Cycle time for workflow?

Only ten more days and I’m back in Corinth.
But I’ve been doing quite a lot of thought experiments re my fans, and the development of a good work flow.

I’m currently pondering the possibility of using the pass through slot to enable a continuous, repeated cutting of a pattern.

Can anyone give me a heads up on the time delay should one just need to keep hitting the print button ?

Let’s assume I have a simple cutting job that takes 3 minutes cutting.
After the completion of the cut, the ‘autofeed’ device outside the forge, pushes a new sheet into the correct position, and in doing so, the cut piece is on its way out. I can’t remember if the pattern is still loaded, or if you have to start again from scratch.
Assuming it’s still loaded, complete with settings, I wonder if the lid has to be raised in order to continue. If not, what is the sequence that gets the gf ready to print ?
I think it needs to scan the material again, process the design(?) then reload, and turn on the magic button.
I wonder just how slow/fast that might be. It might turn out to be a large proportion of the whole cycle.
Ugh.

:upside_down_face:

Basically you’re starting from the point where the design is loaded in the UI. You don’t have to open the lid or anything, but you have to press print in the UI, wait for it to scan, process, and upload, and then the button flashes. How long that takes depends on how complex your design is, how fast your network is, and how the Glowforge servers are feeling at the moment. It’s probably not all that consistent between runs.

3 Likes

Thanks Chris, that’s the sort of pathway that I feared would be the case.
I think I’ll just have to allow double the 'cut time, and fit another operation in there.
Possible a cleaning cycle which might also be proportional to complexity, and therefore cut time as well.
John

If it is three minutes cutting time the material scan will be minimal and the lid does not have to be raised as I have done multiple changes without touching anything. In theory you would just have to hit print in the computer and then the button. In the few seconds it takes me to get from computer to Glowforge the scanning is usually finished.

1 Like

That’s a seriously good bit of news.
Thanks for the heads up !
I can see that it’s going to be difficult to know ahead of time, but it does give me a bit of a push towards getting the ‘external feed system’ worked out .
:upside_down_face:

1 Like

The main thing I would look for in a feed system is accurate measuring. If you could feed precisely eight or ten inches or even 6.37 inches then something like a long banner could be made by putting in all the pieces in one drawing.

The main problem at the moment is knowing exactly what the change is and matching it to the design. I have been running a cut at 20 pews and 500 zooms to see where everything will land but mostly it is pretty much guesswork as you cannot do that more than a couple of times.

1 Like

My current thought is along the lines of an in- and out- feed table, with stepper motor driven rollers, and an optical sensor moving the pieces. Each would have a complete set of parts for one unit, and the pieces would be pre cut (on the forge) to minimise waste.
This, rather than have a repeat design and long sheets of material feeding through.
Seeing a vertical stack of small pieces might be more easily accommodated in the garage than long bits, and just handling might be easier.
:upside_down_face:

2 Likes

In that case you would need some sort of a bar feed as it is about an inch from the slot to where you are able to cut and I have seen the nocut zone change so that not moving anything the cut just made is now an eighth of an inch onto the design and the GFUI says there is no artwork there, I made a single line design to cut off the unused part of a proof grade sheet so the cut area on the piece could be used near the center of the cut area but even PGA sheets cannot be cut top to bottom in a single cut leaving about a 2 inch distance uncut. I rotate the sheet but then again alignment is an issue and I need to do a very light score to see where the cut will fall. And even then I see as much as an eighth of an inch difference between the light score and final cut.

I am doing a 3 1/ 2 hour cut now and it is about a sixteenth of an inch from that fast score which will not be a problem if the various finial cuts lune up correctly.

Something isn’t being done correctly if you’re seeing a variance from job to job. The repeatability of subsequent operations is about .0007” of an inch.

1 Like

While I am experimenting I have been running one thing at a time before moving to the next but if I run what will be a cut at ten pew and 500 zooming will score where I would hope it would cut at full power and 150 zoom but even that can be a bit off. I get the parallax problem but the same cult should land in the same 0lace on the same bit of wood but it frequently looks a bit off.

I think your machine is faulty if cuts from the same file move around.

There is, or at least was, a bug where each run is off by one stepper step but that is less than 1000th of an inch.

2 Likes

Also the time to prepare your job will vary and get longer as your cut job gets more complex. It can be 30 seconds, it can be 5 minutes.

I find the scan/prep process to be annoyingly slow actually, but a: I tend to cut pieces with lots of detail and b: I sit 5’ from my machine so “walking over there” is more like “turning around”.

2 Likes

I might guess that my total cutting ‘length’ could be about 250" - 350", made up of 12 units about 8" x 1/2", with say 6 holes, making up a length of10" on each.
It’s going to be a pretty fast cut, so I might well double up in the art work file., but you can see that the setting up time for the GF will have a big impact on the cycle time.

I had a discussion about this here:

I’d love for them to add a repeat last job feature that would minimize the time between jobs. I hope it’s in the hopper.

3 Likes

Hard to know how long it will take. I will say that the GF handles svgs with fewer paths and nodes in general better than those that have more. I tend to select all my pieces that are the same color and make them one path, to minimize the number of paths I’m sending on larger pieces.

From what you’re describing though this doesn’t sound like a huge piece, but it’ll depend a lot on specifics. 12 rectangles? No problem. 12 highly detailed curved lines with 1000 nodes apiece? You’ll be waiting a minute or 5.

1 Like

@polarbrainfreeze - Thanks for posting that discussion. I should have guessed it would have come up before, and glad to see that it at least progressed to be ‘in the hopper’ .
If it ever comes out it would be a great improvement imho.

:upside_down_face:

I haven’t used the pass through in a month or so, but there were some corner cases where I had to lift the lid to make the software happy. I can’t remember specifics and they may have fixed the bugs. But if you run into it, whatever it is that I can’t remember, you were warned.

Err. … thanks.
:upside_down_face:

1 Like