Arduino Project Upgrades With 2 New Boards

EqualSlash writes "The Arduino Project is releasing two new boards — Arduino Uno to replace Duemilanove and Arduino Mega 2560 to replace the existing Arduino Mega board. With Uno, the board is not just getting a new pronunciation-friendly name but also has a custom-made USB-serial converter to replace the older FTDI chipset, thereby removing the need to install drivers (they now have their own USB Vendor ID). It now has a logo and stylish packaging, and soon will have its own branded web store. A new Ethernet integrated board and a tinkering toolkit will be made available shortly."

Comments

The netduino has been updated as well.

[DAldredge]

http://www.netduino.com/netduinoplus/specs.htm Secret Labs is launching its .NET-friendly Netduino Plus, which adds Ethernet and microSD to a regular Netduino board (which in itself is a sort of high powered, Visual Studio-compatible Arduino, with a 32-bit 48MHz ARM7 processor, instead of Arduino's 8-bit number, but pin compatible with Arduino "shields").

Re:The netduino has been updated as well.

[batkiwi]

Dotnet micro is under the apache 2.0 license.

Fully open source under a certified license.

What are you complaining about?

Re:The netduino has been updated as well.

[arielCo]

Dotnet micro is under the apache 2.0 license.

Fully open source under a certified license.

What are you complaining about?

The toolchain is far from open.

Re:The netduino has been updated as well.

[batkiwi]

If you know what you're doing you can use mono + monodevelop. THere are howto's out there, and they're getting close to having a "1 click install" like arduino is.

So any more problems?

@timothy

[Anonymous Coward]

cool story bro

Re:@timothy

[Anonymous Coward]

Am I the only one to read "Arduino Project Upgrades With 2 New BOOBS" ?

Awesome

[Anonymous Coward]

Just in time for my gibaboo project. Hope the ISA accelerator supports G5 ALG microcode. Is it still using S-100 pinout adaptor? The heat dissipation core was generating impedance mismatch exception.

stupid spacing

[X0563511]

Did they finally get rid of that lame spacing goof?

Re:stupid spacing

[X0563511]

... nope. Buried in the link says they kept the form factor compatible... bah. If you worry about shield compatibility then just sell a cheap shield adapter and fix the damn design!

Re:stupid spacing

[TheGratefulNet]

the official arduino shield spacing is goofy. I avoid it and adapt the 'italy boards' to normal 'point one' spacing that the rest of us sane people use ;)

in fact, you can build an atmel 328 chip with a 16mhz resonator, 1 or 2 resistors, 1 or 2 caps and some wire. its REALLY that trivial. I've built at least 10 on proto boards (perf boards) all soldered by hand pt-pt and they all are nice little embedded systems. wire in the 'classic 6 pin inline' connector to talk to the magic $20 ftdi cable and you are all setup to do programming and test. the perfboard arduino (assuming you have the chip init'd once) costs about $5. the $20 cable is a one-time purchase and has the ftdi chip in it.

if you are not using the shield form factor, you are not bound to the 'funny hole spacing' they picked for their shields.

the shield idea was good. going with something other than .1 spacing was just nutty beyond belief; and I don't mean that in a good way.

Re:stupid spacing

[nickersonm]

in fact, you can build an atmel 328 chip with a 16mhz resonator, 1 or 2 resistors, 1 or 2 caps and some wire. its REALLY that trivial.

You have a goldmine there - I'm sure Atmel would pay a lot to be able to duplicate their IC functionality with a few discrete components!

Re:stupid spacing

[Yetihehe]

You can do it even without resonator, atmegas have internal 8mhz oscillator. I've made some projects with atmegas, never used arduino, but if they have spacing other than .1, it's just design error or plainly dumb.

Re:stupid spacing

[X0563511]

It was a mistake, but by the time it was caught at least someone was already printing shields.

Dumb mistake, and now a legacy "feature"

Re:stupid spacing

[Abdul Jakul]

Great! now I will have to upgrade my motherboard again.. sad thing about being an enthusiast is the sick feeling you get when you are behind tehnologies.. expecially in Motherboards

PS3 controller

[vlm]

From http://arduino.cc/blog/2010/09/24/dinner-is-ready/

more advanced users will be able to reprogram the USB chip to make the board show up as a variety of USB devices (Keyboards, Mice, Joysticks, MIDI etc)

How bout a PS3 controller?

Re:PS3 controller

[dcuartielles]

well as far as I know the ATmega8U2 doesn't really have the memory space to make the PS3 controller, but it is not that far from possible :-) If you manage to make the code, I send you board as a present

New custom chipset...with no drivers?

[Anonymous Coward]

How do you make a new custom usb to serial chipset, with a new vendor id, that doesn't require drivers?

I'm no expert, but have used PL230x chipset cables and FTDI cables, and both required some kind of driver under OS X. I think windows may have included the PL230x driver, but not the FTDI. And Linux includes drivers for both. But if I took the FTDI or PL230x chipset and changed its vendor id, then the driver under linux won't detect it by default.

So how is it that a new custom chipset with a new vendor id requires no drivers? Is there a standard for USB->serial that if you follow you don't need a driver? (and why do most cables not support it if there is?)

just wondering...

Also, since the pl230x chipset is a pain the ___ at times due to inconsistent implementations in cables and buggy drivers, I certainly hope they don't make things worse. FTDI has been the gold standard in my experience.

Re:New custom chipset...with no drivers?

[X0563511]

It probably uses a class that requires no drivers.

Every new thumbdrive out there doesn't need a new driver, and they manage with new VID/PIDs all the time... because they all flag themselves in the Mass Storage class.

Re:New custom chipset...with no drivers?

[Anonymous Coward]

since the pl230x chipset is a pain the ___ at times

Tell me about it. My ___ is still sore as hell. This one time my ___ hurt so much I had to go to a ___ologist to have my ___ examined. Turned out that I had ___ cancer, however it was in an early stage and the cancer was removed.

Re:New custom chipset...with no drivers?

[makomk]

The only standardised USB class for serial-like USB devices is the CDC ACM class. That's driver-free under Linux and Mac OS, but it's not exactly driver-free under Windows - you still have to install a .inf file to tell the driver shipped with Windows to actually load. In practice, for Windows users it's not really better than any of the proprietary driver-required solutions.

Re:New custom chipset...with no drivers?

[cheater512]

Yup they say in TFA that Windows needs a .inf

That must be what they are using.

