Antenna installation and alignment

[AntennaMan]

Try Ellies. They stock the 40mm masts and the brackets.
Also consider using the 50mm Ellies mast instead. The extra 10mm and the slightly thicker material add a lot of rigidity. Only snag is it's usually only available in a 6m length.



Yes, you can often find the Ellies 3-leg brackets and the 40mm masts at Builders.

PS: Before you drill the holes in the wall for the bracket, lay it flat on the floor and check that the mounting surface is level / parallel to the floor. The brackets are pretty crude and most are quite squiff. It's not that obvious when looking at the bracket in isolation, but it makes a big difference on a 3- or 4m mast when erected. Really easy to do by bending the 3rd leg on the ground, but really hard to do once the mast is up and the holes are drilled (spacers etc) - trust me on this one - doh! .



"Coffee here" ?

Thanks for the advice. I'll definitely check the bracket before installation. I'll look into the larger diameter mast, but 6m might be too long for my Micra...

 

[Muttley]

"Coffee here" ?

Great idea!

AntennaMan, I'll have 1 x double espresso thanks

 

[stepper]

Question. Can 2 routers share a single antenna?

 

[sajunky]

Question. Can 2 routers share a single antenna?

Not really. Splitter can be used for receiving-only devices. When one device is transmitting it can destroy receivers of the other device and in reverse.

 

[GforceD]

The LTE tower is down the hill from me so I just need to get my antenna up over the houses in front of me and the tower should be eye level with my antenna.

 

[GforceD]

Question. Can 2 routers share a single antenna?

unlike TV satellite dishes which only receive, internet antennas have to broadcast too so you can't split the feed.

 

[stepper]

Ok thanks guys.

 

[gtsun]

For the first 3 days I was pointing at the Vodacom tower. (Edit: wasn't even a cell tower)

Edit: Telkom mobile sectors are on the MTN tower, so I was only pointing at Vodacom in error, not Vodacom *and* MTN as previously stated. Point at MTN and you're guaranteed to be pointing at Telkom Mobile.

To date I've not had a reply from Telkom Mobile re: antenna direction but I guess I finally found that elusive tower. Your eye kinda gets trained on these things.

Next I'm going to build a drone to map the whole area!

awesome

 

[kalevra]

Hi all.

I'm in Broadacres, JHB. I'm running the E5186 with Telkom Mobile 100GB+100GB.

Just got myself an XPOL-6 today.
Before the antenna setup I was getting between RSRP -97db and -100db and speeds varied from 5Mbps to up 17Mbps.
With the antenna I'm now getting -89db, the speed is on the lower end of the range at 5Mbps.

I've tested without the antenna and speed it still less than ideal (similar speed).

When I was getting the much better speed of 17Mpbs with a RSRP of -97db the router Webui (4 ot of 5 bars) and LED display showed (3 out of 4 ) bars. The strange thing though is now that the antenna is being used I'm on 2 out of 4 bars for the LED display and 2 out of 5 for the web ui (they must share values and overlap at that point or something).

1. Have I been using the incorrect figure (RSRP) to determine my signal stength? I'm confused with the fact that my "better" RSRP should equate to more bars and better speed test results.

2. I think there may be convention issues with the service. If I report these will Telkom Mobile resolve by upping total available bandwidth for the cell tower?

3. Are the other metrics, besides RSRP, important at all, what do they mean?

RSRQ: -10dB
RSRP: -88dBm
RSSI: -59dBm
SINR: -2dB

Note: I have seen at least 3 separate cell ids, the one I normally connected to without an antenna I was also able to connect to with the antenna, as well as another. So I don't believe the issue is related strictly to a difference in cell tower I am connecting to before and after the antenna installation.

 

[Alacrity]

Note: I have seen at least 3 separate cell ids

Can you post the cell ids...

 

[jcheek]

3. Are the other metrics, besides RSRP, important at all, what do they mean?

Have a look at this post for background on what the metrics mean.
The post was written around the B593 ("plain" LTE only), but the principles remain the same for the E5186 and LTE-A, with two notable differences:

• Practically, RSRQ on the B593 maxes out at -6dB. From what I've seen, the E5186 seems to be able to achieve a higher RSRQ, -3dB if I recall correctly. It's not clear to me whether that represents a real improvement in RSRQ, or whether the E5186 just measures it differently.
• The E5186 adds the SINR metric (ratio of (signal) to (interference+noise)), which the B593 doesn't have. For SINR, as with RSRQ, a higher (ie less negative) dB value is better.

