Show Posts
Pages: [1]
1  Community / Bar Sport / Re: A project to prevent a homocide on: October 07, 2012, 01:51:31 am
He he, pretty much the same I would think although I've only got a shotgun so that's not a problem for me.

A little off topic, but last fall I was target shooting at my friend's farm with my 12 Ga. using a slug in a brand new rifled barrel with a new scope. The factory had the scope mounted too far back and I was too distracted (aka dumb) to notice. First time I fired the gun the kickback shoved the scope into my eyeglasses, and the top of the scope cut a perfect 12-stitch semicircle just above my right eye. As if that wasn't bad enough, the shock popped the other eyeglass lens out and tossed it a good 25 feet into the woods, never to be seen again.

Seriously, you can't make this stuff up.
2  Community / Bar Sport / Re: PIC32 Arduino Variant? on: October 07, 2012, 01:33:12 am
I think chipKit is somewhat aided by some microchip-owned slightly-closed-source libraries that they allowed to be used in an otherwise open-source project (low level initialization for the on-chip peripherals, for example.)  ... Also, I think chipKit was better funded; people were able to work on it full-time.
All of the Microsoft libraries, including low level initialization, are well documented and source is freely available. I think Digilent's performance compared to Maple had more to do with your second thought, that Digilent has full time people - probably a whole team - that are paid to do this kind of work full time.

My beef with Microchip is not about the libraries, rather it is the cost of their toolchain. Their compliers are based on gcc and thus source should be available under GPL, but good luck finding something you can actually compile. So you either get the very hobbled free versions (on the order of 2.5x code size), or you pay over $1000 EACH for compilers - one for the 8 bit machines, a different one for PIC24/dsPIC33 and yet another for PIC32. This is a classic case of Microchip now knowing what business they are in. They sell chips, they should give the compilers away. Even Renesas has finally embraced a gcc toolchain, and ARM has had one for years.

This is why I finally jumped ship to the ARM Cortex. The M0 is faster than the PIC 18 at the same price, and the M4 just blows the PIC32 away.
3  Community / Bar Sport / Re: where to recycle IC tubes? on: October 01, 2012, 04:06:03 am
IIRC, they're HDPE and as such easily recyclable. But re-use is definitely better. I'd inquire with the local artists asylum, schools, etc. to see if anyone needs such tubes. Good luck!

Recyclable plastics carry a symbol, a dotted rounded triangle-ish thing with a number inside. Without that, the recycling center can't determine what your material is made of and at least in my area, unmarked or unacceptable material ends up in the garbage bin.

I suspect that IC tubes are not recyclable due to the metallization used to make them conductive, but it's worth a look to see if they have a recycling symbol. Better to re-purpose if you can!
4  Community / Bar Sport / Re: Never fully trust a student and never try to be perfect on: October 01, 2012, 03:47:51 am
Perfection, from a student or otherwise, is an unrealistic goal. I have been soldering parts to PCBs for 35 years and I still make a lousy solder joint now and again. Just a fact of life, I think.
5  Community / Bar Sport / Re: The ARMs race on: October 01, 2012, 03:45:24 am
Performance has been a growing problem for the wildly successful Arduino platform. Many of us want to do things that simply can't be done with the current crop of 8 bit MCUs. For instance, a UAV controller is extremely hard to do in 8 bits, and it is a testament to the Ardu-Pilot guys that they have managed to squeeze as much out the the ATMega328 as they have. Also, ATMega's don't run full speed at 3.3 volts, but nearly every interesting new chip only supports 3.3 volt I/O - 5 volt is a thing of the past.

The newer ARM chips, for instance the Cortex M3 or M4, are not cell phone application processors like the A8. They are designed to be real time embedded controllers, just like the ATMega's were in their time. A M3 can execute code 30x as fast or more than an ATMega328 as a result of the 4x wider data path, C-compiler instruction set compatibility, DSP instrucitons, and internal pipelining. It has much more memory (512K FLASH, 64K SRAM). The M4 adds hardware floating point and memory up to 1MB FLASH and 256K SRAM. Using these would allow us to do much more complex tasks than we can do now, but in a similar form factor. The MBED platform is an example of this sort of approach using the Cortex M3. Use a M4 with a local toolchain, Arduino form factor, shields, and community, and I think you have a killer platform.

The great thing is that much of this complexity will be hidden, as it is now. Arduino is actually a C++ compiler. How many guys who responded to this thread are expert C++ programmers? Probably not many, me included, because there is no need to - the Arduino tools hide that from us so we can concentrate on the project, not on mundane programming details. The Arduino team will certainly do the same thing with the ARM processors, while enabling projects we haven't even thought of yet.

And if you really want to keep using the ATMega boards, don't worry - if you keep buying, they'll keep making!
6  Community / Bar Sport / Re: Whats the deal with Raspberry Pi ? on: October 01, 2012, 03:01:52 am
I have a "first run" RPi and I have found many uses for it as a low power (about 1 watt) headless Linux controller. Running headless you can communicate with it over ssh either with a FTDI UART connection or over the built-in 100mbps wired or USB dongle wireless internet adapters. It is much more capable than any Arduino or PIC for general computation particularly if you need floating or fixed point. The graphics engine can decode Full HD 40Mbps Blu-ray in real time. It is also a fine platform for soft real time use, with analog (A/D) and even PWM. Because it's Linux, you get USB, wireless ethernet and BT, a real file system, robust ethernet networking with SSH/SSL, standard Linux apps, the command line, etc., all without doing any work. It supports SSH/SSL encryption, which Wiznet devices don't do, or at least didn't do when I looked at them a year ago. It even has decent stereo audio which, as LadyAda mentioned on the AdaFruit website, is surprisingly difficult to do on a MCU board at low cost.

For apps that don't require PC-style I/O, the RPi is not a great choice. It's not as good as an Arduino for real time like PWM and such, it has a weird undocumented SOC, and the SD card is dog slow. But for remote sensing / logging / wireless ethernet where you need to process the signal, at $35 it can't be beat. Let's put that into perspective. An Arduino Uno R3 + Ethernet Shield is $75, and an Arduino Uno Ethernet board is $65. I think it would be a challenge to even build an Arduino Uno Ethernet on a homebrew PCB for $35.

I think the closest community supported competitor to the Pi in the market now is the BeagleBone. It has similar memory and performance,  but it has a fully documented ARM 7 (TI AM3359 A8 SOC), no "chip on chip" issues, and most every I/O type you could want. It accepts expansion shields ("capes") of which perhaps 10 are already in production and many more on the way. But it is a different class of device, much like the ARM7 boards someone mentioned earlier. At $89 it costs 2.5x as much as RPi, and it uses more than twice as much power.

RPi is not going to replace Arduino or BeagleBone, and it's not going to replace a PC either. But for me at least this is an extraordinarily useful board at a throwaway price. I think it will also meet its original goal of giving kids something they can hack, at a price that won't break Mom and Dad's budget. I agree it is a victim of its own hype, but I think that will soon abate as people figure out what it is actually capable of.
Pages: [1]