XY home position

Thanks for the explanation. I know that the limit switches in my 3D printer have what look like very inexpensive snap switches:

whereas the mill and lathe have much beefier switches:

Both types work but it looks like the printer manufacturer considered guarding and protection to be more critical than the mill or lathe manufacturer. I haven’t given it a lot of thought but properly installed switches should need very little maintenance and they might be cheaper to replace than a malfunctioning camera.

It’s going to be interesting to see what some of the experienced design engineers here think of the way the GF has been built, both design and fabrication, when it is out in public release. Folks like @rpegg, @jamesdhatch, @jdodds, and a few others.

3 Likes

There’s a lot going on in this thread (teach me to miss a few days!), so I’ll take a stab at the original question, then please split out anything I missed into separate questions.

No.

What we do is put the material in the same place in the bed each time - then just hit the button to re-send the job. [quote=“takitus, post:7, topic:3386”]
I didnt think to check tray wobble at the maker faire
[/quote]

It’s pretty dead stopped. It positively registers four feet in four dimples on the bottom. You just have to be sure to wipe them out if you remove the tray so the tray sits tightly in them.

It’s the opposite. If you mold a ruler into the tray, then the ruler might not be perfectly parallel to the laser head’s line of travel - tiny tolerances could mean it’s just a smidgen off. If you cut it, then by definition it is perfectly parallel to the laser’s line of travel.

One more thing - it’s always super helpful to us when you describe a specific problem you’re trying to solve with an example project that you’re thinking about. Helps us much more than “please implement this feature other lasers have”. :slight_smile:

12 Likes

How are you doing that without something physical to ensure it’s the “same place?”

1 Like

There are also optical endstops that require no physical contact and can be quite accurate.

5 Likes

I might want to put something like that on my lathe to improve referencing accuracy. The mechanical limit switches repeat to only +/- 0.015" or so. Do you have any idea how repeatable switches like the one you pictured might be?

1 Like

Affix a marker or stop to the bed.

2 Likes

Or inductive.

I use these on all my CNCs. Completely Solid state. Since upgrading those machines with absolute encoders, there are nice to have just in case the blank hits the fan.

1 Like

I have a few of those from an Ebay seller but have yet to test them. They were inexpensive Chinese imports so I’m not sure how much to expect from them. Do you have any idea of how well yours repeat?

Before closed looping they worked great. I’ve have 1000s of hours of use on them and a good majority of that where duplex ops with fixtures.

Basically run a job for 4~5 hrs. Pause at the end of the day. Got back in the morning. Re-home and let it go.

Sounds encouraging - I’d like repeatability of 0.0005-in or better. Guess I should set up my own test.

Assuming your machine can hit those types of tolerances then yeah. When I had the encoders in DRO mode vs. torque/correct. I was easily able to get 0.01mm. I literally spent 2 hours jogging the machine to a random point on the work surface then home. Then giggle for 30 sec. then doing it again. It’s the small things in life I guess.

6 Likes

By the way, just to give a sense of what you should expect from your Glowforge, a typical test pattern has the machine draw a box, run for nearly two hours, then put little x’s on the corners of the box. They all land perfectly.

13 Likes

In my experience the optical stops are highly repeatable as long as nothing changes or gets obscured. But half a mil might be asking for a lot, because ultimately they’re taking the analog signal of how much light is getting across the gap and doing a threshold to turn that into a digital signal, so electrical noise, power-supply ripple, ambient light blah blah blah. At least some of them do a quick home, then back off and run much more slowly to get the final position.

Inductive stops also highly repeatable unless the configuration of inductive materials around the stop changes or the electrical situation changes. BTDT.

Which is a longwinded way of saying “Yeah, you really are going to have to run your own tests.” :slight_smile:

4 Likes

Hehehe, those results that have you repeating the operation just for the enjoyment of the giggles it inspires are golden!
It may be a small thing, but it’s effect on the psyche is huge!

2 Likes

My mill is roughly +/- 0.001 in any axis, which is nice at times but re-referencing to establish part zeroes with a 3D probe is very fast, even manually. Faster yet with an electronic probe.

The lathe is much harder, or at least more time consuming, to re-establish part (G54) zeroes and an error of 0.001-in in X gives you a 0.002-in error in diameter, so a more accurate and repeatable way to reference can pay some significant dividends, especially with tendency to accidentally bump the E-stop switches, which forces a re-reference by the control software.

I can identify with your test procedure and the outcome. Giggles are so much better than spewing oaths when something goes belly up.

1 Like

That’s often the case when you are fussy about results so perhaps it’s best to just get it done. Just part of the game I guess. Metal swarf should mostly stay away from the switch locations, at least for the materials I’ve turned so far, so at least that is not likely to be an issue.

1 Like

@dan I think you’re still missing the basic question or I must still be missing something(?) If there’s no physical zero guide on the bed, how would we know the exact size for a guide and where to place it in order to ‘put the material in the same place’ AND at the correct angle?

Kindly clarify this for us.

Thanks

3 Likes

Dan has been a bit busy recently…he might be out of pocket. :slight_smile:

Are you talking about just engraving on something, or do you want to cut it out of larger material? If you’re cutting it out, you just put the new sheet back in the same place, take a quick look to make sure that all the cut lines still fall on the material, and resend the job. Any engraving inside of the cut lines is going to line up perfectly with the cut lines, as promised.

If you are wanting to laser engrave on one or more pre-cut objects that already exist, Dan has recently explained how the Glowforge software will allow for several options here, using engraving on almonds as an example:

I’d personally probably just create a jig, but there are apparently several other options available to us inside the software.

1 Like

I think the issue here is accuracy.

Say I have an object that is already engraved with a specific pattern, and I need to take exactly 3mm off of the top and left sides, anything more or less will ruin the piece. with a home position of 0,0 that is exact, it will be easy to do every time without worry, because we can set up something in the corner to align it with.

If there is no consistent home, to get an exact cut, we will have to make a new jig EVERY TIME. massive pain in the butt. That is to say there is even the ability to use the last home position as the home position on the next cut.

I have to do this on my CNC machine and would prefer not to, but on my chinese laser I cut a corner piece when I first got it so that I know where 0,0 is and I can cut a 4mm piece off of something without a care. Repeatably. Not being able to do this on a glowforge would really take a lot from the product. Its almost like turning a digital machine analog, and that would be a terrible thing for accuracy, and quite a shame

4 Likes

Unfortunately, the mass consensus is ‘I don’t get it- Just make a Jig’

But to do so for every single op is going to be a freaking pain.

4 Likes