Power outputting variation?

Hi. Just to check. Has anyone ever experienced instances with a laser cutter losing power as it goes further from the zero point of the axes?
Will the glowforge have that happen?

Yes. On flying optics configurations like the GF power drops as beam path increases due to various factors (mirror misalignment, amount of smoke in the cabinet, dissipation of the beam, etc…). Plan your cuts accordingly!

edited to add: It would be totally amazing if GF could compensate for that in software but modeling power drop would be really difficult.

Only the GF team can answer that question at this time. But that is one of the reasons the tube is mounted on the gantry, UN-like other makers.
However, @dan, @aeva, or one of the other team members would be the best people to ask about that.

Yes, I have had that happen with an Epilog laser I was using in a Maker space. The closer it was to the 0x0 point, the weaker the laser was. I think it was just a problem with maintenance on the laser, and eventually it got fixed. But for a while, I was doing all my cuts on the right hand side of the bed so that it cut consistently, and avoided the left 8 inches of the bed. It was limiting what I could cut.

Hopefully this will not be an issue with the GF. Although, as I said, it was fixed with some proper maintenance.

:slightly_smiling: I have the opposite issue - my drop off is on the lower right quadrant. For me it’s a combination of beam focus drifting at the extremes of the x & y gantries and smoke residue. I now clean it before every run as even a single run shows an appreciable decrease in delivered power - mirrors get a little dirty but the lens gets really dirty. Lots of folks miss this maintenance. Sometimes I’ll do multiple faster passes vs a single slower speed full cut so I get less fuming/smoke from the material.

The laser tube in the Glowforge is actually mounted on the gantry, rather than being fixed at the back of the enclosure, so it moves with the gantry as it travels along the Y axis (fore and aft). I am guessing this decision was made by the Glowforge engineering team specifically to mitigate this attenuation issue. They have also stated that the optical path is fully sealed. I am struggling to picture how they managed that trick though, given that the tube doesn’t move left to right with the head as it moves back and forth along the gantry. Given these two points, and assuming the beam does still have to travel through the air in the cabinet from the business end of the tube to the head (and therefore potentially through some smoke), it seems to me that the attenuation as a function of head position could only exist along the horizontal X axis in the Glowforge. However, assuming the laser comes out of the right end of the tube, that attenuation has also been mitigated by the strictly right-to left directional nature of the airflow within the cabinet,which would ensure that any smoke generated would very quickly be moved to the left of the head, and therefore out of the beam path.

Here is the post stating the direction of the airflow:

Here is a link to a long video on YouTube of the Glowforge at work at the open house a couple of months ago, but you only need to watch the first few seconds to see the tube moving with the gantry carriage.
of course, now that I look at the videos, it seems clear that the business end of the tube is on the left and there is no window on the right side of the head, so the airflow direction would not be of any help to clear the smoke from the beam path if it does go through the air from the left end.

Knowing what I know now, I don’t think it was a power attenuation, but rather the laser was not focused as well when it was near the origin.

I’m not sure how that could occur, but anything near the 0,0 point came out with really thick lines that did not cut as deep, as if the laser was defocused in that area.

2 Likes

Auto focus is going to ROCK!!!

2 Likes

On the version we saw at NYMF the business end of the tube was on the left along with a naked #2 mirror. Things have certainly changed considerably since then, but keeping the power side of the tube on the right (near the electronics) and the lens side on the left makes most sense.

@polarbrainfreeze has it mostly right. The issue with different power at different spots on the bed is a problem with mirror alignment. Aligning the mirrors is a monstrous PITA, and can often have unexpected outcomes for the inexperienced maintainer.

If you align your mirrors perfectly at one XY location, and never look at the others, you can pretty easily have a minor misalignment (common issue when aligning mirrors at (0,0) position) which is magnified at farther reaches.

But, when you align your mirrors perfectly at a different XY location, again never looking at others, you can have a paired misalignment. In this case (for an example) the first mirror is aiming the beam too high, but the second mirror compensates, so still bounces that too high beam to the third mirror, but even THAT shot could be off, with the third mirror compensating by not being properly aligned anywhere else to get to the focusing lens (so then you are beyond a paired misalignment). Something like this will cause power variation across the entire bed, but also give you some areas which are complete dead spots.

