QuoteYou can. So why hasn't anyone? Even a 100Mhz ARM chip is enough to run a motion control interrupt at ~100KHz and a web server. Smoothie board ought to be able to do it, but I don't know how complete their software stack is. It probably hasn't been done because of the prevalence of arduino's for rep-rap control electronics. They've been good enough that people haven't looked at alternatives,by Polygonhell - General
But you could do all of that with a single ARM uController running a simple Web server and providing motion controlby Polygonhell - General
I have to question why they didn't just pick a fast ARM board with enough GPIO and just expose it, there are plenty of them around, there is a software problem in coming up with some way to do reliable realtime I/O while running an OS with a GPU driver. But I guess it's supposed to be a linux box with an arduino in onboard. The more I see people damage the arduino (usually the ADC pins) the moreby Polygonhell - General
QuoteAfter seeing a local makers group work on their RepRaps--I really don't want to tinker that much--I can still wait. Last shipping estimates for initial orders at currently at mid to late May. Ignoring the design issues pointed out in this thread for a moment. I'm just wondering why you think an assembled printer using many of the same ideas as a rep-rap will involve less tinkering? The reaby Polygonhell - General
The gaps are going to be inconsistent filament feeding, either because the filament is slipping on the hobbed bolt, or because you are printing slightly too cold.by Polygonhell - Printing
Most likely you are losing steps on the X axis, probably the stepper pot is set slightly too low. The kisslicer GCode is probably just different enough to cause this to happen, the speed settings aren't identical for example. It could also be too high a speed set in the slicer/firmware, and possibly something mechanical, though that seems unlikely if you've been printing successfully.by Polygonhell - Printing
Yeah the PRU's are the one saving grace for me on the BeagleBone running linux. Still would like GPU access for No OS but pretty much all the faster ARM CPU's have no GPU access for no OS.by Polygonhell - Developers
akhlut Wrote: ------------------------------------------------------- > I was pretty sold before...now I'm wondering if > cartesian bots have a future. I have 2 cartesian bots based on the Mendel design and a RostockMax printer, I've got a lot of time on all of them. There are a lot of things to like about the Rostock. The obvious disadvantage of the Rostock design is the packaging, it'sby Polygonhell - General
It's a Beaglebone cape. The Beaglebone is a 700MHz Ti Cortex-A8, on an arduino like board for ~$80. I like Beaglebone's, but on the software side you either get straddled with Linux, which is far from Ideal for a motion controller, or you go No OS and lose access to the LCD driver which to me is a large part of the value.by Polygonhell - Developers
I've printed parts well over 100mm square on a delta and the lines are all very straight and as dimensionally accurate as any of my other printers.by Polygonhell - General
I generally use Repetier Host, I like the gcode path renderer since I actually occasionally do see issues with it before I start the print. When printing though I just watch the big temperature graph. What I'd like to know is why all the host software I've used tries to accumulate position distinctly from the firmware rather than just sending an M114 and actually getting the actual position. Eveby Polygonhell - General
Arduino sketches are all basically C/C++ with a wrapper to specify a main loop, so it's no easier. And all they do is rename compile to verify..... I think one of the advantages of ARM is that you could probably get rid of all the edit configuaration.h stuff, since there really isn't a need to optimize the math by having constants defined at compile time. I think the MBed drivers will give you dby Polygonhell - General
I spent some time looking for an ethernet solution for RAMPS/Arduino, the ones I found were all very limited. In addition often the ethernet chip was a faster processsor than the atmel CPU. I'm now of the opinion that ARM boards are the way to go, as well as being faster an LPC1768/9 is cost comparable to the Atmel controllers, has sufficient IO and has built in support for Ethernet among otherby Polygonhell - General
Just looking at it in the video , did you change to a coreXY configuration over a pure Hbot? Just wondering.by Polygonhell - Developers
Sublime Wrote: ------------------------------------------------------- > Does it have lookahead yet? I can not see any > reason to have high step rates without being able > to go around a corner at a decent speed. > > Also when you say > with proper acceleration > > What does this mean? Back when I used Teacup you > always had the acceleration set Really Really low &by Polygonhell - General
CoreXY is a particular motion dynamic, but more broadly it's the same as the HBot style dynamics, one axis is (X+Y)/2 the other is (X-Y)/2.by Polygonhell - General
