Hardware and the Competition


Warning: Several long, multifaceted posts ahead. We may or may not get something going, but you have been warned!

While I count down the hours to my new FW laptop’s arrival, I thought I’d see if I can stir up some discussions.

I’d like to have some constructive, exploratory, NONPEJORATIVE discussions about… hardware and the competition. Compare and contrast how the hardware is different and why, and whether there’s innovation to be had - in either direction.

And not just laptop manufacturers. In fact, most of my next few posts are about tech that’s in laptops - not the laptops themselves. It’s all interesting to me.

Rule number 1: Be nice. Don’t care if company A is a ripoff or company D has lousy tech support. The discussions are about what secrets and clevernesses are in the hardware - not the company that creates it.

I’m kinda old, but I remember when I was just a tyke and my gramma would bounce me on her knee and say “Jakey, my grandson, apple of your mother’s eye. If you remember nothing else I say, remember this: If you can’t say something nice, go read a dozen more posts.”

(None of that is true. The point is the punchline.)

Who am I? Nobody special. I make my living as a contract onsite repair technician - for Dell these days - and I’ve had this gig (one of many) since December 2013. I drive around and do hardware breakfix on laptops, desktops, servers and the occasional printer. I have close to 5,000 service records on this gig, and I figure that 65-75% of those are laptop repairs, so… well, I’ll let you decide whether I speak from experience…

I’ll start with some thoughts in the next few posts, and maybe we can get into some nice meaty dicussions about laptop hardware.


Let’s start with keyboards.

@nrp Really? 50+ screws? Allow me to take a trip down memory lane.

At the start of my gig you could take a Dell keyboard out of a laptop with a spudger. No screws, just bendable tabs at the top. Punch one down gently and start prying from there across the keyboard and you’re gold. To install, connect the cable, lay the thing down, and press. Done.

Then they started with the screws. Just a few - I think the most I’ve ever taken out of older model Dells is 6-7. Little guys, but not too many and their Phillips slots were generous. There’s often a plastic shroud - basically a grid with key-size holes - that needs to be pried up first, but that’s easy. So easy, in fact, that it’s not uncommon to get overconfident and spudge a key into the void…!

I used to work on HPs and Lenovos as well. Lenovo had a cool idea - the shroud was mounted on the keyboard deck, below the keys, and it could slide toward the back a small distance, exposing the screws holding the keyboard to the laptop. IIRC they were captive screws too.

Then they started mounting the keyboards in metal trays. Same keyboards - 6-7 screws. But the tray was held in by 10-12 screws. And around this point the keyboard lost its spot at the top of a repair stack. You had to at least remove the palmrest (what FW calls the Input Cover), then the tray, then the keyboard. And the ribbon cables were no longer connected under the keyboard - they pass through to the internal space, where they usually connect to the circuit board under the touchpad. So you have to remove (not just disconnect) the battery as well.

Another design language showed up in the next generation, in which the motherboard is “on top” of the keyboard, which means basically complete disassembly to get to the keyboard tray.

Nowadays, similar to our FW, keyboards are directly attached to the underside of the palmrest - with tons and tons of teeny tiny little screws that have the world’s shallowest Phillips slots. I blink one too many times and they strip.

(And according to Louis Rossman, some companies don’t bother with screws - they use rivets, or heat stakes, or… (yech) glue. We’re pretty lucky that way. You’re forgiven, Nirav :wink: )

As a technically-oriented consumer I understand why - thinner and lighter. As a computer repair technician I can tell you I’m glad I’m mostly bald - harder to tear hair out that isn’t there!

NB: @nrp I may have missed them, but I suggest you put screw counts on the guide pages - especially for the keyboard replacement. Some of the above-mentioned baldness came from screw miscounts.

1 Like


Will someone PLEASE tell me why we have to have those TINY little coaxial connectors for the antennas??? I’ve watched the diameter of those things shrink to near invisibility. And they’re SO unforgiving! I may or may not have wrecked one today during a repair (honestly not sure it wasn’t broken when I started - and I do own up to my mistakes). The slightest deviation from perfectly round on either side, and it becomes impossible - any further attempts just crush things worse.

You either have to replace the WiFi card (which increasingly is soldered to the mobo) or the antennas (which usually means the top cover - what the LCD panel is screwed onto - or the entire display assembly).

