Rev 0.3 of the 48V IBC board came in! Time to start assembling it.

This version has a few bug fixes like "thin trace where there should have been a fat plane" plus a dedicated low noise power rail for the shunt monitors, some lowpass filtering on the shunt outputs to remove the effects of ripple on the current measurements, and moving the 48V shunt out of the high dI/dt path.

I'm also going to experiment with putting a thermal pad underneath the DC-DC module in order to enhance heat dissipation from the underside of the module to the PCB.

The goal here is twofold: better cooling of the module (where almost all of the board's heat is generated), but also better thermal coupling from the module to the I2C temperature sensor, enabling more accurate self monitoring.

If things go well this should be the final module design, and I can go make a bunch of these to use for powering all of my future large instrumentation / networking projects.

Solder paste print turned out fine. Not perfect, but more than adequate.

On to my old nemesis from board rev 0.1.

I think I have it the right way around now? Pin 1 in the top left?

Placing the 48V current shunt resistor.

One of the prettiest resistors I've ever worked with (Ohmite LVK series). They're a little pricey but work great.

All done and going in the oven.

Back. Board is out of the oven and ready for hand soldering of PTH parts.

Going to try squeezing this 2mm thick thermal gap filler pad under the buck controller to see if it helps thermals at all. In retrospect this probably would have worked better if I had more exposed copper on the top layer to sink heat to, but too late for that now.

@azonenberg that looks like a really neat system! What's the DC/DC module that you're using? Looks a bit like a TDK.

@Nukular TDK Lambda i3A4W008A033V. Based on a TI TPS40170 synchronous buck controller.

Then the rest of the board is just current/voltage/thermal sensing, soft start, bulk capacitors, fusing, connectors, etc.

48VDC in at top, 12VDC out at bottom, up to 72W output capacity.

Made a few small firmware tweaks for the PCB changes and it's working absolutely beautifully.

Characterization setup is giving good data. Next step will be automating just a little more.

And here it is. Fully automated, hands-off characterization of the IBC module in ngscopeclient. There's definitely some rough edges software side to polish up, I had to do some tweaking to make this work and there's still some annoying bugs.

Sweep load from 0 to 6 amps in 250 mA steps then plot input voltage, output voltage on the board, output voltage after cable drop, input and output current, efficiency with and without cable losses, RMS and p-p ripple, and temperature.

There's still some more work to do to make the final firmware (e.g. I2C device mode on the MCU so I can read the voltage/current ADCs from the top level MCU on the main board this thing will eventually be plugged into).

But hardware wise, I'm happy with it. Will probably order parts to make a couple more since I expect this to become a standard component in future designs.

Next step will be continuing with the trigger crossbar design. Hopefully I can get that schematic-complete over the weekend so I can move to layout during the week.