PSU test load

Richard wrote: For a basic test, where we just want to verify that the PSU works properly near its maximum rated load: rated voltage = E rated current = I resistor value = R = E/I resistor wattage = P = E*I = E^2/R For instance, for a 5.2V 35A power supply, R = 5.2/35 = 0.149 ohms. We aren't going to find an actual 0.149 ohm resistor, but rounding up a bit we can find 0.15 ohm. With that, I = E/R = 5.2/0.15 = 34.67A, near the maximum rated current, and P = (5.2^2)/0.15 = 180.27W. Note that we have to bolt that resistor to a good heatsink, and provide force air cooling, to prevent it from burning up. A 0.15 ohm 200W resistor is roughly 2" diameter by 3" long, which is not much volume (or surface area) for 181W of heat. We can buy a chassis-mount 0.15 ohm 200W resistor from Mouser Electronics, but it costs $46, and that resistor probably won't be of much use to us for anything other than testing that particular power supply. We might be better off buying more resistors of a more common value and paralleling them, or buying a variable resistor. With variable resistors (rheostats) we have to be careful, because the power dissipation rating is at the maximum resistance, and for lower resistance we have to linearly derate the power. For instance, we could try using a 225W 1 ohm variable resistor, but when we set the tap at 0.15 ohms, we now have to derate the power to only 33.75W. That won't meet our requirements. Suppose instead we parallel 33 5 ohm resistors. That gives us an overall resistance of 0.152 ohms, and power dissipation of 5.2W each. We can use resistors that cost us $0.93 each. In total, that is $30.69. We haven't saved much money over the single resistor, and now we have to expend more effort in wiring the resistors up to form the load, but it's far more likely that we'll be able to reuse those resistors for a different power supply load in the future. To do "serious" power supply testing, we would need to measure line regulation and load regulation. For load regulation measurement, we can put a solid-state relay in series with some of our parallel resistors, so that we can near instantaneously change the load current (load step), and watch the supply voltage change on an oscilliscope. Lyle Bickley and I built the test loads for the DEC 728 power supplies of the PDP-1. We did not test line regulation and load regulation for step response; we just tested the power supply under four different load conditions, and measured the DC voltage, RMS ripple, and peak-to-peak ripple of each output under each of the four load conditions, with a variac on the input at 105V, 115V, and 125V. Testing the 728 is complicated by the fact that it is a dual output supply, providing +10V and -15V. There is a maximum current rating for each output, but there is also a maximum overall power, such that one cannot draw the maximum current from both outputs simultaneous. We built the load box with a rotary switch to select load combinations of: +10V at 0.2A, -15V at 0A +10V at 7.5A, -15V at 0A +10V at 0.2A, -15V at 8.5A +10V at 2.0A, -15V at 7.0A Eric