of needing the "tickle" voltage adjusted?

-Glen

Glen, thanks for taking a look and commenting.

It's not easy to tell from the image, but the shadowing/border comes on both sides of whatever is engraved. So for example, on the 'G' on the rightmost image, there is a 0.5 mm border on the left of the G, and a 0.5 mm border on the right of the G. That was with X-swing engraving. If I recut this with Y-swing, the border would be on the top and bottom sides of the letter.

I also engraved some crosses and bars in cardboard to examine the shadowing - the border always comes on both sides, with the same amount of offset for both sides, in the direction of the swing axis.

What is tickle voltage? I've not come across that term in any of the LaserWork manuals that I've read.

Mason

Glen Monaghan
09-24-2012, 12:10 PM
If you are seeing the shadow on both sides, I'd say there is an issue of calibrating the bidirectional scans (sort of a backlash problem as you originally suggested). However, if that was the extent of the problem, all the details should be present and just look blurred due to the offset between alternate lines. However, the glitches such as that "tail" I mentioned look like they aren't getting fully engraved which made me wonder if the laser isn't firing right. That made me think of previous posts where people had to adjust tickle voltage to get the laser to fire correctly at edges of the engraving. Never used it myself but the "tickle" apparently it is an extra voltage applied to the tube when not actually firing so that the laser beam forms more quickly when asked to fire after being "off". There are others here who certainly can explain it much better than I.

-Glen

Mason Atom
09-24-2012, 2:55 PM
If you are seeing the shadow on both sides, I'd say there is an issue of calibrating the bidirectional scans (sort of a backlash problem as you originally suggested). However, if that was the extent of the problem, all the details should be present and just look blurred due to the offset between alternate lines. However, the glitches such as that "tail" I mentioned look like they aren't getting fully engraved which made me wonder if the laser isn't firing right. That made me think of previous posts where people had to adjust tickle voltage to get the laser to fire correctly at edges of the engraving. Never used it myself but the "tickle" apparently it is an extra voltage applied to the tube when not actually firing so that the laser beam forms more quickly when asked to fire after being "off". There are others here who certainly can explain it much better than I.

-Glen


hmmm, maybe... I'm not sure how/where to set that in Laser Works. I'll go through the manual again. Thanks for the comments.

Rich Harman
09-24-2012, 4:16 PM
You should first work out the backlash settings by engraving circles. Test the X backlash by starting the circle on the side, test the Y backlash by starting the circles on the top or bottom. You want the start and end of the circles to line up perfectly without a visible jog on the opposite side.

Once you get that perfect then you need to adjust the Scanning Interval which is found by clicking "Config(S)". To adjust the interval I create a series of about ten lines that are a few millimeters apart, like a ruler. Set the scan speed to 200 and the scan gap to .5 then run it. You want a larger scan gap so that you can identify a left-to-right pass from a right-to-left pass. Using a loupe you should see a series of dashes and almost certainly each line will be offset from the next. In the scan interval dialog add a setting for 200 speed and then guess a number for the scan interval. Run it again and see if it is worse or better. My number is .04 for 200 speed. You can repeat the process for other speeds.

john banks
09-24-2012, 4:41 PM
Mason, are you adjusting backlash X and backlash Y in Settings... Other parameters?

The way I adjust is using Config(S)...System setting... General settings tab... Scanning (backlash) ... Add (and setup a few settings for various speeds, I have 0.1mm at 400mm/s, 0.2mm at 600mm/s and 0.3mm at 800mm/s. It interpolates between, and you can put silly values in like 10mm using wide scan gaps as Rich suggested to confirm that it is adjusting and the direction of adjustment.

Mason Atom
09-24-2012, 4:51 PM
Thanks Rich and John. I was just looking through the manual and concluded that I needed to change the settings in Config(S)/Systems/General - scanning (reverse interval) when I read John's post it seemed to confirm my suspicion.

Rich - thanks for the advice on the backlash testing. I've been using a similar technique with scanning crosses, but yours seems more elegant.

BTW - I think that both of you have much less backlash in your systems than I do.

