gen 4/5 electronics

So I was poking around the makerbot wiki and came across these:

[wiki.makerbot.com]

[wiki.makerbot.com]

Gen4 seems interesting simply because the magnetic encoders will indicate the exact position of all the planes and the distance the extruder driver has turned with phenomenal accuracy.

Gen5 seems more interesting to me because it opens the door to zero custom electronics, basically a linux SBC running real time linux. Connect that to something like the pololu stepper board: [www.pololu.com] and something else standard for the extruder and that would both cut the costs and complexity way down.

Any thoughts? I know there are heat issues on the pololu board, any other feedback?

For the gen5 they seem to be looking at some pretty hefty processors - the one they said they liked the best had 64MB RAM!

IMHO, nothing like that strong a processor is needed, and it'll just add cost.

Personally, I think we'll end up with something like the Xduino. The chips range from around $5 to around $15, so it's about the same pricepoint has the atmel's we're using. It has plenty of I/O including built-in USB, I2C, SPI, even CAN on some models. And most importantly it's fully supported by the GCC compiler suite that we're currently using.

It'd be childs play to hook up a SD card reader, simple 2 line 32 char LCD screen, etc. It would have plenty of horsepower and program space to read, interpret and control multiple steppers/heaters/thermistors etc. all at the same time. I don't see where any software we'd want to run on it would ever need 64MB of RAM.

Huh. I just re-read that last sentence and am reminded of Mr. Gates saying something similar. Oh, well. Maybe it's just that I'm such a tightwad. I would love to see the electronics come down to around $50... Price it a HUGE issue to me, and I suspect to many others. So let me rephrase that last bit.

I don't see where any feature that needed 64MB of ram would be anywhere near worth the additional cost.

The only "feature" that I can see that might need it is an embeded RTOS, and that's not a feature except to the guys writing the firmware. It makes it closer to development on a standard PC, and thus easier to write to for your average programmer. Which IS nice, but not worth another $20 - to me, at least. Especially since all the really complex stuff should be done on the PC side anyway.

$10 for the processor, $5 for each stepper driver, $15 for the thermocouple, $20 for sundries. $50. Should be reasonable to do. Yes, I'm ignoring PCB costs, but I'm really hoping those'll go away soon when the reprap can place and connect it's own surface mount chips and such.

But whatever.

Put up or shut up, and since I haven't put anything up, I should probably be quiet. Not that I will.

---

--
I'm building it with Baling Wire

For me the electronics cost $200 from makerbot so the gen 5 would be much cheaper and a lot easier to write much better software for, also something very handy like either wifi or ethernet. If you look closely they are including the real time patches to the linux kernel.

To me the current electronics are a barrier to entry, there is essentially one manufacturer, makerbot, Sure you can build them yourself but most of the people I know (myself included) that want to participant can't even begin to do that or the effort in terms of time makes zero economic sense. If makerbot is out of stock then it also means that those people are totally excluded until they have stock back in. I hate depending on a single manufacturer. The arm9 boards are getting pretty widely available (at least a dozen sources already) and the cost is dropping precipitously, they are pretty much the cheapest way to run almost full blown linux.

In terms of power of processor I have to agree with you about the gates quote, I really don't think there will be a problem coming up with ways to soak up those extra cpu cycles and RAM.

I too am sensitive to cost but for the time it takes to build a board even at minimum wage you could easily make much more than the assembled cost of the board. For me it would have taken at least 25 hours to source the parts, pay for their cost and buy a hotplate, and then actually assemble them assuming I didn't screw them up. I do pretty well but even if I made minimum wage 25*7 = $175 so I don't get it.

The pololu stepper board is also just one manufacturer but I think it could be expanded pretty easily (say sparkfun, etc) if there is demand because is it general purpose rather than reprap specific.

All things being equal I would rather plug the machine into my laptop and not having any more electronics than a super-sized USB cable (say with a lot of i2c or GPIO lines) but I understand why that probably isn't a good idea on several fronts.

I concur about the silliness of the electronics cost and supply issues, which is why I've totally ignored reprap electronics and built my own from whatever pieces seemed suitable.

---

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

As a new 'core' reprap team member, I can say I'm really happy that Makerbot ( among others) is giving a turn-key solution to the electronics for people that want/need it. Personally I'm not using any makerbot electronics at all. ( Arduino Mega with gen2 DC drivers from rrrf.org and my own weird stepper drivers from a local supplier. )

If you don't like makerbots success, feel free to download their open-source eagle designs, and sell your own variants. :-) I think it would be good for everyone!

Buzz.

