A long time ago, a laptop was a basic thing, and you could pretty much run one just by hooking up a power supply to the battery contacts. A modern MacBook is altogether fussier. However, when [Christophe] was stuck in the midst of a 2020 lockdown with no parts available, he found a way to get his damaged MacBook up and running without a battery.
The problem was brought about by a failing battery in the MacBook Pro 13″ from mid-2018, which swelled up and deformed the laptop’s case. Parts were unavailable, and the MacBook wouldn’t run at full speed without a battery fitted. That’s because with no battery present, the MacBook would send a BD_PROCHOT signal to the Intel CPU, telling it to slow down due to overheating, even when the chip was cool.
To get around the problem, [Christophe] used a tool called CPUTune. It allows fiddling with the various CPU settings of a MacBook. He deactivated the BD_PROCHOT signal, and also the CPU’s Turbo Boost feature. This ended the worst of the thermal throttling, and enabled semi-normal use of the machine.
It’s unclear why Apple would throttle the CPU with the battery disconnected. [Christophe]’s workaround got him back up and working again in the midst of a difficult period, regardless. We’ve seen some other great Macbook hacks before too, like this amazing save from serious water damage!
Thanks to [donaldcuckman] for the tip!]
Pretty clear why they would throttle – obviously the worst case current peaks can’t be supplied by the AC psu alone (I think the 13” came with a 60W brick). Max CPU/GPU max display brightness, peak USB load, etc.
Could they have done a pile of extra work to work out which cases could be supported without a battery? Maybe. But that’s not exactly the normal use case for a laptop
I agree on the peak current spikes, maybe it’s just a cheap hack Apple threw in there expecting this to never happen. But to clarify on the purpose of the battery:
It’s my belief that a laptop should run at maxed out CPU and GPU and maintain a full battery when plugged in using the original charger. The charger should be enough to run the laptop, and have enough power left over to charge the battery as well.
Like in a gas powered car, the battery only starts the engine. The alternator provides enough power from the engine to run all the electrical needs, plus have enough left over to charge the battery.
Apple also got excoriated on the iPhone for secretly throttling it as the battery aged, ostensibly to prevent it from browning out but really to push users to upgrade. It totally makes sense that they would pull the same shenanigans on the macbook.
“Ostensibly”? You know how large the current peaks can be there, and how battery internal resistance rises over time? There are no easy solutions. Sure you can pretend the problem isn’t there like some other vendors, but iPhones last a long time (both in support and in value: they hold their value for far longer than other brands, see eBay) so battery aging is a real issue.
Better to have a phone slow down than brown out during a current peak during a 911 call.
Cell phones radios peak at 2A draws (for milliseconds) all by itself. The power spikes are huge, and adding capacitors to ride through it are not really an option unless you want a bigger phone.
Some Thinkpad models have this issue as well– it would hold the processor at a low speed (1.2GHz) if the 65w charger is used and no battery is installed. The solution is to use the 90w charger.
Yep, all T420/430 X220/230 do this. The solution is to run ThrottleStop :) at least the X220/230 run fine on 65W charger and running a program was easier for me than soldering a resistor to fake 90W psu :P
My 2012 Macbook can technically do that, at 100% util it supplies 50W to GPU and 35W to CPU, but the drain of the screen backlight and peripherals could drain a couple of % an hour as the charger only does 85W.
The only time I ever experienced that happening was running Far Cry Blood Dragon with both fans pinned to 100%, dual spinning hard disks, 8ohm headphones, mouse and external backlit keyboard. In real world use it’s very unlikely that you can truly sustain 100% utilisation on both CPU & GPU.
I’m in a similar situation to Christophe, and had used the same method to run the CPU at over 1.2GHz:
I replaced the cells with 18650 holders, but annoyingly the BMS permanently bricks the ability to charge or discharge, when any of the cells go out of the safe range, and the PROCHOT disable technique doesn’t work with no battery plugged in at all. It does make for a very light laptop though, so light it can’t even be opened without holding the base down!
Glad I saw this though, it has given me the idea to just plug the BMS back in with no cells attached at all, or maybe some 3V3 ref ICs connected if required, so I can get higher-than-minimum CPU speeds again!
But even locked to 1.2GHz, the 3rd gen i7 still gives better performance than brand new ‘ultrabooks’ supplied by my uni for class use. I once managed to boot, download & install Fusion360 and get the first CAD excercise started before some of my classmates had managed to even log in! I know it’s the C&C bloatware and network syncing causing those problems, but still seems ridiculous!