Re:New custom chipset...with no drivers?

[Andy Dodd]

"So how is it that a new custom chipset with a new vendor id requires no drivers? Is there a standard for USB->serial that if you follow you don't need a driver? (and why do most cables not support it if there is?)"

As others have said - the CDC ACM class. I believe that class was standardized relatively recently in an attempt to address the fact that there are 23432743290 different USB-serial chipsets all of which required special drivers.

Not sure if XP has CDC-ACM support built in (might have been added in a service pack?), but in general, CDC-ACM is the "new kid on the block" and the legacy (PL230x and FTDI) approaches are still well entrenched.

Re:New custom chipset...with no drivers?

[ZorroXXX]

No, the CDC ACM class is not that new; it has been used for mobile phones for several years. It is at least present in the "Wireless Mobile Communications Devices" document dated February 9, 2007. See http://www.usb.org/developers/devclass_docs/CDC1.2_WMC1.1.zip for full details.

Where's the stream?

[MadGeek007]

The blog post mentions a stream but does not say where it can be found...

Re:Where's the stream?

[vlm]

The blog post mentions a stream but does not say where it can be found...

I think its:

http://www.ustream.tv/channel/world-maker-faire-1

Not a rick-roll, I promise.

Here's one...

[bosef1]

How about some drivers that implement the other serial features of the existing hardware. You know something other than 8N1. It's not like the chipset manual provides functional code snippets to implement those features or whatever...

And we care because?

[SirGarlon]

The summary would be more useful if it mentioned, you know, what the board is for. In case some of us haven't heard of it or something. Yes, I did RTFA. It didn't say either.

Re:And we care because?

[MadGeek007]

It's basically a prototyping system for the rest of us (non electrical engineers).

Re:And we care because?

[Anonymous Coward]

The Arduino system is a development kit around Atmel Mega8 microcontrollers. It makes microcontroller development simpler for those without a background in programming or hardware design by providing easy programmability with a boot loader and an integrated USB to serial converter, ready-made "shields" (pluggable boards with specific functions) and a software environment which abstracts from some of the nastier aspects of microcontroller programming. It has got quite a fanbase in the "maker community".

Re:And we care because?

[zlogic]

It's an awesome tool that makes interfacing with real equipment (lights, motors, sensors etc) easy for a software developer with minimum electrical engineering knowledge and some knowledge of C programming. After it's programmed, it can run on its own without a PC.

Re:And we care because?

[Anonymous Coward]

You can do that with any PIC without any board. It just needs a power supply.

PICKit 3 or similar allow these chips to be programmed very easily. And no, I'm not an electrical engineer. But I know a resistor from a capacitor! :P

Re:And we care because?

[Anonymous Coward]

No, but your mommy does.

Re:And we care because?

[Anonymous Coward]

It is a basic system for prototyping. You have analog inputs, PWM (analog) outputs, digital inputs and digital outputs. Some of the more basic tasks are lighting up LEDs and reading potentiometers. I've used the digital I/O to drive several SSRs to run Christmas lights based off of environmental input which included, trip sensors, filtered external noise, ambient light and control switches that allowed the user to turn everything constant on or run one of several patterns under program control.

I've seen some very well done robots made with an Arduino.

Re:And we care because?

[drinkypoo]

You didn't even use the word "Shield", which is the core of all that is awesome about Arduino.

Re:And we care because?

[Anonymous Coward]

Yiikes!!! What nerd hasn't heard of Arduino? Check it out.... http://www.arduino.cc/

Re:And we care because?

[SirGarlon]

It looks like their page is back up now. Was it Slashdotted or something?

Re:And we care because?

[BillX]

Those with a programming and hardware background, too :-) Once you get used to the pin-numbering abstraction, it's a real suck-saver for quick n dirty / one-off / non-production projects. The entire Arduino project, toolchain and most users' projects are open-source, and it's its own bootloader! One of the Mechies calls up and needs a quick test fixture to cycle a valve once per minute and log a sensor reading for the next several weeks. I can grab someone's microSD-FAT library off the internet and cobble something together in the time it takes to figure out whose desk the PIC programmer is hiding under this week and where the license code for their C (not C++) compiler went.

Re:And we care because?

[Joebert]

I thought I understood what it was for a little better, then you had to go an emphasize the word "shield" and now I'm confused again.

Re:And we care because?

[drinkypoo]

Arduino is a platform based on some minimal hardware wrapped around the Atmel AVR microcontrollers. It makes it very easy to slap the AVR into the middle of a design (especially if you use the Arduino Nano, which is designed to be conveniently breadboard-socketable) and actually get something done with it, like read some sensors or flash some LEDs, or perhaps read some sensors and flash the LEDs based on the input. Indeed, you can load the Arduino bootloader into sufficiently capable AVR processors and use their programming environment, libraries, etc. Arduino devices understand how to read many common sensors and control many common devices, including R/C car servo-motors. All this makes them an extremely common basis for small robotics projects. As well, Arduino devices are extremely inexpensive, and the microcontrollers upon which they are based even moreso.

But that wasn't very hard to begin with, and what's REALLY super-cool about Arduino is the idea of "Shields", which are sub-boards of hardware designed to be interfaced with an Arduino simply by plugging it in, and which are provided with the code necessary to utilize them. Thus, adding some functionality to an Arduino project can take one of two forms: you can build it out yourself and interface it, or you can simply buy someone else's premade "Shield", plug it in, and use their code to interface to the hardware.

Shields generally utilize Free and Open designs and so you get all the source and the full schematic. This means that while you can use Arduino for making finished products, you can also use it for rapid prototyping, then take the (Free, Open) Arduino schematic, combine it with the (Free, Open) schematic(s) of the shield(s) that you used, and send the whole thing off to get made into a permanent PCB for your own product. This in turn permits someone with programming ability but little to no electronics ability to produce finished products with semi-custom hardware.

Shields exist for a broad variety of functions, and generally they add input and/or output to a design. There are shields to control relays, and shields which can be used to determine how much power is flowing through them. There are LED blinking shields, Ethernet shields, sensor-control shields, ZigBee wireless mesh networking, video output, you name it. And anyone can produce any of this stuff from the designs, modified or not. Arduino is open hardware for the masses today, albeit a bit limited in processing power. However, it can always be linked to a more powerful computer; it has RS232 onboard (one of the tricks that Atmel teaches you to do with their AVR when you get the dev kit, in fact) and most designs include a USB to RS232 adapter which both powers the device and permits communication.

