#250

[@vylbird8014]

I think you got the abbreviation wrong. It's Uninterruptable Power Supply.

[@AndreasSpiess]

Generally: You are right. But this module can do much more.

[@connecticutaggie]

@@AndreasSpiess Maybe you should call it a UUPS

[@jafinch78]

Mr. Spiess was clear about using the abbreviation in the title... looks APA format also. owl.purdue.edu/owl/research_and_citation/apa_style/apa_formatting_and_style_guide/apa_abbreviations.html

[@jafinch78]

Excellent details and presentation also. Thank you sir! Sure is far better (including video quality and hygiene) than the Sepic Cuk UPS kit I made: kzbin.info/www/bejne/bJu8pWxphLqLiQ

[@jamesfirth7795]

@@AndreasSpiess he isn't "generally right" he's specifically correct. How is this "universal"? Will it power my home? My car? No, it won't. If you think capable of powering a handful of devices makes it somehow special and "universal" I've got some bad news for you, any old UPS is capable of powering anything with a kettle lead or even more if you branch from the kettle lead to a plug socket, making it a hundred times more "universal" by your measurement.

[@kjur18]

I took a different approach on this topic, I made a small low power module that can charge a battery and power a device separately. for this I used a P channel mosfet (between boost and battery) and extra shottky diode. My module contains TP4057, smaller sot23-6 version of TP4056, some protection designed to cut off at about 3V, boost converter is MT3608 set for whatever voltage you need. When I connect usb lead, charger charges battery, mosfets disconnects boost from battery, and current flows from usb lead to boost part through a diode. As I made this to power up a DSO Shell, it's small, with everything on one side, but I think that mosfet swiching thing can be easily scaled up for higher current demands.

[@arduinosynod1571]

Post the schematic please

[@kjur18]

@@arduinosynod1571 Here it is: easyeda.com/kjur18/Li_ion_charger__boost-55ecc63cf3e34ea082e9d6b1d53dafa5

[@SuperBrainAK]

that is a much better design! you want to eliminate the number of power conversions in your circuits if possible. so the diode sends 5v directly to the loads dropping only 0.5v (could be less with a schottky) where with the power going through a TP4056 you get at most a 0.8v drop (not to mention the excessive wear on the cell when acting as a capacitor). this is always going to be a higher voltage than the battery so the diode in the mosfet will not conduct. Once usb power is gone then the mosfet will conduct the battery voltage directly to the load with very little voltage drop. I just wonder if there is a way to replace the diode with another mosfet, I think the trouble is getting it to turn on and off properly.

[@kjur18]

@@SuperBrainAK I added possible solution for this, it might work. N channel mosfet with gate with a pull down resistor tied to the source connected to a usb power, and drain connected to a boost part of the schematic. This way when there is no usb power, mosfet is turned off, and when you connect power to it, that mosfet should open and conduct without any problems. Probably.

[@HMPirates]

@@kjur18 You can test one if the ideal diode modules you can get off eBay..

[@gtrangoni]

Thanks a lot. I've buyed following your links one shield and two batteries. I've received them now and it's time to test. I'm a fan of your videos and admire your sense of humor. Your knowledge has been of great help for to me and my final career project about BLE beacons for monitoring pacients in healthcare centers. Best regards from the north of Argentina!

[@AndreasSpiess]

Thank you for your nice words. I am glad my 2 cents helped you in you career!

[@gtrangoni]

@@AndreasSpiess Your contribution is very valuable. Thank you! See you in the next episode. Here is summer and fresh ideas are welcome.

[@MagivaIT]

this is great, i really hope you do a follow up video to..... 1) build the auto-disconnect protection 2) daisy chain these together in parrellel 3) build a circuit to use existing powerbanks and treat them as UPS love your channel

[@AndreasSpiess]

The first might become true for an RPi. I already did one in the past (similar).

[@5speedfatty]

Man this is twice now ive been looking at a product on ali and you are doing a video on it. i love you for this man, cant wait to see the results of this video.

[@youliantroyanov2941]

Twice is a low count :) I am on my third dozen already...

[@5speedfatty]

@@youliantroyanov2941 jeeze man, thats alotta projects. just goes to show the point though, the guy with the swiss accent is on it with the DIY electronics and sensors.

[@RTmadnesstoo]

Yeah, I just got mine a couple weeks ago and haven't played with it much yet.

[@NabilTouchie]

Thanks for saving me time money and frustration, I needed 3A 5V and got excited to see this module in a online store but suspicious of the aspects, so your video was really helpful

[@AndreasSpiess]

Glad I could help!

