Jump to content

LTE informal research in South Africa

Recommended Posts

A good friend of mine, who's an Engineer in real life, just made a really good post and I wanted to share it here for further comment.


By now we have quite a few forumites actually using LTE and we've accumulated a number of real results for various users' signal parameters and achieved speeds.


I thought it would be interesting to scrape all this info into one place, see what it looks like on some scatter charts and perhaps try to draw some conclusions. The question : given knowledge of RSSI, RSRP and RSRQ, can we make any useful statements or predictions about likely performance ?

If nothing else, it gives an interesting overview of actual results so far.


First, some preliminaries:


Units & measurements

There are many good sources of info on deciBels, dBm, RSSI, RSRP and RSRQ.

For the sake of completeness, here's my précis version:

  • deciBels, dB and dBm : The amount of power reaching your antenna system is best measured on the logarithmic “Bel” scale, in deciBels relative to a received power of 1 milliWatt (dBm). On that scale, a level of -50dBm - which is a really strong signal - represents a received power of 10nW (10 nanoWatts). Every 3dB fall on the deciBel scale represents a halving of the signal power, and every 10dB fall means a factor 10 reduction. So, a level of -80dBm represents a power level of 10 picoWatts : 1000 times (30dB) lower than the -50dBm level.

    (Incidentally, this is why the apparently small difference between antennas with gains of 8dB and 11dB is actually quite significant : double the signal power.)

  • RSSI (Received Signal Strength Indication) is an indication of the “raw” signal power reaching your antenna system. This includes both the power in the signal you’re trying to utilise, as well as any other signals in the frequency band your receiver is tuned to. Some equipment manufacturers use an arbitrary scale to indicate RSSI, such as 0 to 100% or 0 to 255. Increasingly, however, manufacturers present RSSI in the “proper” unit - dBm as described above.
  • RSRP (Reference Signal Received Power) is an indication of the power in just the relevant parts (the Reference Signals) of the signal(s) you’re receiving. As such, it should provide a better measure of the true incident power of the signal you’re trying to utilise. Like RSSI, RSRP is best presented in dBm. Logically, the RSRP level can never exceed the RSSI level - in practise, it's significantly lower.
  • RSRQ (Reference Signal Received Quality) is derived from a quotient between RSRP and RSSI and measures the amount of “useful power” reaching your antenna, as a fraction of the total power reaching it. Higher values indicate that more of the total signal power you’re receiving is associated with the signal you’re trying to utilise. RSRQ is presented in deciBels.
  • RSSI-RSRP Because of how RSSI and RSRP are defined for LTE, the difference between RSSI and RSRP values (in deciBels) has a theoretical minimum value. This article explains how this minimum difference is around 20dB, and although it probably relates primarily to FDD systems (not TDD ones), the principle of a minimum difference probably applies to TDD as well. Practical measurements support this.



  • Presumably, only received signals in the same frequency band as that to which the receiver is tuned would increase your RSSI without producing an improvement in your own signal performance. This includes signals from adjacent cells on the same network, and other networks using the same frequency. At the moment, MTN, Vodacom and Neotel use FDD in the same LTE frequency band (1800MHz) whereas Telkom Mobile is alone at 2300MHz with TDD.
  • In the context of an LTE connection, “good performance” has a number of dimensions : stability, latency, downlink speed, uplink speed etc. For this exercise we will judge performance purely by downlink speed, which is of course a simplistic view. However, a connection that can consistently achieve a high downlink speed is likely to be good in other respects too.
  • For the moment, I've looked only at results achieved on Telkom Mobile LTE.
What do we expect to see on the scatter diagrams ?

  • A possible correlation between RSSI and speed : the higher the RSSI, the greater the possibility of high speed (neglecting the effect of spurious signals for now).
  • Perhaps a better correlation between RSRP and speed : the higher the RSRP, the greater the possibility of high connection speed.
  • A possible correlation between RSRQ and speed : a larger RSRQ would indicate that most of the signal power incident on your antenna system comes from the signal you’re trying to utilise, and not other irrelevant signals.
  • Mathematically, the deciBel difference between RSSI and RSRP is somewhat analogous to the quotient between those same two quantities, ie to RSRQ. Thus we might also expect to see the same correlation between RSSI-RSRP difference and speed, as we do between RSRQ and speed.
The scatter diagrams and source data are all contained inside the attached PDF document.

