R2C2, a really new generation -- ARM 32bits, 100MHz, runs RepRap at 725mm/s !!

R2C2 electronics for RepRap 3D printers and others machines



Read more here >>


-----------------------------------------------------------------------------

Hello.

My company is developing this electronics and I have a working prototype. It will be commercialized on Ebay (including the releasing of source files) to worldwide 1 January 2011. The electronics and firmware will be released as Open Source that work with / and are fully compatible with RepRap software and others alike.

Advantages of this electronics:
* Just one board to control your RepRap 3D printer (can also control any other 3D printer/laser/milling);
* Plug and play - you will receive it completed assembled and tested;
* Cutting edge technology - ARM microprocessor of 32bits running at 100MHz with DMA, which is about 50x more powerful than actual "official" RepRap electronics and others alike. RepRap "official" electronics uses AVR 8 bits running at 16MHz and without DMA. You will not get any more this problems because of slow AVR core 8bits used on RepRap. If you are looking for hight speed and high quality printing/laser/milling, you need a microprocessor really fast to be able to generate a big number of stepper motor impulses per second - this is our solution;
* Heated_Bed support;
* USB bootloader working as a USB Mass Storage device - you will just need to copy/paste new firmwares to board as if is a USB pen flash;
* Expansion header so you can add support for other functionalities like using a laser to engrave/cut paper, wood, ABS, PLA, etc;
* We strong believe in OpenSource and community development and so we are trying to use much as possible OpenSource tools. All the firmware were developed with OpenSource tools: ARM GCC, ARM GDB, OpenOCD JTAG Programmer/Debugger and Eclipse.

Please read more on RepRap wiki page for updates: HBox_RepRap_Electronics

2 pictures from my prototype:

(Using pololu stepper driver motors from old RepRap electronics)


(While printing...)

While debug the firmware using JTAG -- bringing for RepRap world for the first time the debug for firmware using (OpenSource tools):




Edited 18 time(s). Last edit at 10/21/2011 07:22PM by casainho.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

It really is great to see all the new options coming online.

good luck with this endeavour

This looks promising. I have a few questions:

Will this be a 2 layer or 4 layer PCB?

What chips are you using to drive the stepper motors? Like Pololu or something else?

Will MOSFETs be used to drive the heating elements?

Is the firmware configurable from a separate user interface or does it have to be configured by altering the source code directly, like in the current firmware?

Have you got a rough estimate for the price of a fully assembled unit?

Matthew Wrote:
-------------------------------------------------------
> This looks promising. I have a few questions:
>
> Will this be a 2 layer or 4 layer PCB?

2 layers because 4 layers have higher costs, and 2 layers are enough.


> What chips are you using to drive the stepper
> motors? Like Pololu or something else?

Same as pololu.


> Will MOSFETs be used to drive the heating
> elements?

Yes.


> Is the firmware configurable from a separate user
> interface or does it have to be configured by
> altering the source code directly, like in the
> current firmware?

Current firmware I have is a port of "FiveD on Arduino". Board will have an SDCard so any kind of files like .TXT can be used to easy the machine setup.


> Have you got a rough estimate for the price of a
> fully assembled unit?

Not yet.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Looks very exciting Casainho! Looks I may soon have more electronics kits than intended machines to use them in!

Mike

what compatibility will the firmware have for ARM kits such as the et-arm stamp? comparable in cost to arduino, but about 10x performance, up to 50x performance on float math.

BTW, you may turn heads if you do a bunch of vidoes showing round objects, or making gears at above 32mm/s-64mm/s with '1' micro controller. something easy to do with mach3/emc2 and a 700mhz Pentium, but not realistic on a tiny single arduino micro with serial connection to pc. are you designing serial connection only board?

. the arm processors I've been looking into have incredible math ability, better a/d at 12bits where arduino has 10 bits and appealingly an internal true d/a converter 12 bit resolution accuracy!


Are you going to make your source code open to the public?
do you support sd card as well as the usb?
is your design 3.3v or 5v?

thanks for your hard work!

james

jamesdanielv Wrote:
-------------------------------------------------------
> what compatibility will the firmware have for ARM
> kits such as the et-arm stamp? comparable in cost
> to arduino, but about 10x performance, up to 50x
> performance on float math.