The enclosed optics of the Glowforge will really shine in this regard. You should never mess up the factory set alignment, since you never touch or bump the mirrors themselves.

The reason it is misalignment and not just distance is that a laser beam is collumnated light (Pretty sure I spelled that one wrong…). So all of the light is traveling essentially parallel with one another, which gives you nearly zero loss of power over distance. In a laser pointer this is observed by measuring your spot size at various distances.

3 Likes

That sounds a lot like what I was experiencing. I KNEW I was mostly right :wink:

1 Like

Yes, the effects of poorly aligned mirrors will be MUCH larger than any other factor. But we get power drop even on perfectly aligned machines. It happens. We can argue about why it might be happening but it’s something that hopefully the much smarter than me engineers at GF have resolved. BTW… that collimation can be affected by the mirrors. On every desktop hobby cutter I’ve used (even the super expensive Epilog/Trotech/Universal machines) the spot size is bigger the further you get from the mirror.

edited to add: Most of those machines were fairly old and heavily used so the mirrors might not have been pristine. Having enclosed mirrors might help keep things working properly for a long time. I wonder if the GF folks have tested any of the prototypes for something on the order of thousands of hours of cutting and noticed any change in the optical performance.

If someone with a Beta unit happens to have a fog machine handy, maybe we can get a nice simulated full emissions test going.

Dan mentioned having done a 2 layer of paper engrave, where they burned away the top layer of paper leaving the bottom unscathed. Doing that for the full bed size would be a quick and dirty test of power across the entire frame.

2 Likes

The beam path isn’t fully enclosed - it’s open from the sealed mirrors on the left to the flying head - but the optics are sealed. If I stated otherwise, I misspoke.

We haven’t measured any attenuation across the bed, which is good.

There are many factors that can cause attenuation, including one not mentioned here - the beam diverges slightly, and a larger beam size focuses to a smaller spot size, increasing power density and cutting ability the farther you get from the tube.

But the dominant factor by far is normally poor alignment. We’re solving that for you so you shouldn’t see any issues at all.

4 Likes

There’s a dead easy and cheap way to test this. If you want to be all fancy and rigorous about it you can use an expensive laser power meter. But the poor man’s way is to use some thermal paper. Stick one of these on and hit it with the same parameters at the near/far ends of the bed. Good enough for relative comparison without objective measurements. Will do some tests tomorrow and post pics. Would be great if GF could replicate the tests to verify that there is no significant loss.

Not only did Dan say they did not see any loss, but you must have understood that, as you “liked” his post.

I saw that. Just wondering if they could do a visual test and post the results. 1 picture == 1K words.

Here’s the paper test you mention + the thermal paper marking test I described earlier. Both quick and dirty tests but give a good sense of any relative differences in power at various locations without having to take objective measurements.

First pic is thermal paper mounted in front of the #2 mirror and hit with the same cutting parameters at the near and far ends of the y-axis. I chose to test along the y-axis because that only requires bouncing off the #1 mirror, so alignment is controlled for. It’s basically a straight shot at the paper. Power was calibrated down to where it just gave a nice full mark at the near end, then repeated at the far end. The y-axis travel on this machine is about 600 mm, so just a tad bit longer than the travel along the GF x-axis (where we might expect to see attenuation on the GF). The results don’t really need any explanation.

To test whether the apparent attenuation from the first test actually affects cutting results, I tried the 2-layer paper test you suggested. Power was set waaaaaay down and speed was turned up to 11 to just get through a sheet of thick paper (not sure the exact weight but it isn’t copy paper…no way we could only make it through that on an 80W laser cutter). When the settings were dialed in at the near end to get through the top layer while just barely marking the bottom layer, the cut was repeated at the far end. Near: Go; Far: No Go.

Would be interesting to see pics from similar tests on the GF when the crew gets back from CES.

On what cutter? Without knowing what brand you are doing the test on, it’s kind of useless info.
And are those vector circles or polys? and the scale?