Rogers vs Bell LTE test

[howard]

I remember a years ago when 3G was just hype. I remember how cool it was when I could connect at 57.6kb/s on my Motorola P280 which I could then use to get my laptop online. Years later, 3G is the now norm. In Canada Bell, TELUS and Rogers all have DC-HSPA+ networks that can connect at up to 42mbps. I should point out that 21 and 42mbps DC-HSPA+ is frequently marketed as 4G service even though it’s not really. I could discuss why DC-HSPA+ is and isn’t 4G service but that is well beyond the scope of this discussion.

Now Bell and Rogers have released their true 4G networks which use Long Term Evolution better known as LTE. It eventually promises speeds of up to 150mbps though the version Rogers and Bell are using max out at 75mbps.

So it got me thinking: Who LTE network is faster? Bell or Rogers?

I decided to test Bell and Rogers head to head by taking 2 laptops, LTE modems from Rogers and Bell and using them in a real world setting to see which could download more data in the same amount of time. The real world setting I’m describing is sitting in the back of our very own HC - NO "i"'s car and having him drive around the GTA.

I’ll include more specific details of how I tested, where exactly I tested including the path we took along with issues and pitfalls that arose during the test later.

I split the test into 4 segments: Downtown, East York, 401 and the 427.

The Downtown segment of the test was extremely close. In about 75 minutes I managed to download 12.37GB with Rogers vs 12.33GB on Bell. Yes, that’s GB as in gigabytes. To put this number in perspective I averaged 23.3mbps on Rogers vs 22.6mbps on Bell. It was so close I had to include an extra decimal space. We’ll call this segment of the test a wash. Rogers 1 - Bell 1.

Next up was East York. In 22 minutes I downloaded 3.1GB on Bell vs 6GB on Rogers. That’s 20.5mbps on Bell and a blistering 38.8mbps on Rogers. Rogers 2 - Bell 1.

When we were on the 401 for 45 minutes where I got 7.1GB on Bell and 9.7GB on Rogers. In mbps that’s 30.5 on Rogers and 22.5 Bell. Rogers 3 - Bell 1.

Traffic thinned out on the 427 segment of the test so it only lasted about 6.5 minutes. Bell managed 1.4GB/29.3mbps while Rogers moved 1.3GB/28.8mbps. Rogers 3 - Bell 2.



Honestly, it blew me away how fast these networks are. Between the 2 of them I managed to move over 50 GIGABYTES of data in just under 2.5hrs.



Still, at the end of the day I managed to transfer 29.4GB on Rogers vs 24GB on Bell. That’s an average of 28mbps on Rogers vs 22.6mbps on Bell. I also observed speeds of up to 55mbps on Bell and 70mbps on Rogers. Please read the 'How I tested' section for notes on the maximum network speeds.



How I tested:

To test I actually used 3 modems. On Rogers I used a Sierra Wireless 313U. With Bell I used the same model modem (but with Bell firmware/software instead of Rogers) along with a Novatel U679. The Rogers modem was hooked up to my Lenovo Thinkpad T510 which stuck an Intel 80GB SSD. For the Bell modems Samsung lent me their QX412 laptop computer which also has a 128GB SSD. I chose to use fairly computers with SSD’s because it’s bumpy in the car and I didn’t want to risk losing any data. Both computers were running Windows 7.



I used Firefox to download linux ISO files from 8 different sources. The idea was to max out my connection. When downloading at these speeds it’s important to keep an eye on the download progress because they finish pretty quickly.

Here’s the path we took:

started just west of the cne on lakeshore heading eastbound towards liberty village
rogers disconnected
nb on strachan
wb east liberty st
nb fraser
eb king st
king and strachan
king and winder
king and simcoe
king and bay st
king and jarvis
sb parliment
parliment and mill st
wb lakeshore
sb yonge
yonge and lakeshore
wb queens quay
no rogers LTE on queens quay
queens quay and lower spadina
rogers back on LTE
nb bathurst
bathurst and richmond
bathurst and nassau
eb on bloor st
bloor and borden st
bloor and st thomas
bloor and bay st
bloor and yonge
bloor and ted rogers way
bridge between bloor and danforth

East York portion of test begins

dandforth eldon ave finally have bell lte again
danforth and victoria park
bell no signal again at birtchmount
bell LTE again danforth merged to kingston rd
now rogers no signal, had to remove stick and plug back in to get lte
midland and kingston
kingston and parkcrest
kingston and galloway
kingston and lawrence no bell lte

401 portion begins

wb on 401 now
just past kennedy bell switch to lte
just past warrden bell back to hspa
back on bell lte afer unplugging and plugging modem back in
crossing under don mills on 401
401 and leslie
401 and yonge
401 and avenue
bell back on hspa


427 portion begins

bell back on lte, 427 and rathburn
427 sb to qew
rogers ends at 401 and dixie

End of test


Bell LTE coverage map

If you look carefully on Bell’s LTE coverage map there’s a hole on the Danforth just east of Greektown. Rogers actually had LTE the entire time in that area but since Bell only had HSPA I didn’t record the results. Now some may think that I’m unfairly punishing Rogers for Bell’s lack of coverage but I had a good reason for doing this. Before I ran this test I did many test runs and found that the networks didn’t like to go from HSPA to LTE if I was transferring data. Since I was transferring continuously Bell would never go back onto LTE.


