Why does WiFi suck?

[suprPHREAK]

So we all know that the Spark uses a WiFi connection, and it has limited range and more easily gets interference than digital radio.

Why?

If the Mavic/Phantom (and other RC stuff) can go long range with 2.4ghz radios, why is WiFi so limited? Both are limited by the same power output by FCC or whoever, and I thought all 2.4 channels were the same? In the case if Spark and Mavic, the antennas are the same so you can't even say a higher gain antenna is the difference.

Is it the protocols? Does WiFi use FHSS like Spektrum and others?

Note, I'm talking about when using an RC in all cases. Phone to drone wifi will suck for many reasons. Let's compare apples to apples.

 

[Kilrah]

Protocol/modulation.

WiFi is designed for fastest speed at short range and cohabitation in busy RF environments, so using it as is isn't very appropriate for the purpose but has the advantage of being cheap, small and power efficient since it's so ubiquitous.
The Mavic uses transmission parameters that are optimized for this use but result in more complex hard/software.

 

[Magnetic]

I'm trying to understand this too - I noticed when compared to my Phantom Standard, the live video feed to my iphone is a lot more laggish / choppy with my Spark then my P3S.

I'm using the spark controller, so the only wifi connection is from the controller to my iphone - right? The spark to the controller has it's own non-wifi signal?

How does this compare to my P3S? I don't remember having to fiddle with wifi channels on my Phantom Standard - nor do I think there was an option to do so?

 

[suprPHREAK]

So I was thinking, I think I have a theory:

on non-wifi drones (whether DJI or home made), you have 2 separate data streams: one for sending control, the other for receiving video. Normal RC isnt (usually) a 2 way communication, so the aircraft gets the control, or not. Same with the downlink, you get it, or not. This is why on a Phantom, you can lose video while maintaining control.

With Wifi, there are a bunch of 2-way checks built into the signal. Encode the data, call the recipient, wait for response, send it, wait for response, verify the data was sent, send out a confirmation. I think all that verification creates a lot of points at which signal can be "lost" (not really, but it just didnt pass the check). Might be great when sending computer data, but not so much when trying to maintain control or a flying machine.