July 28, 2010
about 4 years ago
Look at the image again, the image is cut off before it ever gets to 1". Those are centimeters you’re seeing.
It’s a little over 1 and ½ cm long and a little over 1 cm wide, or .78" long and .5" wide.
about 4 years ago
I like the way FEZ did it with the Panda, for the most part. They put the extra IO as solder holes just inside the headers and along the back.
Putting them along the back was probably not the best idea, though, because they had to move the ICSP to do it and it is now not completely pin compatible with all arduino shields (i.e. the Ethernet Shield). I personally think it would be worth losing 3 of the 60 io pins to maintain compatibility, but that’s just me.
about 4 years ago
That’s a fundamental misunderstanding of the RS232 protocol. You have to look at the wire as either the transmit or receive wire, not the individual connections on each end, and you have to look at it from the netduino’s perspective ONLY. That is because the netduino is the Data Terminal Equipment, whereas your PC’s RS232 interface is the Data Communications Equipment. The transmit wire is only for transmitting, and the receive wire is only for receiving, and everything is viewed from the DTE side.
Basically what this means is that TX on the DTE (netduino here) connects to TX on the DCE (your PC or other RS232 com device), and RX on the DTE goes to RX on the DCE. For pin mapping, the TX on your device should always connect to pin 3 (TX) of the RS232 DB9 connector, and the RX of the device should always connect to pin 2 (RX) of the DB9.
So the DTE always transmits over the TX connection, and receives over the RX connection. This means, of course, that the DCE always transmits over the RX connection (i.e. the DTE needs to receive data: we don’t view things from the DCE perspective) and receives over the TX connection.
Why they did it this way, I’ll never know. I’m sure it made tons of sense 50 years ago, but it has been this way forever and it will never change.
The good thing is these connections only use four out of the twenty possible connections in the RS232 protocol. If that were not so you never would have fixed the problem by simply swapping the TX and RX - all of the connections are set up this way.
BTW, it looks like there is room on the bottom to add the USB connection to the back if they use longer pins (perhaps connecting on the bottom instead of the top?). All you need is the width of a header to do it and these nifty SMD headers: http://www.sparkfun.com/products/9015. Alternatively the board could be made slightly long and hang the USB off the back.
Either way I’d expect to see something along these lines soon, Secretlabs and FEZ both seem to be progressing pretty quickly with their hardware designs.
You have to bear in mind the form-factor - it is designed to fit into a 24 pin socket (same as the arduino mini). That means only 24 pins to work with.
So what do you do? Drop 5 IO pins and add USB? Or drop 10 IO pins and add ethernet? How about dropping all IO for both? It’s not nearly as good at controlling stuff as the arduino no matter what you do, and it is intended as a NETMF alternative to the arduino mini. Controlling stuff in a small form factor is the whole point of this little device.
I would expect to see SecretLabs produce a 30-40 pin version of this at some point that does indeed have the USB and/or ethernet pins exposed (perhaps even adapters on the board - since it will have extra room with a longer socket).
It’s compatible with 80% of the shields on the market right out of the box, and that’s a big deal for any modular electronics platform.
A better way of looking at it is “why would they make it incompatible with the hundreds of third party shields on the market when they don’t have to?”
It’s compiled into bytecode, and the bytecode is interpreted on-board. It’s sort of half-way between a native compiled and interpreted language.
The FEZ is exactly the same. .NET is more like Java than Python or Perl (or any other scripting language), if that helps. The C of the arduino gives it inherently better response times even though it’s overall processing power is several times less than the netduino or Fez.
I.e. arduino will flip the LED on quicker, but the netduino will handle your complex calculations quicker and be able able to send a twitter update while it does so.
That’s certainly true, but on the other hand there are many more advanced projects that these devices are much better suited for than an AVR.
For example, RepRap and DIY CNC mills - you have to make a lot of compromises to get them to work with an AVR, and are very difficult. Using a high-level language like C# and the greater processing power of the ARM makes this project even more accessible to the hobbyist. You get to do bigger projects faster, at the expense of being further away from the bare metal.
And what’s wrong with throwing more power at it when I can get a chip that’s three times as powerful for three times the price?
And just FYI, the reason these are almost the same price as the arduino is because the arduino requires a $15 FTDI chip for USB to RS232, whereas these have USB capabilities built in to the chip (and ethernet too, btw, just not exposed on this board). $5 ATMEGA + $15 FTDI = $20 AT91SAM7X512, minus a lot of features. All the savings in the DIY board is in the FTDI chip.
The #1 disadvantage for these .Net Micro Framework devices is response times. It’s running interpreted byte-code (it’s .Net after all), not native C code, so there is about an order of magnitude more latency (1-3ms vs 1-3us) in the response times.
If you need to do any kind of precise timing right now the only real viable options are to add your native code to the firmware, which is beyond difficult (it requires extensive knowledge of the ARM architecture, NETMF internals and various tool chains and whatnot), or implement your timing in hardware, which requires more electronics knowledge but is the significantly easier solution of the two.
On the other hand, these 32 bit processors are significantly more powerful (about 3 times so for the netduino and 4 times for the Fez), and with the exception of precise timings are significantly easier to program given the extensive .Net framework behind them.
All the heavy duty processing work that the arduino just couldn’t handle can be done in the netduino, and the Fez brand (72mhz ARM) should be able to do even more heavy lifting.
What’s more the netduino doesn’t even expose all the features of the Atmel chip it uses. The plus board adds ethernet and an SD card reader for a little more than the cost of an ethernet shield. Expect more boards in the future with more features for not a lot more money. This initial board is geared at direct pin compatibility with the arduino, there is a lot more that it can do.
These little guys are awesome. Don’t expect it to completely replace the arduino in your kit though. Instead let them complement each other, as they should.
Tutorial - Unregulated Power Supply Tutorial
about 4 years ago
In an unregulated power supply there are no voltage regulators and only one output filter capacitor. Basically the wire from the wall goes straight to the transformer, the rectifier turns it into an ugly DC(ish) signal, and the filter cap cleans it up a little.
A linear power supply looks exactly the same as an unregulated power supply, except it has a pair of capacitors on the output side with a voltage regulator in between them. They basically put the voltage regulator you would need to build to use an unregulated power supply inside the supply itself, turning it into a regulated power supply. It’s still pretty simple.
Switched-mode power supplies are a whole other story. They have multiple rectifiers (at least one before the transformer and one after), filter caps before and after the transformer, usually a much smaller transformer, and a voltage regulator in a feedback loop to the transformer. They almost always have heat-sinks as well.
Break open an old PC power supply to see what a SMP looks like.
No public wish lists :(
Forgot your password?
No account? Register one!