Measuring efficiency using our LM2621 demo board resulted in 10-15% lower than the typical efficiency value claimed in the datasheet. Why?
We have found some things on the LM2621 evaluation board that might be of interest. We took some readings on an LM2621 board that was reported to have low efficiency, and we we still got the high efficiency numbers that we had measured on other LM2621 boards. We tried using other methods of measuring the current, another multimeter and a current probe, and got different results. We found out that the first DMM we used was not averaging correctly, and because of the large input ripple on the LM2621, we suspected that the others might not be as well. So to eliminate averaging problems, we used a large LC filter between the lab supply and the board and measured the voltage directly at the board and the current into the LC filter. This gave a much smoother current waveform and all of the meters and probes agreed. With this we got efficiency measurements very close to the datasheet. We believe two things are happening to get strange readings. First, the input current waveform is tough for D