Just to keep things interesting amidst the sea of text, here's a teaser:




How to the read the scatter diagrams

  • In the first chart, the RSSI (in dBm) is plotted on the left axis versus the achieved downlink speed (in Mbps) on the horizontal axis. These are the green squares.
  • In the second chart, the RSRP (also in dBm) is plotted on the left axis versus the achieved downlink speed on the horizontal axis. These are the red squares.
  • In the third chart, the difference between RSSI and RSRP (RSSI-RSRP) is plotted on the left axis, versus the achieved downlink speed on the horizontal axis. These are the pink squares.
  • In looking at all 3 graphs, things get better/faster going from left to right, and from bottom to top. Signal and speed Nirvana is in the top right corner of each graph.
Practical value ranges

  • For RSSI :-85dBm (weak) to -45dBm (very strong).
  • For RSRP : -114dBm (weak) to -72dBm (very strong).
  • For RSSI-RSRP : 29 (bad) to 22 (good).
  • For downlink speed : 5Mbps (poor) to 90Mbps (excellent).
  • Note that 100Mbps is the theoretical maximum for Telkom’s current class of LTE equipment.
Cautionary notes

  • By now I think most of us realise that having good signal conditions is a prerequisite for, but not a guarantee of, high connection speed. The type of equipment at the base station, the type of connection negotiated, the capacity of the backhaul and the number of users connected, all play a role.
  • The signal conditions at the receiver could differ from one occasion to another. This is especially true when using built-in or indoor antennas; less so when using a stable, permanent external antenna system.
  • Even given that your signal conditions are good, the base-station’s “capacity to perform” could differ from one occasion to one another.
  • Practically, there are different ways to arrive at “good signal conditions”. If you live very close to a base-station, and the signal is good anyway, then using just your modem’s internal antennas may get you to a good operating point. Alternatively, if you live far away, you may find you need a high-gain directional external antenna system to get to the same operating point. The results presented are a mix of all situations : everything from internal antennas deployed right next to the base station; all the way up to high-gain directional antenna systems deployed further away.

  • On all three charts, you can notice the gradual slope from bottom left to top right : As signal conditions improve (from bottom to top), so the achievable speed (on average) increases (from left to right).
  • Downlink speeds seem to fall into at least two fairly distinct “groups”. Broadly speaking, the downlink speeds are either in the range of approx 4-40Mbps, or 50-80Mbps. There may be other sub-groupings. These groupings possibly reflect different base-station operating modes, different backhaul constraints, or different classes of LTE equipment, eg “Category 3" 100Mbps equipment versus some lower-speed class(es).
  • Installations achieving speeds above 20Mbps all have RSSI = -75dBm. However, having RSSI > -75dBm is no guarantee of getting high speed : there are cases with high RSSI values and poor speed.
  • Installations achieving speeds above 20Mbps have RSRP = -100dBm. Likewise, installations achieving >60Mbps have RSRP = -90dBm. But again, having RSRP above those levels is no guarantee. Likewise, there are also examples where a comparatively low RSRP (say, below -95dBm) still gets good speed > 50Mbps.
  • Installations achieving speeds above 50Mbps have RSSI-RSRP = 26dB.
  • Looking at the table of values : Above a value of about -8dB, RSRQ seems to provide almost no useful indication of likely link speed. Once RSSI & RSRP are above a very low minimum signal level, the RSRQ quickly seems to max out at -6dB. However we have seen other examples (not included in these results) where a very poor RSRQ (say, -9dB or lower) is associated with poor performance.
Possible conclusions?

(made very cautiously, bearing all the earlier disclaimers in mind):

  • If your antenna and signal conditions are such that:

    - Your RSSI is below -75dBm; or

    - Your RSRP is below -100dBm; or

    - Your RSRP is more than 27dB below your RSSI

    then ... you are probably not going to see more than about 20Mbps.

  • If all three parameters exceed the above criteria, then you have the possibility (not certainty) of achieving speeds anywhere up to about 80Mbps.
Many thanks to all the forum users who responded to my requests for their complete test results!
PDF attachment


PDF note: Tapatalk cannot parse links with the % sign in it. Please access from your browser



Source: MyBroadband forums


A friendly warning. The source forum has a lot of religious and other anti-social content so browse at your own risk.


