This is part four of evaluating a TPS54A20 on an evaluation board. The TPS54A20 is a buck converter with an
interesting topology - it has two phases merged with a switched-capacitor. This isn't a formal road test (I
was given the board by jancumps, nice person that he is) and I'm just doing what interests me and blogging about
it. Disclaimer: some of what I do here is quite sloppy and, to be fair to TI, it shouldn't be taken as a
This time (by popular demand) it's efficiency. I was going to do input ripple next, but then realised that I
don't have a sensible way to measure it.
To come up with an efficiency graph I need to know the power going in and the power coming out. I'm in a hurry,
so this is going to be a bit rough and ready - if my curve differs from TI's, go by theirs because they will
have put a lot more care into it.
Power in is volts times amps. The amps I'll measure with a bench multimeter. The volts I'll assume are the 12V
from the bench PSU. That's not strictly true, because the ammeter drops a small voltage and it will vary with
the current, but it won't make too much difference.
On the output side, my load is 25 strings of resistors which each take 0.4A, so if I remove them one by one I
can adjust the output current by that increment. That's nice because if I tried to measure the output current
the meter voltage burden would make a considerable difference.
Here's my curve. Sorry it's not labelled very well, I never have been any good at getting spreadsheets to
create decent graphs; the axis up is efficiency 0-100% and the axis across is output current 0-10A.
Here's TI's version for comparison
They are a fairly close match. Mine is a bit off, and I'd probably get the sack for being sloppy and taking
short cuts if I worked for TI , but it's good enough for practical purposes.
As jancumps said in the comments to part 3, it's not the best efficiency you'll see in a buck converter, but it's
reasonable given the switching speed and the overall physical size of the circuit. As always with buck
converters, the efficiency gets poorer at lower loads.
At full output (10A) the output power is 12W, the input power is around 15W, and there will be getting on for
3W disappearing in the circuit as heat. That's quite a lot of power to dissipate - not all of it will be in
the chip, there are losses in the coils and capacitor too - so observing the layout guidelines for getting
heat away from the chip is important.