[@aspzx]

Same here. Have you found a solution? This video is 3 years old now so I wonder if there are new options available.

[@ramblurrr]

@NabilTouchie@@aspzx Did you find a battery charger/load sharing solution for 3A 5V?

[@aspzx]

@@ramblurrr No. I can't even remember what I needed it for.

[@wilman_studio]

Great video. The correct way to deal with the issue of charging the battery whilst connected to a load is to run the Vin through a schottky diode to the boost converter bypassing the LiPo charger. The LiPo cell is then connected with P-channel mosfet to the boost converter with the gate of the mosfet also connected to Vin. This will allow the battery to be charge independently of the load but will also switch the battery into play as soon as there is no input power. It's not good practice to sap the load power off the battery charger as you highlighted.

[@AndreasSpiess]

What are the advantages for the application described here?

[@wilman_studio]

Andreas Spiess By drawing power for the load separately from the LiPo charger whilst charging it would overcome the issue you mention in the video where the load of say a Pi doesn’t have enough current to function properly if the battery is charging but at a almost empty state. There’s a fantastic write up by Zak Kemble blog.zakkemble.net/a-lithium-battery-charger-with-load-sharing/

[@thesimbon]

At least you have confirmed my findings about the unit shutting down at around 2A on 5V. Thanks

[@AndreasSpiess]

You are welcome!

[@MartinBgelund]

Thanks for testing this module's UPS capabilities! I bought 4 of these after I saw them recommended on Ralph S. Bacon's KZbin channel, for using them as UPS' on my robot.

[@MartinBgelund]

Ralph's videos where he mentions this module: kzbin.info/www/bejne/bpfMiqmkd6mmeqc kzbin.info/www/bejne/oKKpnWuvmbqtkK8

[@AndreasSpiess]

Thank you for the links

[@deangreenhough3479]

Happy Sunday Mornings😁 another most excellent and worthy subject. As always I didn’t realise the depth of knowledge required to understand this, but I have received another excellent tutorial and a new understanding. I see this video has sparked some real interest in the comments. Hopefully Andreas you can probe into this subject with another video as it’s rather interesting. Happy NY to all the 100k+ subscribers 🏴󠁧󠁢󠁥󠁮󠁧󠁿

[@AndreasSpiess]

Thank you! Maybe there will be some sort of additional video. We will see...

[@stefangroenendijk799]

It's really nice to see that in the transaction history of the UPS on aliexpress the first 17 pages (of 30 total) are from today... they will have to do a lot of packaging there!

[@AndreasSpiess]

:-)) I order quite a few things from them and so far, all was ok.

[@Nicksperiments]

If I had a dollar for every time I successfully bent the pin of an SMD chip, I’d have 0 dollars

[@doggo7567]

lmao

[@anindyamitra5091]

Even if I could I wouldn't xD

[@luluiotrokag3858]

take a slice of mica, melt and put the slice of mica under the pin My personal solution

[@NoHandleToSpeakOf]

This is the module I used for my solar backpack. Bypassed charging circuit by soldering SX01-3A solar controller output to the battery holder. Worked great this summer but now it is winter and the boost converter that is always on drains battery.

[@AndreasSpiess]

As said it always draws at least 0.3mA if you do not disable 5V

[@NoHandleToSpeakOf]

@@AndreasSpiess Yes you did.

[@reanimationxp]

Excellent video as usual Andreas. It always surprises me how you address challenges I've forgotten I already faced, and are able to cover them in far greater detail than I did at the time. Thank you for sharing! I look forward to hearing about your solution for the raspberry pi's extended drain, if you find one :)

[@AndreasSpiess]

The challenges remain the same ;-) We will see when the RPI stuff will come.

[@thekakan]

"Atleast for a particular time or when the sun doesn't shine" Rhymes pretty good :)

[@therandomman2971]

Andreas, you have a large enough viewer base. Why not design a product like this but with the modifications you just mentioned, and then open source the design or contact a chinese manufacturer to create it cheaply?

[@youliantroyanov2941]

That is a freaking great idea. The Andreas brand will rule the market. I for one will be a fan for life.

[@caiopedreira6643]

Professor Spiess, also it would be great if this new product had one more USB input port or even another Micro USB input port for the solar panel. Congratulations for this great class and I hope you design this new product! Thanks!

