Thanks for the info. I didn’t know that. I actually forgot it. I see I clicked the “like” button for the video. I deleted the video link on my comment.
2 Likes
New review at Fast Company:
Bit harsh on battery life, but overall good coverage.
2 Likes
It’s a fair shot. Battery life definitely is Framework’s Achilles’ heel.
2 Likes
I don’t speak Russian language, but it seems this video is published by a popular Youtuber with 1M+ subscribers.
Framework — первый модульный ноутбук. Обзор! (= Framework is the first modular laptop. Overview!)
1 Like
Here is an interesting review collection site I found. There are collected reviews for Framework Laptop.
2 Likes
For german speaking community members: golem.de just dropped their review
Golem.de: IT-News für Profis
I just realize a site I use a lot for reviews reviewed the Framework laptop last year and it’s never been mentioned.
Very favourable review, mainly noting that it doesn’t do well in gaming or 3D rendering.
3 Likes
I didn’t see this one posted yet. " The Framework Laptop Could Revolutionize Repairability. We Hope It Does." The Framework Laptop Could Revolutionize Repairability. | Reviews by Wirecutter
4 Likes
Why would you even compare a 17" with a 13.5"?
2 Likes
This is something I hear a lot regarding newer laptops and especially the Apple Macbook series: the battery life is amazing. You wonder to yourself, “damn, maybe I should get me one of those bad boys. How do they get such good battery life…” and then you see pictures of the insides and realize it’s like 90% battery, and maybe 9% motherboard, because the rest is just daughterboards and wiring.
(This is what the inside of an M1 Macbook looks like, courtesy of iFixit)
I mean, that’s a lot of battery life, but there’s also no room for anything else.
In order to get a sense of the difference in sizes, I overlaid the Framework image provided by rtings over the LG Gram. They are already at a 1:1 scale.
It looks like the LG Gram uses 2 40 Wh battery packs and the Framework uses 4 20Wh packs. It’s hard to get an accurate reading on the battery parameters per unit, such as voltage, because apparently there are more than 20 different SKUs available; I found this replacement part on Amazon, which seems to indicate a general trend of ~7.74V when paired with the OEM rated capacity of 80 Wh. That is half of the rated output of the Framework battery, which operates at a “Nominal Voltage” of 15.4V.
So, yeah, there’s definitely a good reason the Framework’s battery is the size that it is.
M1 is also an ARM SoC which are designed for battery life above all else, because they’re intended for embedded devices.
@Peter_Schofield That was only true up to now, because “embedded devices” was basically slang for “less powerful than x86_64”. Now that Apple has produced an SoC that rivals and even surpasses offerings by Intel, AMD and even Nvidia (when you consider “performance per watt” in the bill of goods), that argument no longer holds water.
Between escalating electricity costs and increases in the cost of silicon due to the chip shortages, power efficiency is going to be more important than ever going forward. After all, it doesn’t matter how powerful your CPU is, if it draws 500W from the wall or lasts 10 minutes on a full charge.
@RandomUser Yes, compromises were made. However, I don’t know if the battery is one of those areas where there is much room for improvement. Short of serious advancements in battery production and capacity density increases, I doubt you could get a much bigger battery than what is already available, at the voltage this one is rated for.
My thinking is, these batteries were chosen specifically because Framework knew that some or most of its users wanted to make use of AMD offerings at some point and that processor family probably has different battery requirements than Intel does. Technologies like Hyperthreading and SMT (Simultaneous Multithreading), for example, aren’t simply semantic changes in technical jargon; the implementation of one over the other carry very real consequences, in terms of how the processor handles tasks and thread schedules.
1 Like
incorrect, ‘embedded device’ means it has to run in an ultra low power envelope because its either always on, it runs from battery for long periods, or both. PowerPC is generally less powerful than x86 but doesn’t qualify as embedded. The term embedded device has absolutely nothing to do with processing power and everything to do with its intended use: i.e making equipment work.
@RandomUser Okay, I should have been more specific in assessment: the team probably had to use this specific battery because A) it is rated for 15.4V and B) it is compact. The Asus Zephyrus G14 is a 14" notebook, but it is also widescreen. If you look at the pictures provided by Notebookcheck, you can see that the battery is almost as wide as the chassis.
Correct me if I’m wrong, but current LiPo battery technology for laptops is near or at its limits: you cannot make batteries denser with existing materials, without introducing major design constraints or potential risk to the consumer. Since the Framework has you remove the keyboard to access the internal parts, stacking cells vertically is not an option. As far as I can tell, the only way to implement a bigger battery safely is to shrink the motherboard.
Note that I’m assuming Framework is using batteries from third parties and not making their own, which is not an unfair assumption. Silk screening text onto existing components is probably a lot easier and way less capital-intensive than developing their own batteries–if that is even an option.
@Peter_Schofield You are correct, I was probably thinking of PowerPC when I made that comment. My error was assuming that ARM has primarily been used in mobile applications (which is how I am most familiar with the architecture) but yes, “embedded devices” is definitely more akin to automated machinery. Smart devices are a relatively recent development.
@RandomUser Okay, then I guess we have to ask Framework if they can disclose such information (assuming, again, that they did not make the battery themselves and that such knowledge is proprietary and guarded by an NDA) and/or try to put in orders for higher-density batteries.
I suppose, if they give us the information about the energy density, the community could also try to source batteries that “sorta” fit into the housing and jerryrig the power delivery system to work with the alternative battery.
I have to admit I’ve actually checked the power draw at idel of my main desktops and down-clocked them. I was running my main rig a 2014 spec Intel 5960X 8core/16 thread at 4GHz. But it’s now running at just 2.5GHz with the power efficiency settings enabled. The rig now idles at around 95w down from 140W+. It’s on for 8 hours a day at least so…
I even down-clocked my GF’s FX-8350 rig (she doesn’t game or do video work, mainly Wordpress stuff).
Ironically enough, ARM actually started out as a general purpose CPU. It was originally used in the BBC micro and Acorn lines of computers (if you’re not British there’s a very good chance you’ve never encountered or heard of either) which were developed with funding from the BBC (yes that BBC) for schools in sort of the 80s and 90s. After that they became almost exclusively used for embedded roles in machinery, set top boxes and the like until the rise of tablets. ‘smart’ and the raspberry pi.
embedded x86 is a thing, believe it or not. it isn’t common but it exists. My high school had a load of custom hardware (dozens of units, designed and built in house) based on a board with a 486 embedded cpu, all of them in near constant use.
upgrade her to a Ryzen, the power savings will pay for it in short order. I have a friend that went from an FX-8350 to a 2700x and it cut his monthly electricity bill in half.