It turns out that Taylor University is indeed using a switching power supply IC The chip they are using to generate the +/- 15V rails is the TPS65130 from TI, and the complex conglomeration of diodes, capacitors, inductors, and resistors coming off the pins of the chip in their poster is a straight implementation shown in the chip's datasheet: http://focus.ti.com/lit/ds/symlink/tps65130.pdf (see page 6 and compare with poster schematic). Most of the similar chips I am looking at also require a few additional low-cost external components, so if we don't have those lying around we may need to look into ordering a few diodes, inductors, capacitors, and a MOSFET too. Due to the new +20V/-20V requirement, TPS61530's output of +15V/-15V will be insufficient for our purposes. Otherwise, its characteristics seem perfect for the space environment and supplying the power to our circuitry as well. I am currently in the process of looking through chips with comparable performance that can generate the desired +/- 20V rails. One possibility is Maxim's MAX1620 (http://pdfserv.maxim-ic.com/en/ds/MAX1620-MAX1621.pdf), which can take an input in the range of 1.8V to 20V and output a single rail, with a maximum of EITHER -27V or +27V. That means we would need two of these chips to get both the + and - rails, but with the small surface-mount package it may be doable anyway. Again, the external components connected require some kind of thought and planning for inductor/capacitor values, diode type, transistor make/model, etc. I think that will be a common theme with these DC-DC converter chips. I am also not sure about the "operating current" referred to in the description on the first page, which is on the order of microamps and may interfere with functionality of the CMS in some way? Overall, this chip looks interesting but probably inefficient for our purposes, so I am going to keep looking for a better alternative. The input voltage range seems impressive though. It is also interesting that the MAX1620 has a DAC, but the resolution is not good enough for the VVS as far as I can tell. Besides the switching circuit used to generate the larger voltage rails, a REF5050 (http://focus.ti.com/lit/ds/symlink/ref5050.pdf) is used to generate the +5V reference - this is a very nice precision reference with low noise and temperature drift. I am thinking this is the way to go for generating those 5V. The 3.3V is generated with the REG101-33 (http://focus.ti.com/lit/ds/symlink/reg101-33.pdf) which is again a good chip for the space environment. As for the 1.55V used during the instrumentation amplifier stage of the CMS, Taylor feeds a simple half-and-half voltage divider to the 3.3V regulated line and follows the divider output with an op-amp buffer. This is a low-cost (or not, since they are using the OPA277UA which may be pricey) approach that seems worth imitating, since we plan on using those instrumentation amp stages in the CMS. If it turns our that different voltages are required for those instrumentation amp stages (there were talks of doing parallel stages to accommodate for the current range?), we could just use multiple divider-opamp-buffer circuits like this. Summary of current thoughts: - We are still in search of an 'excellent' solution for the switching power supply IC; we need something that can take an input around 5V (?) and generate +/-20V rails in the space environment. - We can use Taylor's methods for the rest of the power circuitry - they seem to have selected the optimal chips and approaches for those tasks.