RSRQ: -10dB
RSRP : -88dBm
RSSI : -59dBm
SINR : -2dB

RSSI = -59dBm is a pretty strong "raw" signal. However the 29dBm difference between RSSI and RSRP, and the low RSRQ (should ideally be -6dB or higher) both point towards low signal quality and/or interference.
I don't have a strong frame of reference for SINR on the E5186, but -2dB (?) also strikes me as really bad (ie low). I would expect it to be positive ... ? (see here : +15dB!)

Have I been using the incorrect figure (RSRP) to determine my signal stength?

No, your use of RSRP as a measure of useful signal strength is fine. However ... the other metrics matter.
The signal metrics are the most precise method of characterising signal strength and quality.
Save for what is said below, I can't comment on how exactly they map to the signal bars on the front panel or the Web UI of the E5186.

I'm confused with the fact that my "better" RSRP should equate to more bars and better speed test results.

I don't have personal experience with the E5186, but certainly on the B593, my impression is that the LED signal bars on the front panel are not driven by signal strength (RSRP) alone - other metrics (probably RSRQ) are incorporated.
On my B593, even with a very strong signal (RSRP > -75dBm), I will only see 4 out of a possible 5 front-panel bars if the RSRQ is below -6dB.

Note: I have seen at least 3 separate cell ids, the one I normally connected to without an antenna I was also able to connect to with the antenna, as well as another. So I don't believe the issue is related strictly to a difference in cell tower I am connecting to before and after the antenna installation.
I think there may be convention issues with the service. If I report these will Telkom Mobile resolve by upping total available bandwidth for the cell tower?

The million-dollar question! It's worth a shot at least.
Report the CellID(s) that perform poorly, along with signal stats and speeds to back up your report, and give it a go.
Contention may well be an issue, but as alluded to above, I think signal quality (possibly interference ?) might also be a problem in your case. I would try aligning your XPOL-06 to specific, other stations to see what performance you get.


Without knowing your location, it's difficult to be sure; but it's worth mentioning that I don't see LTE-A coverage for the Broadacres area on Telkom's coverage map, only plain LTE.

(PS: Am I the only one who thinks Telkom has messed up its "coverage map" tool completely with the latest "upgrade" ?
Previously it had an annoying little issue with zooming, but was basically quite functional across all technologies. Now it's just ... an ergonomic disaster )

 

[kalevra]

Can you post the cell ids...

5714949
5820421
5820418

Will post more if I see them.

 

[kalevra]

Have a look at this post for background on what the metrics mean.
The post was written around the B593 ("plain" LTE only), but the principles remain the same for the E5186 and LTE-A, with two notable differences:

• Practically, RSRQ on the B593 maxes out at -6dB. From what I've seen, the E5186 seems to be able to achieve a higher RSRQ, -3dB if I recall correctly. It's not clear to me whether that represents a real improvement in RSRQ, or whether the E5186 just measures it differently.
• The E5186 adds the SINR metric (ratio of (signal) to (interference+noise)), which the B593 doesn't have. For SINR, as with RSRQ, a higher (ie less negative) dB value is better.

RSSI = -59dBm is a pretty strong "raw" signal. However the 29dBm difference between RSSI and RSRP, and the low RSRQ (should ideally be -6dB or higher) both point towards low signal quality and/or interference.
I don't have a strong frame of reference for SINR on the E5186, but -2dB (?) also strikes me as really bad (ie low). I would expect it to be positive ... ? (see here : +15dB!)



No, your use of RSRP as a measure of useful signal strength is fine. However ... the other metrics matter.
The signal metrics are the most precise method of characterising signal strength and quality.
Save for what is said below, I can't comment on how exactly they map to the signal bars on the front panel or the Web UI of the E5186.



I don't have personal experience with the E5186, but certainly on the B593, my impression is that the LED signal bars on the front panel are not driven by signal strength (RSRP) alone - other metrics (probably RSRQ) are incorporated.
On my B593, even with a very strong signal (RSRP > -75dBm), I will only see 4 out of a possible 5 front-panel bars if the RSRQ is below -6dB.



