Broadband speeds on 3

I live near a 3 4G mast and get a strong signal.
With a 3 SIM in my Samsung Note 20 I get typical internet speeds of 60Mb down and 15Mb up.
If I put the same SIM into a 4G USB dongle I get typical speeds of 2Mb down and 15Mb up.

This happens the same with D-link and Huawei dongles, plugged into Windows and Mac devices.
Is this due to 3 controlling how I use the SIM? I pay for a fixed monthly data allowance (not unlimited) and I don't see why they should bother how I use it.

The phone will be running a later, higher performance radio chipset, enabling it to achieve better speeds/ throughputs than the dongles.

For a brief explanation of 4G/LTE radio categories see:
https://www.cablefree.net/wirelesstechnology/4glte/l...

Post deleted by RobertoS

+1

I have Three SIMs in my OnePlus 8 Pro mobile and a Three B311 router.

See my sig for the speeds. Cheap router, two years old. Expensive phone. Previous phone was a Huawei P10 Lite, Not as fast as the OnePlus.

The internals and antennae of the kit on the market are advancing month by month.

I certainly wouldn't be happy with a dongle for any significant use.

(My apologies for the RobertoS post I deleted. I changed my nick/username today. At Preview stage I realised there was nothing in my sig with the new one, and after logging in to copy the relevant part from the RobertoS sig I forgot to swap back to this one)

In reply to a post by RogerGrey:

This happens the same with D-link and Huawei dongles, plugged into Windows and Mac devices.

USB dongles are old technology. Hardly any manufacturer makes them, and they are pretty much all LTE category 4 or earlier. This means can only connect to one 20 MHz frequency channel at at a time, and can only manage the older modes. Your phone is probably LTE Category 18, also with support for 5G networks.

The USB dongle is like driving on an empty motorway with a 1960s reliant robin, and your phone is like driving a brand new Mercedes C class.

You can buy external LTE modems of high category, but they are not cheap. Example the Netgear M1 on Amazon for £279.

It might be easier/cheaper to tether your mobile?

Thanks all for the info. What is strange is that it's only the downlink that is slower on the dongle. The uplink speed is the same.
This was a test to see if it was worth getting a separate contract for the dongle.

I'll do some tests with EE and O2 SIMS to see if the effect is the same - will report back,

I've done some tests with EE and 02 SIMs in the same dongles and the same PC and MAC.
Allowing for the usual vagaries of speed with mobile networks, both those networks are giving very roughly the same speeds up and down.

3 gives faster upload speeds that the others but much lower downloads. It's odd that 3's upload speeds are the same on dongle as on phone, but the downloads are so very much lower.

I have looked at the Netgear M1 and similar devices but they are really over the top. There's the battery to be charged and charger to be carried, and you're paying for the WiFi facility. With the USB dongle it's very much plug-and-play and nothing else to be carried.

So I will stick with the tethering, but I would like to understand the technical reason why 3 has this huge discrepancy,

In reply to a post by RogerGrey:

I've done some tests with EE and 02 SIMs in the same dongles and the same PC and MAC. Allowing for the usual vagaries of speed with mobile networks, both those networks are giving very roughly the same speeds up and down.

That is not surprising as the limitation will be the dongle LTE category. (Most are Category 4). Depending on where you are in the country you could be receiving only 10 MHz of LTE from O2 on Band 20 (800 MHz) and 40 MHz of LTE from EE on two allocations of Band 3 (1800 MHz). Your dongle could only use one of those at time, your phone can bond these transmissions together (known as CA carrier aggregation)

3 gives faster upload speeds that the others but much lower downloads. It's odd that 3's upload speeds are the same on dongle as on phone, but the downloads are so very much lower.

Often that is an indicator of load. Probably lots of people like yourself taken in my 3's marketing that they are the "best network for data" so using it heavily.

I have looked at the Netgear M1 and similar devices but they are really over the top. There's the battery to be charged and charger to be carried, and you're paying for the WiFi facility. With the USB dongle it's very much plug-and-play and nothing else to be carried.

I completely agree, but if you want something of the performance of a Note 20 or iPhone 12 then you need to pay that sort of money.

Your Note phone is capable of using multiple bands, Band 20, Band 3, Band 7, and multiple transmissions on each band, and joining them together. Speeds of 400 Mbps or faster are possible on EE and close to that on Vodafone in the right areas. O2 can't really reach those speeds in most of the country as they haven't deployed enough of their 2300 MHz spectrum. For 3 to reach those speeds in many cases you will need 5G due to the number of customers using.

So I will stick with the tethering, but I would like to understand the technical reason why 3 has this huge discrepancy,

It is all guess work as the radio signal spreads far and wide. If you want predictability use fixed line broadband, not a mobile network.

Quite easily 3 could sell 100 SIMs to people in your coverage area whom all go home and start streaming 4K video through their phones to the TV. This would impact your download speeds, and nothing you could do about it.

I’m getting the opposite on Three. Downloads are fine, uploads are really poor.

Just for clarity, do you mean the speeds when using your phone are lower than when using tethering, or just that download speeds are poor in either scenario?

I’ve seen recently tethering being noticeably quicker than using the phone. Unknown reason why this would happen and suggests Three could be prioritising certain traffic based on the device and it’s network indicators rather than just the APN as the networks can detect tethering traffic as being different to BAU phone traffic.

