HL7588 T suddenly periodically loosing all signal strength on the go


We recently encountered problems with the Nimble Link NL-SW-LTE-S7588 (HL7588) using the A.2.13 version of the firmware where the MBIM USB interface (at+ksubcomp=2) would periodically disconnect from our linux base product when doing mobility testing in a car. Stationary testing on a desk were always successful (no radio unregistration, no USB disconnection overnight). We then saw that in the release notes of A.2.15 that this version introduced a fix for a potential baseband reset problem related to base station handovers with an opened data session, precisely what we were observing. Indeed, upgrading to A.2.15 seem to have fixed the USB reset problem we were seeing on the road. However, in our connection monitoring log we often now also see signal strength drop from 80 %+ to 0% within 5 seconds (obtained from ModemManager in our linux based product which is in return converted from AT+CSQ commands). The radio is still reported as registered and the IP connection is also reported as active but no data can be sent. Our tests are done in a dense urban place where LTE coverage is not an issue and a phone was used as a comparison device). This null signal strength stays like that for a couple of minutes (from 3 to 6 so far) and would jump back up to 80%+ again. This pattern is seen every 10-15 minutes or so when driving around in the city.

Is sudden signal drop like this a known problem with this radio/firmware ? Although it seems to recover at some point, data over our PDP context does not seem to get through during that time.

I would gladly attach a log from our ltemonitor script that shows the different timestamps and the registration/ip connection states overtime as well the signal strength but it seems new accounts can’t do that

Any ideas/support would be appreciated !

Thanks in advance

does latest FW have this problem?
how is the reproducibility?
How many devices are seeing this issue?


Yes the problem repeats itself with the latest firmware we have access to : A.2.15. I currently only have access to two units (two production samples of NL-SW-LTE-S7588) and both seem to show the problem. In order to reproduce this problem, we install our linux based unit with the NL-SW-LTE-S7588 on the dash of a car with the LTE antenna next to the windshield and we drive around the city. Our monitoring script logs the signal quality value provided by ModemManager over time. After a couple of minutes (it varies from 5 to 15 so far) we see the signal quality going from 80%+ to 0% within 5 seconds. After 3-5 minutes, the signal quality goes back up to 80%+. This problem was reproduced the 3 times we did such a mobility testing with either unit


So problem not reproduce at fixed place?

Yes correct, I left the device a couple of hours on a desk (using A.2.15) with signal quality being log periodically and it stays pretty much at the same level for the whole period. The problem seems only be present on the go


Could it be antenna or signal strength problem /


Judging on the results of our test, it’s likely not a simple signal strength problem. We have recently tried using both the primary and diversity antenna on our device and the result is the same. If this would be a signal strength problem we would probably not see such a high RSSI before the drop as it would likely stay at an average value all the time. Also, doing the same test with the device stationary does not show any sign of signal drop over time so if the link is maintained with high RSSI most of the time except when it suddenly drops to 0 the RF path is probably not the source of the problem.

We were seeing the same thing with A.2.13 (on top of USB disconnection). It looks like something was improved in terms of module stability during base stations handover between 2.13 and 2.15 but some unstable behavior remain with mobility performance. We also support other pin compatible LTE modules from Nimble Link on our product (like the NL-SW-LTE-QBG96) and we don’t see such drops with this module so clearly something is off with the HL7588 with A.2.15