QuoteWould this manifest as a high pitched whine in the motor with no movement? It might, basically if the interrupt rate is too high, you start missing interrupts, which reduces the speed and introduces a rather large amount of jitter into the step output, which could stall the motor. But you get the same thing if you run out of torque, run the motor too fast or your acceleration or jerk are setby Polygonhell - RAMPS Electronics
There are many things that limit speeds. It looks like you are using screws to drive the axis, what's the pitch? There is a very real possibility you are exceeding the arduinos ability to produce a step train, in practice about the fastest it can reliably process interrupts is 16K/second, repetier firmware for example processes 2 steps per interrupt for a maximum speed of 32Ksteps/second. If youby Polygonhell - RAMPS Electronics
No I was going to pick one up for testing, but I haven't had a chance to do so yet.by Polygonhell - RAMPS Electronics
The code literally toggles the pulse on/off on sequential instructions, repetier and I think later marlin builds have an optional delay between the on and off toggles. The 4988 just requires a rising edge, as far as I can tell it doesn't need any pulse length, I ran a test over the weekend with a 100MHz ARM board and a 4988, even immediately toggling the signal back off still works even when my sby Polygonhell - RAMPS Electronics
My goal when I was looking at it was to make a viable slic3r for arm based microprocessors with real GPU's like the Pi or the BeagleBone. The issue was the inconsistencies in GLES implementations , and the lack of any openCL drivers on the platform. My code's on Github, all it does is render slices and identify outlines for the tool path, it runs on a Beagle bone, the issue you'll have with otherby Polygonhell - General
I looked at one at one point, there are issues with maximum render target resolution, and extracting usable tool paths from bit maps though other slicers do that effectively.by Polygonhell - General
Ive given this a fair amount of thought, and I'm not sure compensating by adjusting the model is even possible, the problem is that the issue is shrinkage. So lets say I have N layers down and print an N+1th layer, offsetting in or out has negligible impact on the shrinkage, you're still welding a hot layer onto a cold one and when the hot layer shrinks, you are still going to cause it to pull upby Polygonhell - General
They can also wait to see if you make any money and become worth the lawsuit, there is no requirement to immediately defend a patent, and no point at which you are safe until it expires. Practically companies use them for 3 things. Removing competition before it starts, sue a startup you think will impinge your market, defending a patent case even one you are sure is invalid is almost never wortby Polygonhell - General
You want to avoid soldering the hotend resistor, it's usually an area that gets hot enough to erode the solder over time often leading to failure of the joint later, usually when most inconvenient. You should be able to get a decent crimp with a pair of pliers.by Polygonhell - General
You can run Nema17 motors at ~1000RPM with simple pololu drivers. That's IME about as fast as you can run steppers reliably even with expensive drivers. You can't easilly do that with uStepping enabled on arduino based electronics because it requires the interrupt rate be too high, I did some tests over the weekend and it's relatively trivial with a 100MHz Arm board so something like smoothie ougby Polygonhell - General
davew_tx Wrote: ------------------------------------------------------- > has anyone noticed that firmwares generally > support G2/G3 arcs, but I've never seen that in a > Gcode file. > Nothing but G1 moves usually. > > I have plenty of rads and arcs in parts, most are > fixed rads, and not compound curves and splines. > > Dave The problem is all of that informationby Polygonhell - General
Honestly if this is your forst build I think making those hotends work on a reprap is going to be a lot of work. If you look at the KISSlicer thread you'll see how they are used in the Dimension printers and how much additional stuff there is that you don't have.by Polygonhell - General
Is it really necessary to get personal? I'm not American but I've lived in the States for over 20years, trust me there are plenty of assholes and for that matter nice people on either side of the pond. If you really don't like someone or feel slighted I'm not sure what the value is in turning it into a public pissing match.by Polygonhell - General
I just did a quick test Just a square hole inside a square object I sliced it in the Beta I have installed and the paths look correct. The 5mm hole in the middle of the test piece creates lines at +2.75 and -2.75 (0.5mm width requested) which is correct given what was requested. KISSlicer does correctly (assuming you're tool is circular) create arcs for the corners of the inside perimeter, and haby Polygonhell - General