Time and material saving technique

[Doug Griffith]

Hi All,
I wanted to share this technique to anyone who can utilize it. My project requires 1000 of the same item vector cut from .177 acrylic. I approached it 3 ways, timed each sheet (with the lid open), and did the math:

1) I ganged up as many as possible on a 12" x 24" sheet.
Hours to process: 19:27
Parts per sheet: 32
Sheets required: 33
http://www.dogcollarlabor.com/IMAGES/sawmill/single.png

2) I ganged up as many as possible on a 12" x 24" sheet utilizing as many common lines as possible. I then removed any overlapping lines.
Hours to process: 14:06
Parts per sheet: 38
Sheets required: 27
http://www.dogcollarlabor.com/IMAGES/sawmill/group.png

3) I ganged up as many as possible on a 12" x 24" sheet and and then broke the elements into layers. I then optimized path directions and flow from one part to another.
Hours to process: 14:37
Parts per sheet: 38
Sheets required: 27
Layer 1
http://www.dogcollarlabor.com/IMAGES/sawmill/layer1.png

Layer 2
http://www.dogcollarlabor.com/IMAGES/sawmill/layer2.png

Layer 3
http://www.dogcollarlabor.com/IMAGES/sawmill/layer3.png

Layer 4
http://www.dogcollarlabor.com/IMAGES/sawmill/layer4.png

The results:
Method 1 uses 6 more sheets than the other 2. It takes about 5 more hours to process. The 6 material changes take a bit of time for this many little parts.

Method 2 is the fastest but has the worst quality edges because of all the starts and stops.

While method 3 took about an hour to optimize the file, it was almost the fastest and gave better results than just removing overlapping lines.

I am going with method 3.

Cheers

[Mike Null]

Doug

Good job of demonstrating your problem solving technique.

Thanks for sharing.

[Mike Mackenzie]

Doug,

This looks like a perfect application for a dual head this would cut about 30 to 40 % more off of the time.

[Rodne Gold]

That's a good idea for a sticky thread, "various ways of optimisation and cutting cycle times."

Option 1 would have also given you incredible problems re warpage when you cut , the objects are too close.
We once had 1000's of labels to do and by a few methods managed to cut cycle times from over 1 min to about 14 secs an item.

[Bob Cole]

What is the best way to remove the overlapping lines?

[James Stokes]

In x-4 there is a virtual segment delete tool. With it you just click on an overlapping line and it removes it.(That is the one tool that made it worthwhile to upgrade to X-4 ) I think X-3 has it also.

[Doug Griffith]

That's a good idea for a sticky thread, "various ways of optimisation and cutting cycle times."

Another helpful method when cutting a bunch of small parts is to run a strip of masking tape across the cut parts while still in the laser. Then lift out the entire sheet, reload, cut away, and weed out on a separate table.

I removed overlapping lines manually.

Cheers

[Ed Newbold]

Another helpful method when cutting a bunch of small parts is to run a strip of masking tape across the cut parts while still in the laser. Then lift out the entire sheet, reload, cut away, and weed out on a separate table.Great tip. I, like many, spend waaay too much time weeding and cleaning out the machine between sheets of material.

Thanks.

[Richard Rumancik]

Doug, thanks for taking the time to sketch this out.

I see your logic in picking the option you did but I thought I'd throw this out for anybody interested.

I'm partial to option 2 for a few reasons. I like to take a few parts off the machine and inspect them before I make a bunch. By using Option 2, I can make a few, pause the machine, carefully pull out a couple (the masking tape trick is sometimes helpful) and measure them. That way I don't have to make a whole sheet before I have a sample to inspect. Another reason is that any glitch along the way won't destroy a whole sheet of parts. Third reason is that I can monitor how well they are cutting as the pattern progresses down the sheet. My laser does not cut as well in the lower right so I have to ensure that power does not drop off too much. By using Option 2, I can adjust (more or less) on the fly.

Another reason is that sometimes the shape may change (error, or customer change on next run etc). It may take a lot of work to incorporate a design change using option 3. Using Option 2 you only need to update the one entity and then reduplicate. It is fairly quick.

In some cases you may need only half a sheet. Option 3 is good for full sheets only. Option 2 allows you to stop the job anywhere, and even use up smaller scraps of material.

Another idea: when I do a sheet like this, I try to visualize the minimum entity that that makes a logical grouping. Sometimes it is one part, sometimes two back-to-back. To avoid having to erase a lot of lines, I try not to put them there in the first place. See my sketch below (it is a bit crude as I only had the bitmap.) I took your part with two pieces back to back as you did. Then I erased the right line. At this point I would combine segments, set cut order, and set cut direction as needed. This then becomes a group. This group can then be duplicated to fill the sheet. Note that you will need to add one line manually to finish the row. This is pretty simple to do.

Using this method I can plot the first group, pause, and do a dimensional check. You could also do your inspection part on a scrap of material if desired.

Also note that I extended the line on the left fully top to bottom (through the notch). Sometimes doing this will provide better speed and smoother cutting.

