Every iPhone X Is Not Created Equal (pcmag.com)

According to a PC Magazine report that uses data from Cellular Insights, the Qualcomm-powered iPhone X has better LTE performance than the Intel-powered model. From the report: There are three iPhone X models sold globally. Using lab equipment, Cellular Insights tested two of them: the Qualcomm-powered A1865, sold by Sprint, Verizon, and U.S. Cellular and in Australia, China, and India; and the Intel-powered A1901, sold by most other global carriers including AT&T and T-Mobile. (The third model, A1902, is only sold in Japan.) Here in the U.S., we anticipate that the SIM-free model sold directly by Apple will be the A1865, as that's the model that supports all four U.S. carriers. For this test, Cellular Insights looked at performance on LTE Band 4, which is used by every major U.S. carrier except Sprint, as well as in Canada and parts of Latin America. Cellular Insights attenuated an LTE signal from a strong -85dBm until the modems showed no performance. While both modems started out with 195Mbps of download throughput on a 20MHz carrier, the Qualcomm difference appeared quickly, as the Intel modem dropped to 169Mbps at -87dBm. The Qualcomm modem took an additional -6dBm of attenuation to get to that speed. Most consumers will feel the difference in very weak signal conditions, where every dBm of signal matters, so we zoomed in on that in the chart below. At very weak signal strength, below -120dBm, the Qualcomm modem got speeds on average 67 percent faster than the Intel modem. The Intel modem finally died at -129dBm and the Qualcomm modem died at -130dBm, so we didn't find a lot of difference in when the modems finally gave out.

Comments

Range issues

[Falconhell]

Was staying out in the bush this week, both of my companions had service on their cheapo Android phone, my iphone did not, very poor reception at fringes indeed.

Re:Range issues

[Anonymous Coward]

It's a "known issue" - every iPhone with an Intel modem has the same problem. It's why - if you have the choice - you want the model with the Qualcomm modem.

Apple is trying to make Intel modems "a thing" because they want to ditch Qualcomm. (See their recent feuding with Qualcomm.) But Intel modems are terrible. They're slower than Qualcomm's modem (so Apple artificially caps Qualcomm's modems in firmware) and - as you noticed and this article mentions again - they don't work at all at lower signal to noise ratios. They basically only work if you're close to the tower - as soon as you start to move away, they stop working.

This isn't really news - everyone knows the Intel modems are terrible - but Apple uses them anyway because they don't want to have to pay Qualcomm full price for a modem that works.

Re:Range issues

[tlhIngan]

It's a "known issue" - every iPhone with an Intel modem has the same problem. It's why - if you have the choice - you want the model with the Qualcomm modem.

Apple is trying to make Intel modems "a thing" because they want to ditch Qualcomm. (See their recent feuding with Qualcomm.) But Intel modems are terrible. They're slower than Qualcomm's modem (so Apple artificially caps Qualcomm's modems in firmware) and - as you noticed and this article mentions again - they don't work at all at lower signal to noise ratios. They basically only work if you're close to the tower - as soon as you start to move away, they stop working.

This isn't really news - everyone knows the Intel modems are terrible - but Apple uses them anyway because they don't want to have to pay Qualcomm full price for a modem that works.

It's actually the problem. Qualcomm is using their secret sauce of patented tricks to keep working good and fast, because that secret sauce is what keeps people using Qualcomm.

It's why Qualcomm has the monopoly it does and why everyone is trying to ditch them. There aren't many modem manufacturers, and Apple making Intel "a thing" goes way back to the original iPhone (Apple chose Infineon for the modem provider, now owned by Intel).

Performance wise, you're right, Qualcomm rules. It's why Qualcomm charges what they charge for the chips, because they perform great. Power wise is a whole different issue altogether - Apple chose Infineon for the original iPhone because it was very aggressive at saving power - so much so that well, it took down the AT&T network. (At the time, Cingular/AT&T was begging Apple to use Qualcomm modems, the European Infineon ones saved lots of power but were adapted to the European networks that had expanded control channels. Apple felt that the power savings were worth it, and with initial projections of maybe a marketshare of 1% after 5 years, surely one phone can't take down the network...).

