I’ve returned from travel and started working on cleanup for 1 of 4 of my projects… this one I’m most excited about so it’s the chosen one.

Based on my spark up I had a number of errata items to address. More than I’d care to admit; but in my defense this was somewhat by design, since I knowingly pushed the board to fab without running a number of my final checks. Everyone knows that the last 5% of the project is where all the work hides… sometimes its easier to debug in HW than to pour over datasheets for minor oversights.

1) IC+ I1P75Gx – Termination Resistors

We don’t need them parrallel/shunt termination for this PHY. It even says on the front page of the datasheet (BUT LITERALLY NOWHERE ELSE!!!) “built-in 50 ohm resistors for simplifying BOM”. Removing the 50ohm parallel termination resolved the comms issue I was having.

2) USB Ports – Increase Capacitance

I hadn’t read the spec before I did my design (a horrible sin). As a result, the capacitance on my USB ports are woefully out of spec. From the USB standard “downstream facing port Vbus power lines must be bypassed with no less than 120uF capacitance”. So, I’ll add 150uF electrolytics to each port.

3) USB Port Pinout – Footprint Error

I made a custom footprint for an interesting side mount USB connector… I should have triple checked the part. The ports are completely flipped. Instead of VBUS/D+/D-/GND, my footprint had GND/D-/D+/VBUS. To test the circuit for spark up I had to make a disgusting bodge. But it worked!

 4) FTDI – Pinout Error

Another dumb mistake. I had my D+/D- flipped going into my RS232 converter. Nothing a bodge can’t fix though.

5) Clock Stability – Bug

My first board worked fine, but the second seemed to only briefly bootup, then power off after a few seconds. I was able to narrow it down to the clock on the USB to ETH adapter. I had to add a 1M feedback between XI/XO to ensure the clock remained stable (or would it be instable…). I already had the footprint there it was just DNP’d. Easy fix, no bodge this time.

6) Power Switch Variant

To toggle the power to each USB port I’m using a pair of N channel power switches (MIC2536). These were jellybean parts recommended by a reference design so I didn’t give them much thought. This lack of thought led me to buy the first variant I saw MIC2536-2… the enable is inverted on this version, what I needed was MIC2536-1. To continue spark up I just tied the enable to the appropriate rail.