Thanks again,
Mason

Rich Harman
09-24-2012, 5:04 PM
Do not confuse Backlash with Scanning interval. You find the scanning interval settings under Config(S) like John says but first adjust backlash like I have said. You find the backlash settings under the "User" tab which is in the window at the right side of the screen. You will first have to click "Read" to load the values into the window. Do not forget to click "Write" before testing your settings.

john banks
09-24-2012, 5:16 PM
Interesting Rich, you're adjusting backlash on the right hand side, I'm adjusting it by scanning speed under Config(S) menu.

Rich Harman
09-24-2012, 5:22 PM
Like I said, these are not the same thing. Nothing, absolutely nothing you set in the Scanning Interval will correct for engraving a circle where the start and end do not line up perfectly. Adjusting the backlash will correct this.

It could be that adjusting the backlash will not affect a scan interval setting. That does not make sense to me but I have not investigated it. To be safe, first adjust backlash, then adjust scan interval.

To be more clear, backlash attempts to account for very small amounts of play in the drive system when changing directions, speed has no effect. Scanning interval does nearly the same thing but it also accounts for the response time of the laser and the speed. If you look very carefully you can see the effects of belt stretch too. If you engrave a series of vertical lines it is not possible to get each and every one of the lines to align perfectly Most noticeably the first. I attribute this to belt stretch. I think only very young eyes or someone with a loupe could ever notice the difference though.

john banks
09-24-2012, 5:32 PM
I'm confused. All my backlash adjustments are done under Config(S) and setting a reverse interval for different scanning speeds with no adjustment at all in that right hand window. It works great, the only issues I have are with my tube being slow to respond quickly at more than 30% power levels, but on a material that only needs 10-20% power I can engrave bidirectionally at 800mm/s just fine. I set it up by using a 0.5mm scan gap and engraving rectangles with various backlash adjustments at various speeds and then taking a straight edge down the lines to see where the best compromise was at each speed to line up the left and right edges with each line alternately engraving in opposite directions.

I have no adjustment for y axis backlash, but I rarely use y bidirectional rastering and risk missed steps at 800mm/s due to the mass of the y axis.

I'm trying to understand what is different about the backlash settings, I'm rereading "Test the X backlash by starting the circle on the side, test the Y backlash by starting the circles on the top or bottom. You want the start and end of the circles to line up perfectly without a visible jog on the opposite side" in post #8 and am puzzled.

Rich Harman
09-24-2012, 5:51 PM
The word "backlash" does not appear in the settings that you are referring to. It is called "Reverse Interval".

Here is a picture instead of a thousand words;

241707

john banks
09-24-2012, 6:04 PM
I relabelled it as that based on what it did and what the manual said about it IIRC. It was a bit vague though.

I'm not sure what I would achieve by adjusting the backlash settings on the right hand side window? Can you explain a little more about your circle testing? What does the start and end not lining up correctly mean?

john banks
09-24-2012, 6:10 PM
Don't have the RDCAM v5 manual handy, but v3 refers to:

Reverse interval

While bi-directional scanning, for the tension of belt, may cause the edge of graphic irregularity. So

provide backlash setting to amendment it. A specific speed corresponding specific backlash.

Generally, the faster the speed, the greater the backlash.


Add backlash setting

Click button
Add,pop up the following dialog box

After setting
Spped and Reverse interval,then click OK,the backlash value be inserted into the

backlash list

--
It is saying to adjust backlash by speed as I've been adjusting. Perhaps the adjustment that is written to the machine itself on the right hand side window is a static addition regardless of speed, but then it would be duplicating the speed based entries?

john banks
09-24-2012, 6:14 PM
One more thought before bed... could the x and y backlash settings in the right hand window be a compensation primarily for vectoring? Is that what you're getting at when talking about engraving circles? We do have a tiny offset sometimes...

Rich Harman
09-24-2012, 6:22 PM
When engraving (or cutting) a circle the laser must start engraving somewhere on that circle and end engraving and that same point. Ideally, upon close inspection you should not be able to tell where that point is.

If the start point is at a cardinal point - top, bottom, right or left - you stand the best chance of noticing that the start and end do not meet exactly. It is at these points of the circle that the axis must stop then reverse itself. If there is any play in the system the result will be a slight misalignment.

What backlash does - when it comes time for the axis to reverse direction it runs the motor very briefly to take up the slack before getting on with the rest of the movement. It happens in an instant. Without it there would be a very small bit of time where the motor was turning but the laser head was not moving - so the actual position of the laser head would lag behind it's expected position.