Can’t anyone come up with some sort of fitting that will properly align, connect and secure those connectors without me basically having to do it by feel? I have to drag the wire’s connector across the card’s until it feels like it’ll lock in when I push down. 90% of the time I’m right. 90%. I usually get around this by not disconnecting the cables during the repair if I can avoid it.

And once the antennae are connected, FW uses a metal plate screwed down to secure them. I see that with Dell too. I just assume it’s nonconductive, right? I mean, shorting the antennas doesn’t make a lot of sense to me, but I’m no radio tech.

Dell sometimes uses a plastic retainer thingie instead. They also use these little blue push-pin-y type things to hold down M.2 form factor devices. Those would be easier (and less disappeary) than screws, but I think they would be counter to FW’s values - metal recycles, plastic not so much…

I also usually see a little rubber block with adhesive on it, positioned in such a way as to “capture” the WiFi card when I push it down, before I screw it in. The only reason for it that I can come up with is… to hold it down until I can screw it in. Seems unnecessary. At least the adhesive.


Ribbon cable connectors:

Love the horizontal ones - slide the cable in, make sure the two little tabs seat in and pop that lever right down:


Hate the vertical ones - so easy to pop the retainer tab completely out, and sometimes it’s hard to align the ribbon. Honestly this seems like it evolved from the keyboard connector on 8-bit Commodore machines.


My favorite, from a repairability standpoint, is the display connector - the metal one with the swivelling bar that locks it into place. A little fiddly to get the connectors together - you can easily bend things out of alignment - but once its locked in it ain’t goin’ nowhere. Perfect for the back of an LCD panel - I wish it was used more often:


Major kudos on the Input Cover Cable Connector, Framework! Other than no cable (we can hope!), locking all that functionality into a single, easily dis- and connected connector with lots of slack, a strong pulltab and bright eyecatching color is pure genius. SO much better than the three or four ribbon cables I have to slip through to the inside, ensure no kinks, and not forget to attach before I reinstall the battery.

1 Like


Oh this one’s my absolute favorite. Can any of you guess the proportion of motherboards I have to replace JUST because the USB-C port has become loose and creates intermittent connections with the dock? You know, the thing that supplies power, external video, networking… nothing critical.

Why, oh why do we use a flimsy, fragile connector that most users slam cables into, and is mounted on the motherboard? It’s fine for phones, which are light and move easily. But relatively heavy laptops just sit there and transmit that force right onto the mobo.

Also most people let their cable just hang there - no support in any axis, to prevent the plug from slowly stretching the female side as gravity does it’s work. Literally built-in self-destruction.

Over the past couple years I’ve noticed that the USB-C ports now have a metal bracket screwed down over them - usually with longer screws than normal - to try to distribute the force from a cable insertion. Someone agrees with me…

Framework wins a gold medal here, IMO. Not only is all the stress from insertions (of cables) directed to the easily replaced expansion modules, but the structure of the modules and bays almost eliminates any stress on the mobo-mounted connectors - the box will impact the back of the bay before the male puts any force on the back of the female. Absolutely brilliant way to lengthen the life of the motherboard.


Ok, well that’s it for me by way of introduction. I’d love to see the community do something with my remarks (good or bad, it’s all interesting) and maybe come up with other observations. We seem to be a crowd that likes talking about this stuff - let’s!!


1 Like

Oh I have another one. Display bezels.

We all love the magnetic bezel on the FW laptop. I love it even more…

Used to be, bezels were just friction-clipped on. You have to be careful, but it was usually a quick zip around the outside edge with a spudger and the thing came right off.

Now… there’s adhesive. It’s around the inner edge of the bezel, and you HAVE to break it to remove the bezel, like say to replace the panel. Or the webcam.

Only - the adhesive adheres to the panel. Most panels have a foil/mylar coating around the perimeter, and the electronics at the bottom are also covered with a very thin sheet of something.

And it rips up when you pull the bezel. No matter how careful you are, no matter how sharp your spudger is. If someone knows how to remove a bezel that’s glued onto a panel without damaging the panel even just a little, I’d love to hear.

