I ran into this helping someone yesterday too. We'll be doing some troubleshooting and documenting things to try and help address that. Haven't found a pattern to it yet, but you can get around that by going to lightburn settings and disabling the line item for ignoring items outside the work area. I would enable it again when done, but it should get you through the process. Additionally, there seems to also be a scale problem sometimes. The rest of the calculations seem to be working well, but sometimes the scale on either X, Y, or both, tends to make the work area too big and scale output accordingly too large. This can be mathed out manually to accurately scale down by making a square of a known size and then adjusting by the % it is over, scaling it down by that value on the given axis scale (over simplified version of the explanation but best we can do in the comments lol). Hope that helps, and we'll try and submit feedback to the awesome team at LB to try and help improve things as we test and try and find details that may help them. I'm hopeful this should be a quick fix in a future update, and this is still useful as is anyway given that we only have to adjust scale with some quick math and the rest seems to be perfect :) Stay tuned!
@@LaserEverything I am having the same issue with my 70mm lens as well, when I am looking at it the Box extends passed the corners. Not sure if this data is helpful or not.
I also had a lot of problems with my 70 x 70 but I just ran extra boxes and corrected. Mine was wildly skewed incorrectly after the “easy” correction process
I've used this new feature for calibrating 3 lenses, and I agree with the final statement you make (but you gloss over it pretty quick) - after accepting the settings and verifying they've applied, you need to draw a box the maximum size your lens will allow, and mark it, then measure that box. For all 3 lenses I found that the scale factor needed tweaking. eg for my 110mm lens, when I completed calibration, and then tried marking a 105mmx105mm box, it was producing something closer to 103.5 x 107mm or something like that. It showed me that the scale factors for X and Y were not right. I'm sure it was due to errors in measuring but I had the same issue with all 3 lenses I calibrated. Just a suggestion - can you add a bit to the end of your video showing marking a final box, then taking measurements to verify your scale is right, and if not dead on, performing a quick calculation to adjust the scale percentage to get your marked box the exact size you need. I had to do this for all 3 lenses, so others will stumble over this too, and showing them what to do in your video would help some with that final step.
It's easy to fix - lets say your marked box should be 50 x 50 but it marks as 48.5 wide x 52 tall.Go to your Device settings. Look at Galvo 1 and Galvo 2 'scale' percentage values.
To adjust: The formula is: current scale percentage value times desired length divided by actual mark length.
eg if actual mark is smaller than desired (should be 50mm wide but you got 48.5mm wide mark and your scale percentage currently is 91.2329%) The formula is: 91.2329 x 50 / 48.5 = 94.0545% - set this as your new scale percentage for x axis.
If actual mark is bigger than desired (should be 50mm tall but you got 52mm tall mark and your scale percentage currently is 82.3219%) The formula is: 82.3219 x 50 / 52 = 79.1557% - set this as your new scale percentage.
Once you adjust the scale percentage for both axes, try marking the 50x50 box again, and you should see it come out the correct size.@@akbarazadazad1453
Shared this above, but just so you guys see too. I ran into this helping someone yesterday too. We'll be doing some troubleshooting and documenting things to try and help address that. Haven't found a pattern to it yet, but you can get around that by going to lightburn settings and disabling the line item for ignoring items outside the work area. I would enable it again when done, but it should get you through the process. Additionally, there seems to also be a scale problem sometimes. The rest of the calculations seem to be working well, but sometimes the scale on either X, Y, or both, tends to make the work area too big and scale output accordingly too large. This can be mathed out manually to accurately scale down by making a square of a known size and then adjusting by the % it is over, scaling it down by that value on the given axis scale (over simplified version of the explanation but best we can do in the comments lol). Hope that helps, and we'll try and submit feedback to the awesome team at LB to try and help improve things as we test and try and find details that may help them. I'm hopeful this should be a quick fix in a future update, and this is still useful as is anyway given that we only have to adjust scale with some quick math and the rest seems to be perfect :) Stay tuned! Edit: Just seeing@imakethings17 's shared the formula they used to correct scale. good info, you explained it far better than I was able to in text lol
@@ssullivan I'm glad it helped. I'm sure others will have the same issue so as long as they follow those final checks and adjustments it should give them perfect settings.