There is probably a much better way to summarize all of this, it could be a lot shorter, but I've never typed a full missive on what it is. Maybe someday this highly redundant piece of text can be used to train machine language or something :D

Re:And we care because?

[Joebert]

I read your second paragraph, the bit about what shields are unconfused me. I was thinking shielding from electro-bigwordic interference or from eves dropping hackers.

Re:And we care because?

[societyofrobots]

The Arduino, in all honesty, is a technologically obsolete poorly designed budget microcontroller.

There are other controllers, just as easy or easier to use, with much more features for only a little more $. And most of them don't even require expensive bulky shields to do what you need.

Just google my username, as I don't want to spam /. with my financially motivated bias =P

Re:And we care because?

[drinkypoo]

There are other controllers, just as easy or easier to use, with much more features for only a little more $. And most of them don't even require expensive bulky shields to do what you need.

Arduino is not a microcontroller, it is a system built around the Atmel AVR, which is extremely inexpensive for its overall functionality, though certainly you can find something cheaper. Shields give you access to rapid prototyping but you don't need to use them in your final design, or for that matter use them at all. Instead, you can plunk the $25 Arduino Nano down in the middle of a breadboard, and then build the shields' circuit on your breadboard next to it.

Put simply, shields give additional functionality to the hardware-clueless. Further, Arduino offers you a C programming environment, making it still more accessible. Is it the most efficient way to go about building something? Nope. It's targeted at a different market than a bare microcontroller stuck into a piece of foam or a rail.

Re:And we care because?

[societyofrobots]

I think you misread what I wrote?

Anyway, what you are thinking of is called an 'augmented microcontroller', of which I was referring too. The Arduino is a bare-bones budget level augmented mcu. There are much better augmented microcontrollers out there.

As for shields, they aren't nearly as useful as you think - they only exist because the Arduino doesn't have a power bus or a normal 0.1" header spacing.

Here is an example of another microcontroller - notice how the servo and the sharp IR simply plug in without a shield or additional wiring. Also, programming is nearly drag and drop.

http://www.youtube.com/watch?v=FPhMES0PFEw

And the people who don't need the hype

[Anonymous Coward]

...can still get Atmel Atmega8 chips for two dollars a piece and do everything the Arduinos do. These microcontrollers literally need no external hardware other than a power supply.

Re:And the people who don't need the hype

[Anonymous Coward]

Yeah, but they wouldn't be too useful without something connected to any pins other than power.

Re:And the people who don't need the hype

[Yvan256]

That's why I prefer simple, smaller carrier boards instead. They got the ISP header and sometimes a place for a crystal+caps or a resonator and a power supply connector at most.

I had a small batch made for the ATtiny85, and my PCBs only adds one row above and below for the carrier board pins (useful for connecting into protoboards), and 2.5 columns more to the width for the ISP header.

Re:And the people who don't need the hype

[imroy]

That's one of the great things about the Arduino - most of the hardware "magic" is really just the built-in capabilities of the Atmel AVR micro-controller. If you don't need to use a "shield" (daughterboard), there's plenty of simpler (and cheaper) clones for more specific purposes. They just need to be programmed with the Arduino bootloader (which they usually are) and the Arduino IDE will program it just fine.

Re:And the people who don't need the hype

[DwySteve]

Are you selling it? I've tried to make an ATTinyx5 board but... it kinda sucked. I'd be interested in your board.

Re:And the people who don't need the hype

[Yvan256]

I actually have a few, already soldered up and tested. I made a mistake with the hole size for the IC pin headers that go into the protoboard, so I needed to drill those holes to fit the pins. The ISP header holes are just fine, though.

I'll be putting those on eBay in a few days, just search "attiny85" and you should find it (currently only two results on eBay.com, for the IC only).

FYI, I connect to the ISP header via the old STK500, I have no idea if it will work with anything else, but all six pins of the ISP header are wired to the correct pins for the ATtiny25/45/85.

It's very similar to the one from Tinkerlog but with only the ISP header and a small capacitor between the power pins (hidden under an IC socket to take less room). I'd say it's roughly half the size of the Tinkerlog one.

Re:And the people who don't need the hype

[TheGratefulNet]

yup, small carrier boards (boarduino or similar) are what you want to use in deployed units that are in small numbers. for larger runs, you might as well design in the atmel chip and resonator and header and you're done; but small runs might benefit from a carrier board approach.

I don't use 6pin (3x2) ISP headers; I use the 6 pin inline 'ftdi standard' header. if you are doing arduino things, that's the only connector you need; and the bottom 3pins act as the power-in point (2 gnd and 1 for 5v). ie, you program the device using the magic $20 cable, you remove the cable, connect the bottom 3 to power/gnd and the device comes up and runs off that. the other pins are reset and tx/rx which go unused at runtime, anyway.

the 6pin inline is the one to go for for carrier-board style plugins.

Re:And the people who don't need the hype

[John Hasler]

You can even get just the chip (in a DIP package if you wish) programmed with the bootloader for about $5.00.

Re:And the people who don't need the hype

[Anonymous Coward]

Check out: http://make.larsi.org/electronics/ATtinyX5/
You can just order the board from SparkFun's PCB service. It is that simple...

Re:And the people who don't need the hype

[JohnBailey]

...can still get Atmel Atmega8 chips for two dollars a piece and do everything the Arduinos do. These microcontrollers literally need no external hardware other than a power supply.

So how do you get the program into the chip?

Re:And the people who don't need the hype

[makomk]

So how do you get the program into the chip?

With great difficulty these days. Unless you have a computer with a real parallel port lying around, you need a programmer of some sort as well. The nice thing about the Arduino is that it comes with a suitable serial bootloader already installed, and an integrated USB-to-serial interface too.

Re:And the people who don't need the hype

[Anonymous Coward]

Oh come on. USB-to-ISP programmers are cheap and there are lots to choose from. You can build one yourself if you want. You don't need to buy a new one with every controller either, because it's not tied to the application board.

Re:And the people who don't need the hype

[tibman]

I use these: http://www.solarbotics.com/products/kardw/

A nice 10$ kit/board. Pretty easy to solder together.

