USB 5V to 19V

[System Error Message]

With the massive capacity of battery banks nowadays, 99% of them only have USB out, even some of my foldable solar panels too. For example a 20Ah battery pack has more than 60Wh which can charge an average laptop to full but only outputs at usb 5V but can only output more than 3A total over both usb ports (more than 15W). Even my solar panels have multiple usb ports as well each capable of 2A.

So how would i go about to efficiently convert a range of 4-6V DC to 19V DC over multiple USB inputs to provide sufficient amps to power a laptop and charge it?

[floobydust]

Laptops use a lot of power at 19V; 2A to 3.5A so a typical 5V to 19V boost-converter 85% efficient would draw 9-16A!

[Fungus]

Are all the 5V outputs wired together internally?

If so you can just connect them in parallel to a high power DC booster and raise the voltage to 19V.

eg. http://www.ebay.com/itm/201537809116

Laptops use a lot of power at 19V; 2A to 3.5A so a typical 5V to 19V boost-converter 85% efficient would draw 9-16A!

Yep. You'll need to carefully measure your charging current before you can choose a suitable voltage booster.

It probably varies over the charge cycle so take readings at 5 minute intervals (or use a meter with min/max).

If you're buying on eBay then make sure you get plenty of spare capacity. Don't expect it to run at 100% power rating without overheating.

[Ian.M]

The problem is managing the input power so none of the USB charging ports are overloaded.  The most scalable solution would be to use as many boost converters as it takes to do the job, one per USB port, each with  output limited to 19V and input current limited at 2A or whatever current they can draw up to that without USB Vbus drooping excessively.  You'd probably need some smarts in each for that so they don't persistently overload their host power port.   

All their outputs would be paralleled.  Statistically they'll all have slightly different output voltages, so the one with the highest voltage 'wins' and the others.idle or go to low power standby.  When it starts current limiting the output voltage will droop fractionally and the next highest will cut in and start taking up the excess load, and so on till they are all contributing.

However its still an absolutely sh!te idea compared to starting with a 5S bank^* and only having to boost a little bit with a single dumb 5A converter.

* Or a 6S or higher bank but then you need to be absolutely certain the buck converter's crowbar circuit can handle the pass transistor going dead short or the laptop will let its holy smoke out.

[System Error Message]

i dont think you actually read what i wrote. I wrote multiple usb ports, also meaning multiple battery packs.
So if each battery packed provided 5V at 3A, and i had 4 of it, i would have 60W going in which is what a laptop with no dGPU but decent CPU uses. Even my LED projector's power adapter is rated at around 60W.

[Ian.M]

Oh everybody read what you wrote and are pretty much unanimous on the undesirability of starting from 5V when trying to power a 19V laptop.  The problem is efficiency  - with such a low input voltage^*, for high output current you need a lot of MOSFET area, thick copper wire windings, possibly even foil windings to get enough CSA and big ferrite cores to handle the high peak currents without saturating.  Even so efficiency wont be great - I doubt you'll see the high side of 80% (so your 60W (12A) in becomes only 50W (2.5A) out), from it and you have the problem of combining the input supplies without overloading any and without significantly increasing the losses (hence my suggestion above - its much more manageable when each converter only has to deliver up to 450mA.    .

* A laptop car charger that boosts from 12V-14V in to 19V out is bad enough . . .

[System Error Message]

Oh everybody read what you wrote and are pretty much unanimous on the undesirability of starting from 5V when trying to power a 19V laptop.  A laptop car charger that boosts from 12V-14V in to 19V out is bad enough . . . .

If the number of amps in the input wasnt an issue, how practical or possible would it be to step up 5V to 19V? I've seen circuits that step the voltage up or down so it would be interesting to convert one commonly used voltage to another. Battery packs with usb in/out have to convert voltages as lithium is 3.7V nominal and takes 4.2V to charge. How would the circuit look like?

[RGB255_0_0]

You could use a lithium screwdriver charging/power PCB. Many are made for 18V or so and rewire the battery bank with the balancing connections and add a barrel jack.

Or buy a battery bank with barrel jack meant for charging laptops.

[Fungus]

If the number of amps in the input wasnt an issue

It's a major issue.

how practical or possible would it be to step up 5V to 19V? I've seen circuits that step the voltage up or down so it would be interesting to convert one commonly used voltage to another. Battery packs with usb in/out have to convert voltages as lithium is 3.7V nominal and takes 4.2V to charge. How would the circuit look like?

