T-Mobile vs Sprint Signal Strength Question

[Ascertion]

Just wondering....I recently got myself a Nexus 4 and started a prepaid plan on it. I notice that my T-Mobile device always has better than a -80dBm which is really good on HSDPA+. Meanwhile, Sprint (who is on the same tower, although Sprint is on the higher row) yields me around -85 to -90dBm which is weak for CDMA. The panels appear to be facing the same direction (sprint and T-Mobile). Why would T-Mobile be yielding a much stronger signal strength than Sprint at my home?

 

Not dissing Sprint, just curious. Is it because T-Mobile is 1700mhz and Sprint is 1900?

[Mobilesolutions]

Not claiming to be an expert but i suspect T-mobile has the gain set high and increased the amplitude of the signal. I myself wonder how significant a 200 mhz difference is when it comes to RF penetration. And many other variables come into play as well you know!

[Ascertion]

Not claiming to be an expert but i suspect T-mobile has the gain set high and increased the amplitude of the signal. I myself wonder how significant a 200 mhz difference is when it comes to RF penetration. And many other variables come into play as well you know!

This is very informative. Will nv 3g/4g increase the amplitude?

[WiWavelength]

Ascertion, your post contains a couple of errors and misconceptions. First, RSSI of "-85 to -90dBm" is not weak for CDMA2000, not for CDMA1X nor EV-DO. Second, 1700 MHz has nothing to do with T-Mobile received signal strength because the AWS-1 band is 2100 MHz downlink, 1700 MHz uplink. Also, W-CDMA may tend to measure higher RSSI than CDMA2000 simply because the bandwidth is roughly three times greater, and that corresponds to a 5 dB difference.

 

AJ

[Mobilesolutions]

^^^ This is a good guy to have around ^^^

[WiWavelength]

Happy to be here...

 

 

AJ

[tbs123456]

Ascertion, your post contains a couple of errors and misconceptions. First, RSSI of "-85 to -90dBm" is not weak for CDMA2000, not for CDMA1X nor EV-DO. Second, 1700 MHz has nothing to do with T-Mobile received signal strength because the AWS-1 band is 2100 MHz downlink, 1700 MHz uplink. Also, W-CDMA may tend to measure higher RSSI than CDMA2000 simply because the bandwidth is roughly three times greater, and that corresponds to a 6 dB difference.

 

AJ

 

Does that mean W-CDMA is more efficient than cdma2000?

[WiWavelength]

Does that mean W-CDMA is more efficient than cdma2000?

 

Nope, signal strength and efficiency are two different things.

 

AJ

[tbs123456]

Sorry, I've been sick and am kind of out of it. Cdma uses 1.25mhz carriers while w-cdma uses 6mhz carriers, right? So w-cdma stands a better chance at penetrating a wall giving a better signal since more would get through, right?

[Deval]

Sorry, I've been sick and am kind of out of it. Cdma uses 1.25mhz carriers while w-cdma uses 6mhz carriers, right? So w-cdma stands a better chance at penetrating a wall giving a better signal since more would get through, right?

 

Actually if I'm not mistaken, the carrier bandwidth is for simultaneous users (capacity), not distance. That is more determined by frequency and power output.

[iansltx]

To put things another way/explain the RSSI difference, we're gonna need some calculus. Fortunately it's calc 1

 

RSSI is the integral of signal strength over the entire bandwidth of whatever you're receiving. Signal strength is measured in dBm, which is a logarithmic scale (2x the signal is 3db more, 4x is 6db more, 8x is 9db more, 10x is 10db more, 100x is 20db more). So,, all else equal, a 5MHz wide signal will have 4x (6db) more RSSI than a 1.25MHz wide signal.

 

As an aside, WCDMA channels are actually 3.84MHz wide, and CDMA 1x/Ev channels are slightly less than 1.25MHz wide if I remember correctly. However that doesn't modify our calculations here too much.

 

Getting back on topic, this is why RSRP is a better indicator of signal strength for LTE than RSSI; RSRP levels will be lower (because...AJ correct me if I'm wrong...it's the average of integrals over single 15MHz reference subcarriers rather than the integral of all subcarriers in the LTE channel) but you'll get a number that you can do something with without knowing whether your channel width was 1.4MHz, 5MHz or 20MHz.