Re:And the people who don't need the hype

[JohnBailey]

With great difficulty these days. Unless you have a computer with a real parallel port lying around, you need a programmer of some sort as well. The nice thing about the Arduino is that it comes with a suitable serial bootloader already installed, and an integrated USB-to-serial interface too.

Exactly my point.
I got myself an Arduino starter kit last Christmas. First time I ever really tried doing hands on electronics. I actually learned and retained more in a week of playing with it than I did when I was reading electronics magazines regularly.

Re:And the people who don't need the hype

[JohnBailey]

Oh come on. USB-to-ISP programmers are cheap and there are lots to choose from. You can build one yourself if you want. You don't need to buy a new one with every controller either, because it's not tied to the application board.

Oh come on yourself. I'm not the one trying to compare a prototyping board to a chip.

And where do you get the idea that the ATMega is somehow tied to the Arduino boards?

Slashdotted

[iluvcapra]

They must be running their webserver on an Arduino.

Re:Slashdotted

[Anonymous Coward]

The frontpage says they are updating the server.

Re:Slashdotted

[laing]

What a coincidence that their server upgrade began just after the slashdot story was posted...

Re:Slashdotted

[Anonymous Coward]

Yeah, updating their server to an Arduino Uno.

Re:Slashdotted

[dcuartielles]

Well, we did not server upgrade, we did just a layout change :-) and FYI it was planned since 3 months ago that the change would happen at the same time the keynote was taking place in NYC so that there would be a before and an after the talk. If we got Slashdotted is a happy coincidence, but our server is ready to take Slashdot once a day if needed (we do pay for a pretty reliable server). Also as a side note, we registered over 20Gb of page traffic during the weekend, which is about 4 times the normal amount. I guess it is the added factor of the new boards, Slashdot, and so on ... however our server was never over 50% capacity and the downtime for the layout's update brought a total downtime of 35 minutes. We learned the hard way, since we were Slashdotted a year ago and then we had a serious problem :-)

There's nothing wrong with FTDI...

[Gordonjcp]

It doesn't need any funny drivers or anything, it Just Plain Works.

Re:There's nothing wrong with FTDI...

[zlogic]

On Windows you'll need to install a driver in order for USBserial to work. There were no drivers on Windows Update ~6 months ago so you need to install the drivers manually (possibly it's already fixed now). Otherwise the board is useless since it cannot be programmed.

Re:There's nothing wrong with FTDI...

[TheGratefulNet]

the only thing that does not 'plain work' is the rts/dtr hack to reset the cpu.

that, and, well, the ftdi cable is $20 and is needed to program the chip. the ftdi chip is not that much but its a PITA to solder to (fine lead pitch) and the usual solutions (sparkfun) have the board at $15, anyway. cut that out of the picture and things are finally cheap enough to be used by 'anyone'.

anyway, RTS/DTR idea is that you toggle one of those lines for a short while to discharge a small value cap across the cpu's reset line. its a neat idea and the cpu does NEED to be reset VERY prior to the first byte in the download. either you press it 'very quickly' or have software do it.

ftdi uses RTS and some other boards use DTR. that's one problem. and the other is that windows or linux (both) have no easy way to twiddle that bit and needed to call other routines to do the reset and then start the download (via what's known as 'avrdude', the downloader).

if they properly architected instead of hacked this, it would be a big step forward.

also, with regard to ethernet; I posted something about this on the ladyada forum and it did get some commentary. the problem with ethernet/ip is that there is no security (none, not even a tiny bit). for a physical device that can turn things on and off in the real world (including ruining/damaging things) you NEED security. I just firmly believe this. no firewall, no hosts.allow, no nothing. not authentication or encryption or MAC access control, just like TRULY nothing. I find that unacceptable in a real-world device.

given the fact that you can buy $50 things (pogoplug, seagate dockstar, even WRT routers) that run a full linux and IP stack, why hack around with dodgy ip-on-a-chip things (the arduino ethernet shield) when you can front-end the controller (arduino) with an embedded linux board that has proper IP features. as long as the linux board can shake 2 bits it can talk i2c. if it has 2 leds or 2 colors, it can shake 2 bits. there, you have all you need really to talk between the 1board linux plastic router thing and the arduino. AND you can run a real apache, php (etc!) there. even mysql ;)

the controller is great for fast polling of i2c devices and even spi and analog. its really really sucky for things like ethernet and IP and apps that sit on them. just not meant for them and its the wrong tool for the job.

(disc: I develop for arduino and have spent over a year on an embedded project using them).

Re:There's nothing wrong with FTDI...

[Sarten-X]

So Ethernet, at layer 2 of the OSI model doesn't offer the security functions of layer 6?

With all those rules regarding collisions, timing, and even the physical wiring, they didn't even throw in the simple requirement for all Ethernet devices to have a powerful processor, memory, persistent storage, keys, and programming to handle security? Or even the ability to update as new encryption becomes standard?

Those lazy bums must have no idea what they're doing!

Re:There's nothing wrong with FTDI...

[drinkypoo]

While all that you snark is true, today there is no particular reason to do it wrong. Instead of an ethernet shield for Arduino, the arduino becomes an interface board (via RS232 or USB-RS232) for some more powerful embedded system, kind of like a shield in reverse. Except that it can let you use shields.

Re:There's nothing wrong with FTDI...

[Gordonjcp]

On Windows you'll need to install a driver in order for USBserial to work

Well, they can't support every niche OS out there. Anyone with half a brain is going to be doing their microcontroller development in Linux.

Re:There's nothing wrong with FTDI...

[TheGratefulNet]

by 'ethernet' I've meant 'tcp/ip on a chip' kind of service. I wasn't at all being strictly literal.

'ethernet' kind of implies ip level service when used in conversation. I agree its technically wrong (I used to develop/teach networking courses) but when they (the arduino guys) call it an 'ethernet shield' and yet allow you to use the onboard webserver (??) you know its a lot more than just layers 1 and 2.

at some point, I really can envision 'security on a chip' or firewall chips being so easy to just plugin and deploy that they can be integrated in embedded systems. but the arduino ethernet shield (the thing we're talking about) just has no way to have firewalling as a plugin and I think that's a technical mistake that makes the shield kind of useless to me. maybe its good for learning in the lab but its useless to deploy in the real world.

Re:There's nothing wrong with FTDI...

[John Hasler]

So don't use USB at all. The FTDI chip is just a USBserial translator anyway. Works fine with a serial port.

Re:There's nothing wrong with FTDI...

