I have enabled the optional feature for fine time stamping (‘ftime’) in rxpk on the gateway. I am not able to understand what does the ftime counter actually represents.
It says on hal documentation that it is number of nanoseconds since last PPS [0…999999999].
But the values received is observed to increase with time upto 999999999 and then start over. Should the ftime more or less same, since it counts the nanoseconds past the last PPS pulse ?.., which by nature happens every second.
If there is any reference or ideas available to understand it then please do share.
Sounds like a 1 second count to me and if anything different would not be PPS = Pulse Per Second right?!
Each GPS PPS input resets the counter to zero, which then steadily counts up to 999…999 nano seconds at which point next pulse arrives and resets the counter to zero, which then steadily… I’m sure I see a pattern here! Not rocket science even though it uses a Satellite (or 3)
I’m not sure that’s the question you were thinking of asking.
If you mean, how do I use fine time-stamping to enable the ToF for location services, the short answer is buy a gateway that supports it - I’m not aware of any Pi based ones that do - budget for £/$/€1,000+
If that’s not the answer to the question you were thinking of, then telling us a whole heap more detail would be appropriate.
Perhaps there is an assumption, that with the existing hardware, all you need to do is enable fine time stamping and it will work ?
I do recall an instance where Borroz(?) posted some logs of packets received by 3 Gateways that were in the same room. According to the so called timestamps the nodes were many thousands of miles appart.
Thank you ! I think I understood the statement in the documentation now.
I think ftime means that it timestamps the rxpk on the gateway, with the time value past the last PPS. Silly doubt from my side.
Oooooo, good spot. I’d not want to knock WaveShare but I think that’s just a rebadged Seeed SenseCap which strangely Seeed don’t sell standalone. And as much as I love both Seeed and WaveShare for the value kit, I’d not expect much help with getting TDOA turned on & working.
The main reason is they have the resources to pick up batches of a chipset at a good price (or off the back of a passing HGV) and put them on a PCB at a good price as they have lots of young helpers on subsistence wages. But just because the chipset supports something doesn’t mean they’ve made it work with that feature.
So what is the current situation for using fine timestamping etc for Geolocation on TTN ?
A GPS is needed on the Gateway, and suitable hardware, to provide the accurate time sync needed, but does the TTN backend support the feature, assuming there are enough Gateways in an area that can provide the fine timestamping ?
Believe otherway around - think Seeed take Concentrator from WS to build (some) GW’s. WS had earlier 1302 bd IIRC that, like many such card builders, were guided quickly from 1302 to 1303 during the ‘recent’ supply chain issues - might as well as same/similar price, adds lower SF support and also then the fine TS capability plus works with 1250 so no brainer. Think Tony Smith in Oz saw similar move/rational…there were forum posts on this - maybe a year back? Longer? I loose track with pandemic time distortion field!
Not rocket science even though it uses a Satellite (or 3) 
and possibly also then take a Semtech licence to the secret sauce…