update: ok, v16 seems to work.  yay!!

it was not easy, though  I tried freebsd and could not build the IDE.  tried linux (gentoo) and could not get avr tools to work.  tried ubuntu - yay, that worked.  but I do my work on winblows (usb serial and all that) so I wanted to try building the ide from win source.  had to edit the make.sh file a lot (almost all commented-out things had to be uncommented for cygwin).  finally I got an ide that said v16, I loaded my ethernet web client source, built it and it works great now!  no more 'lost packets' (lol).

ethernet now is a LOT better.  before, well, it was useless.

I'm using v15 of the arduino system and I'm also seeing this bug.

I'm connecting out from my wiznet shield to a local apache2 server and when I hard reset the boards, the first connect always works.  then I can do some more connects but not reliably.  when it gets in this state, it still responds to loop() events (I have an IR receiver that detects remote keypresses and that still works fine) but outbound connects just fail.  then I reset the board and all is fine for the next one, etc.

is v15 still in this broken state?

my hack for taking an actual PL2303 from ebay ($5 or less) and tapping into tx, rx, gnd, vcc and DTR (for reset).

yup, reset works, too.  windows does not init the device ('it wont start' in windows terms) but linux sees it just fine and the IDE downloads code over this link perfectly.



unlucky for me, the DTR pin (2th down from top) did not go anywhere convenient (like a big solder pad).  I had to tap onto the chip pin itself.  just barely, too  but it does work.  the wire tapped on the 4th pin from the bottom is CTS and I'm not sure I need to tie that to ground but its commonly done so I did it, too.

some ebay rs232 dongles are not actual PL2303 chips but if you look carefully for ones that call it out by name, you can find them for $5 and less.

its not FTDI but its just about as good and the rest line works fine if you can find DTR or RTS.

I'm using this in an audio project of mine:

http://www.flickr.com/photos/linux-works/3399776581/

a digital volume control for hifi/stereo using a lcd display, IR receiver, rotary encoder, PGA style volume control chip and some input/output device select relays.

the x10/firecracker thing was unexpected but now, when I switch off a main speaker amp and switch on my dedicated headphones amp (yes, they are 2 different boxes) I can power the unused one down via x10 devices and power the other one up the same way (via another device address).

your code came in useful - and it took all of 5 minutes to solder into a spare firecracker (I had no db9's) and download your code to test it.  when I heard a relay CLICK! loud in the other room, I knew it was working

good stuff, many thanks!

/bryan

using the backpack for my espresso machine (lol)


http://www.flickr.com/photos/linux-works/3970170806/


http://www.flickr.com/photos/linux-works/3969400943/

boards are here and I have some photos of the first built unit:

schematic:



and the formal announcement to the audiophile community, via head-fi,  by my friend Ti of AMB laboratories:

http://www.head-fi.org/forums/f6/lcduino-1-i-o-processor-447607/

parts list, visual:

http://www.flickr.com/photos/linux-works/3947809618/

its a very common hammond box, from digikey, mouser, newark.  the audio guys (I do a lot of DIY audio stuff) use these all the time

the top slides off (its metal) and I slipped in some plexiglas.  its not the strongest thing (the thickness I used) but it gives great demo (lol).

the front was done with hand tools and its abs plastic so the finish isn't perfect.  I may have panels made on a cnc and make some kits up, depending on interest.  the neat thing about this kind of design is that you can get away with generic front panels (just square cutout for lcd and round for a single IR sensor).  if you have a big enough IR remote, you have a bazillion buttions there at your disposal for the cost of cutting 1 round hole in the front panel.  big win

a lot of people put DACs and headphone amps and specialized 'audio grade' power supplies on those hammonds boxes; and you do see quite a few with the clear top show-off window.

I'm also working on a board for the SPDIF switch guts.  that will be one of the first 'apps' for the LCDuino infrastructure.

just two sheets of plexiglass would sandwich the board(s), sides open, so you have access to everything.

I can visualize that, I think.

the 6pin download header has to be easy to access and have secure mounting (with all the plugs and unplugs it will see in its lifetime).  

the pot could be a lesser accessible thing since you'll usually find the right spot for your lcd and leave it there.

there's a 3pin header for the IR input sensor.  I didn't want cables coming out of the side since I wanted the lcd module group to fit inside the smallest case that would hold the actual lcd size, itself (plus some room going back).  we kept the width and height of the board to exactly what our model 16x2 had.  my partner suggested we go a bit bigger (to allow mounting of the IR input module, for example) but I really wanted to know that 'if the lcd will fit in the box, so will our module, behind it'.  I didn't want the backpack mainboard being larger than the daughter lcd board.  and that meant that things coming out the side would be impossible for *me* to get to

one proto went into an spdif switch that I built:





I realized that with this popular hammond box, that its still a tight fit with just the lcd by itself.  having stuff come out the sides would not be a good idea for this kind of mounting.

and this is one of the primary use-cases I had in mind.  some hammond or similar box with a minimal front panel.

all the app-specific stuff would go inside either right behind the lcd sandwich (the shield idea, but to the lcd display form factor) or via cables to a horizontal mount board (my spdif switch was like that).

ah!  now I see what you are saying.

my thought was to make good use of the 'internal room' between the boards.  that won't get compressed too much anyway if you use the arduino style female sockets and male pin headers.  all the sockets I've seen are of the same general thickness and the pin headers don't get that much smaller, either.  they are needed to help secure the 2 boards together (you can ignore the screw holes in many installs and only secure the main lcd to the front panel and let this board hang off the top row of pins).  on my proto, I had a bottom row of 3 (say) pins just to keep things tight; but not all lcd's have the dual rows of top and bottom pins for soldering to.

maybe this will help; it was my early prototype before I made the board layout:





I suppose shorter pins and sockets could be used to combine the 2 pc boards and then you'd HAVE to mount the parts only on the back side of the board

do you really think that much space would be saved this way?

You could have reduced the thickness of the whole assembly if you had all parts soldered on one side (the back side, where the LEDs are now).

but then, how does the user SEE the led status if it faces inward toward the sandwich?

similarly, the reset button is a user accessible item; if its inside the 2 board sandwich, its impossible (or hard) to get at.

there's nothing essential (other than the contrast pot) on the back side of the board.  the reset button is not NEEDED for the thing to work and neither are the leds.  I suppose you could argue that the contrast pot could be set once and then left alone; but I'm not sure if we had room 'inside' for the pot so it went on the back side.

yes, that was the intention.

one side faces the lcd (once its assembled) and the other side faces the inside of your box or chassis and that's where I envisioned the 2 leds, lcd contrast pot, reset switch and 6pin FTDI download cable header.  also a 3 pin molex header for an IR TSOP receiver module (since I tend to use IR remotes as my 'multi button pads' these days).

since this is intended to be built by hand (not wave soldered, etc) I don't see a problem with parts being on both sides.  do you?

flip side:

eagle in 'board mode' (no schematic drawing) and then into eag3d and that creates a POVray output.

neat huh?

I just learned about this stuff this week, myself