 

Getting back to the comparison between WCDMA/HSPA+ and CDMA 1x/EvDO, T-Mobile's high-end (DC-)HSPA+ system uses 64QAM modulation in areas of good signal to pack more bits into each Hz of bandwidth (LTE actually uses that same modulation scheme, but with a few enhancements that I won't cover here). The catch with using a more complex, data-heavy modulation is that you need a significantly "hotter" signal to pass data at that rate without getting tons of errors. It's the same reason the WiFi on your computer drops connection speed (say, from 130 Mbps to 78 Mbps) as you get farther away from your router/access point.

 

CDMA and EvDO use simpler modulations to transfer their data, so you can get away with a much lower signal on those platforms and still maintain reliable service (hence why a number of rural carriers went with CDMA gear years back, I believe).

 

As for building penetration, it's harder for a wideband signal to make it through an obstruction intact than for a narrowband signal to do so, particularly if the narrowband transmitter can put the same output power behind that narrow signal as they did on the wide one (increasing RSSI). Particularly if that wideband signal carries a lot of data (high modulation). And, in terms of cell networks, WCDMA is wideband compared to pretty much everything else (remember that LTE is broken into narrow subcarriers).

 

As for frequency differences increasing or decreasing coverage, T-Mobile now has their HSPA+ network on two bands, depending on where you are: AWS (2100/1700) and PCS (1900). PCS is rarer, though some places are covered pretty well by it. At any rate, the downlink of AWS (tower to phone) is on the higher frequency piece of the spectrum, since the limiting factor for transmit power is your phone and not TMo's base station. This means that you're looking at a WEAKER signal, all else equal, FROM the tower, but a STRONGER signal back TO the tower. That said, T-Mobile's NV-esque recent rollouts (except that they already had heavy-duty backhaul to most of their urban sites) are pushing HSPA+ to PCS, decreasing the propagation difference and removing that difference when comparing them to Sprint.

 

One last thing: if you're comparing four EvDO carriers (5 MHz of bandwidth) with one HSPA+ carrier (for these purposes 5MHz of bandwidth) the HSPA+ carrier will win, given near-ideal signal conditions...HSPA+ is a much newer tech than EvDO, ans 21 > 12.4 (4 x 3.1). However I have yet to see a single HSPA+ channel net me more than 15 Mbps in real life, and I've seen EvDO hit 2.6 Mbps in the past. So you're really talking about 15 Mbps vs. 10.4 Mbps of capacity. The difference is that you can't get all of that 10.4 Mbps at once (unless your network is set up for EvDO Rev. B, but in the US it isn't). Hence the move to wider channels.

 

Any more questions?

 

P.S. If I sound like I'm down on either Sprint or T-Mobile, keep in mind that I'm the reason my entire immediate family's phone use Sprint (Ting for them, Sprint proper for me)...and I have a Nexus 4 running on T-Mobile.

[Keitarou]

To put things another way/explain the RSSI difference, we're gonna need some calculus. Fortunately it's calc 1

 

RSSI is the integral of signal strength over the entire bandwidth of whatever you're receiving. Signal strength is measured in dBm, which is a logarithmic scale (2x the signal is 3db more, 4x is 6db more, 8x is 9db more, 10x is 10db more, 100x is 20db more). So,, all else equal, a 5MHz wide signal will have 4x (6db) more RSSI than a 1.25MHz wide signal.

 

As an aside, WCDMA channels are actually 3.84MHz wide, and CDMA 1x/Ev channels are slightly less than 1.25MHz wide if I remember correctly. However that doesn't modify our calculations here too much.

 

Getting back on topic, this is why RSRP is a better indicator of signal strength for LTE than RSSI; RSRP levels will be lower (because...AJ correct me if I'm wrong...it's the average of integrals over single 15MHz reference subcarriers rather than the integral of all subcarriers in the LTE channel) but you'll get a number that you can do something with without knowing whether your channel width was 1.4MHz, 5MHz or 20MHz.