Also see: An introduction to SA carriers

  • Like 6

Share this post

Link to post
Share on other sites

An enjoyable read. We have posted some similar articles and posts in the past. But this does a good job of wrapping up the issue from another perspective.


Robert via Samsung Note 8.0 using Tapatalk Pro

  • Like 1

Share this post

Link to post
Share on other sites

An enjoyable read. We have posted some similar articles and posts in the past. But this does a good job of wrapping up the issue from another perspective.


Robert, the lengthy quote includes an inline link to my RSRP article. It has gone around the world and back again. Ha ha!



  • Like 2

Share this post

Link to post
Share on other sites

Robert, the lengthy quote includes an inline link to my RSRP article. It has gone around the world and back again. Ha ha!



I didn't follow the link. But that's excellent!


Robert via Samsung Note 8.0 using Tapatalk Pro

Share this post

Link to post
Share on other sites

Robert, the lengthy quote includes an inline link to my RSRP article. It has gone around the world and back again. Ha ha!



I dropped that link on my home forum a little while ago as a standard resource for reference. I think the quality of the article speaks for itself. ;)
  • Like 1

Share this post

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • large.unreadcontent.png.6ef00db54e758d06

  • gallery_1_23_9202.png

  • Similar Content

    • By Paynefanbro
      I recently went on an 8 day cruise from NYC to the Caribbean that stopped in Turks and Caicos, Puerto Rico, and the Dominican Republic. My first stop was Grand Turk and there I opted for the free roaming. My S9+ automatically connected to Flow's (Cable & Wireless) LTE network where I received speeds of around 120kbps on average with boosts of up to 150kbps. Something worth noting is that on speed tests, the server prefers to default to Sprint's Miami server as opposed to local servers. Speeds were more than adequate for any amount of web browsing and honestly felt much faster than in reality. It helps that using Chrome will save you data by not loading pictures on certain sites unless you click them.
      In Puerto Rico, I connected to Band 13 on the way into the port in San Juan but once I was in the city, my phone never left Band 41. While the phone was usable, speeds remained significantly lower than what I've come to expect from 3xCA in the mainland U.S. Data speeds peaked at around 25-30Mbos but on average were in the 5-10 Mbps range even on LTE+. Signal remained strong everywhere though. 
      Finally in the Dominican Republic, I entered in Amber Cover which is in Puerto Plata. My phone latched onto a weak Band 2 LTE signal in the port from Altice (called Orange Dominicana in SignalCheck). I had trouble loading pages though. Once off of the ship and out in the open, I had a much stronger signal which allowed me to browse the internet without a hitch. Because it was the last day of my trip, while at the beach I decided to purchase the 24 hour high speed pass for $5. My speeds went from 120kbps to 65Mbps in less than 5 seconds. In some areas speeds were slower, particularly at the port where it struggled to break 2Mbps. Now, back on the boat my phone is flipping between weak Band 4 LTE and overloaded Band 5 HSPA+ from Claro (called Verizon Dominicana in SignalCheck Pro). Here is the difference in speed from before and after purchasing the high speed pass. 
    • By S4GRU
      by Jeff Foster
      Sprint 4G Rollout Updates
      Tuesday, January 31, 2012 - 7:46 PM MST
      Since last fall, there had been talk of a Samsung Galaxy Nexus launching on American carriers other than Big Red. Sprint has finally announced several weeks ago that it is the another vendor slated for release in the U.S. Suffice to say, many of us out there, especially those adverse to heading to Verizon and paying its premium prices, are excited about the impending release.
      The good news is that Google could be working on an updated version of the Galaxy Nexus. It has unofficially been dubbed the Galaxy Nexus Plus. There is much anticipation that it will be released before Sprint turns on LTE this summer. It’s not the first time an OEM has refreshed a device and re-released it to the market place, which works to our advantage. It’s rumored that the new Galaxy Nexus will have either a 1.5 or 1.8 GHz Texas Instrument OMAP4670 dual core processor. This would be a significant upgrade from the 1.2 GHz dual core processor found in the current Verizon version.
      We don’t know anything about official specs, but it’s also rumored to have an 8 MP camera. This is a noteworthy upgrade to the 5 MP shooter on the Verizon model (which has been lauded by many techies). We already know that the Sprint model will come installed with Google Wallet, per previous announcements. Some rumors also point to a beefier battery as well. The phone should have all the other features that’s on the current Galaxy Nexus, so now all we have to do is wait.
      Source: http://androidandme....era-on-the-way/
    • By S4GRU
      by Rick Layton
      Sprint 4G Rollout Updates
      Monday, June 25, 2012 - 4:27 PM MDT
      As technologies advance, the equipment to use the technology must advance as well. With the upcoming release of 4G LTE in our area (Houston), new equipment will be required to be able to use it. Although Sprint will have numerous data devices to handle the usage by the end of the year, only the Sprint Tri-Band Modem will be available at the rollout of the 4G LTE service.
      Due to the enormous dependence my business has on accessing data in a mobile environment, plus the great increases in data speed available with 4G LTE, this makes getting access to 4G LTE imperative to me. I depended heavily on the Sierra Wireless data devices when I started this business 7 years ago for my source of a reliable method of mobile data transmission. This relationship continued on until the release of the original Hotspot with the 4G service in my area.
      At one point, I was so displeased with past models, that I had sworn I would never buy another Sierra Wireless device as long as I live. This conclusion was reached after having numerous issues with previous hotspot models. There were so many problems that it seemed as if the device was never even tested on the networks it was to be used on. Also Sprint actively blocked reviews of the device, likely to not hinder sales in spite of the problems.
      My need for a new device with both WiMax and LTE capability outweighed my outright dislike of Sierra Wireless products. I proceeded against better judgment, and the Tri-Band modem was ordered even though the possibility of getting a substandard unit once again was always at the forefront of my mind.
      On with the show

