Cellular network problems: The Facts

What about SMS's that get delivered...but to the wrong recipient??? Vodacom couldn't help me on that one so I don't know why they are claiming that every network fault is addressed, BS.
 
Roadrunner said:
Just a thought:

They could of course be doing this intentionally behind the scenes to boost their profits. Every dropped call has to be re-initiated again which means you get charged for two calls. Same goes for SMS messages that are not delivered. Most people just send it again.
Click to expand...

This conspiracy theory is as old as the mobile industry. In my past 15 years of in-depth engineering with various operators and suppliers I have never come across deliberate attempts to affect network quality (including dropped calls).

EDIT: negatively affect :) there are continuous attempts to improve it though
 
Last edited:
According to the Vodacom Techie that phoned me to sort out some network problems, Vodacom is a 1st tier internet provider, that means that VODACOM provides all their own network backhaul, at least until the internet traffic needs to exit their network and go overseas... so Vodacom, you can't blame Telkom!

My latest speed test on my supposed 3.6 Mbps connection: http://www.speedtest.net/result/473450516.png
 
superfly said:
Vodacom is a 1st tier internet provider, that means that VODACOM provides all their own network backhaul, at least until the internet traffic needs to exit their network and go overseas[/url]
Click to expand...

Not quite. 1st tier means direct connectivity from the ISP's gateway to the rest of the world (amongst others), not from the ISP to the user.
 
w.m said:
What about SMS's that get delivered...but to the wrong recipient??? Vodacom couldn't help me on that one so I don't know why they are claiming that every network fault is addressed, BS.
Click to expand...

This they can not control because of cross talk with lines. Happens with phone calls and smss'
 
The migration to 3G has created significant problems for all GSM operators, but they won't tell you.

The physical spacing of the original basestations was appropriate for voice-only services at 900 MHz. When 1800 MHz became available, they had to shoe-horn the RF planning for 1800 MHz into the plan for 900 MHz, but it was never a perfect fit. In addition, they had to adjust the planning criteria to take account of data services, which typically require much lower error rates, and hence higher signal levels, and shorter distances between basestations.

When they overlaid 3G, the problems multiplied significantly. For a start, it's a completely different technology, using CDMA rather than TDMA as the basic modulation and access method. As a result, the planning criteria are completely different, typically using signal to interference ratio rather than signal level. It is relatively straightforward to design and build a greenfield 2.1 GHz CDMA network for optimum performance, but overlaying one on a 900/1800 MHz TDMA network is a real challenge. The result is a network with lots of holes that cannot easily be filled.

In terms of basic RF propagation, at 2.1 GHz, running 3G/HSPA, the operators now need something like twenty times the original number of sites, so there is a huge mismatch between the original site spacing and that required for 3G. they've built a lot of in-fill sites, but it's a bit of a patchwork. Basically, to give you a warm and fuzzy feeling, they lie on their coverage maps - 3G coverage is not seamless; it's generally quite patchy, and, in many places, it's not there at all. You're just not supposed to notice because you fall back to GSM, even if you're half way through a call, at least in theory.

Hence, the perception that people have of poor 3G performance is quite real. It would be particularly interesting to see what the call drop rates are on the 3G part of the networks - it's almost certainly significantly higher than the benchmarks or the GSM performance.

A lot of the rest of the issues that are experienced are really a result of congestion on the networks, particularly in dense areas. UMTS uses the same RF carrier for both voice and data, so there is a strong correlation between data traffic load and poor voice performance. This is not the case with some other technologies.
 
Last edited:
ads said:
The migration to 3G has created significant problems for all GSM operators, but they won't tell you.

The physical spacing of the original basestations was appropriate for voice-only services at 900 MHz. When 1800 MHz became available, they had to shoe-horn the RF planning for 1800 MHz into the plan for 900 MHz, but it was never a perfect fit. In addition, they had to adjust the planning criteria to take account of data services, which typically require much lower error rates, and hence higher signal levels, and shorter distances between basestations.

When they overlaid 3G, the problems multiplied significantly. For a start, it's a completely different technology, using CDMA rather than TDMA as the basic modulation and access method. As a result, the planning criteria are completely different, typically using signal to interference ratio rather than signal level. It is relatively straightforward to design and build a greenfield 2.1 GHz CDMA network for optimum performance, but overlaying one on a 900/1800 MHz TDMA network is a real challenge. The result is a network with lots of holes that cannot easily be filled.

In terms of basic RF propagation, at 2.1 GHz, running 3G/HSPA, the operators now need something like twenty times the original number of sites, so there is a huge mismatch between the original site spacing and that required for 3G. they've built a lot of in-fill sites, but it's a bit of a patchwork. Basically, to give you a warm and fuzzy feeling, they lie on their coverage maps - 3G coverage is not seamless; it's generally quite patchy, and, in many places, it's not there at all. You're just not supposed to notice because you fall back to GSM, even if you're half way through a call, at least in theory.

Hence, the perception that people have of poor 3G performance is quite real. It would be particularly interesting to see what the call drop rates are on the 3G part of the networks - it's almost certainly significantly higher than the benchmarks or the GSM performance.