The million-dollar question! It's worth a shot at least.
Report the CellID(s) that perform poorly, along with signal stats and speeds to back up your report, and give it a go.
Contention may well be an issue, but as alluded to above, I think signal quality (possibly interference ?) might also be a problem in your case. I would try aligning your XPOL-06 to specific, other stations to see what performance you get.


Without knowing your location, it's difficult to be sure; but it's worth mentioning that I don't see LTE-A coverage for the Broadacres area on Telkom's coverage map, only plain LTE.

(PS: Am I the only one who thinks Telkom has messed up its "coverage map" tool completely with the latest "upgrade" ?
Previously it had an annoying little issue with zooming, but was basically quite functional across all technologies. Now it's just ... an ergonomic disaster )

Thanks jcheek - thorough and detailed answer - very much appreciated!

I checked yesterday and it doesn't appear the LTE-A has been rolled out in the area - hopefully I'm in for a pleasant surprise in the near future.

I'll report my statistics along with speed test results to TM and report back with the outcome.

My stats this afternoon:
CELL_ID: 5820421
RSRQ: -10dB
RSRP: -89dBm
RSSI: -61dBm
SINR: -2dB

Speed test:
Ping: 48ms
DL: 3.5Mbps
UL: 4.49Mbps

Note: I still need to correct my cable installation as it is short and not allowing me to get above the top of the roof and point back towards Broadacres shopping centre where I think there may be a tower I can connect to. It has fake looks leave on it to make it look like a tree - are those normally exclusively Vodacom or do some Telkom Mobile towers also have the appearance? Could have sworn I saw a Telkom stick on a box beneath the installation which made me thing it may be TM.

Edit:
I just checked Google street view (December 2014 images) of the corner and those Telkom boxes were already there. So they quite possibly have nothing to do with the Cell tower right next to them. Will check it out and post my findings.

 

[jcheek]

My stats this afternoon:
CELL_ID: 5820421
RSRQ: -10dB
RSRP: -89dBm
RSSI: -61dBm
SINR: -2dB

Speed test:
Ping: 48ms
DL: 3.5Mbps
UL: 4.49Mbps

Same comment (about poor RSRQ and SINR), but the fact that your upload speed exceeds your download is also 'interesting'. As my own rule of thumb on Telkom LTE, when everything is OK, upload speed is around 10% of download. In addition to the signal quality issues, a reversed relationship could also indicate excessive contention for download capacity on that station.

I still need to correct my cable installation as it is short and not allowing me to get above the top of the roof and point back towards Broadacres shopping centre where I think there may be a tower I can connect to.

Yup, Step #1 is to find out where your local TM stations are.

It has fake looks leave on it to make it look like a tree - are those normally exclusively Vodacom or do some Telkom Mobile towers also have the appearance? Could have sworn I saw a Telkom sticker on a box beneath the installation which made me thing it may be TM.

The "Big 4" often share highsites, including the camo ones. Have a look here for example.

 

[kalevra]

Same comment (about poor RSRQ and SINR), but the fact that your upload speed exceeds your download is also 'interesting'. As my own rule of thumb on Telkom LTE, when everything is OK, upload speed is around 10% of download. In addition to the signal quality issues, a reversed relationship could also indicate excessive contention for download capacity on that station.



Yup, Step #1 is to find out where your local TM stations are.



The "Big 4" often share highsites, including the camo ones. Have a look here for example.

I've just gone up the on the roof again and removed the old Antenna (was a 800Mhz to 2500Mhz) single Antenna by Term for 3G I believe. Was wondering if that may interfere at all being so close to the other (I doubt it but thought I'd give it a try). No noticeable difference - the metrics changed slightly but could not discern whether this was due to the normal fluctuations seen or to do with the antenna (non point it seems in any case but thought I'd mention).

Armed with some new information regarding the other metrics I directed the XPOL-6 to a different tower (5714949) and got the following connection:
By the time I wanted to copy and paste it had connected to 5714946.

PCI: 236
CELL_ID: 5714949
RSRQ: -8dB
RSRP: -90dBm
RSSI: -61dBm
SINR: 1dB
Ping: 29ms
DL: 10.71Mbps
UP: 4.74Mbps

Still had the speed test result from the other CELL_ID:
CELL_ID: 5714946
Ping: 26ms
DL: 6.07Mbps
UP: 4.20Mbps

From the above it is seems to be clear that the improved RSRP and SINR have made a positive difference. The WebUI is also now showing 3 bars instead of 2 which seems to correlate with the increase in signal quality. Nice to finally understand the numbers better.