Thanks for recommending to take a screen shot of the settings. I went into device settings and nothing changed! I wasn't even able to edit it manually until I deleted the reference to the factory provided .cor file. TL;DR my 10 mm squares now render as 10 mm x 10 mm. Thanks!
Interesting to see the update! However, I have to point out something important - this cannot account for skew distortion. Note, for example, that both skew factors in your example are exactly 1.000. Because you are only measuring the distances, which is fair enough, there is no way of knowing the angle between the two axes. The axes could be severely skewed, but since you’re only measuring distances (in particular, distances 3 & 4), there’s no way of figuring out skew. This is what frustrated my about using corfile2 - it took me ages before I realised there was no way account for skew.
A bloke called Pythagoras disagrees with your assertion. I'm not saying this is doing it, but from the measurements taken you can plot the position of every intersting line in relation to the centre point. If the 12 lines that make the grid are all equal you must have 4 squares inside a square.
@@timg6252 Nope, I'm afraid that's not the case. The simplest way to demonstrate this is to look at the following: imajeenyus.com/temp/no_skew_demo.jpg Pretend that those are two different marked patterns. The pattern on the left is ideal, while the pattern on the right has severe skew. In both cases, the distances between the points (the lengths of the lines) are all 50mm. In both cases, you would measure and enter exactly the same values into either corfile2 or the new Lightburn feature. Yet one situation clearly has skew, the other doesn't. Now, the instructions do say to measure the distance "from each indicated point to the HORIZONTAL center line". In other words, the red line I've shown on the right-hand drawing. That's fine in theory, but in practice the difference between that measurement and the point-to-point distance will be so small as to be unmeasurable by eye with a digital caliper. For example, suppose you have a 2° skew angle, which is significant. If the point-to-point distance is 50mm exactly, then the perpendicular distance will be 50*cos(2°) = 49.97mm. No way is it possible to measure that accurately. Besides, I can easily demonstrate another marked pattern which has skew, yet where all the perpendicular distances are equal to 50mm. I spent several months tearing my hair out over how exactle corfile2 performs its calculations, and finally determined that there is indeed no way for skew correction to be performed, doing the measurements this way. Brief background: the COR file that is produced contains a 65x65 array of "displacement" values for both galvos, which is sent by EzCad to the BJJCZ controller board. The controller receives an uncorrected galvo position from EzCad, looks up the array to find how much it should displace that position, and then sends _that_ position to the actual galvos. Take a look at this graphic: imajeenyus.com/temp/cor_file_random_values.jpg That is a representation of the displacements for each of the 65x65 points in the array - the short black lines show the direction and magnitude of the displacement. The important thing to note here is that the on-axis displacements are always ALONG the axis (the short black lines along the X axis are all horizontal, and those along the Y axis are all vertical). Skew correction can never be performed this way, and I experimented with numerous different values. This isn't to say that the COR file approach itself is bad - it's only corfile2 that's poor. The COR file itself can indeed handle any arbitrary corrections you like - I have written my own script which will perform a 5x5 grid correction based on marking a grid of holes and then measuring the actual coordinates of each. However it's still very much a work in progress. The only reliable way to determine skew would be to measure _diagonal_ distances, but I'm not entirely sure how the calculations would go....
@@timg6252 (I previously wrote a reply to this, but it seems to have not posted) Nope, not the case. I can show you two different mark patterns, one with severe skew, yet both with all distances equal to 50mm: imajeenyus.com/temp/no_skew_demo.jpg
Help needed, when I run this test I found out that the line in the 5-6 area is not burned as sharp as the rest ,any idea how that happens. Thank you Ben
I will try this. However, I am concerned as it appears the distortion I am experiencing is different based on file type. For example, if I mask and flatten a png the distortion of greater than if I create a shape in lightburn or use an svg.
You mentioned in the video you would link another video to leveling the head and troubleshooting, but I don't see it anywhere. I looked through your channel, but wasn't able to find it. Can anyone point it out, please?
Awesome! It just dawned on me that when I did my camera calibration with a Galvo, placing the dot matrix outside of the work area (my lens = 175 but the grid was placed out to the edge of the camera's viewable area) might mess up the camera-to-lens accuracy. Am I right?