[@AndreasSpiess]

I sold hardware in the 1980 when one computer cost at least a million. Then it was fun. Today I will no more touch hardware as a seller. Sorry. Sometimes manufacturers also watch my channel. Maybe they react themselves.

[@zerog2000]

@@AndreasSpiess actually, for the most part, if you design something useful, and at least do a test or prototype run, you will find nearly exact "clones" on your usual far east marketplace sites within a few months. But I do think you should release something of more ideal design under the "Swiss Accent" brand name;

[@gg-gn3re]

@@zerog2000 If he made a circuit and had JLCPCB or anyone else make it, I'd buy it just as support. Like other channels do t-shirts etc

[@SteveWindsurf]

Typically boost converters and LDO can have an enable input. Correct scaling of these enable inputs allows low battery voltage detection and reduction of drain current below your 60mA required for charging deep discharged battery.

[@AndreasSpiess]

You are right. But not here.

[@mwint1982]

"The sun does not always shine." That's deep man

[@AndreasSpiess]

Yes!

[@donpalmera]

I've never seen a Chinese product like this that works all that well and I have a little pile of puffed up lipo cells to show for it. I don't think it's easy to make something that works well with the cheap discrete components making up the charger->protection->regulation path. Richtek has some cheap (~50cents from taobao) parts that can do almost everything and have "power path switching" that switches the load directly to the input so it doesn't fight with the charger. Those would work well for a board like this. Aside from that there are PMIC chips that have the charger, power path and programmable DC-DC supplies for not much more that would probably would well if integrated into module with an ESP32. You'd be able to put sensors etc on their own DC-DC converters and turn them off when not needed, see the state of the battery etc.

[@AndreasSpiess]

There are some very nice chips used in tablets. However, I never saw them on a module.

[@donpalmera]

@@AndreasSpiess >There are some very nice chips used in tablets. Exactly. Something like the AXP203 that's used in cheap chinese tablets would be good I think.

[@makomk]

donpalmera That's how the C.H.I.P. (which is sadly no longer made) did its onboard battery charging, since it was basically a Pi competitor using cheap tablet tech. Unfortunately I'm not sure there's anything like that being made today.

[@donpalmera]

@@makomk The chip was based on an Allwinner SoC. The Allwinner SoCs are always paired with an AXP PMIC... Allwinner make a chip with a 1.2ghz Cortex A7 and 64MB of RAM in an QFP package (hand solderable) that costs about $4. :D

[@gr7480]

Thanks Spiess, i have two with panel solar 5v and a lora32u4II with one18650, i send a message every hour. I installed it 6 months ago and I have never had to change the battery. Thanks, byyyye.

[@AndreasSpiess]

Thank you for your feedback!

[@nemo5703]

Thanks, great video! I'm running MCUs off the 5V USB-A pins, rather than the 5V rails, to utilise the switch. I'm using it as a powerbank, but this would turn off most of the load in your trickle charging scenario as well.

[@AndreasSpiess]

Thank you for sharing your exprerience.

[@danobot12]

Thanks for the video, the clear diagrams and the product recommendations (buying batteries online is scary..).

[@AndreasSpiess]

I agree with the batteries. It got a bit better over the years, though.

[@gte24v]

Really useful, thank you. I will follow the link and buy a few. I use the exact same components (less the boost converter) in some of my projects and think I would generally: 1) remove the boost converter completely and 2) remove two of the LDO regulators as I normally don't need the full Ampere. That way I see a neater board that is easily suitable for small solar projects at a price lower than I can do it by assembling the parts (4056, protection circuit and LDO reg). Thinking back to your previous comparison of these charging boards, the 4056, this was my firm favourite as all of the switch mode versions failed to pass any charge current in dull weather, where this actually does pretty well even down to 1mA. Thanks again, thumbs up. :-)

[@gte24v]

And I have ordered 5 of them so hopefully you will see a few cents from me. Thanks for the hint.

[@AndreasSpiess]

This is why I called it "universal. You can remove things if you do not need them. They are so cheap. Thank you for your support!

[@timh2870]

I don't like the DW01. Wherever possible, I replace it with a Fortune FS312F. It is a drop-in replacement that has much better voltage constraints; it cuts out at 2.9 volts on discharge and 4.25 volts on charge.

[@AndreasSpiess]

Thanks for the tip. I will try it out

[@arnaudd5542]

Hello, I use it to power breadboards (with red and blue lines), it fits the spacing between outputs.

[@AndreasSpiess]

Thanks for the tip! Good idea.

[@Brutaltronics]

Having over 10h of protection time after the power was cut off is pretty good for my MCU projects. Very nice.