[Anonymous Coward]

That's not the point. There's a problem if you provide a bare-bones network connection to a system which is by design dumbed down so that it can and will be used by people who have no programming or hardware development background. When the gadget works, it's done. Security won't even be an afterthought. The things which people build will nevertheless end up in all sorts of applications where they controls actuators which are potentially dangerous or at least have expensive failure modes.

It's like PHP: You can write secure programs in PHP, but the language makes it awfully easy to write exploitable code. When your goal is to reach more people by dumbing things down, then security must be built-in instead of requiring advanced know-how and extra work.

Re:There's nothing wrong with FTDI...

[dissy]

for a physical device that can turn things on and off in the real world (including ruining/damaging things) you NEED security. I just firmly believe this. no firewall, no hosts.allow, no nothing. not authentication or encryption or MAC access control, just like TRULY nothing. I find that unacceptable in a real-world device.

So what you need is to run two networks.

Have your computer LAN, and your hardware HAN.

All your embedded devices, PLCs, PICs, and Arduinos connect to the HAN which has no internet connection.
Take one of the mentioned $50 router devices, or an PC running Linux with two network cards, and use that to bridge between your HAN and LAN. Put all the security in that now-a-firewall computer.

That will protect your real world interfacing devices from random guest computers plugged in your LAN or on your home wifi, or even FSM-forbid an infected windows PC of your own, or even a misconfigured linux system.

And since your LAN should be bridged to the Internet/WAN by a similar firewall/router device, that is two layers of protection between the Internet and your real world controlling devices.

This will limit the attack vector down to a single firewall, and physical access.
Since we are likely talking about your home, or at least one building, you can handle the physical access easily enough by running network cabling as needed, or at least not providing easy to reach wall connectors to jack in (Unlike most LANs, which may have such easy access)

Using a Linux box as the firewall between LAN and HAN will give you the benefit of having a bunch of scripts to control any arduino/pic/plc/whatever from a single place, but the ability to run apache-ssl with more than basic-authentication if you desire. Then you can have web based controls and pretty graphs or pictures for status displays, all over SSL.

I fully agree that security is important and even a requirement to have, but I would rather keep the bulk of that on one machine, and limit all the embedded hardware so only that one machine can reach it.
This keeps the prices of the chips way low, and dumps the hard work of encryption and access control on a general purpose machine more designed for that sort of thing.

If you are super paranoid, you can build a serial cable that only connects the HANs TX line to your LANs RX line, a one way serial port so to speak, and let the devices talk out for logging/status reporting purposes. Leave the control functions to those who have physical access to the console.

Re:There's nothing wrong with FTDI...

[formfeed]

given the fact that you can buy $50 things (pogoplug, seagate dockstar, even WRT routers) that run a full linux and IP stack, why hack around with dodgy ip-on-a-chip things (the arduino ethernet shield) when you can front-end the controller (arduino) with an embedded linux board that has proper IP features.

Exactly, only $25 if you get a deal. But somehow that's not the same for arduino people. I mentioned exactly that point several times to Followers of the Church of Arduino: $25 will give you an ethernet-shield or a zigbee module. The same money buys you a router with ethernet, wifi, and a full ip-stack in linux, plus scripting. Easy to connect the atmega to the serial port of the router.

The Followers of Arduin somehow see it differently: If you add shields to your arduino you built something yourself, if you solder it to a router, you're just using store bought hardware.

Gosh I hate these fundamentalists.

Re:There's nothing wrong with FTDI...

[formfeed]

Those lazy bums must have no idea what they're doing!

I think that sums it up.

You must have thought, that you used irony or something. But if you have to add $50 in hardware so your $3 IC can do what a $7 IC could do natively, you really don't know what you are doing.

Re:There's nothing wrong with FTDI...

[Anonymous Coward]

Isn't the point to simply make network capabilities accessible to the Arduino? Most NICs for PCs don't offer hardware encryption, so why should the board for the Arduino?

First, I doubt it will often really be necessary. Personally, I've been looking at using an Arduino at work, to trigger various forms of robotic entertainment when specific events happen (successful build, overloaded server, etc.). Once someone's on our office network, they already have access to all of the corporate data. Does it really matter that they can make a robot dance?

Second, there's the fact that the Arduino is inherently running custom software. If you need security, you can just add it into your program. If you really need the raw speed of hardware-based encryption, perhaps you can find a separate shield to provide a layer of encryption, or design the shield yourself. There seems to be a fair amount of interest in embedded cryptography already.

If someone's building a project that can access "serious" systems, should they really be relying on the not-fixable-if-wrong security features of their network card? Conversely, if someone's building a project that requires only basic (or no) security, why should they pay for the additional cost of hardware-based encryption?

Re:There's nothing wrong with FTDI...

[Sarten-X]

I wouldn't. I, like many others, don't need encryption, so I'd just be paying $7 for a $3 function. If Wal-Mart has taught us anything, it's that a few dollars price difference makes a lot of extra sales.

With the prices you listed, encryption would have to be needed in 8% of network devices to just break even on price alone. They also must need the exact encryption features offered by the $7 chip, or they'd need the extra equipment anyway. Do 8% of devices really need security? Is my dancing robot (I'm the AC above) or HTCPCP server really that important?

Personally, I'll stick with a cheap insecure shield, and just keep my network secure.

Re:There's nothing wrong with FTDI...

[BillX]

I agree wholeheartedly if you're using something in an industrial setting (besides the obvious safety concern, your robot's dance moves may be a trade secret!), but is that what Arduino ethernet shields really are targeting? As you said, there are already plenty of existing, commercial hardware products for that niche. Personally, I'm not worried about my neighbor hacking in and spoofing data from my homemade garden light meter. (If it came to that, 2048-bit AES won't stop them walking over to put some chewing gum on the sensor...)

Re:There's nothing wrong with FTDI...

[Yetihehe]

Making it yourself is just more fun.

Re:There's nothing wrong with FTDI...

[Anonymous Coward]

Buying a "shield" and downloading a networking library is in no way more "making it yourself" than buying a cheap router and soldering a controller to the serial port.

Re:There's nothing wrong with FTDI...

[Andy Dodd]

WRONG. FTDI has built-in drivers for Linux but not most other OSes. It definately requires drivers in Windows.

The newer USB CDC ACM class, however, has built-in drivers on any modern OS, with the "worst case" situation being Windows where you need an INF (but not actual drivers) to tell Windows that that VID/PID is a CDC ACM device.