I tried this with my 300mm lens but the framing makes it fall way off my bed and making the box smaller makes it fall below the recommended reduced size.
There are other values that handle that for us, with delay and timing values. The LB team had a set of parameter limits on it and has since been updated :)
So these settings are per lens, and not global ? Haven't seen anything where you can pick a particular lens, other than inputting the size in the 1st step
Ideally, you'd set up a "device" in lightburn for each lens size + machine. So if I have 3 lenses for my 50w fiber, I'd say 50w fiber 110 lens, 50w fiber 200 lens, 50w fiber 300 lens, as an example. This way each lens for the machine will have its own profile and be tuned. We do the same for parameter libraries as well, since they change based on lens size and such. Each correction is specific to the lens, like a fingerprint. So for best results, you'd do this process for any new lens setup. We have an updated video coming that includes the whole process, start to finish on a new lens config and tune (updated from this even, its in editing).
That is correct. Note it's only on the current public beta and not available on the normal channel as of yet. When this update releases on the public channel we'll make an announcement about that as well. Otherwise - you'll just want to be sure to grab the current BETA version of the software to access this feature.
Does anyone know what to do when the error message of "one or more shapes are corssing the edge of the machine workspace and will not be sent." happens? Please help!
I ran into this helping someone yesterday too. We'll be doing some troubleshooting and documenting things to try and help address that. Haven't found a pattern to it yet, but you can get around that by going to lightburn settings and disabling the line item for ignoring items outside the work area. I would enable it again when done, but it should get you through the process. Additionally, there seems to also be a scale problem sometimes. The rest of the calculations seem to be working well, but sometimes the scale on either X, Y, or both, tends to make the work area too big and scale output accordingly too large. This can be mathed out manually to accurately scale down by making a square of a known size and then adjusting by the % it is over, scaling it down by that value on the given axis scale (over simplified version of the explanation but best we can do in the comments lol). Hope that helps, and we'll try and submit feedback to the awesome team at LB to try and help improve things as we test and try and find details that may help them. I'm hopeful this should be a quick fix in a future update, and this is still useful as is anyway given that we only have to adjust scale with some quick math and the rest seems to be perfect :) Stay tuned!
Join one of our many support communities I'm sure someone can help. Definitely will need to see some pics. Links to everything in the description, lots of free options.
At 5:56 you mention a video link for leveling the lens to the bed. Can you link that please? I must have a terrible combo of blind/dumb and clueless, as I don't see the link. If its this one: ruclips.net/video/FZT9zOp9Sl8/видео.html than my lens calibration is the problem this vid should help me with. Lasers for knife engraving accurately is akin to hurting kittens and pushing string.
Has anyone got this to work with a 70mm lens? It keeps saying the drawings are out of bounds and it only gives me the dragon, not the boxes.(mac)
I ran into this helping someone yesterday too. We'll be doing some troubleshooting and documenting things to try and help address that. Haven't found a pattern to it yet, but you can get around that by going to lightburn settings and disabling the line item for ignoring items outside the work area. I would enable it again when done, but it should get you through the process.
Additionally, there seems to also be a scale problem sometimes. The rest of the calculations seem to be working well, but sometimes the scale on either X, Y, or both, tends to make the work area too big and scale output accordingly too large. This can be mathed out manually to accurately scale down by making a square of a known size and then adjusting by the % it is over, scaling it down by that value on the given axis scale (over simplified version of the explanation but best we can do in the comments lol). Hope that helps, and we'll try and submit feedback to the awesome team at LB to try and help improve things as we test and try and find details that may help them.
I'm hopeful this should be a quick fix in a future update, and this is still useful as is anyway given that we only have to adjust scale with some quick math and the rest seems to be perfect :)
Stay tuned!
@@LaserEverything I am having the same issue with my 70mm lens as well, when I am looking at it the Box extends passed the corners. Not sure if this data is helpful or not.
i have the same problem but I try the 50x50 lens f63.
