LTE Speeds, signal, SNR

[bmoses]

Here is my scenario that I did some testing with and i need some help explaining what i'm seeing.

 

I have a tower that i connect to at work that i get OK speeds on. I was out checking for B41 installs just now and i went and checked this particular tower. While i was there i ran some speedtests on B25 from the same sector that i connect to at work.

 

I was consistently getting 22 to 24 meg down on B25 while near the tower. My signal was anywhere from -77 to -83 and my SNR was 22+.

 

when i got back to work i tested again outside in the parking lot (maybe 20 minutes later) and was consistently getting around 4 meg down. my signal there was around -93 to -95 and my SNR was around 10.

 

can someone explain to me what is causing an almost 20 meg drop in speeds? (i know user activity could cause it but every time i get close to this tower i get really high speeds and when i'm at work they are significantly lower, so lets just assume that user activity was not the cause). I know i obviously had worse signal and SNR but they were THAT much worse.

 

so what is causing such a drastic reduction in speeds? my first assumption would be the combination of lower signal and SNR but i wouldn't have thought it to have that much affect on speeds. i know its not apples to apples, but sitting inside my office at work on my verizon work phone i'm on Band 13 with a -108 signal and a 3.0 SNR and just got 6 meg down, so worse signal and SNR then my work B25 test yet higher speeds.

 

I was hoping that B41 on this tower would help take pressure off B25/B26 but based on these results i don't think thats really going to help at all.

[WiWavelength]

VZW band 13 is 10 MHz FDD -- twice the RF bandwidth.  That is definitely not an apples to apples comparison.

 

AJ

[gusherb]

Here is my scenario that I did some testing with and i need some help explaining what i'm seeing.

 

I have a tower that i connect to at work that i get OK speeds on. I was out checking for B41 installs just now and i went and checked this particular tower. While i was there i ran some speedtests on B25 from the same sector that i connect to at work.

 

I was consistently getting 22 to 24 meg down on B25 while near the tower. My signal was anywhere from -77 to -83 and my SNR was 22+.

 

when i got back to work i tested again outside in the parking lot (maybe 20 minutes later) and was consistently getting around 4 meg down. my signal there was around -93 to -95 and my SNR was around 10.

 

can someone explain to me what is causing an almost 20 meg drop in speeds? (i know user activity could cause it but every time i get close to this tower i get really high speeds and when i'm at work they are significantly lower, so lets just assume that user activity was not the cause). I know i obviously had worse signal and SNR but they were THAT much worse.

 

so what is causing such a drastic reduction in speeds? my first assumption would be the combination of lower signal and SNR but i wouldn't have thought it to have that much affect on speeds. i know its not apples to apples, but sitting inside my office at work on my verizon work phone i'm on Band 13 with a -108 signal and a 3.0 SNR and just got 6 meg down, so worse signal and SNR then my work B25 test yet higher speeds.

 

I was hoping that B41 on this tower would help take pressure off B25/B26 but based on these results i don't think thats really going to help at all.

Your experience sounds about right for the average 5x5 PCS LTE behavior. I don't see anything odd at all about that. As to why your speeds drop so much, only explanation I could come up with is traffic and interference (which obviously wasn't that bad going by an SNR of 10 so it must've been more because of traffic/users)

 

And yeah, Verizon is twice the bandwidth so in the perfect scenario where all parameters are equal, minus the doubled bandwidth, your speeds should be double that of Sprint B25 on VZW B13.

[bkco14]

I have noticed that at a stop light I would be getting a constant 12-13 meg down, but as soon as I turn the corner it drops to .8 - 1.1 meg. Also, about that tower, it appears the closer I get to it, the worse the speeds are.

[JossMan]

Remember, signal to noise ratio (SNR) is just as important as signal strength.

[Jeff68005]

What was explained to me was the closer to the tower, the larger piece of the pie was mine.  Given the size of the phone, I am not so sure that is true.   I know you have an awareness of some buildings and trees, etc in the way by the time you get to the office.

[bmoses]

thanks for the replies all.

