Hacker News Reader

Comment by patrickg_zill | Hacker News Reader

Little reason to run anything other than ubiquitous Wi-Fi imho.

You can run an ESP8266 or ESP32 chip, or the Realtek RTL8710 chip which is $2 from Pine64, and probably have enough input and output pins to do the job.

View on HN

Completely different use case. LoRa and Sigfox is for extremely low-bandwith, low-power, mostly one way communication (think sensor data).

You pretty much cannot do this with WiFi due to power usage.

It depends what you mean by low power. What is your definition of low power?

The WiFi enabled chips I mentioned have low power modes where you can wake up either periodically or based on an interrupt , with most of the chip turned off in the meantime. I'd have to read the docs to know about power usage in different modes. The RTL8710 claims 30 mA...

With good programming you can get many 8bit microcontrollers into the micro and nano amp range.

Sleeping for 60 seconds, waking up, then sleeping again gets the RTL chip at least, into about 0.3 average milliamps. The 8195 is very similar to the 8710 https://www.amebaiot.com/en/ameba-arduino-power-management/

Yes, but that still doesn't give you years in battery life as with LoRa.

I agree with you for the IoT devices I would create in my life, as I am around WiFi in all of those cases (home, office, etc). However for sensor networks in rural areas this would be a great solution. For example, the water reservoirs where I live report live statistics to a server and I can see in real time what the levels of them are (along with other data, like rainfall, temperature, etc). There's no WiFi in the backcountry!

Wifi has an incredibly limited range compared to these technologies. You'd need a repeater every dozen meters or so, which means running power to all of those repeaters. Meanwhile SigFox can run up to 50km. That's a huge reason to run SigFox instead of wifi.

With your reasoning, we might as well take down all the cell phone towers and just make calls over wifi, wifi is everywhere right?

I wasn't aware that sigfox could travel such distances.

Then again you can do 10km+ with a grid antenna connected to one of the WiFi chips. It might look silly, but it would work.

WiFi is not exactly ubiquitous and it requires an infrastrucutre which is short range by definition. It's also power hungry. If your use case is to place several sensors in a vineyard without having to replace batteries every day you want something like sigfox. Your arguments works if you're a hobbyist hacking together a diy solution (i.e. you value your time 0 dollars) but if you're a company or a more professional user you might value the fact that this works out of the box, is well supported and manufactured to last in the field.

SigFox and wifi aren't really comparable at all. SigFox is ultra low power, ultra long range (up to 50km, they claim) and sends only 12 bytes of data per message.

Right, but SigFox is claiming IoT will mean 500 devices in the typical house. With at least 1Euro/year/device subscription cost, that's a minimum of 500 Euros per year. A typical house has wifi already, so it could be a lot cheaper to use low-power Wifi instead.

The 500-device-home seems to be SigFox's primary target. If so, then they're de facto competing with wifi.

I suspect they're either speaking broadly, or thinking in terms of having sensors in everything, always reporting readings. Their tech doesn't seem particularly suited for interactive devices, because the protocol requires the device to poll the network and see if there are messages waiting for it.

Edit: Also I'm fairly certain that WiFi will choke long before you get close to having 500 devices in an area the size of a house.

Great mentions. In my case, this is appealing especially for the <20mA power consumption for battery-powered nodes to place and leave for 1-2+ years.

Reply on news.ycombinator.com