The ARM LPC1768 I am using have the same characteristics as that you wrote -- much more powerful than Arduino AVR and the same cost. Even it have OpenSource Debugger, which is a must for quick development.


> BTW, you may turn heads if you do a bunch of
> vidoes showing round objects, or making gears at
> above 32mm/s-64mm/s with '1' micro controller.

I do not have all the time to do it, but I am sure community, users as you, may do it when they get the boards on his hands. But I already made some tests comparing my Mendel RepRap electronics and this board -- yes, this board gives more printing speed and better prints mainly on small holes.


> something easy to do with mach3/emc2 and a 700mhz
> Pentium, but not realistic on a tiny single
> arduino micro with serial connection to pc. are
> you designing serial connection only board?

Yes, USB<->Serial done on LPC1768 firmware -- board is full compatible with actual RepRap software but the serial link is much faster!


> . the arm processors I've been looking into have
> incredible math ability, better a/d at 12bits
> where arduino has 10 bits and appealingly an
> internal true d/a converter 12 bit resolution
> accuracy!

Yes, LPC1768 have 12bits ADC.


> Are you going to make your source code open to the
> public?

Sure, when I release the board. I am using a lot of OpenSource libraries like the ones for FAT32 on SDCard -- I am being writing more info and ask opinions on RepRap dev mailing list.

> do you support sd card as well as the usb?

Yes, in the hope to make printer stand alone -- no PC needed during long prints, just like UP printer.


> is your design 3.3v or 5v?

3.3V.


> thanks for your hard work!

:-)

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Hello,

This is indeed an exciting project, and is the answer to many concerns I have, particularly about the limited number of pulses per second that can go in to each axis using the Sanguino. I have some very nice motors that I'd like to get running on a rep-rap like system, but they're going to come in at 1000 pulses per inch, which I think might be too many for the Sanguino to give me a respectable feed rate. Please keep us updated on your progress, I'll be the first in line for one of these!

Regards,

Andy Baker

Kontraptionist Wrote:
-------------------------------------------------------
> Hello,
>
> This is indeed an exciting project, and is the
> answer to many concerns I have, particularly about
> the limited number of pulses per second that can
> go in to each axis using the Sanguino. I have some
> very nice motors that I'd like to get running on a
> rep-rap like system, but they're going to come in
> at 1000 pulses per inch, which I think might be
> too many for the Sanguino to give me a respectable
> feed rate. Please keep us updated on your
> progress, I'll be the first in line for one of
> these!

Hello.

I found that I can't run my stepper motors at more than 100us period each step pulse, for pololu drivers. My Cortex-M3 running at 100MHz, gives ticks of 10ns, so, 10.000 times more the needed for stepper motor pulses.

Since my Mendel have 80 steps per mm on X and Y axis; 1/100us (my motors maximum) = 10.000 pulses => 10.000/80 = 125mm/s. I think my Mendel X and Y axis can't do 125mm/s.

So, the electronics can do much more steps (maybe 10x times more) than my Mendel motors can handle. After that velocity, my motors starts to make noise and do not turn.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

hi casainho,

As far as I know in the firmware of the G code section, acceleration is not done automatically on linear moves, it just starts outputting pulse trains at the indicated veolcity, or blending over the course of the move from starting feed to ending feed, which is a strange, non-standard way to do G-code, but better for the REP RAP. Many, if not all stepper motion control systems have a ramp up and ramp down to get maximum velocity - Any stepper can't start and stop instantly above a certain velocity, especially if they have a load.

Question, are you running your pololus at full step, or microstep? if microstep, how fine? At 16 microsteps, you will of course need 16x more pulses to go 1mm, but microstepping makes motors quiet and happy. For my motors, they'd move Mendel at about 40 pulses per mm, so if the current firmware really can put out 10,000 steps per second, I'll be in good shape.

Faster is still better!

Keep up your good work.

Andy Baker

> Question, are you running your pololus at full
> step, or microstep? if microstep, how fine? At 16
> microsteps, you will of course need 16x more
> pulses to go 1mm, but microstepping makes motors
> quiet and happy.

At 16.


