Meaning of AT%XMONITOR reported PSM values?

[tedhavelka]

Hello Devzone,

I am building firmware for an nRF9160 based device, starting with an ncs 1.6.1 sample app and LTE modem firmware v1.3.1.  I'm in North America, and when testing LTE-M celular connection I can typically find -- or the modem finds -- two or three local celular network providers.

When I issue the AT command AT+CPSMS=1,,,"10101010","00110000", I find the modem remembers the values of this or my latest PSM settings request.  This is expected.  When I issue the querying command AT%XMONITOR, I see a report like either the first or second example here:

%XMONITOR: 5,"","","310260","352C",7,12,"014D3C15",87,5035,39,35,"","11100000","11100000","01001001"

%XMONITOR: 5,"AT&T","AT&T","310410","900C",7,12,"050B8210",391,5110,29,21,"","00110000","11100000","00110000"

In the first case, not one of my requested PSM values appears to be granted.  Am I reading this report correctly?  Are the three "one bits" in third from last and second from last quoted values the bit 8, 7 and 6 which are documented at Nordic Infocenter, command CPSMS and its 'set' syntax?

In the second report from %XMONITOR it appears that my requested active time value is granted by the network provider.  But what does it mean for active time to be granted, but Period-TAU value not to be granted?  How severely does this impact possible low power use of the nRF9160 in, for example, battery based devices?

- Ted

[Hakon]

Hello,

the syntax for the CPSMS command and the XMONITOR command are different and should not be interchanged. The interpretation for the bits 6 through 8 is different for CPSMS and XMONITOR, so you need to treat them differently.

[tedhavelka]

Hi Hakon,

While I am just starting out learning about LTE modem capabilities and AT commands, I am aware that AT+CPSMS and AT%XMONITOR use different syntax each.  I spent a good number of hours reviewing these and related AT commands yesterday, reading both Nordic's documentation and exercising them.

From what I can tell, AT%XMONITOR is a reporting / read type command.  It does not set any values, but rather for some strange reason AT%XMONITOR "set" command reads modem parameters.  Compared with other AT commands which support "set", "read", and at times "test" modes, the %xmonitor command documentation and this command's syntax -- no terminating question mark "?" for read -- seem to break with the convention of most AT modem commands.

My specific question is whether those final three reported values of %XMONITOR, in double quotes, are binary reflections of the PSM and Active Time values which are described bit by bit in the "set command" document section of AT+CPSMS.  The reason I ask is that, I'm learning that not all North America based celular providers support PSM.  In my recent post, ticket 86897 I learned there are a couple of ways after having LTE modem request PSM, to check and learn whether the network provider grants PSM.  I shared the above AT command sequence and results, so I may clearly understand how to interpret %XMONITOR's final three values, for example "00110000","11100000","00110000".

I am not interchanging any syntax as I issue commands.  I need to be sure I understand the results I'm getting after I request PSM.  If we're not able to obtain PSM in our geographies, my team and I are looking at a major redesign if we are to continue on with the nRF9160 part.

- Ted

[tedhavelka]

Hello again Hakon,

I wish to update and clarify my issue regarding PSM settings requests and the nRF9160 LTE modem.  My supervisor has done some research and found celular coverage maps for our greater Portland, Oregon area.  In our metropolitan area we have solid coverage in LTE-M and NB-IoT network provision modes.

Question (1) - when an area has good LTE-M or NB-IoT type network service, does this mean that PSM is supported, or are these network modes separate and unrelated to where PSM is supported?

Here is a capture of the AT commands I issue, first ones to assure LTE modem is turned on:

s1 > at+cfun?
- DEV 0218 - found possible AT command.
modem reply: +CFUN: 0


s1 > at+cfun=1
- DEV 0218 - found possible AT command.
modem reply: 

s1 > at_ 
s1 > at+cfun?
- DEV 0218 - found possible AT command.
modem reply: +CFUN: 1


s1 > at+cpsms=1,,,"00100010","00000101"
- DEV 0218 - found possible AT command.
modem reply: 

s1 > at+cpsms?
- DEV 0218 - found possible AT command.
modem reply: +CPSMS: 1,,,"00100010","00000101"


s1 > at+cereg=5
- DEV 0218 - found possible AT command.
modem reply: 

s1 > at+cereg?
- DEV 0218 - found possible AT command.
modem reply: +CEREG: 5,5,"3558","02BDF917",7,,,"11100000","11100000"


s1 > at+cgdcont?
- DEV 0218 - found possible AT command.
modem reply: +CGDCONT: 0,"IP","globaldata.iot","10.165.91.12",0,0


s1 > at%xmonitor
- DEV 0218 - found possible AT command.
modem reply: %XMONITOR: 5,"","","310260","3558",7,12,"02BDF917",397,5035,41,40,"","11100000","11100000","01001001"


s1 > at%xsystemmode?
- DEV 0218 - found possible AT command.
modem reply: %XSYSTEMMODE: 1,1,0,1

Question (2) - from the above capture of queries via AT command handler, does the syntax for AT+CPSMS= look correct?

Both ways to query the modem for PSM settings -- +CEREG? and %XMONITOR -- report back values "11100000" for both periodic-TAU and requested Active Time settings.  Nordic's AT command documentation states that "111" setting of TAU and active time timer bits means the timers are disabled, in other words, PSM not supported.

Question (3) - from above AT command capture are we correctly querying our LTE modem to find out if PSM values were granted?

- Ted

[Hakon]

It looks from the last xmonitor response that you are getting a valid PSM periodic-TAU value. The response is "11100000","11100000","01001001", and you can ignore the first two since they are zero. The last parameter shows that the periodic-TAU is set to 54 minutes. If you want to confirm that PSM is working on the device you can use the udp sample in the SDK since that uses PSM by default, and measure the current draw.

[tedhavelka]

Hello Hakon,

Thank you for clarifying the last parameter.  The 54 minutes periodic-TAU is different than what I request but I understand that in some cases celular providers will grant different PSM values than we request.  I misread the %XMONITOR report, and thought both PSM timers were turned off.  From above %XMONITOR summary it appears, however, that at least one of timers T3412 and T3324 is turned off.

I am now working to build, flash and test Nordic's `udp` sample app for a Thingy91 development kit.  One new question however:  are eDRX and Active Time timers required to be turned on in order to enjoy the ultra low current power savings of PSM?

- Ted