Rogers LTE coverage map

Traveling westbound on the 401 you’ll notice that Bell doesn’t have any coverage west of Hog’s hollow (between Avenue Rd and Yonge St).

For the route we took Rogers had coverage pretty much the entire time until we were heading westbound on the QEW heading towards Mississauga. There Roger’s coverage ends about 1 or 2 highway stops before Bell does.

The Rogers modem disconnected completely from the network 3 times during the test though I was able to get back on quickly. So it just might be a result of a bad handover.

One feature I liked on the Rogers modem is the Rogers software let’s you tell it to only use LTE. While the Bell has the same modem their software lacks this setting. I didn’t use the Rogers setting.

I noticed that Bell appears to be have some sort of rate limiting feature on their network. A couple of times I noticed the connection would sit at around 40mbps for up to a few minutes at a time. That said I did observe speeds as high as 55mbps on Bell. Rogers by comparison doesn't appear to have any sort of rate limitation. I got speeds of up to 70mbps.

[Treatz]

Nice speeds
Now if the LTE pricing comes down I can replace my bell dsl connection

[frankie5string]

I'm not exactly sure about the Hogs Hollow comment... Somethings not right there.

I live in that area (west of Avenue Road) and have tried Bell LTE here... it does indeed work here. Although I wasn't travelling on the highway, I was stationary at street level just east of the 401 overpass.

Did you guys actually get off those exits or did you just buzz by? Sounds to me like Bell is pointing the LTE network down into the streets, not along the highway.

[Dee-Dub]

One feature I liked on the Rogers modem is the Rogers software let’s you tell it to only use LTE. While the Bell has the same modem their software lacks this setting. I didn’t use the Rogers setting.

while disconnected from the network but with the turbo stick plugged in: tools - options - preferences - hardware - modify - lte only under bearer selection

[howard]

while disconnected from the network but with the turbo stick plugged in: tools - options - preferences - hardware - modify - lte only under bearer selection

Thanks for the tip. It never occurred to me look for the setting when the modem was disconnected from the network.

[HC - NO "i"]

I'm not exactly sure about the Hogs Hollow comment... Somethings not right there.

I live in that area (west of Avenue Road) and have tried Bell LTE here... it does indeed work here. Although I wasn't travelling on the highway, I was stationary at street level just east of the 401 overpass.

Did you guys actually get off those exits or did you just buzz by? Sounds to me like Bell is pointing the LTE network down into the streets, not along the highway.

frankie5string,

We were on the 401 during the rush hours and it is not moving fast at all between 404/DVP and Kipling. It was frequent stop and go in the rain. I also have pretty good idea where all the sites are. So it was not a fluke and actually not quite in line with what Bell is telling about its LTE footprint. But it is okay as HSPA+ still around.

[frankie5string]

frankie5string,

We were on the 401 during the rush hours and it is not moving fast at all between 404/DVP and Kipling. It was frequent stop and go in the rain. I also have pretty good idea where all the sites are. So it was not a fluke and actually not quite in line with what Bell is telling about its LTE footprint. But it is okay as HSPA+ still around.

Well well, looks like I never even got notified that this reply got posted!

I still think you're missing something by just sticking to the highway. I'm a little confused how you can judge coverage of an area based on what you see from the highway. Especially since the LTE sites are pointed mostly towards the subdivisions up here. That would explain why you didn't get very good coverage around the area. Look at the giant tower at 401 and Avenue road - you can count the number of sectors pointing out to the highway compared to those pointing down to the area... the ratio is something like 1:5...

I agree about the coverage... Also, I found LTE to be different at 9am than at 3pm up here. It's kinda like how the Rogers voice network falls apart up here around 4pm... it seemed predictable.

[HC - NO "i"]

frankie5string,

Like I have said, we have pretty good idea of where the sites and sectors are. The point of "staying" on the highway" is NOT performing a industrial drive test by missing the other strategic sites / LTE footprint i.e. Grid by Grid with 50% of all the accessible roads, data gathered by Agilent, R&S analyzers. The point is ALL about driving "2 Howards" from Pickering to Oakville, so he could go home and I could go back on highway again for supper :-p

We were actually observing the downtown core relatively more than the other areas.

[frankie5string]

frankie5string,

Like I have said, we have pretty good idea of where the sites and sectors are. The point of "staying" on the highway" is NOT performing a industrial drive test by missing the other strategic sites / LTE footprint i.e. Grid by Grid with 50% of all the accessible roads, data gathered by Agilent, R&S analyzers. The point is ALL about driving "2 Howards" from Pickering to Oakville, so he could go home and I could go back on highway again for supper :-p

We were actually observing the downtown core relatively more than the other areas.

Ok fair enough...you were just doing a sort of drive-by sample for this. People tend to take the results you and Howard fairly seriously - so when you perform tests like this and say "No LTE service was found in this place" you should qualify it by saying you were just jogging by the area on the highway and didn't actually stop in the area to qualify.

That's all... Otherwise, you guys did pretty good...

[acurrie]

I used Firefox to download linux ISO files from 8 different sources.

So the really important question: Which distros did you get?