Re:There's nothing wrong with FTDI...

[Stele]

Anyone with half a brain is going to be doing their microcontroller development in Linux.

Fortunately I have a WHOLE brain, so I know how to install a (mostly automated) one-time driver on Windows when necessary.

Re:There's nothing wrong with FTDI...

[formfeed]

No no, arduinos are wonderful for controlling things. Especially since you can transfer your prototype to a bare and cheap atmega later. Cheap prototypig platform for cheap embedded controllers. Great.

But if you already know that your project will be connected to the internet or needs any extended connectivity, arduino is not the cheapest way of getting there.

Re:There's nothing wrong with FTDI...

[Anonymous Coward]

Well, they can't support every niche OS out there.

Actually, they do support Linux even though it's completely irrelevant on the market of people who would use an Arduino.

what with the where now

[FuckingNickName]

Is this like the controller I'd connect to the BBC Master 128s as a preteen to control stupid lego toys after I'd done all my scholarship exams (when I wasn't causing mischief on the 8086 with the OMG sound card), only infinitely more complicated because some cuntbucket decided computers are for CONSUMING not PRODUCING so we don't need simple, educational things like User Ports or even Serial Ports any more?

Re:what with the where now

[Anonymous Coward]

It's a board with a microcontroller (CPU with embedded flash program memory and some RAM). You program the flash using your PC. Hello world is typically blinking an LED.

Both consumers and computer manufacturers decided that we don't need user ports or serial ports because USB does the job exactly as needed for the vast majority of consumers and hardware makers.

Almost nobody on either side was thinking about the poor kids who don't have serial or parallel ports when they started becoming dropped from PCs. Happily there's a whole world of USB peripherals to fill the gap.

In summary, rage harder.

Re:what with the where now

[FuckingNickName]

You are on nerd site complaining that someone is "raging" that it is harder than it once was to control external homebrew electronics from a home computer. Why don't you propose that we all get iPads?

I don't know about you, but I have used enough peripherals over the past decade to find that USB is nothing like as reliable as the serial port. It is not just the frustration from a hardware point of view of having to put this huge chunk of complex asymmetrical electronics on any device, but the need to write or supply a complex third-party driver instead of just implementing a simple serial protocol.

And, guess what, almost any USB dongle which supposedly implements a serial port is going to miss out on all the features of a real serial port, using something other than the native 20-year-old full featured and stable driver in OS-of-your-choice... because goodness knows when you're making life difficult for yourself with a complex-but-rubbish bus (and USB is awful when compared to alternatives with similar aims - it's just that Intel have the uncanny ability to spam the world with crap) it would be a waste of money to fully implement what should be a much simpler serial port on the other end.

Re:what with the where now

[Anonymous Coward]

I consider it raging when one uses language like cuntbucket, and speaks almost as if the decline of serial ports is due to malevolence. Maybe I just read too much into your passionate posting.

I wouldn't suggest we all buy iPads. I too like making homebrew electronics. May there always be room for the tinkerer.

I know that many serial USB dongles are crummy, but if the TxD and RxD pins work, that is enough for a lot of homebrew projects. But nowhere did I suggest that the hobbyist use tools that don't fit his requirements, the tool in question doesn't even have to be an RS232 imitation.

If you need something like real time communications, USB won't be up your alley, but if you have that sort of requirement, you are beyond the experimenter in skill and won't be using serial ports with a desktop OS either.

"pronunciation-friendly"??

[Anonymous Coward]

Maybe to those who think english is the only language the whole world speaks.

Re:"pronunciation-friendly"??

[Yvan256]

I speak both french and english, and IMHO "Duemilanove" sounds weird and is hard to say.

On the other hand, "Uno" is much easier to remember and pronounce but will probably get a lawsuit from Mattel.

Re:"pronunciation-friendly"??

[Anonymous Coward]

I speak both french and english, and IMHO "Duemilanove" sounds weird and is hard to say.

On the other hand, "Uno" is much easier to remember and pronounce but will probably get a lawsuit from Mattel.

Because its italian maybe? Duemilanova or 2009. Say it with hand gestures for added effect.

Re:"pronunciation-friendly"??

[kurokame]

Of course it's easier for many non-Italians. Two syllables versus...how many?

Also there's the whole "Duemilanove means 2009 and it's not 2009 anymore" factor, along with the "it's a new product and we've got to call it something" factor.

Besides which, the new name is also Italian. What's English got to do with anything, AC?

it was targeting the enrichment centrifuges

[CreamyG31337]

just read
http://frank.geekheim.de/?p=1189

Re:it was targeting the enrichment centrifuges

[edb]

Although this was posted to the wrong thread/window, it definitely is a very interesting read.

It is under the Apache 2.0 license.

[DAldredge]

http://www.microsoft.com/downloads/en/details.aspx?FamilyID=CCDD5EAC-04B1-4ECB-BAD9-3AC78FB0452B&displaylang=en

The Microsoft® .NET Micro Framework is a platform that enables developers to more quickly develop embedded systems that are smart, securely connected, and easier to manage. With the version 4.1 release, the .NET Micro Framework enables developers to create powerful embedded systems that are more securely connected through a variety of wired and wireless protocols. The 4.1 release is also the first release under the Apache 2.0 license.Version 4.1 also adds support for a broader range of processors with the addition od sample ports for the following Renesas processors and development boards: SH7216 RSK, SH7264 M3A HS64, SH7264 RSK, SH7619 EVB. Additional TCP/IP and SSL support has been added wiht the Open Source distributions of lwIP and OpenSSL.

oops

[CreamyG31337]

wrong window!

Re:oops

[wgibson]

But an interesting read nonetheless. Thank you for the fail-paste :D

.net for dev boards?

[Anonymous Coward]

who cares?

msp430?

[Anonymous Coward]

I'm waiting for my msp430 launchpad, at $4.30 it is dirt cheap and comes with two micro controllers and a usb cable.

Getting Starten on the Cheap

[RAMMS EIN]

Interesting that this story would pop up now. I have recently been thinking about getting into tinkering with microcontrollers. I've always had deep respect for what people used to be able to do with, say, a 6502 and a few kB of code. I think it would be great fun to try my hand at that. However, I have some special requirements that seem to be difficult to meet:

First of all, I would like to interface with hardware I already have. Particularly, video, input, and Ethernet. So it would be really great if I could get a board with VGA out, USB host or on-the-go, and Ethernet, although other combinations are possible (e.g. Ethernet not on the board, but via a USB device).