Unlike a car, a computer can have higher peak currents, and perhaps more relevant, the thermal envelope available for processor/GPU is not great, so in most cases the machine cannot sustain that peak power draw.
But the more annoying case is that that they did not do this “pile of extra work”, probably all of half an hour napkin math.
I would guess they have done this either way(you know, designing a computer require some knowhow and work to get everything playing nice), and then intentionally did not let it run nice without a battery.
Most modern laptops cannot run maxed out CPU and GPU more than a few minutes before thermal throttling. So it wouldn’t make sense to make you carry a bigger power supply just for the few minutes, when battery can provide for the short-time boost.
I totally agree here. A laptop (or tablet or phone) that can run out of battery when it is plugged into the provided or recommended charger is a cheap piece of garbage that isn’t worth paying for. I was under the impression that Macs are supposed to be higher quality, which justifies their premium prices. This tells me that my decision to go with Linux on non-Mac hardware was the right choice. Even the Alienware machine I got for machine learning (that has massive thermal problems that prevent me from using it for that purpose) came with a 240 watt charger that can power it at full CPU and GPU load, and still have power left over the charge the battery, and the price was similar to much less powerful Macs that evidently also come with cheap-[donkey] chargers that can’t even power the lower end hardware their machines come with. Disappointed but not that surprising. Apple couldn’t keep up without Jobs before, and they’ve been heading downhill (albeit starting from much higher this time, and not as rapidly as before) since he died.
Like said, peak power draw; is sometimes twice the standard power draw, for mere milliseconds. Without battery, there’s no buffer, and errors can occur. If I were him, I’d run without intel turbo boost, or limit turbo boost speeds.
I suspect that the answer is far simpler – they’re checking the thermal sensor in the battery, and with no signal, they fail safe and assume it’s hot, so reduce power usage.
I agree. It sounds like how I’d program it. ;)
Depends on what you mean mean by ‘normal use’. A lot of people essentially only use their laptops plugged in either via a dock or the charger. The laptop is just a low power consuming desktop to them. The battery is a convenience not a requirement.
I use an old laptop when I’m traveling and it doesn’t have a battery installed. It’s a 10 year old machine I bought used. A new battery is like 50 USD vs a new used laptop for a couple hundred. It’s just not worth it for my use case.
The battery in my laptop exists mainly for when the power gets unplugged or we have a power outage. When it gets unplugged, I immediately plug it back in, and the battery just prevents it from instantly dying. When there’s a power outage, I immediately start closing things and reducing power usage to the absolute minimum that I need to continue working, and if I’m not working, I shut it down. But I also have it plugged into a battery backup, so I don’t actually need my battery for that, and I could easily make the connector for the charger more secure so it doesn’t come out if I didn’t have a battery.
To be fair, when I bought the laptop, it was for use at college, and not every room had a power socket in the right place, but that was my last year, and now it’s a work machine that’s nearly always plugged in.
Besides consumption it is also possible that battery might have thermal protection (it does) and can simply disconnects power output from the batt. Also drawn battery
(voltage below threshold) might not be charging well enough when running system at max consumption.
Same situation with all iPhones, I believe.
I ran into this with a cheap 3rd party battery in a Macbook. I had to buy a more expensive one from OWC that correctly pretended to be a genuine Apple battery so the CPU wouldn’t run super slow.
One thing I noticed on this particular model of Macbook was the battery was made of several Li-Ion pouch cells and directly under the trackpad it had two flaps in the battery case. Quite the clever design so when the battery failed and the cells turned to “spicy pillow mode” the only place for pressure relief was under the trackpad, which broke under the strain.
Better to crack the trackpad than explode :)
Actually, the idea to throttle the CPU when the INTEGRATED battery failed is also a good idea for user safety. A battery can and will burn if overheated. A failed battery may be unstable, so keeping a main source of heat, the CPU, below cool is actually a reasonable thing to do.
Some Dell laptops use the exact same method to throttle when a non-Dell charger is used. On Windows machines ThrottleStop by TechPowerUP works well to dethrottle!
I don’t understand why “without a battery” isn’t a more common use case for laptops.
When I was in college (long ago I admit) I used laptops with crappy or completely dead and unchargeable batteries all the time! A new laptop was outside my student job budget, batteries were expensive and didn’t last as long as the electronics. I even had one where I removed the dead and worthless battery to save weight.
Was having to plug in a problem?
No.
My crappy laptops gave me the freedom to get out of the dorm and still get my school projects done. I used them in cafes, in the school library and the various dining commons. All those places had sockets I could plug in. Most even had Ethernet!