      The official part number of the Tri-Band Modem is 803S. Along with the modem, I also ordered the SSX7077-V desktop cradle. I had to dig through a lot sites to find the information necessary to make this decision for my business. Much to my surprise, even though I was told the cradle was not available yet, I got a Sprint telesales person who was able to use the part number and find they had it in stock.
      Upon arrival I unpacked the unit and cradle...while holding my breath. The device that came out of the box was a pleasant departure from the previous Hotspots I had owned. Above is a picture of the device as it was shipped with all components. There was a small user guide as well but to get the real instructions the user guide must be downloaded from Sprint.
      Gone was the one piece blow molded plastic case which allowed no air circulation and caused the prior Hotspots to overheat quickly. Although the display is still too small for my aging eyes (it is actually the same display size as prior units) the change to the case makes it much easier to see in the interior of my van where the device will mostly be used.
      In this picture of the front you can see that there is a new button arrangement as compared to the older Hotspots. Also in the picture is the USB cable for use with the charger or to connect a computer, the AC to USB adapter, the battery and the battery cover. I opened the cradle, which was surprisingly inexpensive, and was delighted to find an additional AC to USB adapter which meant the cradle could be left in place without having to move the adapter around.
      As you look at the modem from the side you can see the antenna ports (the covers are open), the USB connector in the middle and the slot for the memory card. The round hole just right of the left antenna port is the reset button for the unit.

      Here is the same view with the battery and cover installed. Notice that the SD card slot is covered by the
      battery cover.

      The opposite side has two switches. The one on the left is a WPS setup button while the one on the right is a slider to mute the unit.

      The unit sits nicely in the cradle and looks to me to be a solution to help keep the USB port for the charger/interface cable from failing. This has been a major issue with the prior Hotspots. The case of the unit also helps support the USB port to take some of the load off of the circuit board.
      It took quite a bit of digging on the Sierra Wireless site to find out that the antenna ports are for the 4G WiMax band only. The cradle contains 2 5dbi omnidirectional antennas to allow full use of the WiMax network architecture.
      Initial testing

