We've created a controlled resistive load. But why?!
We thank you for your attention!
But for real, why? After all, there are many ready-made devices for this role. And even if you do make one, everyone knows that the best option is an active load, which allows one to determine many useful performance characteristics... So, why?
One day, we needed to test a large group of power supplies of different types.
We would use East Tester ET5410A+ as a test load, which was purchased for this purpose. But it turned out that its dynamic characteristics were so poor that some power supplies simply refused to turn on, assuming this load to be a short circuit. At the same time, this device was suitable for sustained tests.
So, we're running out of time and urgently need to find a 600 W controllable load unit to test 24 V power supplies. With no room for another error and a near-zero budget (of course, a zero budget is an exaggeration, but not a huge one), we've chosen a way to solve such a problem with a guaranteed result—a switchable resistive load. We've also decided to make it ourselves.
It seemed convenient if the load block would be controlled by PieBridge and powered by it at 5 volts. We designed a circuit:
In this circuit, the load consists of eight parts, the resistance of which is equal to Rn=1.875×2^n, where n=0..7. These parts are switched using MOSFETs Q1-Q5, controlled by a 74HCT595 shift register.
Heat is removed by forced cooling with a fan, which is automatically turned on whenever the outputs of the 74HCT595 register are in active mode.
The value of 1.875 Ohms for the MSB was chosen to obtain the weight of the LSB equal to 0.1 A and also to ensure that the primary set of load elements is assembled with standard-value 120 Ohm resistors (SQP 5W resistors are used). The load of the two most significant (and most powerful) bits is made of Nichrome 80/20 strips.
SH8KB7 transistors are used to switch parts of the load. Instead of SH8KB7, one can use AO4838 or AO4884, but Q5 may need a heatsink in this case.
The boards were milled:
Assembled the board with triggers and controls:
Put together boards with load resistors:
Fitted all the boards into the box casing:
Installed Nichrome 80/20 strips using high-value resistors:
We wired up the fan to the box and connected it to the Raspberry via PiBridge:
And we're all done:
A small script was written to control the load, implementing both static and dynamic modes of load operation:
####
Note: The script uses GPIO from the PiBridge kit [link].