Where did a dead battery prevent me from using it? Outside? Your going to need an e-ink display for that! Unless maybe you want to go outside at night. That’s kind of a weird place and time to be on a laptop isn’t it?
So now I am an adult with a good job and can afford a decent laptop with a working battery. I’m still not sure there is a good reason to bother with such fussy, short lived and potentially fire starting tech though.. Well.. except that modern USB-C connectors feel a lot weaker than those old fat barrel plugs. I’d hate to bump it and break the socket on the computer. I feel safer using the battery and putting it in a safe place it won’t get bumped to charge.
If only the USB power sockets on laptops were built into removable, replaceable modules. If one goes bad just slide it out and pop in a new one, kind of like a PCMCIA card but much smaller.
1- Adding parts means adding cost. repairability like that (typically) adds cost (i.e. screws are more expensive than glue, and more likely to fail. and increase device bulk. bulk is not a product feature with phones).
2- Failure in USB ports on laptops are uncommon. Some phones DO have a separate PCB just for the connector, because it is a common failure point. Power-in on laptops is probably the most common broken connector, and most laptops do have a separate PCB for that (now). I once replaced the power-in jack (just the jack!) on a HP laptop. It was something like 30 steps and 40 screws to get the thing disassembled to access the jack on the motherboard; that was a pain. Manufacturers do learn, but sometimes the cost savings is worth having fewer parts.
Similar design decisions have gone into the removal of 3.5mm jacks on phones: Without powered headphones (i.e. battery, i.e. Bluetooth or usb-c headphones), you can’t do active noise cancellation, can’t electronically apply calibration correction map to drivers to get better response, electronic beam forming for microphones, and a whole host of new useful technologies that wired headphones just can’t offer. There’s a lot of good reasons to give up on the old tech (i.e. CRTs, spinning hard disks, parallel, serial, PS/2, and etc ports)
3- Batteries are now necessary because peak power is way higher than it used to be. My first laptop was probably 30-40W total. Now you have laptops that exceed 100W. https://ark.intel.com/content/www/us/en/ark/products/series/217837/12th-generation-intel-core-i7-processors.html#@Mobile says max turbo on the HX “mobile” CPUs is 157W, and that’s just the CPU! Power supplies, connectors, etc are optimized for cost and “typical use”, not for the 1% use case. also as others have pointed out, assuming that the battery failed and thus run the cpu at minimum speed is not a bad failure strategy. At least most laptops are reasonable in their battery replacement methods now.
Analogous design decisions: For a “USB” battery pack, deciding not to do charge-through makes the design much less complicated, and your power system needs to only handle the power to charge the battery, not to charge the battery AND power all the output ports. With larger USB-C battery packs now starting at 15W and going as high as 245W (and more to come I’m sure), this is not an insignificant design decision.
I guess the tl;dr is that everything is a design decision. Complexity or increased capacity drives cost up, but how likely is that particular use case? How likely is a phone to be useful without a battery?
The points you make are valid, I’d add one more: the issue of connectors. Customers don’t want huge, fat power plugs because they are clunky. Engineers hate them too, as whatever you do they will put strain on the laptop interiors. But a smaller connector, especially combined with a data port, needs small spring contacts. Small spring contacts can be made to run quite a lot of current, but they get warm (locally, I don’t mean the whole plug getting hot). If you can make sure, that the current status roughly constant over longer time, the contacts stay at a constant temperature. If you follow the load spikes, the contacts will get hot and then cool down quite a lot of times throughout a day. Such plug will age fast and fail sooner, so to be reliable out would have to be bigger. Using battery for peak power allows for a physically smaller and still more reliable power plug – besides all what you already said.
“optimized for cost and “typical use”, not for the 1% use case”
Clearly.
But go back and re-read the post you are replying to.
I wasn’t asking why aren’t designers designing for a niche use case.
I was asking why it is supposedly so niche.
Seriously, where are people using laptops so much that they don’t have access to a wall plug?
Or more to the point, why isn’t there a population population of market worthy size who are using their laptops where there IS power available?
I suspect the number one location they are used is on an office desk! Number two is on a couch. Ok, maybe people are lazy about plugging in. But I bet they already have a cellphone charger with the cord lying on the couch already. What’s a second, larger one for a laptop? They are probably going to plug in when it starts to get low anyway! Is it that much worse to be plugged in the whole time vs only after you realize your battery is low?
Then comes all those places I mentioned; cafe, student commons, library – plugs definitely available.
A couple months ago I went on vacation. There were outlets in the chairs for this purpose at the terminal where we waited for our plane!
