In low- or mid-end phones the baseband is usually integrated in the main SoC. Taiwanese MediaTek is the biggest supplier. Unisoc clearly wants to compete with MediaTek, and Intels modem division has trouble getting into low-end phones as nobody wants complete Intel SoCs in phones. So they agreed to integrate Intels modem as an IP block in Unisoc chips.
In tablets and the few laptops that have mobile data the baseband + RF are usually on a modem module, M.2 or similar. The Librem 5 phone also takes this approach.
Am I not understanding something here?
No, the modem/SoC integration is not a big factor. Having an internal communication inside a single SoC will always be a bit better than going through an external interface, but it's not a significant impact when compared to the total phone power consumption. As other have said, iPhones have an external modem and are fine. Qualcomm insisted on this because they had it, and they're good at marketing.
LTE was a drain at first for the same reason 3G was a drain at first. Cellular standards are designed to have a long life, and are over dimensioned initially. Then Moore law and other optimizations kicks in and make things more acceptable. This is intentional, otherwise the standard would not make the best use of technology over time. And this is something where we need to keep an eye for 5G: Moore law, strictly speaking (higher density for the least cost process. One can still get better, for a higher price) is over. Hopefully 5G will have properly taken this into account.
But then for 5G a big part of the power will be in the RF front-end, with all those antennas. Here too it'll be interesting to see how it evolves over time.
Addressing the second sentence - I have no knowledge of what transpired but my assumption would be a combination of Apple didn't want to have their SOC's fab'd by Qualcomm for the sake of an integrated modem, and Qualcomm didn't want TSMC fab'ing their modem IP into Apple's SOC's, so the obvious solution was a discrete modem which was a product they could offer.
While those point are valid they are not the reason Apple have a separate Modem Chip. The design of iPhone had the Modem literally in a completely isolated system, and act more like a USB modem directly connected to the Internal iPhone for security purposes.
And as other have mentioned the integrated of Modem within the SoC have minimal impact to overall battery life.
It's almost certainly due to Apple wanting flexibility in changing suppliers. Qualcomm was also not going to give them the rf crown jewels either
7, 8 / X All had a version with Intel Modem. XS is, however the first model to ship only with Intel Modem.
I don't think it has horrendous Data Speed. But it is certainly has a different characteristics to Qualcomm Modem. XS is generally weaker with signals, and we are not quite sure if it was the 4x4 MIMO or the Intel Modem. XR doesn't seems to be as bad so a guess would be due to Apple's implementation of 4x4 Antenna. While ALL phones had problems with 4x4 Antenna in their first generation, that includes Samsung and Huawei, you generally expect Apple, being a year late to the tech would have ironed out most of it.
But generally speaking, a Qualcomm Modem would still get better signal than Intel. This is especially a problem in US where the Cell Tower Density is lower as compared to Asia / EU counterpart.
There has also been 3 to 4 Modem Firmware revision since launch, and likely more to come in iOS 12.2, let just hope ( pray ) Intel's next Modem 7660 is actually better. And Apple could actually put a rounding error budget into their Antenna Design, because as far as Flagship Smartphone is concerned, iPhone has the worst one.
Or Qualcomm will relent and work with Apple again, but judging from the Court testimony this seems highly unlikely.
I am surprised there hasn't been more buzz about the degraded antenna performance by the public.
I certainly noticed when I switched from my iPhone X to a Pixel 3 XL. I went to known dead spots, and my Pixel 3 worked just fine. I had just chalked it up to T-Mobile sucking but I guess it was really the phone.
I know where the attenaes are placed versus where your hand is placed can block signal. But is this a similar case, or just a limitation of the tech? I mean 5G uses higher wavelengths, which inherently penetrate weeker into structures, etc. So: what's the problem? I only briefly studied antennae theory in college and have never practiced in theory, so I hope its viewed as an honest question.
AFAIK there is still a sensitivity and LTE capability gap between Qualcomm and Intel.
Kind of sad Apple doesn't think that's worth the extra ~$10 or so on a $1000 phone.
I believe cost is only one part of the reason. The much more important part is, I think, control.
It was my understanding that when Apple started using both Intel and Qualcomm modems that Qualcomm started telling Apple they would have to pay "full price" for patent licensing on the Intel modems, and things deteriorated.
in a way that is free from outside control or influence.
without outside help; unaided.
We all know ho paid for it, so you can keep repeating whatever you want.
It appears the site is run by Milan Milanović, who works for Ookla. He writes extensively on wired and wireless networks.
It's Qualcomm propaganda, like it or not.
It was a very common complaint in youtube and reddit comments anytime a review was posted.
If only more companies realized the same. Not that this means they won't have to protect themselves against China's
relentless IP stealing.
Not a couple of years later. A couple of hours later.
While we are tearing apart our democracy in the us step by step, we'll eventually get a different leader who is more interested in doing useful things and less interested in a cult of personality and less interested in doing any venal thing to get support for his base. Meanwhile, China is stuck with a dear leader for life. The country grows ever more powerful, but also is increasing in it's terrifying Orwellian oversight. I'm afraid that China will show it's possible to watch everyone all the time and crush freedom and then other western authoritarian leaders will follow suit.
It's less propaganda and more like an image salad.
It seems telling to me that someone would make out that one frame was inherently "good" while the other was inherently "evil".
No, the US realized they're in a trade war with China, and companies follow suit on their master's orders on both sides.
And Intel decide to end the partnership? Something doesn't sound right. Or did they get IP from other players?
Now all this makes much more sense. But they had really crappy Modem in the past, as compared to MediaTek in the low end segment. I guess this 5G IP is part of the coming XMM 8160.
The frequencies are so easily blocked by anything that you need transmitters everywhere and receivers everywhere - literally 5G phones will have to have 3+ separate 5G antennaes in different locations to avoid the signal being outright blocked by your hand.
5G devices will practically need line of sight to a transmitter.
The disadvantages you mention only apply to 5G mmWave.
You get diminishing returns even on exponential scaling.
Ofc some possibilities do open, like cloud backup, or even personal data servers, but they are laregly unexplored as of today.
I cant imagine what would 5G bring practical to the phone, bandwidth doesnt seem a killer. Latency is probably acceptable as well. Maybe im too limited.
- feel the internet on our phones is about the same speed as it is now
- pay the same amount each month on phone bill
- have same low coverage areas as we already do
- see a bigger number on the top row of the phone
In terms of the improvements from 4g to 5g, going from 40mbps to 400mbps may not yield super noticeable improvements right now, but you know what will? 5g's other big feature, wired internet levels of low latency. That means things that require instant feedback like remote controlling delivery drones or surgeons performing remote surgeries by controlling a DaVinci robot in the middle of africa from their office in the US.
So no 5g is not hype, the focus is just too much on the speed and not enough on the possibilities opened by 5ms latency on devices connecting miles away from a tower.
I remember getting those speeds back in 2009, ten years ago, when the maximum speed I could get on cable was 3 Mbps.
Main difference, and I guess the real business growth motivator, are the data caps. Back then I had a data allowance in the low hundreds of megabytes which made streaming music and video an expensive endeavor, while the cheapest plans now start at a couple gigabytes.
Cable latencies in the last mile won’t help when your patient is half a world 250ms away due to the speed of light. But it might enable gaming and other real time uses, though.
So one way to guess uses for 5G is to ask:
Is there anybody using those frequencies for low-latency wireless communications ? what is their use case ?