Edited by gary333 (Tue 01-Jun-21 20:13:28)

In reply to a post by gary333:

I’ve seen recently tethering being noticeably quicker than using the phone. Unknown reason why this would happen and suggests Three could be prioritising certain traffic based on the device and it’s network indicators rather than just the APN as the networks can detect tethering traffic as being different to BAU phone traffic.

That's odd, usually the reverse caused by WiFi congestion between the phone and the computer. Some phones support USB 3 tethering to rule this out. A VPN would show if Three was prioritising traffic. They shouldn't be as I gather they gave assurances to Ofcom they would not.

However if you are on a Smarty or other virtual operator using the 3 infrastructure, they do not have to be the same as 3.

In reply to a post by jchamier:

In reply to a post by gary333:

I’ve seen recently tethering being noticeably quicker than using the phone. Unknown reason why this would happen and suggests Three could be prioritising certain traffic based on the device and it’s network indicators rather than just the APN as the networks can detect tethering traffic as being different to BAU phone traffic.

That's odd, usually the reverse caused by WiFi congestion between the phone and the computer. Some phones support USB 3 tethering to rule this out. A VPN would show if Three was prioritising traffic. They shouldn't be as I gather they gave assurances to Ofcom they would not.

However if you are on a Smarty or other virtual operator using the 3 infrastructure, they do not have to be the same as 3.

I agree it's certainly odd behaviour. It could potentially be a network issue where I am as it seems to be flicking back to 3G at multiple intervals in the day, even though I get 2-3 bars 4G if I turn celluar off and on. The above comments were based on tests performed back to back with 3 bars of 4G. I'm seeing it now with 6mbps on phone, 22mpbs on PC tethering to Kcom in Hull (a server I find reliable for speedtest.net).

3's network page doesn't show any issues - but we all know that it's about as useful as a chocolate tea pot.

I also wondered whether the tethering was triggering something on the network to enable CA, where as for some unknown reason the traffic from the phone directly wasn't. Who knows as when it comes to Three all i know is inconsistency is their middle name. Unfortunately O2 and Vodafone in my area are far inferior, however I no longer use 4G service for my internet so it doesn't effect my day to day experience - i just like to play around.

Edited by gary333 (Tue 01-Jun-21 20:45:12)

If you have an android phone you can run cellmapper, and it will show the cell site number for the 4G LTE transmission (eNodeB) and you should see the band you’re connected to. You can then look up that eNodeB on the cellmapper.net website, and you might find out what the mast is transmitting.

Three could be using 5 MHz of Band 20 (800 MHz) and 15 MHz of Band 3 (1800 MHz). You might be lucky and also have some Band 32 (1500 MHz) or some Band 1 transferred to LTE (2100 MHz).

Many phones only use secondary carriers when data is being used, you can sometimes see this in a speed test in steps of increasing speed. Depending on the phone it is possible the device is enabling CA for tethering.

There is also a difference in tethering for upload and download, not all masts and handsets support download CA, and it may also depend on the quality of signal from your handset back to the mast.

So many variables

In reply to a post by jchamier:

If you have an android phone you can run cellmapper, and it will show the cell site number for the 4G LTE transmission (eNodeB) and you should see the band you’re connected to. You can then look up that eNodeB on the cellmapper.net website, and you might find out what the mast is transmitting.

Three could be using 5 MHz of Band 20 (800 MHz) and 15 MHz of Band 3 (1800 MHz). You might be lucky and also have some Band 32 (1500 MHz) or some Band 1 transferred to LTE (2100 MHz).

Many phones only use secondary carriers when data is being used, you can sometimes see this in a speed test in steps of increasing speed. Depending on the phone it is possible the device is enabling CA for tethering.

There is also a difference in tethering for upload and download, not all masts and handsets support download CA, and it may also depend on the quality of signal from your handset back to the mast.

So many variables

I don't have an Android phone, however I can see from the engineer menu within my iPhone X what is going on. The phone is connected to the cell that i expected it to be, which I know has bands 1,3 & 20 but no band 32.

CA is enabled when the phone is tethering and I perform a speedtest. As expected CA is deactivated when no/low data is being used. It's interesting to note (as I don't recall seeing this anywhere else) but the phone is setup to use 3internet APN within the mobile data settings, however when connected by tethering the actual APN it's using is 3hotspot as reported in the engineer menu. The reason this is interesting as i have seen a number of people report the 3internet APN has been flakey recently - thus this might explain the differences in speeds.

I can see the lead band the phone is connected to is band 20 with the phone reporting 5mhz of bandwidth (dl_bw 25) and the CA band being band 3 with 15mhz of bandwidth reported (dl_bandwidth).

Edited by gary333 (Tue 01-Jun-21 22:11:31)

Ugh, different APN for tethering. And iOS won’t let you override that. Engineering mode doesn’t work properly on my 12, guessing the new modem.

Doesn’t surprise me Three are the ones doing this. I’ve not currently got a 3 contract SIM and my PAYG is in my OnePlus. My EE and Vodafone are on iOS and I see the same APN for tethering.

Apple gives the networks the tools to hang themselves!?

Edited by jchamier (Tue 01-Jun-21 22:30:23)

I haven't tested the speeds when tethering, but just from observation they seem to be as fast as using the phone on its own.

I've been using a VPN for all the tests, to avoid any possible selective effects.

I've only used one (free) VPN and it doesn't multistream. So speedtest.net and tbb give poor results.