Computing stuff tied to the physical world

Small oops on the LED Node v2

In Hardware on Feb 23, 2013 at 00:01

The new LED Node v2 has been re-done slightly, by fixing the labels of the FTDI connector on the back side, but unfortunately Mr. Murphy pulled another one on me by making me ignore another issue at the time these new boards were ordered.

The current release of the LED Node v2 has a “disconnected ground plane”, meaning that some large copper areas on the back are not attached to ground!

The circuit is correct, in the sense that the board will function as is, but for better noise suppression and higher range of the on-board RFM12B radio, it’s best to make a small modification. Sooo… time to get out the hobby knife and start scratching!

Here is the back of the LED Node v2, unmodified (the silk screen may differ slightly):

DSC_4371

What you have to do is scratch off the solder mask, without cutting the trace:

DSC_4374

The result will be a bit of bare copper on both the trace and the ground plane next to it. Then, apply a bit of solder to connect everything together. Note that doing so may require quite a bit of heat, as the ground plane is hard to heat up – just keep your soldering iron on the spot until the solder flows freely:

DSC_4375

That’s it. The ground of the circuit and the ground plane are now electrically connected, at least under the RFM12B module.

While I’m at it, here’s another modification – unlike the previous “mod”, this one should not be done in a standard setup, with 12V feeding the LED strips as well as the LED Node itself. This is a change you need only consider if you are using voltages higher than 12V, i.e. for 24V LED strips, or when running off batteries which go up to 14V or more.

The reason for this change, is that the MCP1702 regulator of the LED Node is rated only up to about 13V. Anything higher will damage it, even if applied briefly – I’ve seen several cases of this in the past.

The solution is to cut the trace between the power connector and the LED Node, but keep in mind that you then have to power the LED Node logic from the FTDI connector or one of the PWR pins on the headers at all times:

DSC_4372

Just use a sharp hobby knife to cut through the trace this time:

DSC_4376

So much for hacking the LED Node!

  1. Ah! This could be the same problem I’m seeing at sending commands towards the LED node. I have to send them several times to get them received. Maybe this is because of the bad reception caused by this small problem.

    When will you have some LED nodes available? I’m considering buying two more. The flexibility of them when integrated in a home automation solution are huge.

Comments are closed.