With a dead panel, I suppose it’s no big deal (though I doubt it helps with remanufacturing). But if you’re replacing a webcam? It’s just not that difficult to kill a perfectly good panel.

Again - as a consumer I appreciate the increased resistance to dust and liquid. As a repair tech… (shudder)

1 Like

@Jake_Richter Quite the contrary - from the sound of it, you’ve been in the trenches waist-deep in hardware issues for quite some time. Your experiences are incredibly relevant to the discussion around Framework and it’s goal of repair-ability. Thank you for your comments - I’m going to read them in much greater detail!


@gjason Thank you for your kind words.

So here I am the next morning after all that furious posting, and I realize I forgot to tell you my motivation for doing so.

I have an app (of course) that helps me do my job. It’s the Dell Field Service Mobile or DFSM. It’s actually gotten nicer over the years and really does make my job easier.

Obviously it’s proprietary, not much use to the average user, and only a bit moreso if you’re a non-Dell tech. Can’t get it in the Store - it’s downloaded and installed separately.

Yesterday I got a notification on it about an app Dell released. It’s in the app store where anyone and everyone can download and install. I believe there’s an IOS version as well.

It’s called AR Assistant and it’s an Augmented Reality program for fixing Dell machines. I’ve only checked it out a couple times but it looks absolutely amazing.

For me, of course, it actually slows me down - by the time it gets to the actual AR screen, my repair would already be done. But that’s not the point - I do this for a living. Most of the info is already in my memory, and I regularly update for new models via CBT.

I haven’t really had a chance to dive in deep, but this looks like the future of computer repair support and it looks… really good.

If you can’t tell by now, I’m hardly a shill for Dell. But this - this is impressive…

My goal throughout my career has been to eliminate my own job: Set up self-sustaining (not really, but you know what I mean) networks; inform people about their computers and encourage them to do for themselves; try (with great difficuly) to feed information back to manufacturers about their machines to make them more bulletproof.

This AR assistant could go a long way toward helping me achieve that. I encourage Framework to look into it and see if and how it could be made useful for us.


There are some things that could be done by everyone. Building a desktop computer? Sure, everyone could do that. Surface mount repair? Nope. Not everyone can do that, for the same reasons everyone can’t be a surgeon, or programmer.

That said, laptops are miniaturized desktop computers that are portable. The very act of shrinking something makes it more complicated to build and repair.

USB C for its weaknesses always overcomes with its advantages. Power delivery, high band bandwidth, and high availability on many devices. This means someone can bring one cable and one charger for all of their devices. This is important for minimizing waste and improving longevity of parts.

Most of your observations while true, are necessary evils in this case. Bad flex cable? Framework has guides on how to use it, and you can get a replacement part if you must. This helps them save on the costs of manufacturing and keeping the product attainable by a wider market of customers.

Your USB C port busted? They are $9 bucks and very easy to replace. My guess is, if you have a 3d printer, you could probably print a replacement shell for the busted one, if you were so inclined. I think Framework might even have a system in the marketplace they intend to open to help facilitate. But that is just me guessing.

Not sure I get the keyboard. I’ll take a bunch of screws over tabs that break.

Regarding Dell and the AR app. They are a massive company. Their resources are near infinite when contrasted with Framework. That isn’t to say ideas can’t be garnered from it, but things need to be realistic and reasonable.

Still I think your comments and experience are very important to the on going discussion of right to repair and devices meant to last, instead of being replaced. Thanks!

I second this. I’m willing to compromise on many things, including the 50-ish screws because I understand that it’s just near impossible to get something perfect on your first go. If Framework can figure out how to make a thin keyboard assembly that uses less screws and no rivets or glue, then great. If not, they did the best they could given the time and space constraints and I appreciate that too.


My own perspective on the keyboard as a mechanical engineer:

@randomuser and @CSab6482 both correctly state that those tabs are very likely going to be points of failure for keyboard retention problems. That type of design is extremely easy and cheap to manufacture and assemble, but only easy to repair a handful of times. Every time you bend one of those tabs, you work harden it (making it stiffer), and simultaneously introduce micro (or macro) fractures which will limit its strength. Each successive bend you do is with an increasingly brittle component, with an increasing number of fractures. In other words, You need to carefully ration your repairs and disassembly significantly with such a design, because it’s a time bomb.

