This time I can’t find out a fix at the problems without help.
I had to buy a 4G antenna to could be able going on internet, short story: is missing 60 meters of cable to have FTTC.
I live in a condo made by 3 blocks, every block have 5 to 8 apartments.
So let’s start…
I bought one SXT LTE6 kit and one hAP AC2.
I configured the SXT LTE6 with passthrought on vlan1 also if I can’t have a public IP.
I configured the vlan2 for management so the SXT can take an IP from the same subnet and I can connect directly to make change. (guide found here)
The Cell nearest my home is 700 meters away and from my roof I have a clear los, maybe the cell is a couple of meters lower than my roof.
On the same pole I have a Ubiquiti litebeam, 10-15cm above the SXT. Is not mine.
I have done a lot of test and different configuration, also wrote to Mikrotik support that have sent v022 firmware for LTE6 modem. Nothing change.
SXT is on 6.46.4
hAP AC2 is on 6.47b35
The cell offer B3, B20 and B7. B1 is shutted off by provider (WINDTRE Italy).
I always use only B7 because if I go to CA the performance is a 10% , or more, lower than only B7 alone.
The problems are 2, maybe 3 (LOL).
First: If I use cell lock function seems that after a variable amount of time I can’t have internet. I can’t understand if the LTE modem can’t connect or is stuck. Simply in cellular tab I didn’t find any connection.
I need to remove cell lock with /interface lte at-chat lte1 input=“AT*Cell=0”
Second: Always after a variable amount (hours, days..) of time my modem disconnect from the cell that is near home and connect to a more distant cell that have poor performance and bad signal.
I can’t reconnect again in short time.
This I can’t figure if is my problem or if is the cell that ban me.
In this case I tried to connect to other bands in the cell but is impossible.
any advice?
Any log that I can collect and check ?
Ask the provider about the policies they have for connects from a fixed location. Some providers may not allow you to do that, maybe depending on the subscription you use.
Hi, thanks for the hint.
I already asked and they told me : " We don’t release public ip and for us is not a problems if you use on the same spot".
But, we all know that isp always don’t tell the true.
Maybe operators don’t know the true so I can assume that they ban me from a static location. Right?
Well, when that is what you observe… it could be what is behind it. I remember we had such difficulty when using “GSM dialer” to lower cost of phonecalls to mobile in the past.
(when mobile->mobile was much cheaper than fixed->mobile it was often used as a solution, it was a box connected to the PABX and holding a couple of SIM cards)
You can do only one fast 2CA with B3+B7 OR B20+B1/3/7 who is poor CA.
In future that merge/aggregate/sum speed of 2CA willbe more correct then now.
First: If I use cell lock function seems that after a variable amount of time I can’t have internet. I can’t understand if the LTE modem can’t connect or is stuck.
Simply in cellular tab I didn’t find any connection.
I need to remove cell lock with /interface lte at-chat lte1 input=“AT*Cell=0”
True, I remember that fw v20..23 have that behaviour. v24 have works best. Just sending the at*cell=2,3,… means set the same celllock do re-connection and ca-band still works. But still have problem with that cellLock stuck.
Second: Always after a variable amount (hours, days..) of time my modem disconnect from the cell that is near home and connect to a more distant cell that have poor performance and bad signal.
I can’t reconnect again in short time.
This I can’t figure if is my problem or if is the cell that ban me.
In this case I tried to connect to other bands in the cell but is impossible.
I not know that problem. I can always connect to the cell who works me in default connection.
Cell who was disconnect via cell-monitor can but not must work. This can not work but of you connect at specific CellID then it’s should work.
Right now is back online the B1.
So I got B7@20MHz, B3@20MHz,
When I go to test the single band the best is always B7@20MHz.
Also I noticed that single B7 is better than B7+anything in CA. Is that possible?
Ok. Thank you.
In that post, when you mention RestoreInternet script, where I can find? I search also on google but no results.
Oh, I can stay with v22 or I can ask for testing newer version and help to improve making test?
Currently yes, I hope that will be much better.
In final firmware you should test single Band’s and the CA should do DL+DL aggregation, means DL should be better then only one Band DL.
In that post, when you mention RestoreInternet script, where I can find? I search also on google but no results.