Economically it really only makes sense to have boards made in 50 unit plus orders. I didn't mean to imply that what makerbot is doing is a negative in any, I think it is great and generally a good sign. However for an open source project to so fundamentally depend on a commercial enterprise I think is a bad idea even though makerbot seems at date a very open company. Most of its investors are people tied up in reprap as far as I know which also presents some potentially troubling conflicts of interest down the line.

As I said perhaps you (buzz) can build the electronics yourself but the majority of people I know with interest couldn't possibly. I think an analogy would be if gimp or inkscape required the user to know assembler in order to get running or to hire someone who did... I think the whole "works for me" attitude is really a problem in this community and a huge hindrance towards widespread adoption.

I also recognize that "open hardware" is something still in its infancy and there aren't even really best practices yet.

It was disappointing that the reprap foundation ceased being in the parts business though I understand it.

From an investment standpoint it seems like it would just make more economic sense to spend my time money in redesign so you don't need custom parts. The solution is to just move to standardized boards which are widely available, this seems to be where makerbot is going anyway, I doubt they like being in the custom electronics business all that much either.

How about making the interfaces open (or hmm.. I guess they are already open.. but are they well specified?) and encourage people to publish their own solutions for electronics?
The design itself should be enought modular so it would give temptation to replace parts of it with better solutions. And hopefully publish the new solution.

For example, could the whole exlectronics be replaced with ethernet step motor controllers and make all the CPU crunching on the PC, without any mother board. All needed electronics would be 3-4 step motor controllers connected to the same bus:

[www.probyte.fi]

Edited 1 time(s). Last edit at 02/08/2010 01:30AM by nite69.

nite69: currently, the stepper boards accept standard step/dir signals. The extruder is on an RS485 bus in the makerbot design, but reprap project bypasses that with an I2C connection to avoid timing issues between the xy steppers and the extruder stepper.

Any bus between motors that has variable or medium latency (ethernet, async serial, usb, etc) is a no-go because the motors must remain tightly synchronised for high quality printing.

Feel free to set moves up using whatever bus you like, but then you also need a sync pulse, and you're back to gen1 electronics with a fancier interconnect.

---

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

I think the electronics is more expensive than it could be.

There is a lot to be said for just stepping back from the current electronics.

Looking at what is required defining a spec then designing a solution that fits the requirement.

Part of this spec should be:-

Ease of building electronics for a novice at home.
Ability to obtain parts as a hobbyist.
Cost of the parts.
Design should allow for easy experimentation / expansion. or "Hacking"


The design should allow for a multi tiered user base where there is a basic working system. Which can be upgraded to the cutting edge version i.e. multiple heads colour in afforabe easy upgrades.

For example Mendel Mrk X is the basic starter version Mendel.

MKX++ incorporates new ideas like a bowden extruder, heated bed, higher resolution, faster print speed etc. to get a Mendel MkX++ you reach it by having Mendel starter

The final tier Being Mendel Developer version using cutting edge designs.

I have seen a lot of people start building RepRap that then quietly disappear as they hit the problem of an ever moving target.

Maintaining New RepRapper starter system that's guaranteed to give satisfaction should prevent this.

The Makerbot solution provides the garenteed result at a cost, but its not realy a DIY project its a Kit of bits to build.

---

Bodge It [reprap.org]
=======================================

---

BIQ Sanguinololu SD LCD board	BIQ Stepcon	BIQ Opto Endstop
BIQ Heater Block PCB	BIQ Extruder Peek clamp replacement	BIQ Huxley Seedling
BIQ Sanguinololu mounting	BIQ standalone Sanguinololu or Ramps mounting	Print It Stick It Cut it

My rep strap: [repstrapbertha.blogspot.com]

Buy the bits from B&Q pipestrap [diyrepstrap.blogspot.com]
How to Build a Darwin without any Rep Rap Parts [repstrapdarwin.blogspot.com]
Web Site [www.takeaway3dtech.com]

ARM Cortex M3 looks like the way to go! Because provides more processing power at less price! And ARM have the same Open Source tools PLUS the debugger! And I think debugger is important for such big and complex systems that RepRap is!

I am being working a little with ARM from Phillips NXP and Atmel. NXP have some cheap and nice ARM Cortex M0 and M3.
The M0 goes to 32Kbytes flash and have a version with USB and not USB. The M3 goes to 512kBytes flash and have USB and Ehternet. Both runs at 100Mhz.

Opendous provides FULL Open Hardware based on this Cortex M3 and M0.