On to a deeper issue: I’m willing to bet that those tabs were on beefier, chonky bois with a mighty thicc chassis by comparison. That is to say, back then keyboard rigidity was only important insofar as reducing deck flex. That could be well-controlled by adequate chassis stiffness underneath the keyboard, with no significant other forms of support (adhesive, rivets, screws, etc.). Much to the contrary, on a machine like the Framework Laptop, the keyboard actually must function as part of the structural element to retain not only deck flex resistance, but torsional stiffness of the entire chassis. That means fastening. In this case, Framework uses screws for those fasteners to be removable. To further complicate, the overall thickness (read: flimsiness) of the keyboard assembly necessitates LOTS of screws. Why? Because when the deck does flex, each screw acts as a shear joint. Each screw, being so small, only has a teeny tiny bit of shear strength before it fails. The deck, being so thin, also has locally only a teensy bit of strength before it either buckles (if the screw is stronger than it) or shears (if the screw threads are stronger than it). Thus, to maintain overall rigidity while the top and bottom edges of the deck try to slide against each other, you need LOTS of those mini screws to avoid catastrophic failure, thus maximizing not only the fasteners’ total shear resistance, but also spreading out the number of points that the deck and keyboard are bonded. Think of it like reducing pressure for a given load, you can step on one nail and it’ll pierce your foot, or you can lie down on a massive bed of nails and suffer nothing worse than a bad night of sleep.

TL;DR: It’s not just IC’s/SOC’s with more and more features integrated. Now ultralight machines’ keyboards are being tasked with stuff the chassis used to be solely responsible for.

1 Like

Now, on to the spirit of the thread: none of this is to say that 50+ screws is the only possible answer to this engineering conundrum. Notice how I mention that the problem of making the keyboard a structural element translates to a need for fasteners, not necessarily screws. Screws have lots of great benefits, but equally as many problems. Chief among these problems is that they need to be spun. Round and round they go, wasting your time. Both manufacturers and consumers end up frustrated and wasting their precious time spinning these things around and around, burning untold money in the process. Some manufacturers answered this question by saying “You know what? Screw the consumer, some plastic nubs and a big hot plate, and we can nail down 150 rivets in probably under a second. There’s our money back.” That makes A LOT of sense as a manufacturer if you don’t care about servicing the part afterwards.

However, there are also intermediate solutions. For example, if you can give up the extra space, you could introduce a stamped metal mating plate that acts as the structural member you gave up originally. Or you could increase the parts cost by introducing pins with tight tolerances that fit tightly into mating holes without any clamping (for shear resistance), and just use a handful of larger reinforced screws to give you an equivalent clamping force as the current Framework design. Another solution is to introduce screws with shoulders to improve shear stiffness, if that’s the limiting factor determined after structural analysis, and clamping forces are sufficient with fewer screw counts. Who knows, reduced assembly time might make up for the increase in parts cost in either of these hypotheticals. Each potential solution has its drawbacks though, such as introducing new fasteners, custom hardware, or what have you. The engineering team for each product has to weigh each solution carefully.

1 Like

Sorry for the unnecessary tag, I forgot that it sends the tag-ee a possibly nuisance email… I used it like a footnote/reference out of habit from other types of writing.

Edit: Also, I’d never looked inside an old thinkpad before, but incorporating a drip tray could be a brilliant (and space hogging) way to both introduce water resistance and structural stability. That’s basically the next-level big brain way of doing my suggested stamped metal mating plate.

Oh, awesome! Thanks, guys!

Gotta disagree with you - though I admit I’m being pedantic. Not everyone can build a desktop. Everyone here maybe. Though I’m seeing the occasional post from people who’d prefer not to. Still - valid point - everyone has different skillsets, levels and experience.

Unfortunately I agree with you there.

In the old 4-wire USB days I used to lament “Michael! Steve! What’s wrong with a 1/8” phono? Multiple rings, break-before-make jacks, auto-polarizing, So much easier than making sure the little emblem is right side up!"

(Yes, Michael Dell and Steve Jobs. No, not really - that was for dramatic effect.)

But you know, time passes… In any case, nobody has solved the (well, my) problem anywhere near as effectively as Framework. Those 4 ports get to sit on the motherboard and be treated like the critical components they are, and that makes me happy.