In the morning I'll be sorting out the cable management and doing some tower research to see if I can improve the above. Will post results.

Posted the PCI value up above as well, seems to change according to the CELL_ID that is currently being used. Any idea what it means?

Thanks jcheek - your assistance is much appreciated!

 

[sajunky]

Posted the PCI value up above as well, seems to change according to the CELL_ID that is currently being used. Any idea what it means?

PCI (Physical Cell Identity) identifies scrampling code. It is right that PCI is changing with Cell_ID. If you receive two Cell_ID with the same PCI on the same frequency, it means Telkom did something wrong in network planning. Two Cell_ID with the same PCI would interfere very bad. Source: http://lteuniversity.com/ask_the_expert/f/69/t/589.aspx

 

[kalevra]

Same comment (about poor RSRQ and SINR), but the fact that your upload speed exceeds your download is also 'interesting'. As my own rule of thumb on Telkom LTE, when everything is OK, upload speed is around 10% of download. In addition to the signal quality issues, a reversed relationship could also indicate excessive contention for download capacity on that station.



Yup, Step #1 is to find out where your local TM stations are.



The "Big 4" often share highsites, including the camo ones. Have a look here for example.

Finally sorted my antenna installation but have not seen any dramatic increase in speed. In fact, I have gotten better speeds without using any antenna whatsoever.

My LTE signal metrics with the XPOL-6:
CELL_ID: 5745922
RSRQ: -9dB
RSRP: -90dBm
RSSI: -61dBm
SINR: -2dB

Absolutely shocking Speedtest result this evening (normally get 5Mbps down and up):
Ping: 25ms
Down: 1.7Mbps
Up: 2.32Mbps

Something must be terribly wrong, hopefully contention or issue at the base station they can resolve.

The low RSRQ and SINR are still pointing to bad signal quality although I've improved my RSRP by 10dBm (it was 100dBm without any antenna). The router is still only showing 2 bars signal which correlates with the poort RSRQ\ SINR values.

I'm going to a log a log a call with Telkom Mobile and see if they can offer any assistance.

If I have some more time in the meantime I'm planning on figuring out exactly where the towers are for the +-4 CELL_IDs are that I am able to connect to. I did, what I though was, a thorough check 360 degrees and tested to ensure I get the best signal values. Not sure what more I can do from here!?

In parallel I'll be engaging with a few fibre vendors in the area to see if we can get fibre into our estate, if so I'll perhaps pursue this instead if I'm not able to resolve my poor LTE performance issues.

Anyone got any "magic tricks" I can try to improve the signal DL\UL performance?

P.S. Forced a reconnect and connected to a different tower:
CELL_ID: 5745925
RSRQ: -10dB
RSRP: -89dBm
RSSI: -61dBm
SINR: -2dB

Speed test just as appalling:
Ping: 39ms
Down: 2.03Mbps
Up: 2.33Mbps

 

[rodga]

Hi guys

Can anyone help with tower locations in Pierre van Ryneveld?
I want to install the antenna for the TM uncapped LTE as the speeds/signal is much worse that where I was before

thanks

 

[TheVeteran]

Hi.

From all the experience that everyone has, does the installing of an antenna help reduce ping?
From what I gather is that it improves download and upload speeds drastically but no info on latency.
Any help is appreciated.
Here is a pic of the tower which is around 500m from my house.

Thanks

 

[jcheek]

From all the experience that everyone has, does the installing of an antenna help reduce ping?
From what I gather is that it improves download and upload speeds drastically but no info on latency.
Any help is appreciated.
Here is a pic of the tower which is around 500m from my house.

It depends a bit on what kind of reception you can get, what ping you have now, what sort of ping you need/want, and what network you plan to use. Being that close to a base station - assuming it's the "right" one - you may not even need to bother with an external antenna to get good reception.

The lowest ping time I've ever seen in my own experience - regardless of whether it's obtained with a built-in antenna, indoor externals or outdoor externals - is about 20ms. See here for an example of an improvement from 30 to 20ms using external antennas, and here for a best-case "close to station" result using just internal patch antennas.