[@AndreasSpiess]

Thank you!

[@webslinger2011]

Since I'm playing with two voltages on my DIY raspad (12V for the 10 inch lcd and 5V for the rpi3), I've used 4 li-ion cells in series. Used supposedly "20A" voltage regulators each (for the lcd and rpi3). Planning to try out an XL4015 since it's rated "5A."

[@JensRoland]

webslinger2011 and did you solve the challenge of charging while using?

[@webslinger2011]

@@JensRoland I haven't installed a bms on the battery ( waiting for it to arrive). Need to test an XL4015 on the pi. So I haven't gotten to it yet.

[@mmReefs]

Good video/explanation, listing the basics for a micro-controller UPS. This is something I need for some future projects. Thanks!

[@AndreasSpiess]

You are welcome!

[@easynow6599]

Your video is truly helpful and give valuable information (i couldn't find complete datasheet of V3 or the other shields nowhere).. I just want to mention my problem in case you or someone has any knowledge. I have the V8 model(2 18650 batteries)..and instead of switch there is a button in order to activate the 3.3 and 5v pins on each side.. the problem is that even if there is current need (e.g. a led) that uses these pins..there automatically deactivates after a minute..

[@AndreasSpiess]

Many powerbanks deactivate if not enough power is used. They are useless for low power projects :-(

[@easynow6599]

@@AndreasSpiess but the one you use in the video the 3.3V and 5V pins are always on, even if you draw very little current..right?

[@yazeedmadaeen]

I was under the impression that power banks do not charge and discharge at the same time (9:19) because detecting the battery voltage, and hence state of charge under load can not be made precise without a microcontroller/ammeter on the power bank, A lithium battery has multiple stages of current and voltage changes that when applied incorrectly will ruin the battery, and with no microcontroller, I would like to see the durability of the battery in this charger, many devices i have seen charge lithium batteries as if they were lead acid batteries, but those are usually cheap toys not designed to live for long, whereas in a UPS, reliability is vital. Another thing is that when a battery reaches 4.2, and the charger is providing 4.2, no current limiting is required, as the battery can not sink a current from a voltage equal to or less than it's own

[@AndreasSpiess]

You are right. Most powerbanks do not allow discharging while charging.

[@HIBAW]

Thanks for the video. I also tested two of them. Bad solder joints on both. One dead after a month, and one DOA. I had to repair both, a chip lead, a fused resister, and a usb pin. Ordered months apart. In the real world an rpi cannot run off one. The battery voltage will drop over time. There just isn't enough headroom with 600mah. At least that was my experience on the two I tested. After 4 days, a fully charged lipo which started at 4.2v, was at 3.87v and dropping. So it's not reliable for an rpi without changes. And recovery isn't an option after an outage like you mentioned. For esp's though, it's potentially fine. I considered making one with two usb inputs, and switching. This would leave enough overhead to manage the charging vs discharging separately. That's the design I think I'd like to see. A real ups should switch power sources, not unreliably limit the whole thing.

[@marcrives279]

One dead after a month, with current well below 1A... Deep sleep current was not 300µA all the time. No battery reverse polarity protection (you need to often remove it to make a full reboot). Good idea, although a "v2" would be greatly needed ;)

[@AndreasSpiess]

I have 4 of them. None was DOA, at least :-). Maybe you change R2 as shown to increase charging current to 1A

[@phillkelley7700]

I have two of these that I bought 6 months ago. I use them all the time but not on RPi, just on Arduinos (ESPxxx or ATmega). They are very convenient for breadboard projects and I have never had any trouble with them. The only “problem” I see is that if I change the 18650, I also need to connect the USB-in port to a power source before any power comes out the USB-out port. I have assumed this is some kind of reset mechanism. However, I don’t think I would try to use one of these in a permanent project. For that I would usually go with a board that had a battery port on board. My LoRa gateway is based on a Heltec ESP32 + OLED with an 18650 inside the box wired to the battery port on the Heltec. Most of the time the Heltec gets power from its USB port but switches over to the 18650 instantaneously if USB power goes away. When USB power comes back, it instantaneously switches back to USB power and recharges the 18650. If that doesn’t qualify as “uninterruptible” I don’t know what does.

[@HIBAW]

@@AndreasSpiess yeah I was going to give that a try :) Was super informative. r1 is actually what burned up on the original one I had. So I need to replace that guy already. No clue why it did that or if it was a physical failure.... However an rpi runs pretty much all my home automation. So not starting after a long power failure, is actually worse than the power failure :) I'm working on centralizing my backup power. I've got all the parts, just not all the time to invest. :) Best wishes and thanks again. I enjoy your videos.

