Very interesting bud, I definitely think something is blocking you St home or interfering.
Quick subject, how did you power the modem in the car?
-
Join the MyBroadband community
South Africa’s biggest forum. Discuss, discover, and connect with thousands of members.
-
Question of the Day: How do you normally pay for purchases in stores?
More on SA cell towers
- Thread starter Azimuth
- Start date
-
- Tags
- alandicks antenna cell tower lte
Quick subject, how did you power the modem in the car?
What's weird is that I can see the tower from my roof, when I'm up there with the router my speed test still doesn't jump more them 3mb.
I powered it with my UPS. I cut a kettle cord in half and attached it to a extension adapter
Bloody resourceful. Lol. Maybe we should Google what can interfere with signal.
Lol. A man make a plan. What's why I'm here, I seek advise from wiser men. I find this place more informative than google sometimes.
Last edited:
This is very very interesting jun_tao. I have been following your quest with interest.
Where is the poynting antenna guy, I would love to hear their input on what could be doing this.
If it were me, I would give them a call and ask to speak to a techy savvy person there, not some sales person, and refer them to these results for some help.
I dunno how these towers work, but if those antenna are directional, maybe the ones pointing to your home are just not working.
Where is the poynting antenna guy, I would love to hear their input on what could be doing this.
If it were me, I would give them a call and ask to speak to a techy savvy person there, not some sales person, and refer them to these results for some help.
I dunno how these towers work, but if those antenna are directional, maybe the ones pointing to your home are just not working.
jcheek
Expert Member
Similar posts have sometimes gone into the Antenna installation and alignment thread, but for this discussion, this thread is as good a place as any.
PS: Thanks for observing the thread topics and leaving a link in the B593 thread - nice threadiquette!
If I understand your (excellent!) post correctly, your tests are all of the same station from different locations - correct ?
My aim was firstly to get you to test your "home" station with the B593 up close, to try and establish the station performance with the influence of a home antenna system minimised (which you've now done very nicely). Secondly, my intention was also to get you to try and identify other Telkom LTE stations in your vicinity and test those in the same way as I did here. (You'll see in that post that the 7 stations near me performed quite differently to one another, even with similar signals.)
One question I do have is : In your first pic, where is the station (tower) relative to your three test locations ?
Nonetheless, your repeated tests of the same station show very interesting results! To summarise :
Location / Result
1 RSSI/RSRP/RSRQ=-41/-67/-6; DL/UL=33.57/6.65
2 RSSI/RSRP/RSRQ=-37/-63/-6; DL/UL=19.44/6.42
3 RSSI/RSRP/RSRQ= ?/?/?; DL/UL=2.68/5.37
Before we start, three bits of background info:
- The one "wildcard" variable to bear in mind with the B593s-601 is that you can never be completely certain it's connecting to the same station after power-cycling it, since it doesn't show you the Cell_ID of the connected station. Even after power-cycling it, it will tend to try first to re-connect to the same station it was last connected to. This is more of an issue when you are testing different stations, so in your case I think it actually worked in your favour, ie chances are good it was connecting to the same station every time.
- Even with perfect signals (as you now have), the speed you measure on a given station will always vary from test to test, depending on factors like the station loading (ie how many other users are connected and what they are doing at that precise time). This would be especially true for "busy" stations with high numbers of users, stations tested in peak times, and stations with a high proportion of users on the "uncapped LTE" packages. However, at some point you should be able to measure a "peak" speed that represents the station's true potential.
- A base station doesn't radiate in all directions from just one antenna, it uses several antennas (usually three) to cover all 360°. You can recognise examples of these "sector antennas" in some of the pics in the opening posts of this thread - nearly all are arranged so as to cover 3 x 120° sectors. So as you drive around one station, your B593 may actually be "talking" to different sector antennas.
Now, some discussion:
- The signals measured from the first two locations are both very strong (-37dBm is a h.u.m.u.n.g.u.s. signal!) and of near-perfect quality (RSSI - RSRP = 26dBm, and RSRQ=-6dB in both cases). So I don't think signal quality is a factor at all (which is the whole idea of this test). In fact it's interesting that the highest downlink speed (34Mbps versus 20Mbps) was measured with the lowest signal amplitude (-41dBm versus -37dBm)
- With signal qualities as good as those you measure, one would expect the station to perform at close to the limit for LTE (ie > 80Mbps) if there were no other constraints (like high user count or low backhaul capacity). So I think you are seeing some evidence of an over-subscribed or fairly busy station (or busy time), although 20 to 34Mbps does start to give some idea of what the station's true potential performance is.
