CONFLICT OF INTEREST NOTICE: my employer makes component SMPSs A common problem, often perceived as noise, is when the control loop of the electronic load is operating at a beat frequency of one of the SMPS control loops. I typically use a quality (X/Y-rated) cap from the load test-power terminal to the chassis of the load to reduce/eliminate this problem.
If you are doing noise SMPS noise output measurements, valid data can only be acquired using discrete resistive loads. with a z-match network to the measuring device. EUT connect hardware is very important. In general, radiated noise has never been a problem with any load mfr; but I have had problems with some mfrs that had conducted noise problems at some combinations of current and voltage. At lease one load mfr tests for "load-side" conducted emissions by using the load to sink a pure linear power source. This spec can be difficult to extract from the mfr's documentation. Load-induced noise, for most name-brand mfrs and for my experience, has not been significant for testing either linear or SMPSs, WHEN operating within the rated voltage and current ratings of the load. Remember that the actual "load" element is a linear device, and linear stuff, when properly designed, is very quiet. I suppose that "thermal" noise for some components can be significant. But one should assume a decent load design would take this into consideration... An important spec is the compliance V. At low currents, some electronic loads are not stable when the applied V approaches the rated min input. Finally, much of an electronic load's noise figure can come from user configuration; e.g., constant-V vs constant-I mode, external V control stabilitiy, user connection technique from EUT to load, computer-generated noise (mostly form cheap GPIB cables), etc. If the p.s. "swamps" load-induced noise, it's probably time to evaluate the next p.s. mfr in the queue... R/S, Brian O'Connell Taiyo Yuden (USA), Inc. we are looking into purchasing a programmable DC electronic load tester for evaluating SMPS. The mainframe unit with one module will provide app. 1200W of loading. Features include constant current and constant resistance modes. Agency claims include CE Class A. How is this type of equipment tested for emissions on the input side? Having never used such a device, I am curious to know if we should be specifying a particular noise figure or parameter in our search. Intuitively, it seems to me there would be some small amount of unwanted noise generated by the excitation of the representative loading components. Wouldn't thermally generated noise be worth considering, or is this negligible? Maybe the output noise of most SMPS swamps out typical electronic load noise?