 

Getting back to the comparison between WCDMA/HSPA+ and CDMA 1x/EvDO, T-Mobile's high-end (DC-)HSPA+ system uses 64QAM modulation in areas of good signal to pack more bits into each Hz of bandwidth (LTE actually uses that same modulation scheme, but with a few enhancements that I won't cover here). The catch with using a more complex, data-heavy modulation is that you need a significantly "hotter" signal to pass data at that rate without getting tons of errors. It's the same reason the WiFi on your computer drops connection speed (say, from 130 Mbps to 78 Mbps) as you get farther away from your router/access point.

 

CDMA and EvDO use simpler modulations to transfer their data, so you can get away with a much lower signal on those platforms and still maintain reliable service (hence why a number of rural carriers went with CDMA gear years back, I believe).

 

As for building penetration, it's harder for a wideband signal to make it through an obstruction intact than for a narrowband signal to do so, particularly if the narrowband transmitter can put the same output power behind that narrow signal as they did on the wide one (increasing RSSI). Particularly if that wideband signal carries a lot of data (high modulation). And, in terms of cell networks, WCDMA is wideband compared to pretty much everything else (remember that LTE is broken into narrow subcarriers).

 

As for frequency differences increasing or decreasing coverage, T-Mobile now has their HSPA+ network on two bands, depending on where you are: AWS (2100/1700) and PCS (1900). PCS is rarer, though some places are covered pretty well by it. At any rate, the downlink of AWS (tower to phone) is on the higher frequency piece of the spectrum, since the limiting factor for transmit power is your phone and not TMo's base station. This means that you're looking at a WEAKER signal, all else equal, FROM the tower, but a STRONGER signal back TO the tower. That said, T-Mobile's NV-esque recent rollouts (except that they already had heavy-duty backhaul to most of their urban sites) are pushing HSPA+ to PCS, decreasing the propagation difference and removing that difference when comparing them to Sprint.

 

One last thing: if you're comparing four EvDO carriers (5 MHz of bandwidth) with one HSPA+ carrier (for these purposes 5MHz of bandwidth) the HSPA+ carrier will win, given near-ideal signal conditions...HSPA+ is a much newer tech than EvDO, ans 21 > 12.4 (4 x 3.1). However I have yet to see a single HSPA+ channel net me more than 15 Mbps in real life, and I've seen EvDO hit 2.6 Mbps in the past. So you're really talking about 15 Mbps vs. 10.4 Mbps of capacity. The difference is that you can't get all of that 10.4 Mbps at once (unless your network is set up for EvDO Rev. B, but in the US it isn't). Hence the move to wider channels.

 

Any more questions?

 

P.S. If I sound like I'm down on either Sprint or T-Mobile, keep in mind that I'm the reason my entire immediate family's phone use Sprint (Ting for them, Sprint proper for me)...and I have a Nexus 4 running on T-Mobile.

 

You lost me when you said calculus....

[WiWavelength]

As an aside, WCDMA channels are actually 3.84MHz wide, and CDMA 1x/Ev channels are slightly less than 1.25MHz wide if I remember correctly. However that doesn't modify our calculations here too much.

 

A W-CDMA carrier channel does utilize a 3.84 Mcps chipping rate, but the 95 percent power bandwidth is greater than 3.84 MHz because W-CDMA uses a more relaxed filter. So, 4.5 MHz is a good estimate.

 

Below is a fairly clean RF sweep (besides the multipath fading) that I captured of a T-Mobile AWS W-CDMA carrier. You can see the more rounded shape and wider skirts on the edges, extending roughly 4.5 MHz from noise floor at the low end to noise floor at the high end.

 

 

I will offer a few more thoughts later today...

 

AJ

[S4GRU]

 

I will offer a few more thoughts later today...

 

AJ

 

It's kind of like a teaser for the evening news during Prime Time. "Gun man goes on a shooting rampage at the Oak Park Mall, more at 10 o'clock..."

