Thought I’d give you guys a brief write-up so that other people who are googling for this get a hand up.WhyAfter about 100h of printing the power supply to my board packed it in – I got a nasty smoky smell whilst printing and turned it off in a hurry, then it wouldn’t turn on again. Apparently these boards are known for poor power management and I was procrastinating on a mosfet for the heated bed :/What you needIt turns out that you can get the whole ramps kit on amazon for 35 quid: http://ift.tt/2xeJvVH, you’re also going to need a set of female dupont connectors and a crimping tool, because the z-motor and sensor connectors on the melzi board that comes with the i3 aren’t compatible with the RAMPS board (there’s not enough space to squeeze them in), as well as some heavy cable for connecting the power supply.It helps to have some pliers and a knife handy for prising / squeezing things off the old board and onto the new.Things I noticed whilst doing itI followed this guide to plug in all the cables: http://ift.tt/1NWcPlo Make sure you take a photo of your melzi board before disassembling it because it’s frustrating to try to figure out which cable was which after just unplugging them all at once.Once I had everything unplugged and re-plugged, I turned it all on and installed the Marlin firmware. I used the firmware directly from github on the stock configuration, and I copied this guys’s values into the appropriate places on configuration.h: http://ift.tt/2wkmV1y set the default acceleration to be super low, and I tried loads of different things for the screen but never got it to do anything more interesting than go blank and blue. Fortunately, I have octoprint set up on a raspberry pi anyway, so I didn’t need a screen. I used the octoprint console and M119 (if I remember correctly) to make sure that I had all of the wires for the switches in the right holes and the right way around – basically move the print head over to hit one switch at a time and send that gcode, and check that they all come out right. The first time around I had x and z the wrong way around 0_0ResultsThe very first calibration cube I printed out was within 2% of the size it should have been in every dimension, and the extruder appears to be correct at 111%. The printer is much much quieter and seems to be printing more accurately than before. I’m very satisfied with the results! http://ift.tt/2xeVd2J