This is a any script with that name who will have example, re-enable lte1, set CellLock, do USB Power-reset, do at+reset, etc…
Oh, I can stay with v22 or I can ask for testing newer version and help to improve making test?
All next firmware are better, the best is v24 but I cannot write what will be in next one.
In my opinion the v24 can be shared to anyone. He is not worse then v20
@20Mhz at Both B3&B7 - this should you reach very good transfers.
I hope you reach a good CQI with that signal.
Our r11e-LTE6 is below LTE Rel12 and I mark at yellow max speed at mimo 2x2 - means at one Band you can reach 150Mbps@20Mhz - this give you good possibility.
Really, I hope MikroTik can share newer firmware for our modems who bring us a stability and DL+DL feature from Cat.6.
I’ll hope to have.
I write to support following back my last ticket. I hope that I can do some tests for them.
Oh thank you for clarification
I’ve got always a CQI from 8 to 9 when I go to dowload. In CA I could do 60+60Mbps or only 60?
SINR have some spike down to 9dB, but most of time stay from 14 to 20 dB
CQI and signals coexist together. I hope speak about this with mkx but at forum we not have PM. mkx is a superior person in LTE stuff.
CQI give you a proper modulation what with @Mhz width give you maximal theoretical speed at one Band.
Cat.6 should do DL+DL means max. theoretical ~120Mbps @16QAM
We all know that radio channel conditions change all the time due to various reasons (signal fading, interference, …). In order for eNodeB to maximize throughput to each of connected devices at all times it needs some (more or less) acurate estimate of radio channel quality, so device does measurements according to what eNodeB requests (usually that’s narrow-band measurements around center frequency, other possibilities include whole band or select OFDM tones) and reports back in form of CQI. Based on CQI eNodeB selects radio parameters for next few transmissions: modulation, MIMO rank, OFDM tone range, etc. In radio parameter selection eNodeB might use also other per-device metrics, such as HARQ ACK success rate.
There is correlation between CQI and RSRP/RSRQ … however they are not directly connected. RSRP and RSRQ are measured on special OFDM tones and during certain periods of time … those transmissions are always present (regardless the load on cell), they carry cell idrntity, broadcast information etc. CQI on the other hand involves measuremenrs of all OFDM tones in certain range, including tones used by system and for users.
For CA obviously eNodeB needs radio channel estimations for all frequency channels which are used by certain client device. With event of inter-eNodeB CA (one band transmitted from one tower, another band transmitted from another tower) it is clear that CQI for different frequency channels may not be correlated (even if both/all bands are transmitted from same tower, radio conditions for different frequency bands may not be similar) and client device has to report CQI for all involved cells.
Back to question by @menego: if both serving cells in CA experience same CQI and are otherwise similar (i.e. same bandwidth, same max MIMO, same supported modulation schemes, …), then you would get 60+60 Mbps.
I guess that LTE modem shows CQI for PSC (primary serving cell) which usually has better RSRP and RSRQ than secondary cells, hence one should not be surprised if realised troughput is less than the value in previous sentence. In addition, cell load is likely not the same, but this time secondary cells (usually operating on higher frequencies) are likely less loaded and can offer higher throughput to clients.
In short: things are not simple and are extremely hard to predict. The only hard fact is this: with CA active one should not get lower throughput than without CA when using any of cells. If this happens, then likely LTE modem is to blame.
@mkx, Thanks for answer @menego, Please upgrade to v025 and check this B3@20Mhz + B7@20Mhz.
I think you can have differ speeds depends from CellLock at B3 vs B7. I have.
Check whitch one will be faster and share results in this post.
ciao Fabio,
io invece ho un problema simile al tuo ma non riesco a spiegarlo in inglese
avrei bisogno di aiuto da te ma non sono molto pratico e non so come contattarti… riesci a scrivermi in privato in qualche modo? l’ideale sarebbe via email david.doberdob(chiocciola)gmail.com, così poi se non ti arreco troppo disturbo ci scambiamo i telefoni così posso spiegarti meglio il mio problema con il SXT LTE6 KIT
grazie mille e scusa la richiesta un po’ invadente
David
but at forum we not have PM. mkx is a superior person in LTE stuff.