Secondly, I have virtually no experience with electronics, so I need something that is really easy to get started with. Of course, I am doing this in part because I want to learn, so if it's better to do a few simpler projects first to get the needed skills, I am open to that, too.

Thirdly, I want the device that will be running things to be _cheap_. I am thinking max 20 USD. That's for being able to run some simple software (doesn't need a lot of RAM or ROM, as long as more storage can be added) with video output, keyboard input, and network access. If I need some extra expenses to bootstrap things (e.g. some extra hardware to write the ROM), that's ok, but I want to basically be able to tell my friends "for under 20 dollars, you can get one of these computers and run all this great software, too".

Within these constraints, I would like to get the most bang for the buck that I can get. It doesn't have to be an 6502. If I can get an 68k or an ARM or an FPGA (given enough gates, of course), that would be grand.

I am really excited about the Beagle Board, but that's far outside my budget. I've looked at DigiKey's catalog, and there are many chips there that look promising, but frankly, I'm drowning in information, choices, and unfamiliar terminology for the moment. Perhaps one of the Arduino knock-offs will fit the bill. Uzebox looks really exciting, too. I feel that what I have in mind is out there somewhere, I just haven't found it yet. If someone could help me on my way, I would greatly appreciate that.

Re:Getting Starten on the Cheap

[Anonymous Coward]

First of all, I would like to interface with hardware I already have. Particularly, video, input, and Ethernet. So it would be really great if I could get a board with VGA out, USB host or on-the-go, and Ethernet, although other combinations are possible (e.g. Ethernet not on the board, but via a USB device).

Sure, no problem. Buy a 32-bit microcontroller development board or an FPGA development board with those features. They sell for about $200. Both of them can run Linux too, although you need to set up a soft processor inside the FPGA for that.

Secondly, I have virtually no experience with electronics, so I need something that is really easy to get started with. Of course, I am doing this in part because I want to learn, so if it's better to do a few simpler projects first to get the needed skills, I am open to that, too.

OK, so that rules out the FPGA board.

Everything is hard to get started with. But 99% of the problems that you will run into with a microcontroller development kit are software (as in software on your PC) related.

Thirdly, I want the device that will be running things to be _cheap_. I am thinking max 20 USD. That's for being able to run some simple software (doesn't need a lot of RAM or ROM, as long as more storage can be added) with video output, keyboard input, and network access. If I need some extra expenses to bootstrap things (e.g. some extra hardware to write the ROM), that's ok, but I want to basically be able to tell my friends "for under 20 dollars, you can get one of these computers and run all this great software, too".

You can get 32-bit microcontrollers for about $20. If you want to run Linux on those you do need the ROM and RAM of course. PCB's are not for free either. But yeah, I suppose it's possible.

Within these constraints, I would like to get the most bang for the buck that I can get. It doesn't have to be an 6502. If I can get an 68k or an ARM or an FPGA (given enough gates, of course), that would be grand.

I am really excited about the Beagle Board, but that's far outside my budget. I've looked at DigiKey's catalog, and there are many chips there that look promising, but frankly, I'm drowning in information, choices, and unfamiliar terminology for the moment. Perhaps one of the Arduino knock-offs will fit the bill. Uzebox looks really exciting, too. I feel that what I have in mind is out there somewhere, I just haven't found it yet. If someone could help me on my way, I would greatly appreciate that.

If you can't afford a Beagle board then you can't do what you want within your budget.

You could certainly buy a good 32-bit microcontroller for under $20, but to make it work you would need to design and simulate a circuit for it and then design a PCB and have it made ($100). Then you need to learn to solder SMD components (with your $200 soldering iron). Finally, you would test and debug the whole thing with your $500 oscilloscope and your $150 logic analyzer...

That's why the manufacturers sell dev-boards that work out of the box.

My advice is to get started with Arduino. It won't do the stuff that you want to do, but you will learn lots of useful stuff and hopefuly have fun. It doesn't matter what you start with (FPGA, 32-bit micro, Arduino). The first thing you will do is to blink LEDs. Next up you will probably want to debounce some push buttons. Next you'll probably want to get UART running so that you can talk to your PC. You can have fun doing simple stuff with the Arduino for weeks or possibly months before you will want to move on to projects that the Arduino can't handle. It's better to get started with something that you can afford than to wait for the perfect development board (if it ever comes).

Re:Getting Starten on the Cheap

[tibman]

You don't have to restrict yourself to VGA output or Ethernet connectivity. You can get something like an Ardweeny for 10$, using a cheap LCD for your screen and an Xbee for connectivity. When you get down to the micro sized stuff, you aren't really dealing with "programs" anymore. More likely libraries and classes. If you do want something beefier that plugs into the wall, the beagle board you're looking at would be perfect. Arduino sized stuff rarely uses a full keyboard. More likely you'd have a few buttons to activate the functions you are wanting to perform. Which is really all you need for gadgets.

You can get an Ardweeny here: http://www.solarbotics.com/products/kardw/
But you'll need an FTDI cable (USB->serial) to program it. You can google that and buy the cheapest available.

Re:Getting Starten on the Cheap

[Anonymous Coward]

Given your requirments, the Arduino is probabaly not what you are looking for, but might still be worth looking into.

The Atmega8 series of CPUs are not really meant to support things like video or USB host services. They do have a rich set of capabilities (digital in/out, analog inputs, pulse-width-modulation outputs, Serial, I2C and SPI capabilities), but the chips only have a few KB of storage and very limited memory. They are designed for embedded operation, with user interaction limited to simple things like buttons, and a text-oriented serial LCD.

The Arduino pairs one of these chips with a usb-serial converter chip to provide a connection to the host PC. It also adds a very nice IDE to allow programming of the chip in C, upload code to the unit, and act as a serial console for interaction with the unit. It makes a wonderful environent that can be used for anything from learning basic concepts to some pretty involved stuff. They offer a standardized header arrangment to allow different "shields" to be stacked on the baseboard for expansiton. The Mega lines are based on a chip with much more storage and many more input/output lines. The new units get rid of the dedicated USB to serial conveter and use another microcontroller programmed to do the same job (this gets interesting since you can reprogram that controller to do other things as well).
Basically, they rock out loud... but aren't general purpose embedded computers.