 

i guess i didn't realize that B13 having double the bandwidth would mean double the speeds at certain signal/SNR levels. what i mean by that is, lets say you are at -110 signal and 10 SNR and you are the only one on the sector. based on that signal and SNR the fastest speed you can get on B25 is 5 meg down. having those same signal/SNR conditions on B13 means you can get 10 meg down. so that explains the better speeds on B13 i saw.

 

so bringing the focus back to the B25 scenario i had, is it really that normal to drop ~20 meg in speed with that kind of a drop in signal/SNR? its not like i went from a "perfect" signal down to -115 or worse, i was still at what many would consider a "good" signal. 

[lilotimz]

thanks for the replies all.

 

i guess i didn't realize that B13 having double the bandwidth would mean double the speeds at certain signal/SNR levels. what i mean by that is, lets say you are at -110 signal and 10 SNR and you are the only one on the sector. based on that signal and SNR the fastest speed you can get on B25 is 5 meg down. having those same signal/SNR conditions on B13 means you can get 10 meg down. so that explains the better speeds on B13 i saw.

 

so bringing the focus back to the B25 scenario i had, is it really that normal to drop ~20 meg in speed with that kind of a drop in signal/SNR? its not like i went from a "perfect" signal down to -115 or worse, i was still at what many would consider a "good" signal.

Sinr = signal to noise ratio. If the signal is noisy as hell you can have the strongest network ever but your experience still sucks.

 

Example. 0 sinr and - 80 rsrp. Perfect signal but you will not even get any data since the signal is being drowned out via interference.

 

Sent from my Nexus 5

[WiWavelength]

so bringing the focus back to the B25 scenario i had, is it really that normal to drop ~20 meg in speed with that kind of a drop in signal/SNR? its not like i went from a "perfect" signal down to -115 or worse, i was still at what many would consider a "good" signal. 

 

Too many variables.  We know nothing of the loading on the respective sites or the backhaul provisioning to those sites.

 

At some point, if mobile data transparently works, you just have to let the speed obsession go.  Who cares if you have 20 Mbps or 4 Mbps?  The experience difference for most mobile device usage is almost negligible.

 

And I will add my public service announcement that frequent speed tests unnecessarily load the network and slow down data speeds for everyone.

 

AJ

[bmoses]

Sinr = signal to noise ratio. If the signal is noisy as hell you can have the strongest network ever but your experience still sucks.

 

Sent from my Nexus 5

 

yes i understand that. what i'm trying to find out is was it signal/SNR that caused such a drastic speed drop or not? from what i have seen a 10 SNR and a -93 signal should still be considered a great signal. I would have though i would have still been able to pull ~15 meg down (a drop of 10 meg) at that level.

[lilotimz]

yes i understand that. what i'm trying to find out is was it signal/SNR that caused such a drastic speed drop or not? from what i have seen a 10 SNR and a -93 signal should still be considered a great signal. I would have though i would have still been able to pull ~15 meg down (a drop of 10 meg) at that level.

In perfect condition. Too many variables in any given environment to make a definite guess. Speeds are all theoretical In perfect lab conditions. Real life is always surprising.

 

Sent from my Nexus 5

[bmoses]

Too many variables.  We know nothing of the loading on the respective sites or the backhaul provisioning to those sites.

 

At some point, if mobile data transparently works, you just have to let the speed obsession go.  Who cares if you have 20 Mbps or 4 Mbps?  The experience difference for most mobile device usage is almost negligible.

 

And I will add my public service announcement that frequent speed tests unnecessarily load the network and slow down data speeds for everyone.

 

AJ

 

i'm not asking this because of an obsession of wanting the fastest speed. i'm trying to understand from a technical standpoint what caused that drastic of a speed drop. and yes there are a lot of variables and i am making the assumption that in the 20 minutes it took me to get from the tower back to work the load on that sector didn't increase enough to cause a ~20 meg drop in speed. so that led me to the question of can a signal/SNR drop of those levels really cause that much of a speed drop.

[WiWavelength]

and yes there are a lot of variables and i am making the assumption that in the 20 minutes it took me to get from the tower back to work the load on that sector didn't increase enough to cause a ~20 meg drop in speed.