Starting from simple and cheap, here a LPC1313 Cortex M0 board, without USB (USB connector is just for providing power):


The same board but this time with LPC1343, which have 32kbytes flash and USB. It also have a USB bootloader from factory that presents as a USB flash disk, meaning that users can simple and quickly flash code, much more easy than Arduino!


And a board with LPC1758 Cortex M3:


Features

* based on LPC1758 ARM Cortex-M3 MCU
* Full-Speed USB (12Mbps) and Ethernet (100Mbps) capability
* USB-miniB and USB-A connectors and on-board GPIO-controllable TS3USB221 USB switch for USB Device and Host modes
* all pins not related to USB and Ethernet are exposed on the outer headers
* microSD connector in SPI Mode on the underside of the board
* 10/100 Ethernet based on the LAN8720 PHY
* Cortex Debug Header for JTAG/SWD
* 3.3V 1A+ +/- 2% ST1L05B regulator
* RESET button, NMI button, and USB status LED
* 12MHz main crystal and 32.768kHz RTC crystal
* Dimensions (LxWxH): 3" x 1" x 1"- plugs into standard solder-less breadboard
* completely OpenHardware Design, schematic/layout/assembly files available for Open Source KiCAD. Gerber and other technical files can be customized.
* Design is 2-layers with 7mil/7mil trace/space and minimum 15mil Vias

I had worked a bit in past with ARM9 at 200MHZ with 64Mbytes of SDRAM and etc, and I believe there is no need for such system!!! Also they are not ideal for low hardware control, as when we need PWM, ADC, motor control, etc... ARM Cortex M0 and M3 were made for that tasks, mainly to reach the market for the 8 and 16bits MCUs :-) -- the great are that ARM have Free Software tools and even better than AVR.

Edited 2 time(s). Last edit at 02/14/2010 08:01AM by casainho.

---

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

[www.3dprinting-r2c2.com]

Not sure if anyone has checked out beagleboard...

[beagleboard.org]

Cost is $150...so maybe too high, but very capable board.

The LPC13xx series is cortex M3

The LPC11xx series is cortex M0

Edited 1 time(s). Last edit at 02/13/2010 07:56AM by annodomini2.

Another point 1 key feature of the LPC13 series is they have integrated USB ISP bootloader, which makes the flash of the device behave like a conventional USB memory stick. This is 'programmed' at the factory in the boot ROM. So no messing around with ICSP programmers.

annodomini2 Wrote:
-------------------------------------------------------
> Another point 1 key feature of the LPC13 series is
> they have integrated USB ISP bootloader, which
> makes the flash of the device behave like a
> conventional USB memory stick. This is
> 'programmed' at the factory in the boot ROM. So no
> messing around with ICSP programmers.

I think I did mention that. And yes, that's a great feture that makes it more acessible than Arduino :-)

However LPC13xx just goes to 32kB flash, which can be enough for extruder board but I guess not enough for main board...

---

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

[www.3dprinting-r2c2.com]

casainho Wrote:
-------------------------------------------------------
> annodomini2 Wrote:
> --------------------------------------------------
> -----
> > Another point 1 key feature of the LPC13 series
> is
> > they have integrated USB ISP bootloader, which
> > makes the flash of the device behave like a
> > conventional USB memory stick. This is
> > 'programmed' at the factory in the boot ROM. So
> no
> > messing around with ICSP programmers.
>
> I think I did mention that. And yes, that's a
> great feture that makes it more acessible than
> Arduino :-)
>
> However LPC13xx just goes to 32kB flash, which can
> be enough for extruder board but I guess not
> enough for main board...


Binary size is a component of the code, compiler, instruction set/core design. Just because it currently doesn't fit in an AVR 8Bit micro with 32k, doesn't mean it won't in another microcontroller.

Also there is much room for improvement in the current firmware. Look at Triffid's work on the FiveD for Arduino.

And my own investigation of the system has highlighted many optimisation points.

Example: A 32Bit Microcontroller will typically perform 16 and 32 bit math operations much more code efficiently than a 8Bit one.

> Binary size is a component of the code, compiler,
> instruction set/core design. Just because it
> currently doesn't fit in an AVR 8Bit micro with
> 32k, doesn't mean it won't in another
> microcontroller.
>
> Also there is much room for improvement in the
> current firmware. Look at Triffid's work on the
> FiveD for Arduino.
>
> And my own investigation of the system has
> highlighted many optimisation points.
>
> Example: A 32Bit Microcontroller will typically
> perform 16 and 32 bit math operations much more
> code efficiently than a 8Bit one.

