A few weeks ago, two US network operators have launched 5G in the Chicago area, just when I happened to be there for a week. While I don’t have a 5G device that supports US deployments yet, I expected that, perhaps as a side effect, there would also be a significant LTE deployment. However, I was rather underwhelmed about the data rates I got. While I staid in Evanston, about 15 km North of Chicago, I could only get 5-15 Mbit/s from one network operator in most places I measured while I could at least get 10-25 Mbit/s from another network operator in that city. As that is very little compared to data rates I get when I travel to other parts of the world I started to investigate a bit.
This post focuses on the operator from which I could only get 5-15 Mbit/s on LTE. Let’s call the company operator 1, this is not about naming and shaming here. According to Cell Mapper, I was only about 350 meters away from the serving cell site (I compared the PCI from the web page with the PCI in my trace tool, so I’m certain) and around 600 meters away from the cell site in the other direction. So it’s a clear serving cell scenario.
The first reason I only got 5-15 Mbit/s on the 5th floor of the hotel I staid with only trees but no buildings between me and the cell site was, that the signal strength was well below -105 dbm. Very low for that distance, so perhaps I was at the border of two sectors of the cell site. The cell has served me on eUTRAN band 4 on 1.7 GHz with a channel bandwidth of 15 MHz. I’ve also observed other channels that my device actively used in the area, namely 10 MHz on band 2 (1.9 GHz), 5 MHz on Band 12 (700 MHz) and 5 MHz on band 66 (extended band 4). These bands and their bandwidths are also listed in SIB5 broadcast messages by the cell so that made sense.
When adding all these channels together, operator 1 has deployed 15+10+5+5 = 35 MHz of LTE in the area. Wow, 5 MHz channels, I can’t remember having seen such narrow channels in Europe anywhere. But nevertheless, why did I only get such ‘low’ speeds despite 35 MHz on air? A closer look revealed that the network never attempted to aggregate the carriers. Now that was rather strange so I had a look at which capabilities my high end smartphone made for Europe and Asia (but not US) with all bells and whistles reports to the network. Again, no naming and shaming here, so I will keep the maker and model to myself.
During the attach procedure the network requests a report for the support of the following bands with those in bold actually being on air:
band 2, 1.9 GHz, 10 MHz observed
band 4, 2.1/1.7 GHz, 15 MHz observed
band 5, 850 MHz, NOT observed
band 12, 700 MHz, 5 MHz observed
band 46, TD unlicensed 5 GHz, NOT observed
band 66, Extended AWS, 2.1/1.7 GHz, –> extended band 4
band 71, 600 MHz Digital Dividend US
An interesting side note: The base station wants to know if the 600 MHz band is supported by the device and even more interesting, if the device supports the 5 GHz unlicensed band for LTE! Furthermore the network asks for a report of carrier aggregation combinations in the downlink and uplink:
The UE then reports support for quite a number of bands, including band 2,4,5,12 and 66. In other words, it supported all bands that are used by operator 1 in the area. However, and that was the big surprise to me, it did not report even a single carrier aggregation combination for any of those bands. ZERO, ZIP, NADA!? So even if the network wanted to use carrier aggregation, it could not. I’m not sure at this point if this is a bug in the European modem firmware of my device or if it is really not capable to aggregate carriers in those bands. Something to be followed up.
Overall, I see three main reasons for the low speeds I experienced: First, there are only 35 MHz of spectrum deployed, which, at least from a European point of view is not much, as operators there have typically deployed 50-60 MHz of spectrum in urban areas and base stations are also much closer together than the inter-site distance of 800 to 1000 meters I experienced in Evanston. Secondly, there was quite a bit of traffic on the network, I could usually not get the maximum number of downlink resource blocks assigned. One can’t blame the operator for that directly, though. And finally, my device did not offer the network the option to aggregate carriers. With that feature, my data rates would have at least been double those experienced, which really shows how much this feature is needed these days.
4 thoughts on “A Look At LTE Deployment North of Chicago – Part 1”
Hi Martin , there are quite a few observations here that needs mroe understanding.
Devices ,from another operator , have specific combinations that can be aggregated even if they support several bands individually.
Your device limits up to 3CA in DL
but You need to look at UE capability message to see, which 3 bands are allowed to aggregate.
You mentioned that you signal on Band 4 was -105 dBm, what was your Signal-noise-Ratio (SNR) – that impacts DL speed more than RSRP.
If Band 4 was -105, you must have seen Band 12, but if your device did not allow the combination of B4 + B12 , then it wont attempt to aggregate.
>Devices ,from another operator , have specific combinations that can be aggregated even if they support several bands individually.
Yes, my device is optimized for EU and Asia CA combinations and none were reported for the US frequencies. A pity. Yes, -105 dBm for band 4 was weak, another reason why the connection was relatively slow. But non-availability of CA combinations were the main reason.
Which phone (model etc) did you use to get no CA? Also which tmous US mnc? Judging by “requestedMaxCCsDL-r13: 3” in perfect UE condition it would have been max 3CA anyway.
There are alot of carrier policy rules in phones for band and ca supported. Not always are those rules best optimized for the network connected.
Don’t want to discuss phone model. And it is as you say and as I said in my post: The UE is optimized for EU and Asia and does not report any CA combinations for the US. A bit of a pity.
Comments are closed.