A lot of the rest of the issues that are experienced are really a result of congestion on the networks, particularly in dense areas. UMTS uses the same RF carrier for both voice and data, so there is a strong correlation between data traffic load and poor voice performance. This is not the case with some other technologies.
Click to expand...

Umm... come again...? English maybe? :D Just kidding. :p
 
caps said:
With the way calls drop on my side and MTN claim only 1% of calls drop, then I'm the one who makes up that 1%. What are the rest of you guys complaining about???
Click to expand...

That means I must be on the other end of your dropped calls, ie. that whole 1% occurs between us!
 
Interesting post ads. I think the difficulty with site accusation seems to aggravate the problems associated with a network which requires more sites, like the higher frequency 3G network.

I can however confirm that MTM has in March released statistics showing very similar dropped call rates on 2G and 3G. The statistics are in fact nearly identical.

ads said:
The migration to 3G has created significant problems for all GSM operators, but they won't tell you.

The physical spacing of the original basestations was appropriate for voice-only services at 900 MHz. When 1800 MHz became available, they had to shoe-horn the RF planning for 1800 MHz into the plan for 900 MHz, but it was never a perfect fit. In addition, they had to adjust the planning criteria to take account of data services, which typically require much lower error rates, and hence higher signal levels, and shorter distances between basestations.

When they overlaid 3G, the problems multiplied significantly. For a start, it's a completely different technology, using CDMA rather than TDMA as the basic modulation and access method. As a result, the planning criteria are completely different, typically using signal to interference ratio rather than signal level. It is relatively straightforward to design and build a greenfield 2.1 GHz CDMA network for optimum performance, but overlaying one on a 900/1800 MHz TDMA network is a real challenge. The result is a network with lots of holes that cannot easily be filled.

In terms of basic RF propagation, at 2.1 GHz, running 3G/HSPA, the operators now need something like twenty times the original number of sites, so there is a huge mismatch between the original site spacing and that required for 3G. they've built a lot of in-fill sites, but it's a bit of a patchwork. Basically, to give you a warm and fuzzy feeling, they lie on their coverage maps - 3G coverage is not seamless; it's generally quite patchy, and, in many places, it's not there at all. You're just not supposed to notice because you fall back to GSM, even if you're half way through a call, at least in theory.

Hence, the perception that people have of poor 3G performance is quite real. It would be particularly interesting to see what the call drop rates are on the 3G part of the networks - it's almost certainly significantly higher than the benchmarks or the GSM performance.

A lot of the rest of the issues that are experienced are really a result of congestion on the networks, particularly in dense areas. UMTS uses the same RF carrier for both voice and data, so there is a strong correlation between data traffic load and poor voice performance. This is not the case with some other technologies.
Click to expand...
 
These stats only address one of the complaints, i.e. dropped calls. There is no statistic supplied for calls abandoned due to poor quality, or inability to initiate a call due to "network busy" or having to redial repeatedly because you get the message "the subscriber you have dialled is not on the XXXX network" only to have the connection successful after three or four attempts.

And the non/late delivery of text messages on MTN was quietly avoided by focusing on the Idols issues. There is no question that MTN has had some serious problems with their SMS offering over the past few weeks.
 
arf9999 said:
These stats only address one of the complaints, i.e. dropped calls. There is no statistic supplied for calls abandoned due to poor quality, or inability to initiate a call due to "network busy" or having to redial repeatedly because you get the message "the subscriber you have dialled is not on the XXXX network" only to have the connection successful after three or four attempts.
Click to expand...

It's called Call Setup Success Rate.
 
It actually is a huge difference.
Click to expand...

I assume it's quite a significant difference, number wise. Thousands of calls must be transmitted each day. It was just funny to read the percentage difference of 1%.
 
swb said:
I assume it's quite a significant difference, number wise. Thousands of calls must be transmitted each day. It was just funny to read the percentage difference of 1%.
Click to expand...

It's one of the dangers when you work with rolled-up numbers. You need to be very sensitive to small changes. 0.1% can mean something big is going down (literally!).
 
1%? My ass... i DEFINATELY have more dropped calls than 1 in a 100! In fact i would say i have difficulty with at least 10% of my calls be it terribly quality, or dropped calls.
 
Bobb said:
1%? My ass... i DEFINATELY have more dropped calls than 1 in a 100! In fact i would say i have difficulty with at least 10% of my calls be it terribly quality, or dropped calls.
Click to expand...

Remember the 1% is spread over the total subscriber base.

Radio communications is not a uniform structure over the total country with everyone experiencing exactly 1% dropped calls. :)

So, you could see 10% and it will roll up into that 1%.

Best you report it to your network for investigation.
 
I've noticed the same in the past month. I rarely get a call to stay connected more than 60 seconds - and that is at all times of the day or night not just peak times.

I often dial 2-3 times to connect to anyone as the line is just dead. At least these calls are unpaid!

I think this is about making money. Every dropped call is a redial. And they can't be measured unless the authorities demand to know how many times the user called the same number 4 times within a few minutes for less than 30 seconds each time!
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X