search

jetty840 / Sailfish-MightyBoardFirmware

Sailfish, faster than a Marlin
118 stars 75 forks source link

Idea: Some global variables #138

Closed Dogm closed 9 years ago

Dogm commented 9 years ago

What if do some refactoring and place some board preferences as global variables?

I mean this:

<==========> // microstepping is 1 / (1 << MICROSTEPPING) ... // 4 for 1/16 // 5 for 1/32 ...

define MICROSTEPPING 4

<==========>

make define from scons script? like: "stepdrv=a4988" A4988 = 4 A4982 = 4 DRV8825=5 ... or just fixed set "stepdrv=4"

What do you think?

Dogm commented 9 years ago

This affect "firmware/src/MightyBoard/Motherboard/boards/(mb_folder)/Configuration.hh" files

dcnewman commented 9 years ago

Using 1/32 microstepping is pointless with these printers. On the 16 MHz, 8bit AVRs the firmwares all quickly turn 1/32nd stepping back into 1/16th stepping by double stepping. They do this because they are not fast enough to execute the necessary number of steps per second. If you want to effectively use 1/32 microstepping you should be using an ARM processor. Additionally, you cannot see any visible difference in print output between 1/16th and 1/32nd microstepping with FDM 3D printer.

Dogm commented 9 years ago

sad 8(

dcnewman commented 9 years ago

You have to keep in mind that one of the three key goals of sailfish is to get the highest performance possible out of the 8bit, 16 MHz AVR. To accomplish that, hard assumptions are made about the underlying electronics and hardware including steps/mm for X, Y, A, and B around 100 steps/mm and Z around 400 steps/mm. Also the approximate build dimensions. If you start making a 1 mm Z axis (Jetguy) or high resolution stepping (e.g., 1600 steps/mm in Z) then issues can arise with the fixed point arithmetic. Issues arise because the fixed point math can start overflowing. So, our position has always been that if you want to do high resolution stepping or very large build volumes, then use Repetier or Marlin which both use (software) floating point arithmetic. (And are about 5 - 6 times slower in the planner as a result.)

Other two key goals of Sailfish is safety and pressure/relaxation advance.

Dogm commented 9 years ago

Understand...

Dogm commented 9 years ago

Also ... what do you think about more powerful chips and progressive languages? i mean what about try something like this https://www.ghielectronics.com/catalog/product/543 https://www.ghielectronics.com/catalog/product/530

Cpu: 120-240 mhz Language: C# Code: 1,3 ... 2,8 Mb Ram: 10 ... 13 Mb ...

Idea: do 3d printer software with this devices...

Dogm commented 9 years ago

This assembly cost 150$ with graphic lcd display (tinker kit)

total cost - not so much... this allow to do more interesting devices with huge community of c# developers. p.s. i think Arduino has smaller community even dot.net/c# developers..

toybuilder commented 9 years ago

C# is not for real time embedded systems. On Aug 12, 2015 12:09 PM, "Dogm" notifications@github.com wrote:

This assembly cost 150$ with graphic lcd display (tinker kit)

  • need board with 5-6 stepper drivers (can use any available interface) and just power source ~10-25$

total cost - not so much... this allow to do more interesting devices with huge community of c# developers. p.s. i think Arduino has smaller community even dot.net/c# developers..

— Reply to this email directly or view it on GitHub https://github.com/jetty840/Sailfish-MightyBoardFirmware/issues/138#issuecomment-130414617 .

Dogm commented 9 years ago

Does it matter if the reaction rate is clearly higher than in the real-time system?

toybuilder commented 9 years ago

Yes. It does matter. I won't rehash the discussions, but if you look around the various 3d printing groups, you will find plenty of past threads on this issue.

The second answer on this link ( http://stackoverflow.com/questions/3762081/using-c-sharp-for-real-time-applications) also captures the issues well. On Aug 12, 2015 12:22 PM, "Dogm" notifications@github.com wrote:

Does it matter if the reaction rate is clearly higher than in the real-time system?

— Reply to this email directly or view it on GitHub https://github.com/jetty840/Sailfish-MightyBoardFirmware/issues/138#issuecomment-130417401 .

Dogm commented 9 years ago

I understand difference of both architectures... may i ask of real problem with soft for 3d printer with non/real-time code? if theoretically make such a system...

Dogm commented 9 years ago

sensors? - it can have buffered with other chip currents

Dogm commented 9 years ago

stepper drivers - it can update faster than delay of 1/64 microsteppings (for sample use standalone chip with stack of real time commands)

Dogm commented 9 years ago

i just dont understand which of sub-logic must be only real time explain to me please..

Dogm commented 9 years ago

Also, if so antipathy witch non-real time languages and/or garbage collectors just can use more powerful chips like https://www.sparkfun.com/products/9713

dcnewman commented 9 years ago

As I stated before, a key goal of Sailfish is to get the most performance possible out of 8bit, 16 MHz AVRs. Sailfish isn't a firmware to port to a different architecture: it's not meant to be ported. And this isn't the place to be discussing new processors for 3D printing firmwares. I do not know of a good place for such a discussion. Perhaps 3dprintingtipstricksreviews?

Dogm commented 9 years ago

I open discussion thread https://groups.google.com/forum/#!topic/3dprintertipstricksreviews/aZtrkl9KPLE