Is a regulator necessary?

Question

Sorry for the provocative subject line. I was looking at using one of those integrated ESP32 + SX1276 + SSD1306 boards (there are a few that have popped up) in a battery powered application. I would like to be able to use the Wi-Fi / Bluetooth features and the LoraWAN features, but probably not at the same time or in the same application even.

I looked up all those part datasheets, and by my reading they should all operate down to 1.8V. Someone asked me, "could you just run it off of two AA batteries?" Let's not worry about questions of battery life for now, and treat that question as "is that a plausible thing to do without fundamentally compromising basic performance?"

Electrically, it seems plausible to wire battery power to the 3.3V header pin, bypassing the module regulators. But my gut said that radio performance might suffer from unregulated power, and that using a 1.8V regulator between the battery and board would be a better idea, but I wasn't satisfied with my confidence in giving that answer, which is what brought me here.

So what are the pros and cons of bringing a regulator to the party here?

Answer 1

If they run all from the same battery, and no analog circuitry involved, no problem. The LoRa has its own built-in regulator if the module has a wide-range power spec.

No, I don't see why it shouldn't work.

The con of a regulator is size and heat created on one spot. Well, the latter even may be a pro if you have to include a heatsink of some sort anyway.

Answer 2

I'm inexperienced with microprocessors, but know a few things about them as loads. For an acceptable voltage range that small, it's easily possible that the device simply operates over that range. Efficiency may vary. If the device processes in bursts, you should probably use a 2V regulator at least so voltage drop from current spikes doesn't cause performance issues or dropouts.

Some microprocessors can operate both more efficiently and at higher frequencies at lower voltages, so long as voltage (including volt drop due to current) stays within spec. Depending on your load, this may amount to a loss of 1/3 of your voltage (Depending on the circuit and any existing regulator built in, if the 1.3V above 2V are just going to waste, that's more than 1/3 of your power, and could justify a 70-80% efficient voltage converter.

Yeah it should probably work with just a pair of batteries, but depending on the wattage, I'd consider throwing in an extra battery and also using a regulator. If you have some fresh batteries around and some deader batteries, you could hook them up to the circuit and measure current under the same load, and as long as you have two or 3 datapoints you should be able to get an idea of how much the change in power is affecting efficiency.