The power supply for the CMP3000 controller was failing and needed some work. It had a crowbar circuit across its output, to protect the expensive & sensitive downstream control circuitry. Unfortunately, it would kick in once in a while, sometimes causing the power supply to burn itself up. It wouldn’t even blow the fuse – just cook itself to death.
Everyone else was busy, so Lorne gave me the job. I dove into the TL494, its operation, and how our circuit worked. Amazing! I mean, I’d been working with simple linear regulators for years, but the idea of discontinuous conduction being able to transform voltages, was a revelation to me.
First, we improved the heat dissipation capability of the main switch transistor by increasing its heatsink area (as I recall, we just stacked an extra one on). That addressed the immediate problem of circuit damage.
Then, we reduced the fuse rating, but worked hard to ensure that no nuisance fuse operations would occur.
Lastly, Lorne introduced me to the concept of foldback current limiting. I actually got the circuit operation backward the first time, and sent it into a serious overcurrent, ha ha. Then I got it right. Lorne helped me get it fine tuned, honestly, but I was quite enthused about the result. No overheating, no fuse blowing, and now fully short circuit proof!
At the same time, I did an parametric analysis of the crowbar circuit, and came up with a better way to trip it with more accuracy & repeatability.
With these changes made, the CMP3000 supply was solid for years, until we later switched it to a MOSFET switch…. someone else did that 🙂