[@AB-pb8oo]

@@HIBAW why then not consider ol' good 12v lead battery if it is home automation and size/weight aren't an issue? Use tiny one, like 4.5AH. Controllers can be bought cheaply at firms that sell fire alarm / security hardware

[@nosafetyswitch9378]

NIce video. I have found that a low profile boost converter in combination with TP4056 ic for charging (and associated protection chips) is the best combination to use when powering something. Specifically, I personaly choose to boost the voltage to 6.4V and feed that to the LDO of the uC i happen to be using as I have found that they are always very good in regulating the 5V bus at exactly 5V +- a couple mV. This is critical if using a hall sensor like ACS712 which needs exactly 5V to produce the 2.5V middle voltage.

[@AndreasSpiess]

Good solution for your special need. Not very power efficient, I assume...

[@BlueyMcPhluey]

I think this video title is somewhat misleading. I didn't watch it for a long time because I thought the knowledge would be specific to the board, but this information is very useful and generalisable (particularly the map of power options at the beginning)

[@AndreasSpiess]

My videos always should have a little bit more info than absolutely needed ;-)

[@mcflapper7591]

good idea. thought about the 18650 case as UPS but didn't get around evaluating it. thanks

[@AndreasSpiess]

You are welcome!

[@ravishanker9014]

Happy New Year to Guru G...As usual very informative for me...Power supplies are backbone of any system...A good power supply enhance the reliability of system...Thank you... Namaste.

[@AndreasSpiess]

This is maybe not the best. But one of the cheapest ;-)

[@JohnMacGuy]

I know this is an old thread, but given the SuperPower project was cancelled, I wanted to try this out on an ESP32. Some amazon reviews of this board say there is an interruption of power when the charger power is disconnected. I didn't see an mention of this in your video, plus it would not be an UPS it it did this. Just checking.

[@AndreasSpiess]

Mine did not have this effect. But I know that there are many different variants in the market.

[@sethrd999]

For young players a good general overview of the do's and donts when relying on a non mains power source, I would just add that you characterize your devices ( main mcu/sbc, sensors ) thoroughly aswell as your systems duty cycle. I liked the nod to the 'non deep sleep' note on your consumption numbers. I actually purchased a couple of solar charging boards from a EE on hackster IO after long conversations and alot of research, I was ok with the price which was a little more than these from China, but I had the added bonus that these were actually fielded in the real world power both esp8266 and esp32 boards of various types and sensors without issues. Just as a note also for bigger iron ( Pi, BeagleBones etc etc ) you might do better with a hat/shield/cape as these are specifically targetted for these systems over being more generic, trade off is always going to be cost / your budget. Remember one thing though 'You get what you pay for'. Again great video Andreas, as always great food for thought!

[@AndreasSpiess]

I have such a hat and it did not work properly. I had to „Pimp“ it :-(

[@dikl2689]

What about us older players who don't know much. ;^)

[@sethrd999]

@@dikl2689 Thats why we have Andreas ;)..

[@HMPirates]

I am trying to do the same thing. What I figured is to use the tp4056+dw01 module for the battery charging and a MT3608 boost converter for powering the pi/mcu. I need to figure out a way to sense the battery voltage on the actual device(pi/mcu) to trigger a safe shutdown when the battery gets low. A simple comparator ckt can take care of turning the device on when the battery voltage gets above 3V.

[@AndreasSpiess]

I made a video about that.

[@angelneb]

Great detailed description! Will look into using these in future projects!

[@AndreasSpiess]

Thank you!

[@Mrcaffinebean]

Wow great timing, I’m in need of just such a thing! BTW I noticed a few battery bank makers are now offering charging and output at the same time.

[@AndreasSpiess]

Good to know.

[@davidkitts8017]

WOW!! I have two of these in my junk drawer, guess I better get them out an do some testing. Thanks.

[@AndreasSpiess]

:-))

[@MeisterQ]

Thanks for the Video! Its a really nice thing if you dont need to watch out for LowPower. My ESP32 Modules only need 8µA while in Deepsleep. This regulators and LEDs would take to much energy.

[@AndreasSpiess]

My favorite for low power still is LiFePo.

[@MeisterQ]

@@AndreasSpiess Thats what im using to sometimes. Even on Li-Ion with the HT7833 you are having a good LowPower performance

[@dannyjensen4954]

I love this video. Such a great explanation. Some of the links are not working but I can find elsewhere. Excellent.