I also had a lot of problems with my 70 x 70 but I just ran extra boxes and corrected. Mine was wildly skewed incorrectly after the “easy” correction process
I've used this new feature for calibrating 3 lenses, and I agree with the final statement you make (but you gloss over it pretty quick) - after accepting the settings and verifying they've applied, you need to draw a box the maximum size your lens will allow, and mark it, then measure that box.
For all 3 lenses I found that the scale factor needed tweaking. eg for my 110mm lens, when I completed calibration, and then tried marking a 105mmx105mm box, it was producing something closer to 103.5 x 107mm or something like that. It showed me that the scale factors for X and Y were not right. I'm sure it was due to errors in measuring but I had the same issue with all 3 lenses I calibrated.
Just a suggestion - can you add a bit to the end of your video showing marking a final box, then taking measurements to verify your scale is right, and if not dead on, performing a quick calculation to adjust the scale percentage to get your marked box the exact size you need. I had to do this for all 3 lenses, so others will stumble over this too, and showing them what to do in your video would help some with that final step.
I am facing the same issue after lightburn new future, it marks a square 2mm off X side compared to Y side
for example, it marks 50 x 50 as 52x48
It's easy to fix - lets say your marked box should be 50 x 50 but it marks as 48.5 wide x 52 tall.Go to your Device settings. Look at Galvo 1 and Galvo 2 'scale' percentage values.
To adjust:
The formula is: current scale percentage value times desired length divided by actual mark length.
eg if actual mark is smaller than desired (should be 50mm wide but you got 48.5mm wide mark and your scale percentage currently is 91.2329%)
The formula is: 91.2329 x 50 / 48.5 = 94.0545% - set this as your new scale percentage for x axis.
If actual mark is bigger than desired (should be 50mm tall but you got 52mm tall mark and your scale percentage currently is 82.3219%)
The formula is: 82.3219 x 50 / 52 = 79.1557% - set this as your new scale percentage.
Once you adjust the scale percentage for both axes, try marking the 50x50 box again, and you should see it come out the correct size.@@akbarazadazad1453
Shared this above, but just so you guys see too.
I ran into this helping someone yesterday too. We'll be doing some troubleshooting and documenting things to try and help address that. Haven't found a pattern to it yet, but you can get around that by going to lightburn settings and disabling the line item for ignoring items outside the work area. I would enable it again when done, but it should get you through the process.
Additionally, there seems to also be a scale problem sometimes. The rest of the calculations seem to be working well, but sometimes the scale on either X, Y, or both, tends to make the work area too big and scale output accordingly too large. This can be mathed out manually to accurately scale down by making a square of a known size and then adjusting by the % it is over, scaling it down by that value on the given axis scale (over simplified version of the explanation but best we can do in the comments lol). Hope that helps, and we'll try and submit feedback to the awesome team at LB to try and help improve things as we test and try and find details that may help them.
I'm hopeful this should be a quick fix in a future update, and this is still useful as is anyway given that we only have to adjust scale with some quick math and the rest seems to be perfect :)
Stay tuned!
Edit: Just seeing@imakethings17 's shared the formula they used to correct scale. good info, you explained it far better than I was able to in text lol
@@imakethings17 This worked perfectlly, thanks!
@@ssullivan I'm glad it helped. I'm sure others will have the same issue so as long as they follow those final checks and adjustments it should give them perfect settings.
Still useful to have lightburn update explained like so, continue to produce such videos content , it is a pure delight
Thanks for recommending to take a screen shot of the settings. I went into device settings and nothing changed! I wasn't even able to edit it manually until I deleted the reference to the factory provided .cor file. TL;DR my 10 mm squares now render as 10 mm x 10 mm. Thanks!
Glad it was helpful 🙌
Great video as always! Thanks for the feedback on the maximum speed limit. I've increased it to 5000mm/s for the final release.
You are literally the best, thanks man! We appreciate every bit of what you guys do
So glad we can do this now!
Saaaaaaaaaaaaaaame!
1.5 has been great in the betas. Definitely a huge update in functionality.
Agreed!
Yeah I did this today, it set my center point to the upper right hand corner after doing this.
