Hacker News Reader

Comment by mschuster91 | Hacker News Reader

> That means that the lower limit of the supply voltage is about 5V but the upper limit is much less than 40V. That's because the input capacitor is rated at 25V, so I would not suggest running it on more than 20V.

Even that will stretch it without rework. The higher the voltage difference, the higher the heat losses - at the minimum you'll need a heatsink, at higher voltages I'd even go for active ventilation.

View on HN

If the total power consumption at the supply is the same or lower, how is the voltage regulation circuitry generating more heat?

The problem is the voltage drop. The bigger voltage difference it has to handle, the hotter it gets as its the loss from "internal loss current x voltage drop" is converted into hat.

The heat has to come from the input power though. If I'm pumping 5W at 12V, or 5W at 20V into the device, with it idling, presumably the output from the devices voltage regulator circuit is the same voltage, and the downstream components are consuming the same wattage (let's call it 1A at 3.3V downstream, so 3.3W), the both the 12V and the 20V input would have 1.7W of heat-loss. The article shows that as voltage goes up, power actually drops, which would imply that it is producing less heat.

That only applies to linear regulators.

The router on this page uses a switching regulator. Switching regulators have a relatively constant power draw over the entire input voltage range.

Yeah, I think they’ve mixed up linear regulators and buck converters. Buck converters are often more efficient with a larger voltage drop because the switch can run with a lower duty cycle.

Thanks, I never knew why. All I knew is whenever I need a buck converter in my DIY, novice PCBs, I just toss in an off-the-shelf preassembled one, with a small LDO on my actual PCB to drop the last bit and smooth it out.

Reply on news.ycombinator.com