[@AndreasSpiess]

Unfortunately I cannot keep all links op-to-date :-(

[@ursmii_]

nice part, but as you stated, not usable as uninterruptable power. It's a fact that UPS are more expensive in case of "extended design" The other problem of solar charging, at least in our country, is, that the solar cell most of the time delivers not enough voltage/current to fire the LiPo-Charger (most of them need 4.8V). There are interresting products with the AEM10941 Solar Harvesting IC, but there output current is (badly) limited to 80mA

[@AndreasSpiess]

I would use it for situations where you have to protect against short outages up to a few hours. Then it should work ok. Concerning Solar. You are right. Here we need quite big panels to get enough electricity. Harvesting might be something for really low power devices. I have a harvesting chip here and on the to-do list...

[@GRBtutorials]

6:50 Cut the PCB trace instead. It's easier and you don't risk breaking the chip.

[@AndreasSpiess]

Maybe you try it . Will not be easy.

[@damedaE90]

@@AndreasSpiess or lift one side of the IC, put Kapton tape in between the leg and the pad and resolder the other legs... Funktioniert gut und schnell 👍

[@RatthaninWartcheeranon]

Though this is an old vid but helpful, wanna choose the right UPS for my ESP32 that runs 24/7 for Home assistant projects even during power outages. all I have to do is to make sure ESP32 and other modules do not consume more than 600 mA for automatic charging recovery. Thank you. ♥

[@AndreasSpiess]

My pleasure!

[@psykopat81]

​@@AndreasSpiessnot 600 ma, 60 ma, right ?

[@psykopat81]

not 600ma, 60 ma, right?

[@AndreasSpiess]

@@psykopat81 no

[@nonamecukubik]

Thank you for this wonderful analysis, will buy a few when i start to work on project again.

[@AndreasSpiess]

You are welcome!

[@honigwachsreuben]

Would love to see the PSB01012B reviewed. And the challenges with 3V/5V-mix projects.

[@AndreasSpiess]

The PSB01012B seems not completely different. Just other chips for the charging and the booster. And no 3.3 volts.

[@robywankenobi32]

I made a solar powered Adafruit feather lora board by using a car phone charge module SMPS/DC to DC, 20 volts in from solar panel, 5 volts out to feather, then the lipo charger uses that to charge the battery :)

[@AndreasSpiess]

Good idea!

[@robywankenobi32]

@@AndreasSpiess Thanks :)

[@stompreaper]

This was very timely, I just had a couple of these delivered.

[@AndreasSpiess]

:-)

[@brenobertone2856]

Love your reviews!! They are very detailed and I learned a lot watching them

[@AndreasSpiess]

Thank you!

[@rowifi]

Simply disabling the boost chip can be a problem. Although the boost chip stops switching, the inductor now provides a dc path from the supply to the device.

[@AndreasSpiess]

You are right. But here it worked.

[@eliinthewolverinestate6729]

Been working with Stubblefield coils. Found that I was doing it wrong. I was running current same way and creating stationary magnet field in Stubblefield coils. Therefore not getting any extra current from emf. Using some of the coils to move the magnet fields was the solution. Electricity is as simple as galvanized rod and cooper rod in dirt. With super capacitors we don't need batteries. But need resistor to keep from frying stuff. Have a pi zero running off 3 earth battery cells and a buck/boost converter. It runs my WiFi continuously and will keep doing so till corrosion eventually kills it.

[@horrorcoder]

Hi Andreas, greeting from Italu and nice video, a lot of useful informations here! I'm a bit disappointed that the boost IC consume so much, I'll try for sure to disconnect it

[@AndreasSpiess]

Good idea!

[@NiHaoMike64]

For 3.3V, it's worth looking into using LiFePO4. Its voltage range of 3.0-3.6V is pretty much a perfect match for 3.3V logic.

[@AndreasSpiess]

I did this long time ago and made a video ;-)

[@altamiradorable]

Thanks Andreas ! Always precise like a swiss watch !! Very thorough analysis !

[@AndreasSpiess]

You are welcome!

[@steverussell1215]

Did I miss something Andreas? Your final summary bullet says, "The maximum sustainable current is only 1A".... where does this come from? I don't think I saw it mentioned in the video. Otherwise, this was a very interesting piece on these units. Well done.

[@AndreasSpiess]

It comes from measuring it. I do not remember if I mentioned it before.

