Glowforge release

[Lee DeRaud]

Again- I don't know any of this at all. I'm just playing devils advocate to point out how it's possible they're not doing any actual path planning or "motion control" on the machine end. I don't think it's an ideal solution at all, I just like thinking about how things may work

Got it. The thing I keep banging my head on in that regard is the assumption that GF went with a cloud-based motion planning/control setup to (1) leverage cloud server power and (2) make the hardware simpler. The first seems a bit pointless: it's a fairly trivial problem for anything worthy of the name "computer" built in the last decade. And, based on the hardware they ended up stuffing into the beast, the second didn't happen.

[Bert McMahan]

I totally agree. I'm not sure what "power" their microcontroller is lacking in; motion planning isn't exactly the most processor intensive task.

[Art Mann]

I think Glowforge is trying to do for laser cutting what Paint-By-Number kits did for aspiring artists. They are purposely removing control and decision making so that craft persons can successfully complete simple canned projects. That is one reason why the system is as dumbed down as it is. I believe that, if it survives, it will not be very popular with professional engravers, any more than paint by number is to professional artists. Another possibility is that Glowforge wants to use the same philosophy as CarveWright CNC routers. These machines aren't compatible with any other machines or software and require you purchase all enhancements, tooling and replacement parts from them. That is a great profit model if you can convince people to buy into it.

[Scott Shepherd]

I wonder if the extremely long run times and very slow rastering is a result of moving the motion control to the cloud. Ever watch videos of them rastering? It's painfully slow for rastering.

[Bill George]

If they were smart and thinking ahead what they should have done is just use the "cloud" to create the gcode files. Kind of like VCarve does for the router people. The post processor could prepare the gcode, send it back to the machine as a complete file and the machine controller does the work. Takes the hobbyist out of the picture, he just uploads his project and gets the gcode sent to machine memory. Set the materials, and hit the Go button. Simple.

[Dave Sheldrake]

I wonder if the extremely long run times and very slow rastering is a result of moving the motion control to the cloud. Ever watch videos of them rastering? It's painfully slow for rastering.

I noticed that Scotty, slow.....for anything other than hobby use the speed is simply worthless

[Scott Shepherd]

I think the basis of it is based on a mistake. When the story is recounted, Dan says to his electrical guru "Why can't we do it in the cloud" and his guru stares blankly and doesn't know. The answer is simple. Because it's the wrong way to do it. You don't just ask a question, get someone saying "I don't know" and then base your entire company on that idea unless it's proven to be a solid solution to a problem that exists. Right now, no problem existed. NC machines have been around since, what, the 50's? Motion control isn't new. Steppers, servos, open loop, closed loop, it's all been out there for 50 years or more. And people do mind blowing stuff with it. I remember when CAMM systems started to become more powerful than the CNC controls. Prior to that, complex surfaces weren't able to be done. I recall running software on a UNIX based system that could create point to point commands for some really complex shapes. We were starting to get G-Code programs with 100,000 lines of code and the machines were just choking on them. 20 years ago, High Speed Machining came along (mainstream) and you were seeing motion controllers doing mind warping speeds for rapid and for cutting, especially in aluminum. The technology is there and companies have spend billions of dollars over 1/2 a century to fine tune it and make it the most efficient. I seriously doubt that a "What if" comment off the cuff negated all the work all those companies have done for decades. All the engineers, all the R&D, all the applications were machines had to process a lot of data very fast, and not one company said "Hey, you know what would be a great idea? Let's rip the motion control out and put it behind an internet based system"?

There's a reason for that. Because it's not a viable, smart solution for the engineering of a system.

That's my opinion.

[Bill George]

There's a reason for that. Because it's not a viable, smart solution for the engineering of a system.

And that's why the GlowForge will fail in its present form. The concept of a easy to use, almost fool proof unit for folks who do not want or need to learn, they just want it done is a good one.

[Kev Williams]

Well gee, this thread's 2 year birthday was yesterday, and nobody noticed...

I have to marvel at all that's happened in the past 2 years.

And what hasn't...

[Jerome Stanek]

Well gee, this thread's 2 year birthday was yesterday, and nobody noticed...

I have to marvel at all that's happened in the past 2 years.

And what hasn't...

Yeah the people that have received their units are having lots of problems just like the K40's.

[Bert Kemp]

Yep but we all new that was going to happen. I read everyday about the problems people are having.

Yeah the people that have received their units are having lots of problems just like the K40's.

[Dave Sheldrake]

The alignment issue with the overhead camera is a reasonably easy fix but it seems nobody at GlowForge has any idea of how to do it. (it's the same problem that affects Galvo's and has been compensated for many years)

Just yesterday I was asked by a GF owner why all their items are out of alignment when they use the camera on their new machine

[Bert Kemp]

Heres the latest one just showed up
Her Glowforge:Stuck on "Needs Calibration" , have rebooted everything with no movement at all. Refresh bed image won't refresh, have centered laser head so logo is directly under camera. Lights are on but no noise and no movement from laser head.
Everyday more problems.

[Keith Downing]

I think the basis of it is based on a mistake. When the story is recounted, Dan says to his electrical guru "Why can't we do it in the cloud" and his guru stares blankly and doesn't know. The answer is simple. Because it's the wrong way to do it. You don't just ask a question, get someone saying "I don't know" and then base your entire company on that idea unless it's proven to be a solid solution to a problem that exists. Right now, no problem existed. NC machines have been around since, what, the 50's? Motion control isn't new. Steppers, servos, open loop, closed loop, it's all been out there for 50 years or more. And people do mind blowing stuff with it. I remember when CAMM systems started to become more powerful than the CNC controls. Prior to that, complex surfaces weren't able to be done. I recall running software on a UNIX based system that could create point to point commands for some really complex shapes. We were starting to get G-Code programs with 100,000 lines of code and the machines were just choking on them. 20 years ago, High Speed Machining came along (mainstream) and you were seeing motion controllers doing mind warping speeds for rapid and for cutting, especially in aluminum. The technology is there and companies have spend billions of dollars over 1/2 a century to fine tune it and make it the most efficient. I seriously doubt that a "What if" comment off the cuff negated all the work all those companies have done for decades. All the engineers, all the R&D, all the applications were machines had to process a lot of data very fast, and not one company said "Hey, you know what would be a great idea? Let's rip the motion control out and put it behind an internet based system"?

There's a reason for that. Because it's not a viable, smart solution for the engineering of a system.

That's my opinion.

I would be willing to bet that the "A-ha!" moment had nothing to do with hardware, or engineering, but instead the fact that they realized that they could seamlessly control the end user.

Why not set yourself up to be able to sell the machine for $3,500 in the future with a "nominal" $99.95 monthly fee for access to the Glowforge cloud?

At least that's where I see this going. That is unless too many people burn their houses down before they get that far. In which case, I guess a corporate on/off switch would also come in handy...

Just look at what Apple just admitted to yesterday: throttling the performance of older Apple phones to induce upgrades (though they claim it was to avoid older devices crashing).

[Bill George]

When shopping for a new router I briefly considered CarveWright, very briefly I might add. Everything is 2x or more the cost and most of the time it has to come directly from them or it won't work.