      The initial testing of the unit looks promising. The antennas in the cradle for 4G WiMax actually seem to get 3 – 5dBm gain in all conditions tested. The new unit has the ability to search the other bands for signals while staying connected. This allows less downtime between band changes. I notice a lot less disruption when switching bands.
      This unit has better reception on 3G and 4G WiMax than the previous hotspots and even the U600 USB modem I use as well. 4G WiMax is able to connect quickly even at 10% and the cradle has improved stability of WiMax and decreased ping times. For a short time I had access to Sprint 4G LTE as they were testing the towers in my area. The speeds were incredilbly faster. A 10% 4G LTE signal averaged 8.12Mbps download and 1.85Mbps upload. An 80% signal was able to get 35.8Mbps down on my best test and 22.1Mbps up.
      The upload speeds was very unexpected, and much higher than Sprint LTE smartphone devices have reported. This is likely due to much stronger transmit capabilities of the hotspot. I also discovered that when the modem is tethered the cable limits the bandwidth to approximately 20Mbps total speed. It will be interesting to see how it works in the 12 to 14 hour days of hot Houston Weather.
      First week in the field
      The Tri Band Modem got pressed into service a little quicker than planned, as my main unit went down with a bad transmission and the U600 USB modem with a Cradlepoint that was in this unit appears to have been damaged by the wrecker’s radio which runs on the edge of the WiMax frequency at 5 watts. The units have been sent in to determine cause of failure and for repairs but I think next time I will make sure all electronics are powered off before getting that close to a transmitter (OUCH!!).
      I am running the same routes in a rental van with the Tri-Band Modem that I normally use the other units on. There is less downtime in the signal gaps I am familiar with and areas where I have had signal problems in both 3G and 4G WiMax are much improved. I have yet to encounter any more 4G LTE signals but am looking forward to the service coming online soon. The unit seems to be running hotter than I would like with a fully charged battery but is actually cooler that the previous Hotspots. The temperature is supposed to soar over the next few days without the cloudiness we have had this past week. So it will be interesting to see if the overheating problems of previous models still occur.
      Week 2 – The True test
      The unit is getting worked really hard this week with temperatures outside up near 100 degrees. The GPS is useless with this kind of sun load as the unit will overheat if left in direct sunlight (as the instructions state) in about 20 minutes. The good news is that this is about twice as long as my original Hotspot will last. How anyone can make a unit that requires a clear view of the sky for GPS but can’t handle sunlight is beyond comprehension. A quick check of the Tri-Band’s temperature specs shows that the unit is only rated for 95 degrees. The prior Hotspot was rated well above the century mark but couldn’t even handle 90 degrees for any length of time. The crappiest laptop on the market will handle 105 degrees plus all day long. The true test will be my afternoon calls when the temperatures are high. Battery life has been about 8 to 9 hours which is far better than the prior Hotspots.
      The unit started overheating one afternoon. I can’t say I’m a bit surprised at that, but what is surprising is that it will run steadily as long as the air temp is below 98 degrees. This is a first for Hotspots as they always overheated well before the rated temperature spec. The bad news is the crappy overheat shutdown doesn’t turn off the unit before damage starts to occur, nor does it turn the unit off completely.
      Removing the battery cover seems to help air circulation and overheating some. The button lights are flickering after one overheating but the unit seems to be working fine other than this. It will be interesting to see what happens when it really gets hot here.
      According to the specs 4G LTE takes the least amount of wattage to run so it may not overheat as fast when using 4G LTE. I had the chance to try the modem in the old school 3G EVDO mode as one of my locations is 40 feet underground and that is all that is available at this location. I shut the unit down after 30 minutes as the unit was so hot you could barely handle it even though the temperature underground is around 70 degrees. I would not recommend trying to use this for any length of time if you want the Tri-Band to not overheat!!
      My Opinion
      Although Sierra Wireless has made some major improvement in the 3rd generation Hotspot, this is still a unit for the casual user. It is not designed to handle heavy use or outdoor summer temperatures for any length of time. It will be going in my climate controlled cabinet to protect it from the heat next week. I will let you know how it works when the temperature stays below 85 degrees. The improvements in connectivity, reception and stability are worth the investment. As long as you know and adjust your usage for the limitations of the unit.
    • By pyroscott
      Sprint Nextel revealed their second quarter 2012 corporate earnings in a conference call to their investors today and S4GRU was covering for news on Network Vision.
      Network thinning of the iDEN network is complete, taking 1/3 of Nextel towers off air. The Nextel network was built to support 20 million subscribers, but was only supporting 4.4 million subscribers, so it could easily be thinned without [much] noticeable change in street coverage. Sprint also converted 60% of the Nextel subscriber loss into their Sprint subscriber base. Interestingly, they stated that Verizon has been the biggest poacher of subscribers leaving Nextel, grabbing 50% of former subscribers in the last 4 1/2 years. In that same timeframe, Sprint has grabbed 25%, AT&T 20% and T-Mobile 5%.
      On the Network Vision topic:
      4 additional cities will launch, including Baltimore, by the end of August.*Edit* Cities were disclosed VIA press release following the conference call. They are:
      Baltimore, MD Gainesville, GA Manhattan/Junction City, KS Sherman-Denison, TX  
      Over 2,000 sites are currently online with 12,000 sites to be online by the end of the year
      Network Vision towers are seeing 10-20% additional voice minutes usage per tower, overnight after activating Network Vision. This will equal roaming savings for Sprint, and ESMR will only increase that savings.
      CEO Dan Hesse confirmed that Sprint will be releasing the Motorola Photon Q "in the very near future." It will be a QWERTY slider "with robust business and consumer features." It will also be sporting world phone capability.
      Several hundred Network Vision sites are waiting for backhaul, and will turn on when the backhaul is installed, several hundred more sites have birds nesting on them and Sprint won't be able to turn them on until the birds leave, according to the conference call.
      Sprint sold 1.5 million iPhones during the quarter, even though other carriers saw slowing of sales with rumors ramping up that the new iPhone would support LTE. 40% of the iPhone sales were to new customers. They also stated that iPhone customers require less customer support and are expected to churn less than customers on other phones.
      Mr. Hesse confirmed that Sprint is not looking to change plans in the near future.
      Things are looking up for Sprint. This quarter saw their highest ARPU and their lowest churn rate to date. They posted a larger loss than Q1, but beat their revenue goals for Q2. For more detailed financial information, check the source link below.
      Source: http://investors.spr...spx?iid=4057219
      http://finance.yahoo...-141200985.html -Thanks to S4GRU sponsor marioc21 for finding this link!
    • By lilotimz
      Ericsson RRUS31 B25 + RRUS11 B26
      These are the newest and greatest remote radio units to come from Ericsson. 