Apple is pissed off at Qualcomm because Qualcomm's fees are based on the retail price of the phone - if you make a $500 smartphone and a $1000 smartphone, use the exact same chip, the $1000 phone will cost you double in Qualcomm fees even though it's the same, and that's where Apple disagrees. Qualcomm knows they can charge it because they are performant, and they are pissed Apple cripples Qualcomm chips to Intel performance.

In short, yes Qualcomm are faster. Why, it's because Qualcomm has secret patent stuff they either are not sharing (or don't have to share) in order to make LTE faster. Apple is trying to get Intel going - either because their chips have better battery performance, or more importantly, trying to get a viable alternative to Qualcomm on the market and break the Qualcomm monopoly. There are few modem manufacturers out there - Qualcomm owned basically the lion's share of modem chipsets, then there's Intel and I think MediaTek.

Anyhow, of any cellphone manufacture rout there, perhaps Apple is the only one that could make a viable alternative to Qualcomm, which is why companies like Samsung are cooperating with Apple. (Qualcomm is also why Samsung phones in the US don't use Samsung processors, but instead use Qualcomm processors)

Re:Range issues

[misnohmer]

What is "secret patent stuff"? In the US and most if not all other countries,all patents are publicly available to anyone who want to read them.

Re: Range issues

[Richard_at_work]

Ironic considering Apple charge app developers a variable exorbitant fee for a fixed cost service for each app sale...

Field Test

[JBMcB]

That's interesting. Get them out in the field. Sensitivity is not the only measurement that matters. In some cases, it's the least important measurement. Noise rejection and selectivity are usually more important in urban environments when you get crazy ghosting from signals bouncing off of buildings. Unless you have an insanely expensive RF capture and playback setup, you aren't replicating that in the lab.

Re:Field Test

[JBMcB]

I used to field test satellite radios in vans, it wasn't that difficult. We used a lot of the same gear they use in field testing cell phones. Nowadays they have apps that can do it for you - constantly send SMSes and count the failures, attempt to download something and check throughput, etc... You put three or four of each phone under test in powered cradles in the back of a truck and drive them around running the test apps. The trick is to run the same route over and over again to build up your test data. For control you have another set of phones stationary somewhere along the test route.

Re:no two iphones are equal?

[Anonymous Coward]

Created equal, every iPhone X is not.

They are all equal, but some are more equal than others.

Learn boolean algebra, dammit

[ClickOnThis]

The correct phrase is "not every iPhone X is created equal." That means some are equal, but not all.

TFH says "every iPhone X is not created equal." That means each and every one is different for all the others. That is incorrect.

Sorry for ranting, but every time I see this mistake, I die a little.

Re:no two iphones are equal?

[ClickOnThis]

Created equal, every iPhone X is not.

Posted to slashdot, Yoda has!

Cluster Beowulf of Natalie Portman hot grits in Russia Soviet, down pants imagine you, hmm?

Anyone test battery life?

[blindseer]

I seem to recall this dual sourcing of chips causing a significant difference in battery life in previous iPhone versions. Did anyone test if the batteries last as long on both variants?

I don't care much of the data rates on the phone's modem. WiFi is easy enough to find for when I want to avoid data charges and need speed. All I need the modem for is maps, e-mail, and some web browsing so I'm not concerned if the data is 50Mbit or 150Mbit. What does concern me is if the phone eats up its charge.

This speed testing is still interesting, but not what concerns me much. I'd just think that if they went through this effort of testing that the battery life testing would not have added much to their efforts. Given battery problems in the past I'd think that would be something many others would be curious about as well.

Don't we hear the same story

[SlaveToTheGrind]

about just about every model? Apple doesn't single-source-source modem chips; as a result, they get minor but measurable differences in performance between manufacturers. Shocking, really.

I paid $400 for a MIL-SPEC Kyocera phone

[Khyber]

It's a fucking tank on steroids. It may not be as FAST as the latest iPhone (it's a 2016 phone) but it gets signal where the iPhone will not (and on the same carrier - Verizon, out in the Hauser Geode Beds in middle-of-nowhere Imperial County. Closest facility? A state prison almost 20 miles away.) Hell, this phone is designed to withstand solar radiation, which means this thing should fucking work in orbit, and probably even on the moon. Oh, and it works through gloves up to 3mm thick.

Enjoy your shit new hardware that can't even hold a candle to the usability of older hardware.