More on SA cell towers

[atomcrusher]

Here's a Telkom tower on the outskirts of Swellendam, WC

I posted the photo in another thread in June this year

 

[SteveO]

So the antennas without RRU's are GSM only?

Are you saying microwave is added for a separate LTE backhaul?

RRU is used for any technology and almost any frequency band, so its not a direct indication. Each site will have a specific setup and configuration, so its difficult to make a statement saying "GSM only"

Microwave(MW) will be used based on what is most cost effective. If there is possibility to have 1Gbps fibre at one site, the operator may choose to use MW to the next 3 or 4 sites as its often more cost effective and easier to install than fibre.

 

[Azimuth]

Mostly its just the addition of LTE RRUs here in SA as most of the LTE rollouts are being done using 1800 or 2100 mhz.

Please expand on your "mostly" statement? This is what led to my assumption in my last post.

Also, LTE on 2100MHz??

 

[SteveO]

Please expand on your "mostly" statement? This is what led to my assumption in my last post.

Also, LTE on 2100MHz??

Cell C used 5mhz of their "4Gs" spectrum to trial LTE.


Vodacom and MTN carved some of their GSM 1800 spectrum to free up 5 or 10mhz for LTE.

Depending on the site configuration they could have:
1. Stuck the radios in the cabinet (if there is space) and then inject the LTE into the existing feeder system along with the GSM 1800
2. Use remote radios (RRU)mount the RRU at or close to the antenna and then inject it

Solution 2 was often the easiest or most effective as there was no need to add in extra cabinets and other equipment.

That said, each site will have a different configuration. So what may work in one site may not work 100% in another.

 

[Azimuth]

Sweet, awesome explanation! I wonder when Cell C will finally launch their LTE service? They've conducted an insanely long trial.

I have partial LOS to a tower 1757m away from me. I'll have to get creative with installing an LPDA for that one!

On that note, will the smaller camouflage type towers receive LTE upgrades? Usually the antennas are quite sparse with one microwave link. Sometimes there isn't even a BTS shelter.

 

[SteveO]

They cant afford to rollout LTE. No spectrum and no cash!

1.7km should be fine for MTN, but it depends on how much the site has been tilted and what the area is like, dense trees buildings etc.

Also, with LTE, LOS isnt 100% necessary. Sometimes multipath reflections off certain surfaces can give you better results.

Not sure on those towers. They can always remove a "branch" to add an antenna

 

[Azimuth]

Sorry, the 1.7km link comment was for a Cell C base station.

I already connected to both MTN and TM at 1622m away with stellar results.

 

[jcheek]

Yes, thats a RETU - Remote Electrical Tilt Unit
Allows the antenna to be tilted remotely from the NMS (Network Management System)

What is the use case for this?

Basically saves a lot of time in terms of optimising the network. Without a RET a rigger would need to visit the site, adjust the tilt manually (involves climbing the tower and in some cases shutting off the site) so safe to assume a day to have this done. With RET its literally 5 minute activity with no downtime.

Sorry, I meant from the consumer side (your above post is still plenty useful!)

Only a guess on my part, but I would think that adjusting the amount of tilt allows the network operator to adjust the size (range) of a specific cell or sector : tilt up (towards horizontal) to enlarge it; or tilt down (towards ground) to shrink it. I would guess this allows them to optimise the "borders" between adjacent cells to ensure smooth handover.

 

[jcheek]

An active GPS antenna.

Another example of a GPS antenna on a base-station.
Not the best choice for antenna placement ... about 120 degrees of sky missing

 

[jcheek]

You are right in that the TMA almost negates attenuation losses from the "long" coax (feeder) However it only amplifies the UL signal. What is being achieved with a TMA is to allow the (weaker) signal your phone is sending to site to be amplified before it is attenuated via the feeder. TMA is essentially a LNA and not a PA, which is used in the RRU/RU

That makes perfect sense - thanks, SteveO !

 

[Azimuth]

The reason i compile this is not just because of signal issues- its to look out for you in the first place- if you have a tower in your area thats sporting 400kbps download and 1.2mbps upload? Its a congested tower and needs to be reported- extra transmission is needed on the site! What would be even better is if a new tower could be constructed

Voldemort, what is the process for reporting a congested tower? I have tried to do so on two occasions but it seems like fruitless task. Are there some keywords one has to use or special department to specifically ask for?

Is the MTN Coverage Query Template the right form to use?


SteveO?

 

[SteveO]

Voldemort, what is the process for reporting a congested tower? I have tried to do so on two occasions but it seems like fruitless task. Are there some keywords one has to use or special department to specifically ask for?

Is the MTN Coverage Query Template the right form to use?


SteveO?

Congestion is difficult. It could be RF issues, baseband (processsing) capacity or transmission.

The best is to contact the operator via something like the MTN coverage query template. Do it weekly as these guys get a lot of complaints and many a time its just subscribers doing silly things or perhaps a network outage due to some unforeseen event.

An upgrade will only be given the go ahead once it goes past the planner for that area. The planner will look at all the issues and then advise on what needs to be done.