[@steverussell1215]

@@AndreasSpiess.... Thanks.

[@salvobellino3490]

Hi Andreas, thank you for this video! I'm new in DIY electronics, just a basic question, how the 5v 2a pins in the PCB can be easily used? What type of connector should I solder/use)

[@AndreasSpiess]

This question is too general. I would say: Any connector which fits your needs and the dimensions of the holes. Most people probably would start with the small pins used on all MCU boards and cut 2 of them. Or solder two wires instead of a connector because it is more reliable.

[@adityakushsolan]

Hi ! Really informative video. Can you please try to solve the problem of not charging if battery goes below 3v and make a video about that.

[@AndreasSpiess]

Thank you! So far I have no such plans.

[@adityakushsolan]

@@AndreasSpiess Okay. Maybe give some hints about the possible solutions if you think about it, maybe in future. thanks

[@GregTank]

Thanks for your amazing videos as always!

[@AndreasSpiess]

My pleasure!

[@xnademolicious]

Have you seen any of these modules which allow use of LiFePO4 batteries? They seem to be a much more robust chemistry.

[@AndreasSpiess]

No

[@MohitGupta-ju7us]

Thanks for very informative video Andreas. Please do you have suggestions for someone who is starting learning IoT. If you can point to some learning resources and suggestions in IoT kits etc.

[@AndreasSpiess]

Iot is a wide area. Google is a good source. And, if I believe my viewers, also this channel. Most important: Search a project you want to do.

[@ymisetz]

Interesting video. I was looking for something like that, so you saved me a lot of time. Thank you Andreas,

[@AndreasSpiess]

You are welcome!

[@Ultrazaubererger]

In my opinion, the reason many power banks can not charge/discharge at the same time is that they use one chip to do both (with one inductor). Others may do it because the heat would be excessive (some get very hot even in standard operation).

[@AndreasSpiess]

I do not know all designs.

[@lancebenson8400]

Thanks for the very informative post. I'm trying to run a $4US LED (normally 3 AAA batteries) off of one of these modules. I have different varieties of LED. With one of them, with a USB LCD display of volts and amps in between, I get a reading of 1.55A, and it cuts off after 4 seconds and I have to remove and reinsert the battery. If I am providing power to the module, it continues running, but after several minutes, the LED dims and the display turns off. Flipping the switch off and on brings it back to 1.55A for a few minutes, then dim again (output USB voltage has dropped to 2V6). With a different LED, it shows 1.33A and runs for several minutes (no mains charging). With mains, it runs for longer--but then burned out some of the LEDs and dropped to 0.041A. Oh, well. One of the same type as the first LED has run on a 5V mains charger now for several months continuously. I'll next try one of the charging battery holders with 2 18650s. Hope they come out with an iteration of this module which incorporates your suggestions and also delivers the current which it advertises.

[@AndreasSpiess]

I did not have users with more then one Ampere in mind as the battery would depleat very quickly.

[@offgridsweden]

Awesome video Andreas. This video was very helpful. Have an amazing Sunday, Andreas from Off Grid Sweden 🇸🇪

[@AndreasSpiess]

A little late: Thank you and have a nice evening!

[@offgridsweden]

@@AndreasSpiess No problem man. Have a great evening

A other way would be, to connect the charger to the usb input and the output to the 5v in/output of the raspberry pi. That would charge the batteries if they are lower then 3v.

[@AndreasSpiess]

True.

[@geischt23]

Another very interesting video, thanks! Did you try connecting a ESP32 with deep sleep phases? Many COTS powerbanks just shut off when a MCU goes into deep sleep, I only found very few that keep providing enough energy to keep the device alive in its deep sleep state.

[@AndreasSpiess]

This one should not switch off. For sure not on 3.3v

[@geischt23]

@@AndreasSpiess perfect, thanks. Already ordered some for experiments anyhow :D

[@hansfortov]

@@AndreasSpiess Tested it with deep sleep for days now and does now switch off when connected via usb

[@TaiViinikka]

@@hansfortov Thank you for reporting the deep sleep results! Did you mean "Does now switch off" OR "Does not switch off" ?

[@jochen_schueller]

@@TaiViinikka we will never know :)

[@northshorepx]

Thanks Andreas. I learned some interesting things here. I wonder if it's possible for you to feedack to the manufacturer of the board about the trickle current resistance so that they could perhaps add it by default? I know that there are 100s of these suppliers and most will not care but just a thought. Anyway thanks for this tutorial.