Interesting to see the update! However, I have to point out something important - this cannot account for skew distortion. Note, for example, that both skew factors in your example are exactly 1.000. Because you are only measuring the distances, which is fair enough, there is no way of knowing the angle between the two axes. The axes could be severely skewed, but since you’re only measuring distances (in particular, distances 3 & 4), there’s no way of figuring out skew. This is what frustrated my about using corfile2 - it took me ages before I realised there was no way account for skew.
I manually measure diagonals on ezcad, but you are right!
A bloke called Pythagoras disagrees with your assertion. I'm not saying this is doing it, but from the measurements taken you can plot the position of every intersting line in relation to the centre point. If the 12 lines that make the grid are all equal you must have 4 squares inside a square.
Pythagoras, triangle = diagonals... :)
Skew is about making a parallelogram, al lines are equal length and it does NOT have to be square.
@@timg6252 Nope, I'm afraid that's not the case. The simplest way to demonstrate this is to look at the following:
imajeenyus.com/temp/no_skew_demo.jpg
Pretend that those are two different marked patterns. The pattern on the left is ideal, while the pattern on the right has severe skew. In both cases, the distances between the points (the lengths of the lines) are all 50mm. In both cases, you would measure and enter exactly the same values into either corfile2 or the new Lightburn feature. Yet one situation clearly has skew, the other doesn't.
Now, the instructions do say to measure the distance "from each indicated point to the HORIZONTAL center line". In other words, the red line I've shown on the right-hand drawing. That's fine in theory, but in practice the difference between that measurement and the point-to-point distance will be so small as to be unmeasurable by eye with a digital caliper. For example, suppose you have a 2° skew angle, which is significant. If the point-to-point distance is 50mm exactly, then the perpendicular distance will be 50*cos(2°) = 49.97mm. No way is it possible to measure that accurately. Besides, I can easily demonstrate another marked pattern which has skew, yet where all the perpendicular distances are equal to 50mm.
I spent several months tearing my hair out over how exactle corfile2 performs its calculations, and finally determined that there is indeed no way for skew correction to be performed, doing the measurements this way. Brief background: the COR file that is produced contains a 65x65 array of "displacement" values for both galvos, which is sent by EzCad to the BJJCZ controller board. The controller receives an uncorrected galvo position from EzCad, looks up the array to find how much it should displace that position, and then sends _that_ position to the actual galvos.
Take a look at this graphic:
imajeenyus.com/temp/cor_file_random_values.jpg
That is a representation of the displacements for each of the 65x65 points in the array - the short black lines show the direction and magnitude of the displacement. The important thing to note here is that the on-axis displacements are always ALONG the axis (the short black lines along the X axis are all horizontal, and those along the Y axis are all vertical). Skew correction can never be performed this way, and I experimented with numerous different values.
This isn't to say that the COR file approach itself is bad - it's only corfile2 that's poor. The COR file itself can indeed handle any arbitrary corrections you like - I have written my own script which will perform a 5x5 grid correction based on marking a grid of holes and then measuring the actual coordinates of each. However it's still very much a work in progress.
The only reliable way to determine skew would be to measure _diagonal_ distances, but I'm not entirely sure how the calculations would go....
@@timg6252 (I previously wrote a reply to this, but it seems to have not posted) Nope, not the case. I can show you two different mark patterns, one with severe skew, yet both with all distances equal to 50mm: imajeenyus.com/temp/no_skew_demo.jpg
Help needed, when I run this test I found out that the line in the 5-6 area is not burned as sharp as the rest ,any idea how that happens. Thank you Ben
Hey Ben, hit up the link in the description for our free discord or facebook group and we can check it out!
Hello witch video shows the process of alingning lens with workspace??
I will try this. However, I am concerned as it appears the distortion I am experiencing is different based on file type. For example, if I mask and flatten a png the distortion of greater than if I create a shape in lightburn or use an svg.
Great video and directions. So simple this way! Thanks Kyle
Glad it was helpful!
What height shout I have my 175 laser. It’s not marking on the black construction paper
this is so awesome. I cant wait to get this for my machines that I bought more lenses for!
Hope you enjoy it!
You mentioned in the video you would link another video to leveling the head and troubleshooting, but I don't see it anywhere. I looked through your channel, but wasn't able to find it. Can anyone point it out, please?