 

You never know.  Loading fluctuates by the minute, even the second.  Because of multipath, signal, too, can vary in locations just feet apart.  All of these factors affect speeds.

 

That is why I caution against speed tests.  One speed test provides just a snapshot of one time and location -- that may never be repeatable at a different time and/or location.  By the time you have run that speed test, the moment it reflects is gone.  Only frequent speed tests at many locations give an overall perspective.  And therein lies the problem.

 

AJ

[Yuhfhrh]

i'm not asking this because of an obsession of wanting the fastest speed. i'm trying to understand from a technical standpoint what caused that drastic of a speed drop. and yes there are a lot of variables and i am making the assumption that in the 20 minutes it took me to get from the tower back to work the load on that sector didn't increase enough to cause a ~20 meg drop in speed. so that led me to the question of can a signal/SNR drop of those levels really cause that much of a speed drop.

 

Somebody could have also been running a speed test that same moment as you. There are too many variables.

[S4GRU]

Signal Quality is extremely important. If you have been ignoring that, you probably have experienced lots of unexplainable anomalies in your observations. Noise causes the signal to get drowned out. Causing delay and errors. The lower the signal quality, the slower the throughput, regardless of signal strength.

 

It's kind of like if your friend in the room is talking to you at 60dm loudness. If there is a lot of background noise, you may not hear all of what he says. You may even have to ask him to repeat some things, causing a slow down in your conversation. This is what happens with data with poor signal quality but good signal. Other things are causing interference. RSRQ and SNR give you some ideas of what's going on with signal quality beyond signal strength.

 

In general, the stronger the signal, the better the ability for the signal to get above the noise floor. Using a spectrum analyzer shows things like signal strength and noise floor. They help you to better see and analyze what's going on when you can see it on a graph. But I have seen really strong signals get covered by noise.

 

Also, wider carrier channels (10MHz+) handle noise and signal quality distortions better as they have far more subchannels to spread out across. Noise may not go across the whole channel uniformly. The narrower the channel, the more likelihood that the entire channel is affected.

 

Using Tapatalk on Nexus 6

[Flompholph]

i'm not asking this because of an obsession of wanting the fastest speed. i'm trying to understand from a technical standpoint what caused that drastic of a speed drop. and yes there are a lot of variables and i am making the assumption that in the 20 minutes it took me to get from the tower back to work the load on that sector didn't increase enough to cause a ~20 meg drop in speed. so that led me to the question of can a signal/SNR drop of those levels really cause that much of a speed drop.

You may want to search about qam snr and cqi. I haven't seen where there is a standard. It seems different oems use different code rates until they start to see errors then adjust. The better the snr the more bits/hz. 20 to 10 is about half so your going to have about 15 Mbps drop there. The graphs I seen showed ~20 snr as good enough for the highest code rate.

[Jeff68005]

Just a thought...would the number of devices in the area create or contribute to the SN signal noise interference?

[S4GRU]

Just a thought...would the number of devices in the area create or contribute to the SN signal noise interference?

Not on the downlink channel.

 

Using Nexus 6 on Tapatalk

[bmoses]

sorry to bump an older thread here. been doing some more testing and observing to this situation and i'm convinced that its some kind of signal/interference issue that causes slow speeds at my work. i'm reporting that in the sprint zone app so hopefully they take a look at it but i have a thought/question to run by the experts.

 

so in this hypothetical situation lets say person A is the only one the tower/sector so they have the full bandwidth available to them. but because of the signal strength and SNR the max speedtest you can get is 5mbps down. then person B comes along and is right next to person A. they are they only people on the tower/sector and have the exact same signal stats. Will person A and person B share that 5mbps download bandwidth? Or would they each get their own 5mbps slice?

 

hopefully that makes sense and the reason i'm asking is we basically have this situation at my work. whenever i get close to the tower that covers us i get 20 to 27 meg download speeds, but back at work its usually 3 to 5 meg down. we have 5 people at my work on sprint and i was curious to know if we were all fighting eachother for that 3 to 5 meg down or not?

 

thanks!