- I don't think the massive drop from 20Mbps to ~3Mbps is likely to be caused by an obstruction, unless there's a humungus obstruction (or some other form of interference) between you and the station. This is borne out by the fact that you get similar (poor) performance from right up close, as you do from home.
1. You measured the lowest downlink performance from this position.
2. This is (more or less?) the same direction as that from which you would connect from home - right ?
Nonetheless, my interpretation of your results would be :
- The signals from the locations for Test 1 and 2 are perfect.
- The station's true speed potential is at least 20-35Mbps.
- The sector antenna(s) - or some part(s) of the transmission infrastructure - involved in Test 3 may be somehow faulty. This theory would be proven nicely if the signal parameters from the Test 3 location were sub-standard, however your "home" figures - assuming the same sector antenna is involved - were actually also pretty good.
- Be careful of the Wi-Fi thing .... it's another 2.4GHz radio link (just saying lol)
- Go back and repeat the test at location #3 (especially signal parameters).
- Try to locate other Telkom stations in your area and get an idea of how they perform from close up.
- Do at least 2 or 3 tests from each location to establish the extent of test-to-test variation.
- Try testing in likely off-peak times (like the dead of night - not great from a security perspective when it involves driving around with a laptop, I know)
Great testing, and great posts, by the way - well done!
Last edited:
I did a close as jcheek did. I have only to viable towers in my area.
The first one gave me 4 signal bars on the router. Download speed where also not that great. RSSI and the rest of the numbers were similar as doing the test at home.
I switched towers and bam 5 bar signal on router. Downloads where more consistent. Meaning 78mbps versus 70mbps. (Band) At home the band was larger.
From what I heard or the message given to me you are locked to Telekom infrastructure.
I also can say both towers had Telkom presence.
You also need to be mobile when doing these test from you car in front of the tower.
Hope this helps.
Lol, thanks for following, Telkom have gotten back to me asking for my address and affected number, I think its their standard response
but its something for now.
Hi All (its me again) haha
@jcheek I'm so appreciative of your input and the time you're taking to write those posts so THANK YOU very much AGAIN!!
And all of those who've had some input.
Before we start, I went back to location 3 to get the stats here they are: RSSI: -56.0 RSRP: -82.0 RSRQ: -6.0
So after being burnt at the taste of Joburg, I decided to redo (almost) the tests again WITH a cable this time. Lol. All these tests are done after 5pm, I did about 3 speedtests per tower, 40 seconds apart, power cycled the router before I went to the next one.
Driving around I found two other towers near me one is quite a distance away though, none the less I did a test there as well.
Tower 3 ( In President Hoerskool shooting range)
Stats: RSSI/RSRP/RSRQ = -69/-96/-6
Speedtest: ping/DL/UL = 41/29.02/1.02
26/29.77/1.31
26/29.75/0.87
Tower 2 (Witspos Mail Centre)
Stats: RSSI/RSRP/RSRQ= -47/-74/-6
SpeedTest: ping/DL/UL= 11/50.73/7.03
26/43.25/7.05
26/57.73/7.06
Tower 1 (Pretty much the same location as in the previous post)
Stats: RSSI/RSRP/RSRQ = -39/-65/-6
Speedtest: ping/DL/UL = 26/28.75/6.02
42/35.09/6.40
26/38.78/7.18
So then a though came to my mind, why not do a test every +/- 70m going from the tower to my house. As far as the road would let me.
Test 1:
Stats: RSSI/RSRP/RSRQ = -58/-84/-6
Speedtest: ping/DL/UL = 41/38.20/6.25
26/40.77/6.35
26/35.65/7.25
Test 2: You can see the stairs from pic in Test 1 and the tower.
Stats: RSSI/RSRP/RSRQ = -42/-68/-6
Speedtest: ping/DL/UL = 26/41.40/7.46
26/35.59/5.89
26/39.98/6.75
Heres where is dies a sad death LOL
Test 3: Those trees on the left are the same tress on the right in Test 2. The tower can be seem just to the left of the dish.
Stats: RSSI/RSRP/RSRQ = -60/-86/-60
Speedtest: ping/DL/UL = 42/2.36/5.44
11/3.01/3.97
Then I went back to the same place as Test 1 just named it Test 4
Stats: RSSI/RSRP/RSRQ = -44/-71/-6
Speedtest: ping/DL/UL = 20/43.91/5.99
31/50.63/6.67
20/64.12/5.66
Test 5:
Stats: RSSI/RSRP/RSRQ = -53/-79/-6
Speedtest: ping/DL/UL = 20/3.86/7.09
Sadly I couldn't do anymore testing my ups died after that. More or less an hour of testing.
On the up side Telkom have gotten back to me wanting my address and affected number, which they say they will forward it to their mobile team. Lets hope something happens!!
Thank you very much guys
@jcheek I'm so appreciative of your input and the time you're taking to write those posts so THANK YOU very much AGAIN!!
And all of those who've had some input.
Before we start, I went back to location 3 to get the stats here they are: RSSI: -56.0 RSRP: -82.0 RSRQ: -6.0
So after being burnt at the taste of Joburg, I decided to redo (almost) the tests again WITH a cable this time. Lol. All these tests are done after 5pm, I did about 3 speedtests per tower, 40 seconds apart, power cycled the router before I went to the next one.
Driving around I found two other towers near me one is quite a distance away though, none the less I did a test there as well.
Tower 3 ( In President Hoerskool shooting range)
Stats: RSSI/RSRP/RSRQ = -69/-96/-6
Speedtest: ping/DL/UL = 41/29.02/1.02
26/29.77/1.31
26/29.75/0.87
Tower 2 (Witspos Mail Centre)
Stats: RSSI/RSRP/RSRQ= -47/-74/-6
SpeedTest: ping/DL/UL= 11/50.73/7.03
26/43.25/7.05
26/57.73/7.06
Tower 1 (Pretty much the same location as in the previous post)
Stats: RSSI/RSRP/RSRQ = -39/-65/-6
Speedtest: ping/DL/UL = 26/28.75/6.02
42/35.09/6.40
26/38.78/7.18
So then a though came to my mind, why not do a test every +/- 70m going from the tower to my house. As far as the road would let me.
Test 1:
Stats: RSSI/RSRP/RSRQ = -58/-84/-6
Speedtest: ping/DL/UL = 41/38.20/6.25
26/40.77/6.35
26/35.65/7.25
Test 2: You can see the stairs from pic in Test 1 and the tower.
Stats: RSSI/RSRP/RSRQ = -42/-68/-6
Speedtest: ping/DL/UL = 26/41.40/7.46
26/35.59/5.89
26/39.98/6.75
Heres where is dies a sad death LOL
Test 3: Those trees on the left are the same tress on the right in Test 2. The tower can be seem just to the left of the dish.
Stats: RSSI/RSRP/RSRQ = -60/-86/-60
Speedtest: ping/DL/UL = 42/2.36/5.44
11/3.01/3.97
Then I went back to the same place as Test 1 just named it Test 4
Stats: RSSI/RSRP/RSRQ = -44/-71/-6
Speedtest: ping/DL/UL = 20/43.91/5.99
31/50.63/6.67
20/64.12/5.66
Test 5:
Stats: RSSI/RSRP/RSRQ = -53/-79/-6
Speedtest: ping/DL/UL = 20/3.86/7.09
Sadly I couldn't do anymore testing my ups died after that. More or less an hour of testing.
On the up side Telkom have gotten back to me wanting my address and affected number, which they say they will forward it to their mobile team. Lets hope something happens!!
Thank you very much guys
Last edited:
Speedster
Honorary Master
I received my Telkom Uncapped LTE package this morning and have it all set up and rumbling along pretty sweetly now. How do I go about finding which tower it's connecting to? Any suggestions would be appreciated.
LowVoltage
Member
- Joined
- Mar 1, 2005
- Messages
- 15
- Reaction score
- 0
Really great work with all the speedtests and such. Very informative to hear other's feedback.
I recently signed up for the TM Smartbroadband uncapped wireless blah blah and it's giving me grey hair.
