Incorrect transmit frequency

[Suretronic]

Gateway receives uplink ok at 865062500 but then transmits an ack at 865620000 when it should be the same as the Uplink frequency.

Could be misinterpreting the ack request from the LoRaWAN server as the same end node communicates with an 8+1 channel gateway (mcard on RPi) using the same router, exact same configuration, with no problem, i.e. 86506250 up and down.

This router is Orne Brocaar's LoRaWAN server, latest version.

Here's some output from the ESP8266 gateway via telnet:
rxpk 0176c600a020a6f42f19ff48 message {"rxpk":[{"rssi":-91,"stat":1,"modu":"LORA","rfch":1,"tmst":1590172565,"datr":"SF12BW125","lsnr":7,"time":"2017-10-28T22:54:58.558062Z","codr":"4/5","data":"gAEBAQGACgAPSO+NYWcB0v7UWQ==","freq":865.062,"chan":0,"size":19}]} length 235
txpk {"txpk":{"imme":false,"tmst":1591172565,"freq":865.062,"rfch":0,"powe":27,"modu":"LORA","datr":"SF12BW125","codr":"4/5","ipol":true,"size":12,"data":"YAEBAQEgFgBuDfmj"}}
txpk_ack {"txpk_ack":{"error":"NONE"}}
transmitPkt 16598 12388 -1 865620000 192 112 2 64 27 12

If I'm reading this correctly the LoRaWAN server request is 865.062 but actual transmission is 865.620.

Spectrum analyser indicates these frequencies are indeed being transmitted.

Probably a while since you've looked a this! So any suggestions would be appreciated.
Cheers, Ron

[Suretronic]

Further.
It works ok with frequency 866200000, which indicates either a calculation problem or problem parsing the server request.
Fix would appreciated if poss. Cheers Ron

[JaapBraam]

This is a bug in https://github.com/JaapBraam/LoRaWanGateway/blob/master/src/LoRaWanGW.lua#L191

[Suretronic]

Hi JaapBraam
Apologies, missed seeing your response until now. If it isn't fixed in that line yet (no chance to check yet), it shouldn't be too difficult to sort out. Much appreciated ... and the gateway is great work by the way, particularly with such an innovative approach to handling multiple spreading factors. Best wishes, Ron.

[jrw77]

My solution for line 191 (split into lines and indented for better readability):

  local fix=data:sub(5):gsub('"freq":(%d+)[.](%d+)',
    function(d,f) 
        local h=f; 
        while h*10 < 1000000 do 
            h=h*10 
        end; 
        while #f > 0 and f:sub(1,1) == '0' do  
            h=h/10; 
            f = f.sub(2) 
        end; 
        return '"freq":'..(d*1000000+h) 
    end)

[JaapBraam]

Looks good!

If it works as expected, please create a pull request to fix the bug.

[jrw77]

Actually, I´d love to, but first I have to get my gateway to receive its first package. I am struggling with creating a node that would send a package. My current approach is to use the same code base (second set of nodemcu/D1 + sx1276, Ra-01) and basically call

function send(mydata)
    padlen = (4 - (#mydata) % 4) % 4;
    
    pkt = { codr="4/5",
            data=mydata,
            freq=CONFIG["GW_FREQ"],
            ipol=true,
            modu="LORA",
            powe=27,
            rfch=0,
            size= #mydata + padlen,
            tmst=tmr.now()+8000,
            datr=CONFIG["GW_SF"]..CONFIG["GW_BW"]
         };
    radio.txpk(pkt)
end

This does send something, which I can perceive as a chirp in a receiver tuned to that frequency, but my gateway (configured to the same frequency and spreading factor) does not receive it. Anything obvious that I overlooked? Does radio.txpk(pkt) send a packet format that the gateway ignores, as it seems to come from another gateway (would make sense to reduce ping-pong interference)?

[Rubensen]

DIO pins configured correctly ? I expend hours last week with exactly this issue with a Wemos board...

[jrw77]

When I shorten the pins to 3V3, the correct interrupt handler gets called (rx timeout on DIO0, rxpk on DIO1, which does not find data available). I don´t think I damaged the sx1276 outputs by that, but it helped to check the wiring to the nodemcu.
I am using 433175 kHz, as my module is for that freq band, do you think that is an issue?

[Kingwulin]

When I shorten the pins to 3V3, the correct interrupt handler gets called (rx timeout on DIO0, rxpk on DIO1, which does not find data available). I don´t think I damaged the sx1276 outputs by that, but it helped to check the wiring to the nodemcu.
I am using 433175 kHz, as my module is for that freq band, do you think that is an issue?

I guess you are a Chinese, I meet some question on the function of detecting multiple SF,would I have a talk with you over QQ or wechat (: