Why no boot without battery in some portable electronics?

[Pete]

Hi folks

TL:DR;  Why is it many laptops are designed to let you boot them from wall power without the battery but other smaller devices like smartphones never seem to be designed for this?

I am struggling to find this out via Google.  I found someone suggesting it is due to that smartphones are designed to charge by USB and that they would not be able draw sufficient amperage during boot.  However I am sceptical about that idea because I have found it is the same for older mobile phones charged via DC [barrel] plugs.  I am wondering if it is a commercial decision based on planned obsolescence or perhaps a commercial decision based on some technical complexity, e.g. designing in two power delivery modes on already cramped boards.

Just to give you an idea of my lack of knowledge; I am a software engineer who only studied electronics as a teen in school but I regularly watch the EEVBlog and enjoy other electronics channels like BigClive and ElectroBOOM so have a sort of superficial and limited understanding of certain concepts.

I have now had a couple of phones brick themselves due to battery death and I would feel less annoyed about it if I knew there was a sensible reason I cannot power them without going via the charging circuitry/logic.

[Renate]

Well, mostly it's being over cautious.
Yes, Android phones do take a lot of current on booting.
If you're connected to some USB that you don't know if it can really deliver over 100 mA you might have a problem.

For hacking, I've often taken apart batteries and salvaged the battery protection module.
This usually has the thermistor, ID resistor and/or some sort of controller.
You can connect USB power through a dropping diode into where the battery was.

This sort of thing is useful when you have a current monitor plot.
Then you can tell how/whether it's booting up.
I'll post a plot later.

[tooki]

I think it’s because the instantaneous (i.e. pulsed) current draw can be quite significant, even if the average current draw is not. So by using batteries that can provide that pulsed current, they avoid needing to have the entire power subsystem be capable of so much current, adding cost and weight.  Instead, you just use the battery for this, since the device has a battery anyway.

For example, with a class D audio amplifier (which uses pulses to create the waveform), I once measured 8A current spikes (on a meter that captures peaks of at least 1ms duration), while the average current was well under 1A. In testing, a 1A power supply was able to keep it happy when I put a large capacitor in parallel, but without the capacitor, it would cut out occasionally.

P.S. Planned obsolescence is 99% a myth. People confuse it with “engineering down to a cost”, which has very different motivations, and with design tradeoffs, like choosing thinness and water sealing over a battery compartment. That and people simply not understanding that some things wear out.

[ThomasDK]

I think it’s because the instantaneous (i.e. pulsed) current draw can be quite significant, even if the average current draw is not. So by using batteries that can provide that pulsed current, they avoid needing to have the entire power subsystem be capable of so much current, adding cost and weight.  Instead, you just use the battery for this, since the device has a battery anyway.

This is it. The radio power amplifier can take pulses of 2-4A (in GSM mode, maybe 1A in LTE mode) which is more than most chargers can deliver. On top of this you have the main processor booting the operating system. While newer quick-charge type chargers might manage, you still get the inductance of the charging cable to overcome. It could in principle be solved by a couple of big capacitors, but these are expensive and bulky - No point doing it when you have the battery there already.

[tooki]

And that also reminded me that in many phones, some activities use more sustained current than many chargers can provide. The classic example of this is GPS navigation in the car, which will be using GPS full time with the screen on bright, with sporadic data, probably playing music too. This easily exceeds the 1A charging that was standard for a long time, meaning that even with a 1A car charger, it will still be draining the battery! (Just more slowly than if the charger were not connected.) This didn’t really get alleviated until phones got the ability to do fast charging.

[atmfjstc]

Incidentally, some (all?) Macbooks have a similar limitation. They will boot without a battery, but they will be stuck with drastically reduced performance (25%-30% of normal). The official justification is the same: running the machine at full tilt requires so much current sometimes that the adapter can't cut it and the battery needs to be there to shoulder the burden (presumably this refers to instantaneous not average current requirements).

[langwadt]

I think it’s because the instantaneous (i.e. pulsed) current draw can be quite significant, even if the average current draw is not. So by using batteries that can provide that pulsed current, they avoid needing to have the entire power subsystem be capable of so much current, adding cost and weight.  Instead, you just use the battery for this, since the device has a battery anyway.

This is it. The radio power amplifier can take pulses of 2-4A (in GSM mode, maybe 1A in LTE mode) which is more than most chargers can deliver. On top of this you have the main processor booting the operating system. While newer quick-charge type chargers might manage, you still get the inductance of the charging cable to overcome. It could in principle be solved by a couple of big capacitors, but these are expensive and bulky - No point doing it when you have the battery there already.

yeh, much of the hassle in making something like a USB gsm modem is supplying the several amps of peak current for the radio while staying below the 500mA usb limit

[angrybird]

Tooki and ThomasDK are precisely correct for the smartphone case - The inrush current during LTE/GSM radio activity drives this behavior.  Due to the resistive and inductive nature of a charging cable you can not get the required input out of a wall charger.  This starts to be less true with PD devices in high power mode but I'm not aware of any trend in the handheld power space to support these radios without the battery.

[Pete]

Thanks for all the replies.

If the reason for relying only on the battery is peaky current from things like radio and audio amplifiers then it is a shame smartphone vendors do not offer some sort of "limp mode" where anything not strictly required to do diagnostics or data recovery is disabled.  If some already do, it'd be nice to know which ones.