I get top level RSSI RSRP RSRQ values (-41,-69,-7) and sometimes I can speedtest up to 50Mbps down.
My first problem is I get almost nothing on the uplink. (0.6Mbps). But this might be a distance from tower issue (1km, obscured).
My second problem is that one moment I will pull 30, 40, 50 Mbps and then a minute or two later I can't even load google and get speeds less than 1Mbps. A short while later it's back up and available, and then it's **** for a few minutes to hours. Constantly all over the show. It's actually so all over the show there's no way to draw any sort of pattern or trend.
The fact that this happens significanlty more after 6pm and sort of clears up around 1:30am leads me to believe the biggest problem Telkom is facing is backhaul. The fact that it happens hard and fast on Sunday evening at 8pm makes me believe even more so. I think they have dumped us onto the crappy uncapped free-for-all ADSL backbone with deep packet inspection devices and various throttling and shaping and left us there to rot, without building any additional capacity.
So they've already oversaturated their network, and they still intend to sell a further 148 clients in my neighbourhood? Hmmm.
That being said, I will certainly do a similar "as close to tower" as possible experiment and post feedback.
Oh, and btw, the little rocket ships on the map here:
http://www.telkom.co.za/today/unlimited-broadband/
Those are their towers.
I recently signed up for the TM Smartbroadband uncapped wireless blah blah and it's giving me grey hair.
I get top level RSSI RSRP RSRQ values (-41,-69,-7) and sometimes I can speedtest up to 50Mbps down.
My first problem is I get almost nothing on the uplink. (0.6Mbps). But this might be a distance from tower issue (1km, obscured).
My second problem is that one moment I will pull 30, 40, 50 Mbps and then a minute or two later I can't even load google and get speeds less than 1Mbps. A short while later it's back up and available, and then it's **** for a few minutes to hours. Constantly all over the show. It's actually so all over the show there's no way to draw any sort of pattern or trend.
The fact that this happens significanlty more after 6pm and sort of clears up around 1:30am leads me to believe the biggest problem Telkom is facing is backhaul. The fact that it happens hard and fast on Sunday evening at 8pm makes me believe even more so. I think they have dumped us onto the crappy uncapped free-for-all ADSL backbone with deep packet inspection devices and various throttling and shaping and left us there to rot, without building any additional capacity.
So they've already oversaturated their network, and they still intend to sell a further 148 clients in my neighbourhood? Hmmm.
That being said, I will certainly do a similar "as close to tower" as possible experiment and post feedback.
Oh, and btw, the little rocket ships on the map here:
http://www.telkom.co.za/today/unlimited-broadband/
Those are their towers.
Consider that high level of signal can saturate receivers. Also give few minutes before starting speed test for self tuning between tower and device.
I don't like powering from UPS, it generate lot of interference, in a limited space it can be important. Better is to use DC-DC converter plugged to cigarete lighter socket.
LowVoltage
Member
- Joined
- Mar 1, 2005
- Messages
- 15
- Reaction score
- 0
I have confirmed physically by driving around, and with Street View. It's relatively accurate, imo.
jcheek
Expert Member
Hi LowVoltage !
Some questions:
- What LTE device are you using ? B593, E5186 or other ?
- What antenna(s) are you using when you measure these ? An XPOL-0006 (as supplied by Telkom as part of the deal ?), or the device's internal patch antennas, or the included flat whip anetnnas ?
Those are certainly very powerful signals (both RSRQ and RSRP are high), but the combination is not ideal. From the body of practical experience gained on the B593 and other LTE-related threads, there's good evidence that an RSRQ of anything less than -6dB (on the B593, that is) does not bode all that well for a stable connection or high downlink speeds. The large-ish difference between RSSI and RSRP (28dBm), and the low uplink speed are also indicators (uplink not usually congested).
RSSI-RSRP ≤ 26dBm seems to be the start of the "sweet spot"; and an uplink speed of roughly 10% of the downlink speed would be a more average result.
The figures almost suggest some sort of local signal degradation or interference, IMHO.
That said, I do take your point about performance being better outside of peak hours.
Are your signal stats better when you experience greater performance and better stability ?
Or do your signal stats stay much the same ?
First time I've seen Telkom release base station info in public. It's for sure not a complete listing of all stations, but it's nice to see !
