ATMEL Mega1284P evaluation board avalible

.. the soldering the microsd on the bottom side is not a big deal, the added value of having there both sdcards is much bigger, though. BTW, he can leave the both sd card traces on the top as they are now, and to add the new one for one microsd on the bottom. So everyone may do it manually afterwards, when required. Remember an additional card_select signal from somewhere then..P.

The 5V regulator needs to be isolated, the pad is 5V while the ground polygon is .. ground.
The other regulator takes 5V down to 3.3V and I don't think actually gets hot.

Putting microSd on the bottom - I was thinking reflow as skyjumper said, and only having one or the other in use.
I don't see any reason it couldn't go on the bottom, will have to see what the part looks like mechanically & ease of hand soldering. Good suggestion pito.
I was thinking most folks would opt for the microSD from the # of earlier comments against the SD, but that was the socket I wanted to use so cards could go right into a PC with no need for an adapter (just more pieces to lose!).

Okay. MicroSd is on the bottom -that was pretty easy.
Oooh, rats - forgot the SS line ...

Okay, its not gett a seperate SS line - that is gonna be one huge pain in the ass to route seperate SS thru the level translater.

I did it! Got the uSD on the bottom of the card, and seperate buffered SS line going to it.
Saved it as a seperate version, in case my review of the uSD socket shows it to be difficult hand solder to the bottom of the board.

.. great! Now you may backup one sd-card to another one, on fly! Or you may run a Bitlash script from the first one and store the data to the second one..

  1. What would be necessary, though, is to have the detection of cards insertion working.. Do we have it?

  2. There are issues reported with other similar boards as the sdcard operation is concerned - current surge - as the sdcards may take a quite a lot of current, thus they reset the mcu or introduce some instability.. The recommendation is to place a ceramic! 10u or 22u smd capacitor as close as possible to each sd-card's Vcc and GND pins, maybe an inductor into Vcard rail (4u7 or 10uH) would help as well..

PS: do not fab too many pcbs in the first batch - it would be a miracle to have the very first pcbs error free..
-Pito

The hardware design for the SD card Bob used is known good. It works on my board beautifully. The chip detect and write protect pins will have to be checked manually by your software. The SdFat library does not check those.

Hmm, card detect, hadn't considered that. Both cards have CD & WP in parallel still.
Looking closer, the uSD doesn't seem to have CD/WP pins.
So, yes, I do have Card Detect & Write Protect connected, but only on the SD card, not the uSD card.

Current surge is covered by the separate 3.3V regulator that is good for 400mA.

Suggest "change isolate" of the bottom-layer GND fill to 10mil or so. It's currently "0", which means the polygon will obey only the design rules for clearance (5mil!), which is pretty tight to have over as many spots as the polygon approaches traces.

Smaller battery? The next size down is 16mm diameter (cr1620 and compatible.) A lot less capacity than a 2032, but plenty to backup a RTC. Or use a two-pin wire connector and have the actual holder off-board. Foam taped to the CD card holder or something...

One of the things I do to save space is change the "elongation" of oval pads in the "shapes" part of the design rules. That will shorten the pads, and you can pick a value that gives you some of the advantages of the larger pads without being quite so huge.

Routing of +5V near "L" led and r6 is weird.

I find the 5V trace near the top connector pins uncomfortably close to the pins, and it doesn't need to be that close.

Move C9 and C14 slightly leftward.

Center the RESET signal between the pads of the JTAG connector.

Here's a gerbv snapshot highlighting the areas where I made some changes. The changes near the pads are mostly from changing the elongation to 70% (from 100%)

Umm, okay, I'll look into those.

  1. change isolate - where do I find that?
  2. battery - that is the size I want.
  3. my library editing skills are crap. I am going with those as that is what the duemilanove uses and I know the shield pins will fit.
  4. L/R6 - I flipped R6 around so the 5V feeds thru straighter now, no more back & forth.
  5. 5V trace moved down some.
  6. C9/C14 - moved a scooch
  7. Reset centered. Also centered/spaced out a few other signals in that area.

Do you know about itead DRC files here? I can't seem to save, open, anything.
http://iteadstudio.com/store/index.php?main_page=product_info&cPath=19_20&products_id=175

Elongation - found that, that is cool!

  1. change isolate - where do I find that?

It's in the "change" menu, or you can use "info" on the polygon and change it in the pop-up window.

Do you know about itead DRC files here? I can't seem to save, open, anything.
http://iteadstudio.com/store/index.php?main_page=product_info&cPath=19_20&products_id=175

".rar" files are an alternate archive format, like .zip. There are assorted things that will unpack it; I have Stuffit (paid) and "Stuffit Expander" (free) on my Mac and they seem to work fine.
Inside the DRC .rar file is an "ITead_rule.dru" file that can be read using the "Load" button of the DRC sub-window...

Could you do me a huge favor? Unzip those & send to me?
robert @ crossroadsfencing.com

I can't get windows vista to download them.

Got the spacing/isolation changed to match the top layer. Gotta go fix trace placements now to get ground plane coverage back in some big areas now.

"A4-SDA" does not appear to connect to the CPU. (the signal connected to the CPU says "A4-D18")

(found while considering: Do you need to keep the mapping up the Analog pins the way it is? That part of the board would get a lot simpler if you swap the analog connector end-for-end. It would mean a different pin-mapping file on the SW side, and I don't know how common it is to use those pins together as a port (it would be portC on an Uno instead of portA, anyway.))

Similar but more complicated issue with A5-D19 and A5-SCL ?

EDIT: OOPS. Never mind. I just realized that what I thought was a peripheral connector is actually a jumper block!

I have always thought the analog port would have made a lot nicer routing also if the pins had gone the other way.

Take a look at the clean up:
http://www.crossroadsfencing.com/atmega1284_Rev15_itead.brd
http://www.crossroadsfencing.com/atmega1284_Rev15_itead.sch

And now I'm really off to bed! Thanks for the help.

..cr1220 would be more elegant, as I wrote few times.. :stuck_out_tongue_closed_eyes:

Okay westfw and pito, after a nights sleep, I am willing to add another mounting hole for a smaller coin cell battery holder, wired in parallel with the existing, if you can help me with a symbol for it.
You can see space is a little tight in that area. Options are to use the existing Bat+ hole and add a new Bat- hole, or use the existing Bat- hole and add a new Bat+ hole.

Ah! I see it is surface mount - that will be easier to add, there are not too many bottom traces in that area.

Rats! I went thru a lot of effort to make sure no thru parts would interfere with the battery holder - then stuck a thru hole +5V regulator right in the middle of it! Back to the drawing board - or, forming the TO-220 leads. Really wanted thru hole for mechanical strength tho to support addition of a heat sink.

Is there a way to make eagle only show the +5V traces? I am concerned that with my multiple revisions to the regulator area that I have +5 routed in multiple paths to some of the same parts.