I wouldn't like to build such a circuit myself and I wouldn't use a USB battery pack to do it. Any USB battery pack has a 5V booster inside it and that booster is part of the problem. Much better to start with the raw LIPO battery.

Best way to do this: Search google for "19V LIPO battery" and buy something prebuilt. Yes they exist.

Alternate method: Get a second battery for your laptop.

[Ian.M]

<=please reread above, I've added more background=>

If the number of amps in the input wasn't an issue, how practical or possible would it be to step up 5V to 19V? I've seen circuits that step the voltage up or down so it would be interesting to convert one commonly used voltage to another. Battery packs with usb in/out have to convert voltages as lithium is 3.7V nominal and takes 4.2V to charge. How would the circuit look like?

Trivially, I've done that from 5V in to 12.5V, 22.5V and 25V (switch selectable) at up to 50mA out for an EPROM programmer.  It was over 20 years ago, but IIRC it was just a TL497 (500mA peak switching regulator chip), a bobbin choke,and a few caps and resistors and took about 1 sq in of PCB area.

For higher power levels you'd want something more modern, optimised for a higher switching frequency (to reduce the size of the choke) probably with an external MOSFET for the pass transistor and an external Schottky diode for the catch diode, as its easier to get the waste heat out of discretes with nice large heatsink tabs.

However the TL497A is still in production in PDIP and with its typically low switching frequency (variable, about 30KHz max boosting from 5V to 12V) is forgiving enough to be able to get going on a solderless breadboard, and T.I did a fairly good design guide for it, so IMHO its a reasonably good introduction for a total novice to the world of switching regulators. 

Just remember there are much newer, much better more efficient ones on the market now, but they typically are SMD in packages that have to be reflowed and require very careful PCB layout due to the much higher switching frequencies (can be >1MHz) and currents involved so aren't at all breadboard friendly as its impossible to make an effective breakout board that doesn't also carry the magnetics, decoupling caps and external pass transistor and diode (if used).

[System Error Message]

Trivially, I've done that from 5V in to 12.5V, 22.5V and 25V (switch selectable) at up to 50mA out for an EPROM programmer.  It was over 20 years ago, but IIRC it was just a TL497 (500mA peak switching regulator chip), a bobbin choke,and a few caps and resistors and took about 1 sq in of PCB area.

For higher power levels you'd want something more modern, optimised for a higher switching frequency (to reduce the size of the choke) probably with an external MOSFET for the pass transistor and an external Schottky diode for the catch diode, as its easier to get the waste heat out of discretes with nice large heatsink tabs.

However the TL497A is still in production in PDIP and with its typically low switching frequency (variable, about 30KHz max boosting from 5V to 12V) is forgiving enough to be able to get going on a solderless breadboard, and T.I did a fairly good design guide for it, so IMHO its a reasonably good introduction for a total novice to the world of switching regulators. 

Just remember there are much newer, much better more efficient ones on the market now, but they typically are SMD in packages that have to be reflowed and require very careful PCB layout due to the much higher switching frequencies (can be >1MHz) and currents involved so aren't at all breadboard friendly as its impossible to make an effective breakout board that doesn't also carry the magnetics, decoupling caps and external pass transistor and diode (if used).

anything with 20Ah or more that ouputs 19V costs around £100.seems like itd be cheaper to build my own. since im using from lithium ions, how best to approach this? The batteries do need to be recharged too and i want to stack many batteries in parallel.

[Ian.M]

You are suffering from the fallacy that you think you can build a one-off with comparable features to and cheaper than the average retail price of a commercial product.   "I cant afford a diamond ring for my girlfriend - I know what: I'll mine my own diamonds and gold, and make it myself and save lots of money!" 

Even if you get good quality batteries for free you are unlikely to be able to beat that £100 price to build and test the electronics and you could easily end up out £1000 with nothing working to show for it. It only takes ONE malfunction to fry the laptop you are intending to charge . . . .

Unless the journey of discovery is the major purpose of the exercise^*, you'll be better off flipping burgers or washing windscreens till you can buy a 19V out powerbank off the shelf.

* and in that case you'd better start small with trying to boost the voltage to charge something low-current and not particularly valuable.  e.g. build a boost circuit from 2xAA to charge a £1 PoundLand single 18650 cell USB powerbank as you'll learn just as much from burning up £5 of parts as you will from burning £50 of parts.

[System Error Message]

You are suffering from the fallacy that you think you can build a one-off with comparable features to and cheaper than the average retail price of a commercial product.   "I cant afford a diamond ring for my girlfriend - I know what: I'll mine my own diamonds and gold, and make it myself and save lots of money!" 