I did this recently for a site in Lonehill. However the planner knows me and I pretty much demanded he upgrade the capacity on the site It still took 3 or 4 queries on the MTN website and about 2 months of waiting.

 

[Azimuth]

Heh! Excellent, I'll do as you suggested. I've been seeing some funny things happen with TM and MTN LTE.

One more question on this topic if I may, what does "upgrade the capacity" entail exactly?

 

[Azimuth]

SteveO or Voldemort, I've read about Self-Organizing Networks (SON), where the overloaded cell offloads to neighbouring cells, but I have no idea if this tech is used in SA?

Edit: I see there are many aspects of SON so feel free to comment on one/all of them.

 

[AntennaMan]

View attachment 139748
Another example of a GPS antenna on a base-station.
Not the best choice for antenna placement ... about 120 degrees of sky missing

This is not really such a big problem as it seems. Let's look at your comment about 120 degrees of missing sky. What the GPS antenna actually sees can be demonstrated as follows: Take a tennis ball, cut it in half and throw away one half. The piece left over is pretty much what a GPS antenna sees in an open environment (the other half of the tennis ball being the ground and everything below the horizon). Now take a nail and drive it through the tennis ball at some arbitrary point. Cut off the piece of tennis ball behind the nail when looking at it from the centre of the tennis ball. Not such a big piece, is it?

From recent tests I did in an open environment, the GPS receiver should detect anything around 12 - 14 GPS satellites at any time. You need a minimum of 4 for a 3D fix. Statistically the chance of 10 satellites being in the piece of sky obscured by the tower is extremely low. I think from a batch of 14 possible satellites, you might lose 2, maybe 3 satellites. There are still enough to get a very accurate fix.

 

[Azimuth]

Great post! Now between you and jcheek, please explain why a GPS on a fixed site is necessary? What's wrong with just submitting the co-ordinates when the site goes live? I'm clearly missing something.

 

[SteveO]

I think its for A-GPS.

Not 100% sure on this one.

If its a TM site, then the GPS is for TDD lte, as it needs perfect time sync to work without causing interference.

 

[Azimuth]

Explains it right away! Didn't even think of that. That's brilliant info right there, thanks SteveO. I suppose a follow-up post on the time guard interval for TDD is now in order. (I'll post a link, don't worry)

 

[jcheek]

Now take a nail and drive it through the tennis ball at some arbitrary point. Cut off the piece of tennis ball behind the nail when looking at it from the centre of the tennis ball. Not such a big piece, is it?

That depends on the relative sizes of the antenna and the nail, and whether the antenna is directly behind the nail or not! To a small antenna, that mast is one big, badass nail

As a semi-related aside : I believe one of the interesting effects in the early days of vehicular GPS was the signal challenge presented by doing a U-turn. The GPS receiver - in its usual position just under the metal roof of the car - went from seeing one half of the upper half of the tennis ball, to seeing a completely different quarter, with a different set of satellites.
These days the vehicular receivers are a LOT better, so much so that most of them no longer even have a port for an active external antenna.

From recent tests I did in an open environment, the GPS receiver should detect anything around 12 - 14 GPS satellites at any time. You need a minimum of 4 for a 3D fix. Statistically the chance of 10 satellites being in the piece of sky obscured by the tower is extremely low. I think from a batch of 14 possible satellites, you might lose 2, maybe 3 satellites. There are still enough to get a very accurate fix.

All true. In the early days of GPS when Selective Availability was engaged, using an active GPS antenna made a world of difference in getting stronger signals from as many satellites as possible to reduce the SA error. I had the experience of seeing a "full-house" on a 12-channel receiver many times as a result.

In any event, I doubt that positional accuracy is too big an issue in this application .. hopefully the base-station is not moving enough to need a really accurate position fix. If it is, we've got other problems ...

Apart from getting an accurate time-signal into the bargain, I suppose the main advantage of having a GPS receiver at the base-station is just convenience, really. Plug it in and the equipment gets the right info without anyone having to enter mast coordinates manually. I was interested to see it used in more than one base station.

 

[Azimuth]

Explains it right away! Didn't even think of that. That's brilliant info right there, thanks SteveO. I suppose a follow-up post on the time guard interval for TDD is now in order. (I'll post a link, don't worry)

TDD FDD Duplex Schemes

While FDD transmissions require a guard band between the transmitter and receiver frequencies, TDD schemes require a guard time or guard interval between transmission and reception. This must be sufficient to allow the signals travelling from the remote transmitter to arrive before a transmission is started and the receiver inhibited. Although this delay is relatively short, when changing between transmission and reception many times a second, even a small guard time can reduce the efficiency of the system as a percentage of the time must be used for the guard interval. For systems communicating over short distances, e.g. up to a mile or so the guard interval is normally small and acceptable. For greater distances it may become an issue.

The guard interval required for TDD will comprise two main elements:

- A time allowance for the propagation delay for any transmissions from the remote transmitter to arrive at the receiver. This will depend upon the distances involved, but it takes 3.3 microseconds to travel a kilometre, 5.4 microseconds to travel 1 mile.
- A time allowance for the transmitter receiver to change from receive to transmit. Switching speeds can vary considerably between equipments but can take a few microseconds.