P.S. Planned obsolescence is 99% a myth. People confuse it with “engineering down to a cost”, which has very different motivations, and with design tradeoffs, like choosing thinness and water sealing over a battery compartment. That and people simply not understanding that some things wear out.

Perhaps better said as 99% misunderstood because planned obsolescence is very much real -- a product designer in my office once showed me something he was asked to design such that it have to be mounted on a plastic bracket... his specification stated that the bracket must break after a single usage and the motivation was that the client's customer (a gov't organisation) would have to buy tens of thousands of those things.  Granted, in the smartphone battery example it is probably more like "tacit obsolescence" because Sony didn't see fit to ship the phone with decent protections against over-discharge.

[tooki]

Incidentally, some (all?) Macbooks have a similar limitation. They will boot without a battery, but they will be stuck with drastically reduced performance (25%-30% of normal). The official justification is the same: running the machine at full tilt requires so much current sometimes that the adapter can't cut it and the battery needs to be there to shoulder the burden (presumably this refers to instantaneous not average current requirements).

I think it's not all models. IIRC, early MacBooks could operate without a battery. (Though it's highly inadvisable, since without the battery, it's extremely top-heavy, and it's obviously very easy to unplug the MagSafe power plug, which is obviously bad if there's no battery power to fall back on!)

And yeah, it'd be due to the instantaneous current, since the AC adapter is capable of supporting the average current of simultaneous full operation and charging.

[tooki]

P.S. Planned obsolescence is 99% a myth. People confuse it with “engineering down to a cost”, which has very different motivations, and with design tradeoffs, like choosing thinness and water sealing over a battery compartment. That and people simply not understanding that some things wear out.

Perhaps better said as 99% misunderstood because planned obsolescence is very much real -- a product designer in my office once showed me something he was asked to design such that it have to be mounted on a plastic bracket... his specification stated that the bracket must break after a single usage and the motivation was that the client's customer (a gov't organisation) would have to buy tens of thousands of those things.

No, it's a myth 99% of the time, as I said, and your example is precisely an example of the remaining 1%. (I'd also argue that if it's not supposed to be a consumable or a sacrificial breakaway part, then it's an example of outright fraud.)

In the vast majority of products, manufacturers want it inexpensive, but not fragile, because they're responsible for the warranty fulfilment. Not to mention that if products consistently lasted only as long as the warranty, the company's reputation would ultimately reflect this. (The fact is, most products still generally outlive their warranties significantly.) If companies tried to make them last, say, 2 years instead of a natural lifespan of 5, then natural variation would mean that they'd be pushing a much larger "tail" of the curve back into the warranty period, which could end up costing them a fortune, not to mention that people burned that way would be much less likely to buy from them again.

Granted, in the smartphone battery example it is probably more like "tacit obsolescence" because Sony didn't see fit to ship the phone with decent protections against over-discharge.

Natural obsolescence is a thing tons of people don't understand. Like… technology progresses, and it's unreasonable to expect it to old stuff.

As for the batteries, given that Sony is one of the pioneers of LiIon, and a giant in that industry, and never a cut-rate company, the chances that they did too little is, well, nonexistent. Chances are, what you're considering the lack of decent protections is in fact those protections doing their job.

[angrybird]

There has been some discussion of offering this "limp" mode due to the desire from some car manufacturers for the "phone as a key" implementation in their cars.  The concern is that if the user's phone dies, they now cannot start the car with their phone until it has been charged by some amount.  This would require some cooperation from the phone manufacturers and/or changes in in the way the "dead battery" case is handled by the phone.  One of the problems with this coordination is that historically the car manufacturers have cycled their designs every 3 years, while the handheld market is more like 6 months.  This is why the car technology has always lagged behind the handheld technology by years, or more!

[tooki]

I know that newer iPhones have a limp mode like that for transit passes and the like. I assume it'd be trivial to expand it to car keys.

(And yeah, my understanding is that car electronics have historically been frozen at about 3-5 years before the release date of the car. I assume it's been reduced a bit these days, but probably to maybe 2-3 years.)

[angrybird]

It still lags quite a bit, a big problem is the typical 3-year automotive design cycle.  To attempt to compensate for this, OEM's have tried to become "forward looking" in their RFQ's but in many ways this is "guessing" as they are attempting to target what the requirement will be years in the future.  This has driven many conflicting and very difficult requirements for the Tier 1 and this has resulted in a continuous negotiation between the OEM and Tier 1 on project goals as the infotainment system designs move forward.  Often, the C sample design has little resemblance to the A sample.  The negotiation often involves the Tier 2 (thru both the Tier 1 and OEM) who are more closely connected to the handheld industry, who have better insight into what the actual market trend will be.  All of this negotiation coincides with the other difficult requirements for modern automotive applications - EMI, safety, power consumption, cost - It's an incredibly exhausting market space.  I have no envy for my many friends at Tier 1's worldwide - They are super heroes in their own way!

[Renate]

I wanted to grab a plot on some Android startup.
Immediately to hand I only had a Nook, an OMAP 3621.
It was the only thing I could find without looking that had wires on it.
Nope, this is not representative, but you might get some idea of what a startup looks like.
This is the current draw out of the "battery".
The first peak is plugging it in, the second peak is a (too short) power on press, the third peak is booting.