Every circumstance is different and Option 3 obviously worked in this case.

[Doug Griffith]

Doug, thanks for taking the time to sketch this out.

Thanks Richard,
I would like to see other tips/techniques to make our lasering lives more efficient and proffitable. I also agree with Rodne that something along this line would make for a good "sticky".

Cheers

[Eric Allen]

That would make for a great sticky, I'm sure every time I spend the afternoon agonizing over optimization to cut head motion that there must be a better way:)

[Rodne Gold]

Here's a few tips
Use colours to gang or cluster items , for example if you have a large space between 2 items , its often faster to do each seperately then let the laser scan back and forth, make one one colour , the other another and the laser will do 2 entities.


Do NOT be afraid of dropping dpi , often the difference between 600 and 300 or 500 and 200 dpi are negligable to the eye but will save you a fortune in time.

Continuous borders are often faster by using vector engraving of multiple offsets. Do the border as a vector , use corel to make 4 or 5 "inlines" spaced as you want and then vector these and raster the stuff inside the border. Often vactoring out of focus will give a far wider line width and require fewer "inlines" (called offsets in cad) to acheive a given engraved thickness.

Ideally you want to optimise speed when lasering , so when determining settings always start at 100% speed and vary power, only drop speed when you run out of power or your laser doesn't do high speed well.

IMHO... NEVER do multipass lasering , this leads to huge problems in most cases , if you cant cut thru/engrave in a single pass , then rather pass on the job. Multipass vectoring does not eject melt or debris from the kerf and often you just waste power getting thru this to the next pass, sometimes you will never cut thru regardless of the no of passes, misregistration can be a big issue as the part can shift or the motion system does not return to the exact start point , even a few 1/100th of an inch is problematic. Multipass engraving just burns over existing dross left by the first pass and can result in remelts or excessive heat affected zones.
This is apart from the time overheads of 2 or more passes. There is a case for doing 2 passes and thats a clean up , often materials like pex have a crusty ridged look when engraved deep , a polishing pass at high speed reduces this.

Large solid areas of engraving can be done a lot quicker if you drop dpi radically and engrave out of focus.

If you want speed , never use vectoring to outline engraved lettering etc, the vectoring of the outlines will take massive amounts of time.

If your driver does not do optimisation , then use the object manager in corel to do it , ie reorder the parts in the object manager so the laser doesnt skip around.

Often sent artwork is overly complex , it sometimes saves laser time by you redrawing it , especially vector art.

On smaller runs , its often more time effective to do items on a one by one basis than trying to make matrix type jigs and setting em up. The engraving time might be shorter on a matrix , but often setup time is way longer and more complex.

On long thin type engravings , set em up to be done horizontally rather than vertically.

On step and repeat multiples , its often quicker to do column by column than row by row , depending on your laser.

On small runs with logos, you will often save a lot of time by using photograv on less than stellar jpg type artwork then redrawing it , on screen the job might look scrappy , but the lasered output will look just fine.

Always bear in mind that your idea of good quality might be a magnitude over your customers , spending time getting it to your standards which might be total overkill will not endear you to your customer who would most likely be happer without getting the nth degree and paying for it.

Try quote and price on a job/item basis rather than time , often you leave money quoting on time , a $1 label taking 10 secs should not be charged at 1/6th of the minute rate. The reason for optimising speed is not to pass a saving onto your customer , it's to enable you to do more jobs in that time.

NEVER underestimatge time taken for placement , loading, unloading unpacking , repacking etc. As much as you optimise laser time , this too is significant.

Make sure your laser itself is optimal , like check alignment , focus distance etc, if you putting out max power , you can often do the same job quicker. If your 25w laser is only putting out 18 w due to misalignment , scratched optics etc , you can lose 25% of your time and get bad results.

[Doug Griffith]

Always bear in mind that your idea of good quality might be a magnitude over your customers , spending time getting it to your standards which might be total overkill will not endear you to your customer who would most likely be happer without getting the nth degree and paying for it.

Great tips Rodne.

I am most guilty of the above. It is easy to get caught up in accuracy for parts that are either cosmetic or so small that the kerf of the laser makes high accuracy virtually impossible. I try to educate customers and even get them to change their design if it will save me time and them money.

While speaking of accuracy, drawings meant for CNC output do not account for the thickness of the laser. The CNC postprocessor sets the tool diameter, etc... Files that are laser cut need to have the lines offset 1/2 the thickness of the beam outside or inside depending on the cut. Scaling the entire drawing the thickness of the beam will throw all the dimensions off.

Another tip:
For small text, vectoring hairline characters may be faster.

Cheers

[Rodne Gold]

Im sure someone more experienced than I would be able to write a Corel script

Select relevant objects
Select contour/inline/outline distance
Select side to offset - inside or outside
Go!

I do know there is a corel script out there that allows a set of selected objects to be "remembered" without grouping them........
Otherwise , surely if you put insides on one layer, outsides on another and scaled each , it would also work? Expand outsides , shrink insides?
There is a big problem with contours and offsets, not everything can be offset.