For my motors, they'd move Mendel
> at about 40 pulses per mm, so if the current
> firmware really can put out 10,000 steps per
> second, I'll be in good shape.

As I said before, I need 80 pulses per mm because: on X axis for example, the gear have 8 teethes of 5mm each, so, 40mm for a complete rotation. My motors take 200 pulses for 1 rotation and multiplying with the ratio of 16 from pololu, it takes 200 * 16 = 3200 pulses for one rotation. So 3200 / 40 = 80 pulses per mm. And 40 / 3200 = 0.0125mm = 12.5um of resolution (theoretical).

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

I can get those step rates with my 25MHz 16 bit micro so it should be a walk in the park with an Arm.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> I can get those step rates with my 25MHz 16 bit
> micro so it should be a walk in the park with an
> Arm.

What step rates? can you please give numbers?

ARM maybe be faster than yours 25MHz 16bits, however this ARM will have others tasks as USB, SDCard, FAT, etc... so it needs all that power.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

I think my fastest axis is an M8 threaded rod drive driven by a Pololu x 16 driver at 10mm/s.

So that is 10 / 1.25 * 16 * 200 = 25600 pps. It is also doing 100 Mb Ethernet. It don't think it would be a problem doing an SD card instead.

---

[www.hydraraptor.blogspot.com]

nophead, any issues with axis skipping steps and interpolation? for example G0 X40 Y11 ? (max feed-rate)

the issue i see is floating point math is needed to ensure the y axis turns at a constant federate instead of stopping and starting every few steps unevenly spaced .even spacing of the y steps versus x steps. then again this is mitigated by micro stepping at 16x, can you do it at the same speed?

if you can do this with the gcode line i gave above and not have the y motor skip that would be interesting. thanks. either-way, I'm exited about arm. Also an admirer of your work nophead.

thx.
James

Those figures are for a lead-screw z-axis. X and Y are always slower because I use belt drive (or multi-start lead screws in the case of HydraRaptor). The fastest rates I have tried on X and Y is 100 mm/s with Pololu drivers. That would be 100 / 40 * 16 * 200 = 8000 pps. I can do any XY move at that speed without skipping.

I use the normal Bresenham algorithm and have never found a problem with the jitter on the slow axis. At high speeds the inertia smooths it all out anyway. Having said that I have always used micro-stepping drivers from day one apart from the z-axis on HydraRaptor.

I use 256 step acceleration curves to get to 10mm/s on a studding lead-screw. That is 0.1mm.

---

[www.hydraraptor.blogspot.com]

uSDCard arrives

Now there is an uSDCard that stores "config.txt" file with configurations like steps per mm and maximum feed rate of each axis. Now you can configure your 3D printer just by editing a txt file - no more need to build the firmware sources and program the microcontroller. This uSDCard works also as board internal memory and in future will store GCodes files, letting printer working standalone without the need to be connected all the time to a computer;

My current "config.txt" file:

STEPS_PER_MM_X = 80
STEPS_PER_MM_Y = 80
STEPS_PER_MM_Z = 6400
STEPS_PER_MM_E = 36
MAXIMUM_FEEDRATE_X = 1800
MAXIMUM_FEEDRATE_Y = 1800
MAXIMUM_FEEDRATE_Z = 60
MAXIMUM_FEEDRATE_E = 40000
SEARCH_FEEDRATE_X = 120
SEARCH_FEEDRATE_Y = 120
SEARCH_FEEDRATE_Z = 60
SEARCH_FEEDRATE_E = 1600

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

jamesdanielv Wrote:
-------------------------------------------------------
> nophead, any issues with axis skipping steps and
> interpolation? for example G0 X40 Y11 ? (max
> feed-rate)
>
> the issue i see is floating point math is needed
> to ensure the y axis turns at a constant federate
> instead of stopping and starting every few steps
> unevenly spaced .even spacing of the y steps
> versus x steps. then again this is mitigated by
> micro stepping at 16x, can you do it at the same
> speed?
>
> if you can do this with the gcode line i gave
> above and not have the y motor skip that would be
> interesting. thanks. either-way, I'm exited about
> arm. Also an admirer of your work nophead.
>
> thx.
> James

My firmware can do that without floating point, it's really not that much different to bresenham. Just a matter of finding the next time an axis will step instead of finding which axes should have stepped - all the same numbers, just handled a little differently.