 

Robert via Nexus 7 with Tapatalk HD

[WiWavelength]

It's kind of like a teaser for the evening news during Prime Time. "Gun man goes on a shooting rampage at the Oak Park Mall, more at 10 o'clock..."

 

Hmm, good idea, I could use a teaser clip. Get David Koeller on that, pronto.

 

 

AJ

[WiWavelength]

Okay, I am back and have an exercise for Ian...

 

As for building penetration, it's harder for a wideband signal to make it through an obstruction intact than for a narrowband signal to do so, particularly if the narrowband transmitter can put the same output power behind that narrow signal as they did on the wide one (increasing RSSI).

 

I would make the argument or even assertion that -- all other factors equal -- W-CDMA voice can provide broader coverage and deeper building penetration than can CDMA1X voice or GSM voice. Why? The hint is GPS. How can I still get a GPS lock even through the roof of a structure when the satellites are fully 12,000 miles overhead?

 

AJ

[iansltx]

Didn't realize GPS was spread over a wide bandwidth. Silly me.

[tbs123456]

Thanks for the info! A lot of it is kind of over my head right now(asking questions to learn it though). Can I suggest maybe writing an article for the front page of what was explained here? Like using ascertions' original question as an example? Like providing all the info here, but some more "dumbed"(dont know what other word to use) down examples?

[WiWavelength]

Didn't realize GPS was spread over a wide bandwidth. Silly me.

 

I am not sure whether you are being sarcastic -- my detector is not always accurate -- but your answer is spot on. Low bit rate, high bandwidth spreading gain. To varying extent, the same holds true for both CDMA1X voice and W-CDMA voice.

 

And just to be clear, I respect your expertise. I believe that you have at least one engineering degree, while my higher education is mostly in the liberal arts. Any math, science, or engineering that I bring to the table is basically self taught and should be taken with a grain of salt.

 

So, I did not pose a question to bust your chops. I mostly wanted to break up what could become a long lecture into some smaller, more manageable, interactive questions that would be easier for all interested members to follow.

 

Now, maybe I am reading far too much into this. My empathy response is not that accurate, either. I might be a Replicant...

 

 

AJ

[iansltx]

I am not sure whether you are being sarcastic -- my detector is not always accurate -- but your answer is spot on. Low bit rate, high bandwidth spreading gain. To varying extent, the same holds true for both CDMA1X voice and W-CDMA voice.

 

And just to be clear, I respect your expertise. I believe that you have at least one engineering degree, while my higher education is mostly in the liberal arts. Any math, science, or engineering that I bring to the table is basically self taught and should be taken with a grain of salt.

 

So, I did not pose a question to bust your chops. I mostly wanted to break up what could become a long lecture into some smaller, more manageable, interactive questions that would be easier for all interested members to follow.

 

Now, maybe I am reading far too much into this. My empathy response is not that accurate, either. I might be a Replicant...

 

 

AJ

 

No offense taken. No sarcasm meant either. I actually composed that reply on my phone without fact-checking, but I guess my gut instinct was right I'll try and use <sarc> tags (why aren't they an HTML standard?!?) if I'm being completely ridiculous.

[WiWavelength]

So, all other factors equal -- frequency, power, modulation, coding -- could we agree that, say, an isolated ~8 kbps W-CDMA voice channel has a potential spreading gain advantage over an isolated ~8 kbps CDMA1X voice channel?

 

AJ

[iansltx]

So, all other factors equal -- frequency, power, modulation, coding -- could we agree that, say, an isolated ~8 kbps W-CDMA voice channel has a potential spreading gain advantage over an isolated ~8 kbps CDMA1X voice channel?

 

AJ

 

Yep, we could.

[Ascertion]

Glad I made this thread. Was very informative. Thanks s4gru community for being so knowledgeable.

[SpenceSouth]

Glad I made this thread. Was very informative. Thanks s4gru community for being so knowledgeable.

 

So many knowledgeable people here. We just have to know what to ask!

[WiWavelength]

Glad I made this thread. Was very informative. Thanks s4gru community for being so knowledgeable.

 

Hey, there is still more to come. I will have that for you in the morning. Now, back to David with weather...

 

 

AJ