Last edited:
LowVoltage
Member
- Joined
- Mar 1, 2005
- Messages
- 15
- Reaction score
- 0
It's the B593 and the pre-supplied XPOL, although I've used varying configurations in dozens of tests by now. Surprisingly, my most satisfying initial results were with the default flat whip antennas, but that was only the first few hours.
Great feedback, thanks. Technically my distance from tower is outside the published coverage area. Not by much, and fortunately I'm six floors up so line-of-sight to the tower is fairly good. Obviously for uploads the transmitter in the B593 is only as powerful as it is, which will impact upload speed. That being said, Day 1 saw uploads of 6/7Mbps (with flat whips). Both down and up were good speeds. which I haven't seen since.
When I'm not mucking with the antennae, they are absolutely stable. The speed was varying wildly, but the signal was rock solid with no changes.
I did pick up on another forum thread that during the pro-rata period (a period I was not even aware of) it is possible that bandwidth gets capped after a certain quantity used. I'm lending myself to that idea as since the beginning of the month I have not had nearly as much variation in speed and connectivity on the whole has been a bit more consistent.
I'll keep monitoring it though, because I can't afford to be locked into a crappy contract for 24 months if the solution is not stable.
Thanks for the feedback!
jcheek
Expert Member
Noted, thanks for the confirmation.
PS: You may find the PDF in this old post of some interest. That was however in the "early days" when there were a lot fewer TM LTE users and certainly no 'uncapped' ones.
Not an unusual scenario. The coverage maps are compiled for "typical" indoor conditions and being up on the roof (or 6 floors up!) changes the game completely from a signal perspective.
Absolutely!
On TDD LTE systems, the down:up bandwidth ratio is configurable and Telkom are rumoured to run their TDD links at a 3:1 (down:up) ratio. However I've very rarely seen better uplink performance than about 10:1. That's one area where the XPOL-0006 should help enormously, ie making the best possible use of the B593's comparatively paltry (relative to the base station) transmit power.
Interesting!
Pleasure. Good luck!
Hey guys,
Bit of an update on my situation. Two guys from ATC South Africa came to my place to do some tests with their own equipment to check my speed problem. They tried with their router, then router+ antenna, the stats improved but the speed didn't. He said I should be getting at least 35Mbps with 4 bars. They took a lot of screen shots put them into an excel doc
Nifty software they have, was able to see which tower I was connecting to, and all the towers in the area.
He says it looks like it's a faulty sector on the tower pointing in my direction. Just my luck. Lol
Was also telling me yesterday they went to guy who was complaining that his 75Mbps was too slow and they must improve it to 100. Wtf?? People must be happy with that they get. Told me Telkom won't be calling that guy back.
So now it's just a waiting game to see when it's gonna get fixed. Hopefully sooner than later as it affects everyone in the sectors direction.
Thanks to everyone for their input about my problem.
Cheers.
Bit of an update on my situation. Two guys from ATC South Africa came to my place to do some tests with their own equipment to check my speed problem. They tried with their router, then router+ antenna, the stats improved but the speed didn't. He said I should be getting at least 35Mbps with 4 bars. They took a lot of screen shots put them into an excel doc
Nifty software they have, was able to see which tower I was connecting to, and all the towers in the area.
He says it looks like it's a faulty sector on the tower pointing in my direction. Just my luck. Lol
Was also telling me yesterday they went to guy who was complaining that his 75Mbps was too slow and they must improve it to 100. Wtf?? People must be happy with that they get. Told me Telkom won't be calling that guy back.
So now it's just a waiting game to see when it's gonna get fixed. Hopefully sooner than later as it affects everyone in the sectors direction.
Thanks to everyone for their input about my problem.
Cheers.
Last edited:
jcheek
Expert Member
Hey jun_tao
That's interesting! ATC have popped up on a few occasions in these threads, but this is the first time I've heard of them actively getting involved in fault diagnosis on behalf of one of their highsite clients (Telkom). I'd be very interested to see what they manage to achieve.
Great feedback, thanks for coming back, and good luck!
Hey jcheek,
This is the first place I'll let know when I come right.
When I looked at the spreadsheet that he was filling in, I saw a tabs for each mobile operator so I can only assume they are getting more involved. After he was done with the testing he did phone Telkom with the results. So at least something is happening.
Have a good evening!