This will probably spawn a spirited discussion:
None of the below intends to deter anyone from making improvements or building alternate approaches. Its just that in this discussion I have “been there done that” so I fill compelled to poke the beast…
First off, I’m fully prepared to learn from someone with a different background. I have an EE background, but only months of thinking about the K40, mostly on my own,
I would never endorse a laser or CNC machine in which main power is turned on any automated way. Once the machine is turned on and enabled by a human (after ensuring the machine and room is in a safe state) the software can control its job operation. It also must have an estop that is mechanical. The dangerous parts of these machines (laser and spindle) must not start without mechanical interlocks closed.
That’s reasonable. Electrical safety standards do require a manual disconnect as well. UL, IEC, and other country standards specify that this can be done by the electrical socket disconnect below X power (I’d have to look up X, but the K40 is less than that). I’ve always thought that was silly, and that “polite” design made this be at least a hard two-wires disconnect - a switch. But there are cases where once you have the machine enabled manually where it would be very nice to disable AC power to sections of the machine. Hence the second AC switch inside the manual disconnect.
EPO switches are a good idea, but putting them on the machine likely to generate the fire is something to be thought about. I come from a formerly major computer company, where the EPO is mounted hand-high next to the exit door, so you can slap it as you’re running from the roaring fire. In my mind, the EPO switch for the K40 might best be mounted from 3 to 10 feet away from the machine, perhaps right next to the fire extinguisher so your movement away from the fire leaves you an efficient way to both kill the electrical power while deciding on fight or flight as you flee the machine. I am by no means a safety expert.
The smoothie and Lightburn folks spent multiple years getting a controller and software combination that works for the K40. I tested and used all these configurations on the way to the current state. I think the K40 controller and associated software needs are solved!
May well be. As I said, I’m new at this. I will need some experience before any real discussion of those deeper issues.
What is the advantage of a PI when C3D and Lightburn already provide; control panel, Gcode machine & job control which includes camera & bed movement?
In my baby steps approach, wifi connection to the machine. Do C3D and Lightburn let you do that? If so, maybe none. I’m aware of the advantages of the software operation of the laser movement and so on that users have posted. Just haven’t tried them, so I’m ignorant of the details. I see the Pi as an overseer, that I can reach out for ready to burn jobs from the graphics machine.
External devices like air, evac and water pumps can also be controlled with the C3D although I do not want my water pump ever to be off. In my machine, an aux panel controls all AC to the air assist and evac pumps but its built from simple relays switches & LEDs.
I agree, the pumps and fans should not be off. It’s more like killing any possibility of the laser going on when a pump/fan/flow fail is sensed, a second level of stoppage for those times when for example, you told the pump to run, but it’s inlet hose is kinked, or the fan motor spins but the impeller spins on the shaft. It amounts to feedback on the desired result, not on starting the auxiliaries up.
The things I was thinking of kit-ing in my earlier post was about the simple, critical but illusive interlocks, tube protection etc. I doubt the newbies that we have been talking about would be that willing to spend the $$$ and time to install anything that requires much wiring and if they do the support is probably a nightmare.
No disagreement here. I’m still thinking my way through it.
I have thought many times that an add-on embedded monitor was essential and in fact, started down that road a few times. I stopped after I found easier, safer and cheaper ways to do each of them in turn. Then again I love building embedded controllers so I still think this would be a fun project if it included bed controls.
That’s another thing requiring thinking. A Pi suitable to the things that are needed for the K40 is $35, and you might well be able to use a Pi Zero W AT $10. A 5" Pi display that fits the existing control panel cutout is ~$40, and an old HDMI monitor might be less, although external. In the role of a com link and overseer, it might be a fit. K40 whisperer runs native on the Pi, although it brings its own limitations as a software package, probably much more limited than the more-extensive packages. But it might well be a usable cost point for people satisfied with the limitations of the M2 Nano controller. Think of this as one step up from the base system for the limited budget, with some added overseer/monitoring functions and native communications. Different price point.
I even got deep into the notion of just building a better machine from scratch but realized the world did not need another Full Spectrum or Glowforge and full safety compliance is expensive in design and approval. It also became clear that the folks that want these for hobbies do not want to pay much more than a K40.
Building a better machine from scratch? Not for me. And I totally agree that the hobbyist market is cost- and attention-limited. Hence the cheaper Pi package might hit a useful step up. Especially if you go for an low-end overseer role, leaving off the display.
All in all adding over-temp protection, DVM on the pot and interlocks to a stock machine gives the new K40 user most of the basics needed to get good to excellent results.
Agree. I originally just programmed a PIC to monitor that kind of thing. Left off when a nominal-price Pi offered both wireless communications and the possibility of expansion to higher function.
Doubtful??? Ask @NedMan what upgrades he has done enabling him to create his amazing work :). Last I checked he still has a nano!
Not doubtful. As I said, I started with a PIC, and was seduced by the Shiny Thing of comm, display, and expansion without having to write the detail code inside the PIC.
No argument, just a baby steps approach in the design space.