Awesome! It just dawned on me that when I did my camera calibration with a Galvo, placing the dot matrix outside of the work area (my lens = 175 but the grid was placed out to the edge of the camera's viewable area) might mess up the camera-to-lens accuracy. Am I right?
I tried this with my 300mm lens but the framing makes it fall way off my bed and making the box smaller makes it fall below the recommended reduced size.
So good, thanks for the info guys!
Our pleasure!
Great video, could talk about taper warp?
We'll absolutely be covering that too, in a dedicated video :)
I imagine the speed cap is to keep an accurate calibration and keep vibration down?
There are other values that handle that for us, with delay and timing values. The LB team had a set of parameter limits on it and has since been updated :)
My lens not parallel working table. Tell me please, how to fix it ?
We have a great video for that. ruclips.net/video/JwWV6Hki5Dk/видео.html
before you do this test you need to make sure your lens is in focus correct ?
100% yea. Always in focus
So these settings are per lens, and not global ? Haven't seen anything where you can pick a particular lens, other than inputting the size in the 1st step
Ideally, you'd set up a "device" in lightburn for each lens size + machine. So if I have 3 lenses for my 50w fiber, I'd say 50w fiber 110 lens, 50w fiber 200 lens, 50w fiber 300 lens, as an example. This way each lens for the machine will have its own profile and be tuned. We do the same for parameter libraries as well, since they change based on lens size and such. Each correction is specific to the lens, like a fingerprint. So for best results, you'd do this process for any new lens setup. We have an updated video coming that includes the whole process, start to finish on a new lens config and tune (updated from this even, its in editing).
Not seeing that feature in the Lightburn 1.5.01 version
Be sure you have a galvo device selected, this feature will not appear when using a gantry device.
So does this mean i dont need to import and delete drivers and swap back and forth? That the only reason why i havent swapped to lightburn yet
That is correct. Note it's only on the current public beta and not available on the normal channel as of yet. When this update releases on the public channel we'll make an announcement about that as well. Otherwise - you'll just want to be sure to grab the current BETA version of the software to access this feature.
Awesome I may finally be able to use lightburn
Does anyone know what to do when the error message of "one or more shapes are corssing the edge of the machine workspace and will not be sent." happens? Please help!
I ran into this helping someone yesterday too. We'll be doing some troubleshooting and documenting things to try and help address that. Haven't found a pattern to it yet, but you can get around that by going to lightburn settings and disabling the line item for ignoring items outside the work area. I would enable it again when done, but it should get you through the process.
Additionally, there seems to also be a scale problem sometimes. The rest of the calculations seem to be working well, but sometimes the scale on either X, Y, or both, tends to make the work area too big and scale output accordingly too large. This can be mathed out manually to accurately scale down by making a square of a known size and then adjusting by the % it is over, scaling it down by that value on the given axis scale (over simplified version of the explanation but best we can do in the comments lol). Hope that helps, and we'll try and submit feedback to the awesome team at LB to try and help improve things as we test and try and find details that may help them.
I'm hopeful this should be a quick fix in a future update, and this is still useful as is anyway given that we only have to adjust scale with some quick math and the rest seems to be perfect :)
Stay tuned!
eta of final release?
I don't think I'm up to date on that information but when we have news we will definitely share it.
My leaser source. Jpt 100 Walt. Marking. With shadow?? I am in big problem. Help me if you can ?
Join one of our many support communities I'm sure someone can help. Definitely will need to see some pics. Links to everything in the description, lots of free options.
No matter what size I set my working area of the lens, it makes the same size square when I frame it.
It’s not saving to my device settings. Has anyone else had this issue?
As clear as mud, I have no idea what you are doing this for!
Ok? Do you have a laser, are you using Lightburn?
At 5:56 you mention a video link for leveling the lens to the bed. Can you link that please? I must have a terrible combo of blind/dumb and clueless, as I don't see the link. If its this one: ruclips.net/video/FZT9zOp9Sl8/видео.html than my lens calibration is the problem this vid should help me with. Lasers for knife engraving accurately is akin to hurting kittens and pushing string.
Way of the Dragon.
🐉
Still useful to have lightburn update explained like so, continue to produce such videos content , it is a pure delight