Rev0 first boot

[tnt]

Hi,

So a fedex enveloppe showed up today with a rev0 board. Inside it was a sturdy cardboard box containing the boardu itself, a SD card and a psu (not a fan of the 'wavy' title on the sticker though :p).

The first impression is that the board is small, very small and very dense. It appears at first a bit bigger than a credit card but in fact the PCB is an exact match. It probably comes from the non-rounded corners and the fact some connectors protrude a little from the PCB (so that they could line up with a case external face)

So after reading the quick-start, I tried booting it.

The PSU uses a US plug (but it supports 240v input), so I first had to find an adapter. Even after doing so, the barrel connector doesn't seem to exactly fit the connector on the board and wiggling it causes the board to flash and sometime even get into weird states (like red led on, not even sure what it means). So in the end I just wired it up to a lab supply by soldering to the power input pads, that made the boot much more reliable.

It does get HOT, really. Right now as I'm looking at it, it draws 1.66A at 5v. During some phase of the boot it can rise as high as 2.1A. So I setup a small fan blowing over it, but even then the epiphany is much warmer than on the prototype.

Once booted, I could login to it as linaro/linaro and run the test. Unfortunately, I get

"ERROR: C_epiphany is different from C_host !!!"

Not sure what's wrong yet, I'll have to investigate deeper.

Cheers,

Sylvain

[aolofsson]

Sylvain,

Fantastic!

1.) Again, most of the power issues have been fixed and new boards should run significantly cooler in gen1. That being said, most rev0 boards have been running at 1.32As consistently. That should come down to 1A for gen1 due to the redesign of the power system.

2.) The prototype getting hotter is a problem we have been fighting since day one due to the package. The smaller the board gets, the worse the problem has gotten(most of the heat is being sucked through the board). Here are some good news for gen1:
a.) The new Epiphany package should give 20 deg /Watt theta-ja. This is MUCH better than our old prototype package.
b.) We have beefed up the ground copper plans to pull heat away from Epiphany more effectively.
c.) We tried to make nice clamp pads around the mounting holes in case someone wants to pull some of the heat off the board through those. We did note that those pointy metal standoffs can give you a nice burn ring on your finger if you leave your finger on them (and there is no fan on..)

3.) The barrel connector was terrible, we have already switched it out with something much sturdier and more standard on gen1.

4.) In terms of the Epiphany failure, that is not good. Two suggestions, let the board cool down, run the fan, and possibly try again? It was definitely working correctly before we shipped it so I am a little perplexed..if it persists we'll start debugging it. The good news is that the test at least runs until completion. (there's a lot of things going right for that to happen..but no cigar of course)

5.) The PSU was tested to work with that board, but we have found that 3 out of 4 supplies bought on amazon market failed to give anywhere close to 2A, even though that's how they were sold. With the new wider barrel connector, there should be more robust solutions available from places like sparkfun/adafruit etc.

Thanks,
Andreas

[Gravis]

[tnt]

[aolofsson]