Having said that, I'm still working out how to do acceleration properly.

Even without temporal step, inertia will smooth out all the jitter in all but the most extreme situations.

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

Triffid_Hunter Wrote:
-------------------------------------------------------
> My firmware can do that without floating point,

And my code is a port of "FiveD on Arduino" from Triffid_Hunter (and others developers) -- so, my code is a kind of "FiveD (on Arduino) on ARM" :-)

But I added floating point to some places... I don't see the need to go without it on ARM - there are almost no limitations when compared to Arduino AVR.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

hey casainho,
any news on your electronics? Im looking forward to buy and test them...

cheers
Till

Board design is being finished. In 2 weeks I should have on my hands the first boards to test. I will post here on that time, pictures and more technical details.

We are taking a good deal of time to make a robust and high quality board. I believe the board will also be used for making laser engraver/cutter, delta robots, etc.

Thanks.

Edited 2 time(s). Last edit at 03/04/2011 02:36PM by casainho.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Casainho:

I too am looking forward to using your board, if you need test and development help (hardware of firmware) let me know. Any Idea when the first ones will be available and what do you project the cost to be?

Bryan

bblazie Wrote:
-------------------------------------------------------
> Casainho:
>
> I too am looking forward to using your board, if
> you need test and development help (hardware of
> firmware) let me know. Any Idea when the first
> ones will be available and what do you project the
> cost to be?

Thanks.

We are busy working to get the boards available for everyone, which should happen soon.

Right now, we have 2 prototypes which we are fully testing. Due to the power of the ARM 32 bits @ 100MHz, we want to go ahead from actual RepRap and having the stand alone printing - we have a uSDCard working on the board and we will need to reuse/adapt the firmware for stand alone printing, something everyone will be able to help on it after we release the board (mainly porting the code from Makerbot??)

We do not have a price yet, we are in talks with commercial/buys guys.

Thanks.

Edited 1 time(s). Last edit at 04/08/2011 11:51AM by casainho.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Hi Casainho,

We have a few machines we would like to test your electronics on (also experience in ARM 32 as well) so I hope we can help you in some way soon.

We are keenly waiting on the release of your boards, any way we can get one to test sooner?

We are in Barcelona, Spain and may have a lot in common.

cheers,
David

davebass Wrote:
-------------------------------------------------------
> Hi Casainho,
>
> We have a few machines we would like to test your
> electronics on (also experience in ARM 32 as well)
> so I hope we can help you in some way soon.
>
> We are keenly waiting on the release of your
> boards, any way we can get one to test sooner?
>
> We are in Barcelona, Spain and may have a lot in
> common.
>
> cheers,
> David

Hello.

We don't have the boards produced yet, however we have a working prototype and we are busy working for production the boards. We are in talks with a very well know 3D printer, to change from RepRap/AVR electronics to our ARM boards.

I will contact you with a PM. Thank you.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Any updates? I'm really looking forward to this board and thinking about using it in my upcoming nutless Mendel build.

Carmiac Wrote:
-------------------------------------------------------
> Any updates? I'm really looking forward to this
> board and thinking about using it in my upcoming
> nutless Mendel build.

Hello.

Yes, see project site: [www.3dprinting-r2c2.com]

And we are running the RepRap Mendel @ 200mm/s... we will post a video soon.

Thank you.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

Really, I only have two questions. When and how much? I'm ready to put in a pre-order as soon as possible.

Carmiac Wrote:
-------------------------------------------------------
> Really, I only have two questions. When and how
> much? I'm ready to put in a pre-order as soon as
> possible.

We hope to soon have the pre-order, maybe in 2 or 3 weeks. We don't have the price until we have produced boards, please wait a bit more.

Thanks.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]

We finally have the boards available, this time for developers and beta testers. I updated the original message with the regarding information.

You can simple go here http://www.3dprinting-r2c2.com/ and read about it.

Thank you.

---

---
New cutting edge RepRap electronics, ARM 32 bits @ 100MHz runs RepRap @ 725mm/s:

[www.3dprinting-r2c2.com]