      The new Ericsson RRUS31  B25 should be fairly distinctive compared to the earlier RRUS11s and now the RRUS12s being deployed by ATT and Verizon. One of these new RRUS31s can do the job of two earlier RRUS11s thus reducing deployment costs for Sprint and complexity in deploying new sites and making it easier for users to spot as there are now 4 jumpers coming out of one RRUS31 rather than two from each RRUS11 that Ericsson originally deployed. 

      All future deployments will be utilizing the new Ericsson RRUS31s. In addition Ericsson are sending crews to their original deployments and swapping out older RRUS11s for these new RRUS31s due to the aforementioned fact that one RRUS31 can do the job of 2 RRUS11s. Weight savings will be significant at sites where there are 4 or 5 RRUS11 B25s that can be replaced by one or 2 RRUS31s. The Ericsson RRUS31 deployment project is known as the 65 Mhz Project. 


      Ericsson RRUS11 B26 top and RRUS31 B25 bottom


      Ericsson High Capacity / 4x4/2 MIMO Deployment
      Note the additional antenna + PCS radio.
      Previously Ericsson utilized additional PCS radios and used RF combiners for high capacity setups where they utilized three or more PCS radios. This new setup will utilize a completey new antenna + radio set just like Samsung and run 4x2 MIMO on the LTE antenna / radio set. 



      Ericsson RRUS11 B25 [EOL'd] and B26
      A standard Ericsson Network Vision 1.0 site with 3 RRUS11s where two are dedicated to PCS and one to SMR.  

      This type of setup is no longer deployed or utilized in new sites. Existing sites will be slowly converted to newer RRUS31 B25 via the Sprint 65 mhz project. 

      Ericsson NV high capacity site [EOL'd]
      3 or 4 PCS RRUs are present for a total of 4 or 5 RRUS11s per antenna. 



      Close up of Antennas

      Ericsson cabinets 

      All credit to those who took the photographs. They know who they are!
  • Posts

  • Recently Browsing

    No registered users viewing this page.

  • Create New...