Yikes… there is a major bug in the current JeeNode USB v2 board!
The on-board MCP1703 3.3V has been connected with Vin and Vout reversed!
The effect is that the +3V pins on all the port headers will have approx. 4.7V instead of the intended 3.3V. This in turn means that all plugs connected to these headers will have a higher voltage than should be.
The on-board logic is not powered by this regulator but gets its power from the 3.3V pin on the FTDI chip, so the JeeNode USB v2 itself works just fine (my weak excuse for why I didn’t catch this problem earlier…).
Here is the schematic – with the problem staring me in the face, now that I know about it:
Here are the printed circuit board traces:
And here is the board with the voltage regulator connected the wrong way around, as it has been shipped until yesterday, i.e. Feb 1st 2010:
If you have a faulty unit, there are a number of ways to fix this:
- Send it back and I’ll fix the voltage regulator for you.
- Remove the voltage regulator and add your own from a PWR pin to a +3V pin.
- Unsolder and re-solder the VR flipped, as described below.
To reiterate: this bug only affects plugs connected to the JeeNode USB, not the JeeNode USB itself. If you don’t use the ports, or don’t mind having a higher voltage on the +3V pins, then you could just ignore the issue.
Here’s how I’ve fixed this on existing JeeNode USB’s, including ones still in stock. First, I used solder wick to remove the regulator (careful not to lift the pads):
The next step is to mount the VR back on its side. The pins which need to be exchanged are:
One way to do this, is to turn the VR on its side and solder pins 1 and 2 back on pads 1 and 3:
As last step, I added a thin wire from what used to be pad 2, to the pin now sticking out on top (was pad 3):
As I said, I’ll be happy to do all this fixing for you if you send me your JeeNode USB v2.
It ain’t pretty (real bugs never are) – but it’ll at least let you continue to use the JeeNode USB v2 as intended.
All JeeNode USB v2’s shipped from now on will have this fix, until an improved “v3” is ready. As always, “new copper” (i.e. a new PCB) will take a few weeks – it’s at the top of my list to fix and resolve this for good.
With many thanks to castello for drawing my attention to this bug on the forum.