=================================================================
“Adding parts means adding cost. ”
Yes, yes. I know. Adding 5¢ to the cost for me as a consumer is nothing but multiplied by a million units becomes larger. If consumers wanted a reliable device it would be a selling point and so worth it however 99.9% or more would never think twice of such a thing so it would actually be a loss.
Consumers seem to be the weakest link in the market system. Clearly Apple gets this.
“Failure in USB ports on laptops are uncommon.”
I didn’t really have any problem with USB ports.
“Power-in on laptops is probably the most common broken connector”
Exactly!
“and most laptops do have a separate PCB for that (now).”
News to me, awesome though, I’ll watch for that. Actually, can you go tell that to the people over at NexDock.com?
“Similar design decisions have gone into the removal of 3.5mm jacks on phones”
Not a fan of that decision either. Although it’s growing on me since most headphone plugs don’t fit through the holes in most protective phone cases anyway. So annoying!
“Batteries are now necessary because peak power is way higher than it used to be.”
Um… what point were you trying to make? You think your battery can handle higher peaks than your wall socket? Just how brown is your power?!?!
“you can’t do active noise cancellation”
OMG, life without active noise cancellation on a laptop! How would one check their email and do homework?
Seriously though, that’s not a problem that matters to every person in every situation. Does it even matter to a majority? Is every laptop owner either a telecommuter or a YouTuber now?
And.. you can always add a USB mic if you need something powered. How did we get talking about audio sockets though?
“assuming that the battery failed and thus run the cpu at minimum speed is not a bad failure strategy.”
How do you figure that? If the battery is failed but the machine is running then clearly it is on wall power. Why slow down the CPU? I guess I can see it if the battery IS present but weak.
Seriously though, there isn’t an available GPIO available anywhere on that motherboard that could be used to detect the presence of external power? Ok, that would take a diode and a couple resistors… and not even paying 2¢ for a feature consumers aren’t demanding.
Yah right. I’m sure there is a charge controller chip already in there that is already talking to the CPU via I2C or similar and already could tell it there is power attached with zero additional hardware. We aren’t talking about 2¢ of hardware multiplied by millions of devices, we are talking about $5 of firmware code multiplied by 1 and probably even re-used in multiple models of laptop!
“For a “USB” battery pack, deciding not to do charge-through makes the design much less complicated”
Wow, really all over the place on topic aren’t we? Well, yes, I would expect charge-through to be only available on the more expensive USB battery packs. Are you trying to imply that the charging circuitry on a hundreds if not thousand dollar laptop should be comparable to a $5 USB battery pack from the corner convenience store? Come on, demand more for your money!
“How likely is a phone to be useful without a battery”
To most people… not very. But we were talking about laptops! There are many points where modern phones, laptops and tablets converge and should be talked about as one. This is not such a point.
> If the battery is failed but the machine is running then clearly it is on wall power. Why slow down the CPU?
To reduce the risk that the battery catches fire. See the photo, the battery in Apple devices is pretty much a naked pouch, so if it fails (and inflates) it is time to keep things around it cool. Note, that the battery is NOT user removable, so if the laptop does not detect a battery it has to assume that the reason for this is that the battery failed totally and the battery protection shut it down cold. You may dispute the design decision to integrate the battery as a non-removable part, but the way I see it they made a decision A and then they have to make decision B too.
My work-provided Lenovo laptop works without observable throttling when connected to power and with no battery installed, but a) its power brick is 230W (two hundred thirty watt) and the whole thing weighs a ton and b) the battery can physically be disconnected and so the laptop has a way of telling it has been removed.
“Consumers seem to be the weakest link in the market system.”
Aptly put, sadly this is also my impression.
But I never formulated this thought so succinctly myself.
And i agree on most points, a nice post.
However I see no technical reason to not have charge through in powerbanks. There is power going in to the battery, and power going out, and the net difference of this is what the battery sees. Cases such as the load being greater than the input power or vice versa should be covered already by charge and discharge protections.
Question here is, why can’t the stupid computer run without a battery by default?
Because their batteries handle power draw surges, voltage sags, etc to reduce load on power management circuits, charger load/complexity, etc. Devices are a lot more complex than most people understand.
I’m running my Late 2012 batteryless Macbook Pro at normal fan and OS speed and without any screen dimming. I’ve also used no software to achieve this. The successful hack is to cut the connector board from the extracted dead battery but ensure the 2 thermal sensors and their wires remain in place (cutting only the -+ copper terminals to release the cells from the board). Then screw the board back onto the motherboard socket, tape down the thermal sensors somewhere in the space the battery used to be in and you’ll have a perfectly functioning Macbook.