32 bit Arduino for reprap $50

I think for now, we don’t have to worry too much about developing the next generation electronics, it’s come to us:

An Arduino with 32-bit 72MHz ARM Cortex M3 chip.
[leaflabs.com]

With 32-bit processing power, we can do PID control for the extruder and heated bed without the need for a separate Extruder Controller.
And storing Gcode file in the SD Card is also a possibility…

Porting firmware from current boards should not be a big problem. I hope

I wish it had PWM hardware as well. It would make a great robot-on-a-board or Underware ROV-on-a-board if it did. Running 6 to 12 PWMs in software, even with fast interrupts, could get messy. On the other hand, 16 or even 32 channel external PWM chips exist with I2C interface, so that would be easy to add.

However, Tony, the PID can easily be done in the ATMega or Aurdionos that we currently have, it is not difficult to do if you are programming in C. It is a nice step up to have direct USB programming, but even that is not a biggie. Extra DMA channels to add an SD card interface is great, if ti does not use up too many I/O pins.

Mike

i like the idea of it, but for some reason it seems like a little overkill. not sure why though

---

[mike-mack.blogspot.com]

“I wish it had PWM hardware as well.”
- I think it has 15 PWM pins at 16-bit resolution.

“…the PID can easily be done in the ATMega or Aurdionos that we currently have, it is not difficult to do if you are programming in C.”
-The point here is not about how easy but how FAST. Having a poor 8-bit controller do floating-point math, G Code processing, PID control… I think you border on microcontroller abuse
If we use two 8-bit Arduinos, total more than $100, why not use just one 32-bit cost only $50? I would select the cheaper one even if it is “overkill”

The 32 bit Arduino can be found here:
[leaflabs.com]
· STM32F103RB: a 32-bit ARM Cortex-M3 microprocessor at 72Mhz
· 20KB RAM and 128KB Flash
· Dedicated USB port for programming and communications
· 37 General Purpose pins
· 16 Analog inputs at 12-bit resolution
· 15 PWM pins at 16-bit resolution
· Dual I2C and SPI ports
· 3 USART Serial ports
· 7-channel DMA
· Low power and sleep modes
· Integrated LiPo battery charging
· Dimensions: 2.05″x2.1″ (5.21x5.33cm)
· Starting price: $50
· Extensive documentation

STM32F103RB datasheet is here:

[www.st.com]

the only issue hardware wise that i can see is the fact that it is 3.3v and all the stuff i have seen so far is for 5v. i have read some people are worried that the 3.3v would pick up allot more noise too

---

[mike-mack.blogspot.com]

Quote

With 32-bit processing power, we can do PID control for the extruder and heated bed without the need for a separate Extruder Controller.
And storing Gcode file in the SD Card is also a possibility…

Both is possible with the real Arduino, as you probably know. The separate Extruder Controller is some artefact introduced with Gen3 electronics and already in the process of going away.

BTW, the name "Arduino" is trademarked.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

... separate extruder controllers are needed, if you use more than one ... but this can be solved by a serial bus too ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Tony Wrote:
-------------------------------------------------------
> “…the PID can easily be done in the ATMega or
> Aurdionos that we currently have, it is not
> difficult to do if you are programming in C.”
> -The point here is not about how easy but how
> FAST. Having a poor 8-bit controller do
> floating-point math, G Code processing, PID
> control… I think you border on microcontroller
> abuse


True. If I were worried about process speed, I would do the PID in scaled integer (fixed point) arithmetic, or even in assembly code. The algorithm is simple, it is tweaking the 3 coefficients to match the hardware that is tough. I have already done this in floating point on an ATtiny controlling geared motor to steering linkage system, so I am sure an ATMega can handle the load, too. Where I think the great processing power of the ARM chip comes in is doing image analysis from a built in camera for self calibration or scanning 3D objects or verifying the quality (slipped X or Y coordinates fault) the current build piece.

Mike

Mike wrote:
“If I were worried about process speed, I would do the PID in scaled integer (fixed point) arithmetic, or even in assembly code. The algorithm is simple, it is tweaking the 3 coefficients to match the hardware that is tough. I have already done this in floating point on an ATtiny controlling geared motor to steering linkage system, so I am sure an ATMega can handle the load, too. Where I think the great processing power of the ARM chip comes in is doing image analysis from a built in camera for self calibration or scanning 3D objects or verifying the quality (slipped X or Y coordinates fault) the current build piece.”

- Thanks Mike! How far you would go to stay with an 8 bit controller helped me understand why an 8-bit Arduino Mega costs $65 while a 32-bit Arduino-compatible costs $50 and another 32-bit board with built-in SD card socket only $39.99.

Tony,
I should also add that I have never used an Arduino of any type. I have the STK500 development and programming board, so instead of buying $65 Arduinos, I buy $8 dollar ATMegas. So for me, it is spend $8 for another ATMega, of which I already have about 2 dozen laying around waiting to be used, or $50 a pop for something I haven't needed yet. I only mention the Arduino because many people are more familiar/comfortable.

That still applies to the switching to 32 bit Arm for $50 from one's 'comfort base' of an Arduino. On the other hand, if all you have done is PICs and ATtinys and now you need to crank it up a notch, and are not well dug in to the Arduino culture, then jumping straight to a 32bit processor with lots of expansion, I/O, and storage capabilities makes perfect sense.

I may get one of these supper boards to learn and experiment with, just as soon as I have finished developing the perfect glass extruder nozzle, built my repstrap, made the RP parts for and built my Mendel, started work on my Open Air Reprap project, built test controller for a single stepper drive and heater, developed a program to read one or more USB joysticks and convert the movements to motor commands for my uderwater ROVs, and started work building next years killer ROV, as well as wire up, insulate, organize my garage workshop and learn to use the lathe-mill-drill. Maybe a 64 bit version will be available by then!


Mike

lol so that should be next week some time. i think...lol

---

[mike-mack.blogspot.com]

Greetings all,

I have one of the first batch of leaflabs maple boards.
FYI, I posted a how-2 for the original leaflabs bootloader and LINUX [forums.leaflabs.com]

From what I understand, their newer boards with an updated version of the bootloader is considerably easier to get talking, and also supports Windows, LINUX and MacOS. leaflabs is local to me (Cambridge, MA), and they seem like a good bunch of people -- when I was eager to get my hands on one, I was able to pick it up from them after work. Another significant benefit is that they're putting their software (and perhaps even schematics and board art) out as open source.

---

Larry Pfeffer,

My blog about building repstrap Cerberus:
[repstrap-cerberus.blogspot.com]