Hmmm, ok. But 32kbytes flash for RepRap main board? where there is SD Card and FAT drivers?

I really like that LPC1343 board from Opendous, because it's FULLY Open Hardware and very simple, cheap. Looks much much better to me than Ardunio (unless on Arduino API which are very well known).

I wish I could mill that LPC1343 board on my CNC... but that LQFP of 0.5mm pitch is hard. I can solder it easily on board like that from Opendous, but I guess I can't mill such board...

But if LPC1343 don't have the space we need, we can go to LPC17xx, since the package is the same LQFP 0.5mm pitch.

---

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

[www.3dprinting-r2c2.com]

If you want a cheap LPC1343 look at LPCxpresso, their £20 and come with free dev tools and a full JTAG.

Ok its not the open solution, but it will help with development.

annodomini2 Wrote:
-------------------------------------------------------
> If you want a cheap LPC1343 look at LPCxpresso,
> their £20 and come with free dev tools and a full
> JTAG.
>
> Ok its not the open solution, but it will help
> with development.

Are you from their marketing team? If we have better Free Software and Open Hardware solutions, why would we go with that closed commercial product?

Are you aware of the problems and disadvantages of closed products?

Edited 2 time(s). Last edit at 02/15/2010 06:13AM by casainho.

---

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

[www.3dprinting-r2c2.com]

casainho Wrote:
-------------------------------------------------------
> annodomini2 Wrote:
> --------------------------------------------------
> -----
> > If you want a cheap LPC1343 look at LPCxpresso,
> > their £20 and come with free dev tools and a
> full
> > JTAG.
> >
> > Ok its not the open solution, but it will help
> > with development.
>
> Are you from their marketing team? If we have
> better Free Software and Open Hardware solutions,
> why would we go with that closed commercial
> product?
>
> Are you aware of the problems and disadvantages of
> closed products?

No I do not work for them, I was attempting to offer a cheap solution.

The dev tools are GCC based and the IDE Eclipse based.

I am currently looking into making a OpenOCD compatible firmware for the JTAG unit. It operates from USB DFU and is not saved in the device.

I understand the need to keep the IP open source to allow other users to re-create, but you will see that the boards you have listed have the new 10pin cortex m3 JTAG connector, same as supplied on the LPCxpresso JTAG.

If you keep the library sets to CMSIS and newlib and not use the Code Red libraries, you will be able to develop the firmware for this system and also have a very cheap JTAG to support that development.

There is no need to use the LPCxpresso board itself as a part of the design, but the firmware developers will require the JTAG to understand problems in the code.

It offers a cheap development capability.

I just posted an ASM framework to GITHUB, (HIDAVRGcode) which indicates that most of what is needed for a hardware controller (hPod) will fit into a mega88. Or a series of mega88s.

The direction this branch is taking, is more to move the controller core to a real time Linux kernel. Not unlike what EMC2 has evolved into. The advantage of using something like a PC103 is direct access to the internal processor bus. The future of the generic parallel port being suspect.

This may be a better one click solution. Much easier to close and control into a product which has sale value.

-julie

I've added [[Generation_4_Electronics]] and [[Generation_5_Electronics]] to the wiki.
[objects.reprap.org]
[objects.reprap.org]

It may be good to give Generation 4 its own forum, since we'll see a lot of activity using those boards.
What do you guys think?

The direction this branch is taking, is more to move the controller core to a real time Linux kernel. Not unlike what EMC2 has evolved into.

If we work hard at it, we might reinvent EMC2.

SebastienBailard Wrote:
-------------------------------------------------------
> I've added [] and [] to the wiki.
> [objects.reprap.org]
> onics
> [objects.reprap.org]
> onics
>
> It may be good to give Generation 4 its own forum,
> since we'll see a lot of activity using those
> boards.
> What do you guys think?

I think there should be available money for a group of people be able to buy PCBs to experiment.

There are a lot of ideas, but someone needs to go ahead and prototype a board... -- a board sensible to cost and DIY connections.

If RepRap accepts donations, isn't there any money to invest on this next generation of electronics? -- maybe we could use BatchPCB services, where we could prototype there and in the end users can buy them directly the PCBs!