I don't know of any fully functional embedded sytesm in the $20 range. The BeagleBoard might well be the sweet spot in capability vs price right now. You might also look into hacking a chumby or getting the "Chumby Hacker Board" from www.adafruit.com (also a good place to look at arduino-related stuff).

As an alternative, you might look at getting something like an arduino and simply using your host PC as your user interface. This would let you experiment with a dedicated microcontroller capable of interfacing to sensors and other hardware while maintaining your normal computing environment for developing your user interface code.

Re:Getting Starten on the Cheap

[RAMMS EIN]

Thanks! The Ardweeny looks really interesting.

Re:Getting Starten on the Cheap

[RAMMS EIN]

``You can get 32-bit microcontrollers for about $20. If you want to run Linux on those you do need the ROM and RAM of course.''

Right. Being able to run Linux would be a nice bonus, but isn't really what I'm thinking about. What I have in mind is more like a "build your own C64" kind of project, except that it doesn't need to be a C64 and doesn't have to have a case. Just the board will be fine, but I do want to be able to be able to use it like a desktop computer, so keyboard (preferably USB or PS/2) and video out (something I can hook up to a PC monitor or TV, preferably VGA), and, if at all feasible, Ethernet.

``You could certainly buy a good 32-bit microcontroller for under $20, but to make it work you would need ... <lots of stuff and skills>

That's why the manufacturers sell dev-boards that work out of the box.''

Right. I'm not against having to spend some money to get started, but, after I get the first one working, I would like to have the cost of an additional unit under $20.

``My advice is to get started with Arduino. It won't do the stuff that you want to do, but you will learn lots of useful stuff and hopefuly have fun.''

That has the sound of good advice, and I think I should take it. Just because it isn't the board I'm looking for doesn't mean it won't be great fun and a great learning experience. Thanks!

Re:Getting Starten on the Cheap

[Eil]

First of all, I would like to interface with hardware I already have. Particularly, video, input, and Ethernet. So it would be really great if I could get a board with VGA out, USB host or on-the-go, and Ethernet, although other combinations are possible (e.g. Ethernet not on the board, but via a USB device).

Bear in mind that the core of Arduino is a microcontroller. It's meant for small to mid-size projects. Although it has a lot in common with what we think of as a "computer" today, you won't be turning it into a proper computer without a lot of hacking and peripherals. In many ways, it's better than a computer because a computer is extreme overkill for so many things where you want your project to be programmable but don't need the sheer amount of raw horsepower that a microprocessor offers.

There are tutorials on using Arduino to control LEDs, relays, servos, motors, LCD displays, and so on. You can connect virtually any kind of sensor to it thanks to its analog inputs. And there are "shields" that plug right into the board so that you can get started with them right away without soldering or breadboarding. The new Uno board comes with a second, smaller microcontroller on-board which (as I understand it) can be programmed to be any USB device. Various hackers have even programmed Arduino to do VGA output with a minimum of additional supporting circuitry, although I imagine that it doesn't have many clock cycles left to do anything fancy on top of that.

Secondly, I have virtually no experience with electronics, so I need something that is really easy to get started with. Of course, I am doing this in part because I want to learn, so if it's better to do a few simpler projects first to get the needed skills, I am open to that, too.

You're in luck, because there are a lot of Arduino tutorials that assume little in the way of previous electronics experience. (I highly recommend ladyada's Arduino tutorials.) You can do quite a lot just by internalizing a few basic rules (like, always put a resistor in series with an LED and its power source), but eventually you'll hit a point where you need to understand voltage, current, resistance, and the various laws that dictate how electronic circuits operate. The good news is, the Internet is bursting with information and tutorials thanks to the ever-growing maker/hacker movement.

Thirdly, I want the device that will be running things to be _cheap_. I am thinking max 20 USD. That's for being able to run some simple software (doesn't need a lot of RAM or ROM, as long as more storage can be added) with video output, keyboard input, and network access. If I need some extra expenses to bootstrap things (e.g. some extra hardware to write the ROM), that's ok, but I want to basically be able to tell my friends "for under 20 dollars, you can get one of these computers and run all this great software, too".

Arduino pretty much set the bar for cheap electronic tinkering and prototyping. A few years back, development kits for microcontrollers started in the $100 range. Arduino offers 95% of what they did for less than half the price. A basic Arduino kit will run between $30-$35. (There are Arduino boards that cost less, but they might have a few less features or aren't physically compatible with Arduino add-on shields.) If you want a hackable under-$20 computer with a large software library, I'd suggest buying a Gameboy Advance off of eBay.

Within these constraints, I would like to get the most bang for the buck that I can get. It doesn't have to be an 6502. If I can get an 68k or an ARM or an FPGA (given enough gates, of course), that would be grand.

Adafruit sells an Arduino-like ARM Cortex M3 board for $40. I don't know of any e

Re:Getting Starten on the Cheap

[drinkypoo]

Right. Being able to run Linux would be a nice bonus, but isn't really what I'm thinking about. What I have in mind is more like a "build your own C64" kind of project, except that it doesn't need to be a C64 and doesn't have to have a case. Just the board will be fine, but I do want to be able to be able to use it like a desktop computer, so keyboard (preferably USB or PS/2) and video out (something I can hook up to a PC monitor or TV, preferably VGA), and, if at all feasible, Ethernet.

You can do this stuff with Arduino, except for VGA out. You'll get more of a C64 experience with a component video output anyway :)

^^ Informative post above ^^

[arielCo]

So any more problems?

Nah, I'm good. Hardcore OSS advocates fear a lawsuit against Mono and I'd be happier if that were entirely impossible, but I don't see MS tarnishing their recently improved image because of a project that benefits them. So, given equal capabilities and ease I'm happy :)

I'd mod your post Informative if I hadn't posted already

Ethernet, security

[Anonymous Coward]

for a physical device that can turn things on and off in the real world [...] you NEED security.

Very good point indeed.

given the fact that you can buy $50 things [...] that run a full linux and IP stack, why hack around with dodgy ip-on-a-chip things [...]

I can think of at least one very valid reason: power consumption. It makes sense to make those devices which spend most of their lives just waiting for something to happen as power miserly as possible.

As for security -- we'll have to think hard about this. Sure I won't try to cram a complete SSL layer (or [gasp!] IPSEC) into an ATMega, but for some cryprographically backed UDP challenge-response trickery, the resources might be just there?