Even if you get good quality batteries for free you are unlikely to be able to beat that £100 price to build and test the electronics and you could easily end up out £1000 with nothing working to show for it. It only takes ONE malfunction to fry the laptop you are intending to charge . . . .

Unless the journey of discovery is the major purpose of the exercise^*, you'll be better off flipping burgers or washing windscreens till you can buy a 19V out powerbank off the shelf.

* and in that case you'd better start small with trying to boost the voltage to charge something low-current and not particularly valuable.  e.g. build a boost circuit from 2xAA to charge a £1 PoundLand single 18650 cell USB powerbank as you'll learn just as much from burning up £5 of parts as you will from burning £50 of parts.

Im not trying to build one much cheaper, im trying to build a couple a little cheaper. I expect it'd cost a bit but not as much, reason being i need more capacity than what is sold. I plan to use lithium ions as they're cheaper per capacity and size/weight doesnt matter to me.

I already have spent the £1000 before for the tools. I have the tools but im not sure what components i need entirely and i've been running simulations on DC to DC voltage boosting. I could easily set up a test circuit to take 19V, some load to simulate the required amps and a decent oscilloscope to look at the output voltage waveform at various loads. The problem is that i dont have a clue on how voltage boosting and joule thieves work and i want to build decent circuit that outputs well and implement some good safety as well. A lot of the devices i see that are available arent exactly practical. Charging them takes forever, some include an AC inverter which would make it less efficient and a number of them i would question the numbers on the back of the device.

So im not suffering from any fallacy, i want to build one as its something that'd be useful to me and at the same time do more electronics. Im a beginner in electronics but i want to learn more about the electrical side of things. Getting an arduino and hooking up components isnt difficult but i also want to learn to build circuits that will reliably and safely power other DC devices.

back to the original question, Each 5V charger could have 2 USB ports to output a total of 3A or more at 5V. The problem is that many USB chargers like to implement qualcomm's quickcharge which reduces the voltage for more amps but im not sure if this feature requires there to be a chip. So if i had 4 USB in ports, each port using a cable that connects to 2 USB out ports to get more amps from each charger, and a smart chip to check that the voltage is within a certain amount given the amps, would they need to be seperately balanced after or could they be input into a single boost converter? Should i run 2 boost converters in parallel and balance the output to reduce the strain on the components?

If i use batteries, is it better to have 10 lithiums in parallel, or 5 pairs of 2 series connected lithiums? What are the best ways to balance batteries, multiple inputs in parallel and if needed outputs if having multiple boost circuits in parallel to reduce heat and component strain (possibly for better efficiency)?

What is it that determines the efficiency of the circuit? Does a wider voltage gap reduce efficiency? Higher amps and temperatures?

If i use batteries, How would i recharge them given different voltages? Would i need another circuit or can i reuse the boost circuit?

What sort of safety suggestions and methods to look at?

[Kevman]

i dont think you actually read what i wrote. I wrote multiple usb ports, also meaning multiple battery packs.
So if each battery packed provided 5V at 3A, and i had 4 of it, i would have 60W going in which is what a laptop with no dGPU but decent CPU uses. Even my LED projector's power adapter is rated at around 60W.

Couldn't you just wire them in series? You'd get 20v at 3A then. 

I'm not really sure how the DC-DC converters in them would take to that, but I don't see any grounding issues if you wire your circuit properly and have a small circuit to disconnect power when one goes flat (likely 4 dropout detectors- one on each input).

[System Error Message]

i dont think you actually read what i wrote. I wrote multiple usb ports, also meaning multiple battery packs.
So if each battery packed provided 5V at 3A, and i had 4 of it, i would have 60W going in which is what a laptop with no dGPU but decent CPU uses. Even my LED projector's power adapter is rated at around 60W.

Couldn't you just wire them in series? You'd get 20v at 3A then. 

I'm not really sure how the DC-DC converters in them would take to that, but I don't see any grounding issues if you wire your circuit properly and have a small circuit to disconnect power when one goes flat (likely 4 dropout detectors- one on each input).

How would you wire usb battery packs in series? If you connect the output together you could end up passing too much voltage that you could blow the circuits in them or trigger any protection they have.

[Ian.M]

Wiring USB battery banks in series is a fairly horrific idea.  Good quality packs will be protected against such abuse and at worst will trip their protection circuits and brick thmselves, but cheap Chinese clone packs may not have effective protection circuits. Even with a high current Schottky diode across each to prevent polarity reversal, it sounds like a potential recipe for an incendiary bomb.  Paralleling them without load balancing is also high on my dumb ideas I'd rather watch with binoculars than close up list.   

If the device you want to power has a universal input power supply (typ. AC/DC 100V - 250V), its tempting to simply design and build a 40 cell series pack giving you 144V nominal, 120V at the 3V/cell discharge limit.  That's enough to run an efficient 120W laptop PSU for 1H per AH of an individual cell, at full output.  If the laptop is fully charged and idling you'd get considerably longer.

Of course the cells will need balancing.  This can be handled on a per cell basis with a cheap low pin count MCU which must have an ADC and internal bandgap reference and work over the range 2.7V to 4.5V, driving a MOSFET dumping excess charge into a resistor.    Each balancing MCU would have a comms link to the MCU of the cell above it and the cell below it in the chain so you don't have to handle any high voltage differences when the master MCU is checking the pack status.   

The master MCU would monitor the pack status (including all cell voltages) and load or charging current and disconnect the pack on over-current, at the full charge and full discharge voltage limits or if a cell failure is detected.,or on loss of comms with the balancing MCUs.
It would also provide a PWM output to fine trim the charging circuit's voltage limit, and monitor the charging voltage.  Probably the easiest way of powering it without unbalancing the pack is to give it its own LiPO cell, which obviously wont need anywhere near the capaacity of those in the main pack.  If it looses power the pack is disconnected until its put back on the charger.

The charger would just be a CC/CV PSU from *whatever* energy source - probably AC mains, but  boosting from the bus voltage of a solar system could also be an option.   The master MCU can turn it on or off and trim the CV mode voltage limit so it can hold the string at the correct voltage during the saturation charge.

For safety the pack should be assembled in blocks of no more than 10 cells, (to keep the max terminal voltage under 50V) and arrangements made for each block to disconnect for maintenance befoore the cells can be accessed.  The balancing circuits would stay with the block.  If any cells fail, they can be manually removed from the pack and bypassed so the pack can be reconnected with a lower output voltage until they can be replaced (subject to the trim range of the charger).

[System Error Message]

If the device you want to power has a universal input power supply (typ. AC/DC 100V - 250V), its tempting to simply design and build a 40 cell series pack giving you 144V nominal, 120V at the 3V/cell discharge limit.  That's enough to run an efficient 120W laptop PSU for 1H per AH of an individual cell, at full output.  If the laptop is fully charged and idling you'd get considerably longer.

Of course the cells will need balancing.  This can be handled on a per cell basis with a cheap low pin count MCU which must have an ADC and internal bandgap reference and work over the range 2.7V to 4.5V, driving a MOSFET dumping excess charge into a resistor.    Each balancing MCU would have a comms link to the MCU of the cell above it and the cell below it in the chain so you don't have to handle any high voltage differences when the master MCU is checking the pack status.   

The master MCU would monitor the pack status (including all cell voltages) and load or charging current and disconnect the pack on over-current, at the full charge and full discharge voltage limits or if a cell failure is detected.,or on loss of comms with the balancing MCUs.
It would also provide a PWM output to fine trim the charging circuit's voltage limit, and monitor the charging voltage.  Probably the easiest way of powering it without unbalancing the pack is to give it its own LiPO cell, which obviously wont need anywhere near the capaacity of those in the main pack.  If it looses power the pack is disconnected until its put back on the charger.

The charger would just be a CC/CV PSU from *whatever* energy source - probably AC mains, but  boosting from the bus voltage of a solar system could also be an option.   The master MCU can turn it on or off and trim the CV mode voltage limit so it can hold the string at the correct voltage during the saturation charge.

For safety the pack should be assembled in blocks of no more than 10 cells, (to keep the max terminal voltage under 50V) and arrangements made for each block to disconnect for maintenance befoore the cells can be accessed.  The balancing circuits would stay with the block.  If any cells fail, they can be manually removed from the pack and bypassed so the pack can be reconnected with a lower output voltage until they can be replaced (subject to the trim range of the charger).
[/quote]

what if i want to use 25 pairs of cells in parallel? So they'd be 6-7V and have an MCU for each pair rather than every single cell? My laptop has a 240W adapter and a 99whr battery. I could use high gauge wires and higher grade components to handle the amps too.

[Ian.M]

25 series connected parallel pairs only gives you 75V at 3.0V/cell end of discharge and that's not enough for most universal PSUs.   35 pairs would work but lower voltages stress the PSU more as it has to draw more current to maintain the same output power hence my choice of 40 above..   

The reason for using one MCU per cell (or parallel group of cells) is the extra cost and complexity of making it monitor several cells in series.  If you only connect it to one cell, it doesn't need a regulator and doesn't need potential dividers and buffer amps to measure the cell voltages.  The balancing MOSFET can be driven direct from an I/O pin without level shifting and the comms links are each a single transistor and resistor, so the total is 8 components per cell balancing circuit,  1 MCU, 3 transistors, 3 resistors and a decoupling cap.  Also the quiescent current can be as low as the MCU's lowest sleep mode current with I/O triggered wakeup.

You'd do all your early development work with a 4S test pack outputting a nominal 14.4V, as that's enough for a proof of concept of the upwards and downwards comms link for each balancing MCU and for the master MCU at the bottom and the high side pack isolation switch at the top.

p.s. Please fix or remove the quote in your reply #16 above.

[System Error Message]

25 series connected parallel pairs only gives you 75V at 3.0V/cell end of discharge and that's not enough for most universal PSUs.   35 pairs would work but lower voltages stress the PSU more as it has to draw more current to maintain the same output power hence my choice of 40 above..   

I meant the other way round, 2 cells in series, 25 of those connected in parallel giving 6-7V per pair and 25 of those in parallel giving both the amps and capacity, then boosting them to 19V DC, switched with a transistor to give as DC like voltage as possible and connected to the laptop rather than the PSU.

So if this was the case, does that mean i'd need 2 MCUs, connect 25 cells in parallel, another 25 in parallel and put them in series?

PSU may be 240W but the laptop will accept various. It does detect the wattage of the PSU attached so i wonder if it does that the same way that phones do or if there is some chip that tells it.

[mikeselectricstuff]

Or just buy a power  bank specifically designed for the  job
https://www.amazon.com/Intocircuit-26000mAh-Capacity-Portable-External/dp/B00BB5VQCE

[Ian.M]

Sounds like a mess - you'd need to balance each pair individually.  It would be a lot simpler to parallel more individual cells and then build series strings.  Also, as high currents make everything *MUCH* harder you'd want four in series for 14.4V nominal into a boost converter or seven or eight  for a minimum of 21V at full discharge into a buck converter.

The laptop charger ID is yet another hurdle to overcome - you may need to buy a dead charger and tear it apart to find the ID scheme and how to emulate it.

[System Error Message]

Or just buy a power  bank specifically designed for the  job
https://www.amazon.com/Intocircuit-26000mAh-Capacity-Portable-External/dp/B00BB5VQCE

Power output is the problem. My laptop only gives me 10 hours at best, 8 hours during normal use and 2 hours gaming. It also charges up very fast for its 99whr battery. That battery cannot output enough amps for it to be practical enough to use over long distance travel. So during normal use it would have to reliably provide voltage at the required amps. Since the battery charges within 3 hours, that would mean the batteries must provide 10A alone just to charge the battery.

I could run more simulations on different voltages to see the efficiency of boosting the voltage and compare the various battery setups. I guess the only choice is to either run all in parallel from 3+V or have pair up 4 or more to get better efficiency. My main concern is just the safety from the number of batteries and amps.

[Buriedcode]

Sorry if this has been mentioned before but... whilst you seem set on making your own external laptop charger (nothing wrong with that really) perhaps its more of a usage problem. 

Laptops, when running off battery will purposefully under clock everything obviously to save power and extend battery life.  But laptops that are fast enough to be responsive - or play games which are very graphics/cpu intensive, are as we all know power hungry.  When the charger is plugged in that provides power to the laptop, as well as power to charge the batteries.  It also shifts up to a higher power profile - which can be tweaked for CPU throttle, screen brightness etc.. but for the most part it will *always* use more power when the external power is provided. 

So you're taking a battery pack, boosting it to 19V, which is then going to be stepped down to charge the batteries, as well as stepped down to the nominal battery voltage for the actual laptop power, *and* the laptop will draw more current.  I realise if one wants longer battery life - there really aren't many options, so building an external 'power pack' seems logical, but one would need a battery pack much larger than the original to make much of a difference at all.

That leads me to think other solutions, be they more or less practical. 

 - A spare battery or two?  Laptops give plenty of warning that the battery is dying, even automatically backing up and saving, and if the laptop is reasonably powerful, shutting down and booting will take <10 seconds. Making the whole change of battery take <30 seconds.

 - Not playing games when its on battery?  I understand the need for portability but I never understood those who used laptops for gaming.  Battery life, weight and size (portability) all go against performance which games need.  This is why 'gaming laptops' aren't really laptops at all - just semi-portable flat desktops that can burn legs and weigh a ton.

 - Having another tweak of the power profile. Uninstalling unnecessary processes, reducing graphics acceleration etc.. 10 hours sounds like an excellent battery life for a laptop so it sounds like it is well optimized (I rarely see laptops - not tablets or notebooks - go over 6 hours even when idle).  With 8 hours for normal use being fantastic.  I'm not sure you can expect anything more? 

Ultimately its all a compromise, and I commend you for thinking about an external 'pack' but I can't help thinking its a complicated and somewhat expensive solution to the problem all large laptops have. For one extra charge (another 8 hours use) I'm sure solutions already exist.  For multiple charges its going to get big and expensive, and a ballache to make.  I'm sure there are situations where it would be handy - like being miles away from any mains or vehicles, but in those instances do you need a powerful laptop?  I have to ask, what sorts of things do you need the laptop for when you're in remote locations? (genuine question, I'm curious).

[System Error Message]

That leads me to think other solutions, be they more or less practical. 

 - A spare battery or two?  Laptops give plenty of warning that the battery is dying, even automatically backing up and saving, and if the laptop is reasonably powerful, shutting down and booting will take <10 seconds. Making the whole change of battery take <30 seconds.

 - Not playing games when its on battery?  I understand the need for portability but I never understood those who used laptops for gaming.  Battery life, weight and size (portability) all go against performance which games need.  This is why 'gaming laptops' aren't really laptops at all - just semi-portable flat desktops that can burn legs and weigh a ton.

 - Having another tweak of the power profile. Uninstalling unnecessary processes, reducing graphics acceleration etc.. 10 hours sounds like an excellent battery life for a laptop so it sounds like it is well optimized (I rarely see laptops - not tablets or notebooks - go over 6 hours even when idle).  With 8 hours for normal use being fantastic.  I'm not sure you can expect anything more? 

Ultimately its all a compromise, and I commend you for thinking about an external 'pack' but I can't help thinking its a complicated and somewhat expensive solution to the problem all large laptops have. For one extra charge (another 8 hours use) I'm sure solutions already exist.  For multiple charges its going to get big and expensive, and a ballache to make.  I'm sure there are situations where it would be handy - like being miles away from any mains or vehicles, but in those instances do you need a powerful laptop?  I have to ask, what sorts of things do you need the laptop for when you're in remote locations? (genuine question, I'm curious).

Spare battery or 2? Not possible as its inside the laptop. I dont expect to be carrying a screw driver and operating on my laptop every 8 hours not to mention charging it would be a problem too.

I already use a tweak thats how i get 8 hours of battery life for normal usage.

I already know how many watts the laptop uses maximum. If i combine the CPU,IGP,GPU by itself will be around 160W (combining TDPs), 10W-20W more for the screen any other things, 30W to charge the batteries.

The battery pack can be combined with a solar panel and a super capacitor array to handle peaks, i could have different common voltages out rather than selectable voltages like what current ones have and i could have multiple DC out, a huge pack (seriously check the price of lithium ions). Size and weight dont matter for someone who already has a large and heavy laptop. What really limits the size is the wattage limit for flying. 99whr is the legal limit for a single battery pack for flying. I could easily have a single circuit and multiple packs for it to draw from.

All computers by default save power regardless if on battery or not. They save even more power on battery but thats determined by how comfortable you are on turning off things like wifi, lowering screen brightness and so on. I already undervolt the CPU and IGP.

Im not doing this because it is practical, im doing this to provide myself portable 19V DC and learn more about the electrical side of things. Im not doing this because the market is too expensive or that i see one and want to build my own, i just so happen to have a bunch of tools and components sitting around to try things. I also want to find some decent use from my projects so definitely not using lead acid batteries as a cheap battery option for portability.

[xtraveler]

Hi System Error Message and pals'

so actually you can do it,
You need a powerbank/ external battery charger with usb/microUSB in and 19v DC out, (like this for example:
 https://www.aliexpress.com/store/product/Vinsic-30000mAh-Quick-Charge-3-0-Powerbank-2-4A-Dual-Output-With-Type-C-Port-External/835329_32792426820.html), and you can charge your laptop via the 19DC port and charge the powerbank via foldable solar panels (or battery banks).

I know it is not the cheapest solution, but affordable and SAFE'