How to compare power consumption of 433 MHz and 2.4 GHz?

[muhkuhns]

When i google for power consumption 433mhz 2,4ghz and similar most results tell me that 433MHz consumes less power... but when I look at specifications of explicit chips/modules I can't seem to find a 433MHz module that use less power than the 2,4GHZ module..

For example those two nordic chips:

www.nordicsemi.com/.../nRF51822

6.3mA - TX at -4dBm

www.nordicsemi.com/.../nRF9E5

9mA Radio TX at -10dBm

The SUB-1-GHz chip seems (at least it seems to me) to loose the duel. Less signal, more power use. Where's my mistake?

As I don't need a long range nor much throughput, power consumption is all that matters to me. The less the better. a few years out of a CR2032 would be ideal.. (Every 5 or so seconds a measurement via SPI from a sensor, and then send the two 24bit values via bluetooth)

[awneil]

You might want to take a look at this: www.ganssle.com/.../ultra-low-power-design.html

[mnhs]

One of the advantages of 433 Mhz is the free space path loss. It is about 15dB less with 433 Mhz when compared to 2450 Mhz. Thus, you theoretically achieve about 5 times more range with same output power and receiver sensitivity (source - playing with this calculator www.qsl.net/.../jsffield.htm ). But as written in your question, the range is not your priority.

If your only priority is the battery life, I would recommend to go for nRF51xxx and ANT-like protocol or advertise the data as a Bluetooth beacon (send non-connectable advertising packets). These types of communication do not need any communication between transmitter and receiver, I would call them "passive" protocols. Also care about the battery you use. You shouldn't use standard cheap CR2032, but some kind of special one with low self-discharge current. Then you should be able to achieve several years of battery life.

Note: Please be aware that this approach will offer you extended battery life, but it is quite dangerous for critical applications. The data cannot be properly encrypted without two-way communication and so the communication can be sniffed and also hacked quite easily.

[muhkuhns]

Thank you very much for this detailed answer! It really helped a lot! I need to talk back wether we need encryption or not. If I was thinking about using Gazell for it.

[mnhs]

No problem ;-). For some applications, it is not needed to encrypt the data, but only to encode them. One approach I used was like this (my implementation was more complicated but to demonstrate, I chose this easy example):

1. Generate a random number with the internal RNG
2. Subtract this number from all bytes in your payload
3. The last byte of your packet payload will be the randomly generated number

When you receive the encoded packet, you just add the last byte of packet to all other bytes and you recover the original payload.

It is still hackable, but when you use more difficult type of encoding, you make it significantly harder.

[muhkuhns]

I will take this into consideration! But most likely I am going to encrypt it using a PSK. (Well let's see how much that will drain...)