You’re absolutely right. And since they appear to have just released it I’m sure it’s been in development for a long time for lots of money.

But that “ideas [can] be garnered from it” phrase is where I was going. One cool thing about tech is that once it’s invented, future iterations are easier. I’m certainly not expecting tomorrow but it would make me happy to see Framework have a disruptive way of delivering repair information.

Oh, I hope you don’t think that was a criticism. If anything, I prefer dozens of screws to an alternative that prevents me from a repair…! I was just pointing out the progression over time (and yeah, how it’s made my job a bit more “interesting” over that time).

The keyboard is getting great reviews, from what I’ve seen, and I’m looking forward to checking it out personally.

(Does quoting a user’s @name cause another notification? If so, sorry…)

Yep, and earlier today I was going to make 2 points, but I’ll retract them here:

  1. How many times are you going to replace the keyboard? (Uh, Jake, that’s not the point. This is about sustainability, repairability and upradeability. Replacement count doesn’t matter).

  2. The tabs are on the keyboard, so replacing the keyboard gets you a whole new set of tabs. (But I can’t swear that’s true, I honestly can’t remember the last time I replaced one of those types of keyboard so I can’t say for certain where the tabs were.)

Tell it, brother! With each successive generation of laptops I repair, I see various parts of the innards being put to slightly different, increased use, especially physically. The keyboard used to just sit there - now it’s a major structural piece. Progress…

Also not a radio tech here, but I know a few things. These are coaxial cables, so there’s a tiny conductor, dielectric (white plastic) spacer/insulator, metal shielding (foil or braided), then the final insulator. The connector is metal on the outside, but that’s connected to the shielding.

So if the FW metal plate is conducting and touching the connector, then it’s actually connecting the shield of the coax cable to the case, which is probably actually helpful.

Also yes those connectors are tiny and annoying, but it’s hard to find connectors that are thing enough to fit in laptops, yet certified for 2.4 GHz. And 5 GHz is harder. And I heard the AX210 even supports 6 GHz these days.

Radio stuff is basically dark magic, especially at these higher frequency. I wouldn’t expect easier connectors in the future.

Also I didn’t even know the connector on m.2 wifi cards is different (smaller again) compared to mPCIe cards. TIL.

Of course. The shielding. How dumb am I - after all these years it’s so obvious. Thank you.

I guess it’s not all that critical a thing because I don’t always see a metal plate - on most older laptops and some newer ones the card’s just held down with a screw. Maybe moreso as speeds increase?

The cards are also in desktops and in those cases (no pun intended) if there’s a retainer it’s usually made of clear plastic.

Yeah it was slightly stressful at the bigger diameter - it’s a whole bunch of fun now…!

Even worse (or is it? see below), they’re soldering WiFi cards to the motherboards now, just like the CPUs and RAM.

On the one hand, a more stable base makes it a little easier to connect the dang things. Trying to orient the connector and hold the card AND push the thing down can be frustrating. Sometimes the available slack on the antenna leads is measured in millimeters.

On the other hand, if the connector gets messed up, you may have to replace the LCD back cover (or wherever the antenna lives) and/or the motherboard, not just a much cheaper WiFi card.

On the other hand, making the WiFi card part of the motherboard means that maybe the connector could be changed - integrated into the display cable (though I suppose there would be signal issues with that), or… I don’t know. Just something better. Especially due to the reason above.

I’m not even necessarily asking to change the connector - I’d just like some sort of jig or alignment tool or something to help prevent crushing the thing. Maybe the metal plate could have some “vanes” or something that steer the connector to the right place, and just sorta putting it all together and screwing it down connects it? Yeah even as I type it I’m thinking how fragile that would be…

Well I just learned a lesson. In front of a customer - always my favorite scenario.

So the venerable BIOS battery is going away. Apparently the job of supplying power to the CMOS settings now belongs to the main battery.

It might just be resistance to change, but this doesn’t seem like the best solution.

I’m no electrical engineer, but–

(Actually, let me get this outta the way now: I’m no expert on anything. I’m a Jake-of-all-trades - I know something about a lot of things, but not enough of any one thing.)

And anyway, why are we using battery-backed memory in 2021? Don’t we have flash?

I must be missing something - not uncommon…