---

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

[www.3dprinting-r2c2.com]

casainho Wrote:
-------------------------------------------------------
> SebastienBailard Wrote:
> --------------------------------------------------
> -----
> > I've added [] and [] to the wiki.
> >
> [objects.reprap.org]
>
> > onics
> >
> [objects.reprap.org]
>
> > onics
> >
> > It may be good to give Generation 4 its own
> forum,
> > since we'll see a lot of activity using those
> > boards.
> > What do you guys think?
>
> I think there should be available money for a
> group of people be able to buy PCBs to
> experiment.
>
> There are a lot of ideas, but someone needs to go
> ahead and prototype a board... -- a board sensible
> to cost and DIY connections.
>
> If RepRap accepts donations, isn't there any money
> to invest on this next generation of electronics?
> -- maybe we could use BatchPCB services, where we
> could prototype there and in the end users can buy
> them directly the PCBs!


BatchPCB is fine for U.S. based users, but there is a world outside the US and paying $20-30 shipping for a $5-10 pcb is ridiculous.

> BatchPCB is fine for U.S. based users, but there
> is a world outside the US and paying $20-30
> shipping for a $5-10 pcb is ridiculous.

Thats true. Do you know other better options? I never used a PCB service... and I am from Portugal, Europe.

---

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

[www.3dprinting-r2c2.com]

I think there should be available money for a
group of people be able to buy PCBs to experiment.

The RRRF exist(ed) to organize and fund research like that
but I think the guy running it may have moved on to other things.

You could email Adrian or Zach and ask about a set of boards for
people like the Heated Bed team guys? They are very busy people,
so I don't know what will come of it.

Edited 2 time(s). Last edit at 02/18/2010 05:40PM by SebastienBailard.

---

-Sebastien, RepRap.org library gnome.

---

Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!

SebastienBailard Wrote:
-------------------------------------------------------
> I think there should be available money for a
> group of people be able to buy PCBs to
> experiment.
>
> The RRRF exist(ed) to organize and fund research
> like that
> but I think the guy running it may have moved on
> to other things.

Makerbot sells and then can have money to invest, but, as work really a lot of time, I guess they can have some money but not the time :-)

Their heated bed solution/product seems very poor when compared to NopHead and others ones, which were very well reported and documented.


> You could email Adrian or Zach and ask about a set
> of boards for
> people like the Heated Bed team guys? They are
> very busy people,
> so I don't know what will come of it.

I would first define a path for RepRap. Mendel is stable and looks great. In what would developers/hackers should invest their next energies?

I think there sould be a poll and a discussion to define were we should focus. I would say that heated bed is a must add on module for RepRap.

Maybe electronics also, but why? to get lower prices of finals total electronics?

And I guess support material would be very important! and a laser head for engraving and future work on powder-based stereolithography printer!

And after we get a path defined for RepRap next works, then we could go to shops as Makerbot and ask money/or parts for development. Or maybe could be better to make a poll fo money, asking for donations, but for that, money usage should be very well justified!

---

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

[www.3dprinting-r2c2.com]

The thing with makerbot I see happening is that it makes absolutely no sense for them to sell parts and kits for just electronics or subsystems from a business standpoint. As long as they can continue to sell out of complete kits each batch it is pointless to sell part kits. To do so is just taking away from the higher margin complete kits and fostering 'competition' from other projects. Given their explosive continuing growth I would doubt they will have electronics kits in any substantial stock ever again though would love to hear differently.

I think in the past the kits were a way to unload surplus inventory, with their growing volume that isn't a problem anymore, in fact I would bet it is a distraction from their core business that isn't worth their while.

Right now reprap is completely stuck, new participants can't get assembled electronics or printed parts. Contrast that with makerbot which will be moving 1,000 kits a month or more in 12-24 months time. If you want to get a reprap what do you have to do today? Buy a makerbot... oops.

goinreverse Wrote:
-------------------------------------------------------
> Right now reprap is completely stuck, new
> participants can't get assembled electronics or
> printed parts. Contrast that with makerbot which
> will be moving 1,000 kits a month or more in 12-24
> months time. If you want to get a reprap what do
> you have to do today? Buy a makerbot... oops.

You are right, I totally agree with you. However people are selling to other the repraped parts, but that must be a very small number compared with that 1,000 kits of Makerbot a month.

I have nothing against Makerbot, they even help me in the crucial task of getting electronics, or I wouldn't be able to build my RepStrap.

RepRap Mendel is not good for build and assembly commercially, that's why Makerbot went with other solution, which is perfectly understandable.

---

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

[www.3dprinting-r2c2.com]

Hi, this is Ryan from Pololu.

The heat issues you mention with the A4983 are not an issue with the board, but an issue with the way people are driving it. Some people were stepping the board with a very low duty cycle with too little current limiting. This leads to current spikes of over the 2 A per coil spec. of the A4983. If you stay within the specifications there will be no heat problems, and if you want to exceed them you can use a heat sink.

- Ryan Mulligan