Mason Atom
09-25-2012, 9:19 AM
John and Rich,

Thanks for the very interesting discussion. I was originally trying to solve my problem using the menu that Rich showed in his attached image. I believe that this setting only refers to vector cutting, because it makes *zero* difference when scanning (that was where I was originally trying to fix my problem). I haven't had any issues with circles (well, if you look under a microscope there is probably a 10ish micron offset), but the naked eye wouldn't see it. I arbitrarily changed the backlash settings in that menu and did some circle tests and made them disconnect, so my impression is that menu item only refers to vector cutting.

John, it turned out that my scanning (reverse interval) list was pre-populated with grossly miscalibrated numbers. I'm not sure if it came shipped that way or somehow became corrupted, but in any event the numbers were waaaaay off. I was engraving at 350 mm/s and the config file had -0.24 mm set at that speed, which accounted perfectly for the .48 mm offset I was seeing in my bidirectional scans. I did a series of tests on various cross patterns and under a microscope it was looking like about a 20 micron offset when cutting at 400 mm/s. So I deleted all the gibberish from the scanning (reverse interval) list, entered and saved 0.02 mm, and now the bidirectional scanning is making beautiful cuts.

Thanks so much for all of the help John and Rich.

Mason

Rich Harman
09-25-2012, 8:49 PM
The Backlash setting gets programmed onto the main board of the machine whereas the Reverse Interval does not. I suspect that when scanning the backlash setting is "in effect" but it may not be noticeable on something scanned. It must be tuned by vectoring where it's effect is easily seen.

I ran a quick experiment. I rastered a pair of vertical lines using my Backlash (0.06) setting. Then I ran it again after changing the Backlash to 0.2. I moved the lines to be directly below the first set. The difference was quite noticeable. Both times the Reverse Interval was the same.


241780

Mason Atom
09-26-2012, 8:24 AM
The Backlash setting gets programmed onto the main board of the machine whereas the Reverse Interval does not. I suspect that when scanning the backlash setting is "in effect" but it may not be noticeable on something scanned. It must be tuned by vectoring where it's effect is easily seen.

I ran a quick experiment. I rastered a pair of vertical lines using my Backlash (0.06) setting. Then I ran it again after changing the Backlash to 0.2. I moved the lines to be directly below the first set. The difference was quite noticeable. Both times the Reverse Interval was the same.


241780

Interesting experiment. Also weird in the sense that when I enter backlash settings through the 'user' tab (that you pictured above) they make absolutely no difference in the scans for me. I've used ridiculously high values up to 1 mm with no change in the quality of scanning. At first I thought the settings weren't getting written to the machine, so then I changed the acceleration down to 1000 mm/s, along with the 1 mm backlash in X and Y, then wrote those to the machine. It was definitely written, because the scanning slowed down markedly due to the lower acceleration, but the observed backlash in the scanning didn't change at all. So it seems to me that there is something wrong with that setting in my software. I think I'm going to do a re-install. I don't know why part of the user config would work but part wouldn't?

Gerhard Koekemoer
11-06-2012, 12:30 AM
My first time on SMC and hope I can help.
Seeing the picture I can see that the so called backlash is not just on the ends of the job but also visible inwards which means that you are not getting the full beam but might be getting a scattered beam in these areas. I would suggest that you clean the optics properly and that you might have a flat/bad spot, the scatter light will not give 100% focus. I have seen this happening on cutting a circle would be out of focus on one side but in focus on the other side. Also make sure the beam passes through the center of the nozzle otherwise scatter beam will also give similar problems. There is always a slight movement on the area where the beam passes through when changing direction on the cut/scan.
2. Replace the machine parameters and the vendor parameters available from the manufactures as this will solve the speed problems.
3. Get a copy of the new version of RDCam6 from your supplier. It works on Corel X5,X6 and Adobe Illustrator.

axs karaci
11-25-2016, 6:50 AM
hello
did you solve the problem
i have a problem just like yours348204348205 as you can see from the pictures.

i own a chinese kh5030 white and blue, 50 w tube(brand new) and a 60 w psu(brand new)
the problem started after i changed those parts.the "ghost" apears on x swing , uni, and y swing and uni....difference is that on x-uni or y uni the marks appear only on the left side, the right side, top side or bottom side
also it is visible with naked eye as i exeed 60mm/s
thank you
regards axs