[@bardenegri21]

They don't do that because the charge current you get might be too high for some cheap 1200mah or similar 18650 s people might put in there. It would be risky so they played it safe

[@AndreasSpiess]

Some of them watch my videos...

[@myounges]

Thank you and happy new year Andreas!

[@AndreasSpiess]

Happy new year to you, too!

[@kokoscom]

Andreas Happy new year! THank you for the video review - I sq this in one of your previous videos - Already got it - never had time to test it! So.. its great that you did this review!

[@AndreasSpiess]

Happy new year to you, too!

[@stamasd8500]

I can vouch that the black hole is real... Not only the holiday expenses, but the first week of January is when I also have to pay all of my professional license fees for the upcoming year, or else I'm not allowed to work this year. :( This product looks interesting, just ordered a couple for testing. I mostly use Arduinos so I'm not too worried about discharging the cells too profoundly. Modding SOIC8 chips this way is tricky, those pins are fragile. If I were to attempt it, I'd lift the pin only slightly, place a layer or two of Kapton tape between the lifted pin and the pad underneath, solder a wire to the pin and then the second wire somewhere along the trace away from the pad. And then glue the wires to the PCB with epoxy close to the solder points to avoid them getting yanked out by mistake.

[@AndreasSpiess]

I am sure I would be more cautious the second time ;-)

[@3dgussner958]

Great mini UPS, let's hope the company watch your video and change the board with switching off the FP6298 and not just the USB output port. On Aliexpress the 5V Amps rating are changed to 2A in the description, bit closer to your load test :-)

[@AndreasSpiess]

Interesting that they changed the description...

[@pratikbhagat9028]

AFTER E.O.V I DIRECTLY SUBSCRIBED . BEST VIDEO AND THOROUGHLY EXPLAINED.

[@AndreasSpiess]

Welcome aboard the channel!

[@pratikbhagat9028]

@@AndreasSpiess Sir, Lithium ion polymer 4000mah, The charging module tp4506 with dw01 battery over discharge protection, The booster stepup step down circuit 3.3volt constant output, An Esp32wroom, The 2amp/5volt DC wall socket 230ac adapter Will this step simultaneously charge the battery as well as the esp32 and can this setup work as ups

[@pratikbhagat9028]

Sir, I was expecting an expert advise regarding the question

[@pfrischknecht]

Awesome video .. AGAIN. Really like your stuff brother... Howdy from North Carolina...USA

[@AndreasSpiess]

Like your area. I had a customer (Syngenta) in Greensboro and still remember Kitty Hawk.

[@95rockforever]

Very good video !!! Do you know if it is possible to connect two of those shields in parallel?

[@AndreasSpiess]

It should be possible. But maybe you will not get double the output power.

[@riccardopet]

What about adding a diode after the battery and a connecting the boost/ldo directly to the power supply? Would not this solve all the problems while charging and "discharging" at the same time? (provided that the psu is good enough).

[@AndreasSpiess]

Maybe you make a drawing and include the fact, that a diode has a voltage loss of about 0.7 volts. Then you see if you find a way which works.

[@AdityaMehendale]

The most elegant fix to the trickle-lockup would have been if the boost-convertor had an under-voltage lock-out (UVLO) for limiting discharge to a safe voltage-threshold. Draining batteries below 3.5V doesn't increase run-time by much.

[@AdityaMehendale]

...Wait a minute! The DW01 datasheet mentions "Overdischarge Release Voltage" as 3.0V (+/-0.1V) --> This means, that a discharged battery won't be re-enabled until >2.9V (which is SOMETIMES above the trickle-threshold of the TP4056) - Matter of binning?

[@plemli]

Using a voltage divider on the EN pin would probably accomplish an UVLO. I'm too lazy to check if there is sufficient hysteresis to guard against cycling the output around the UVLO voltage.

[@AndreasSpiess]

The under voltage protection kicked in at about 2.6 volts. And it does not cut-off the TP4056

[@AdityaMehendale]

@@AndreasSpiess Oei.. bummer :( That's far below the datasheet-spec too..

[@RobertSzasz]

A 2C "UPS" type supply with USB C in and say, 2A 5v output would be a great thing for the hobby market.

[@MaxC_1]

luckily those have existed for quite a while or you can just use a standard 18650 holder with one of the 2.1/2A 5V USB C + USB A charging boards that you can find for pretty cheap that are used for battery banks to do the same

[@psykopat81]

​@@MaxC_1do you have a link ? this type of UPS ?

[@TheHamzeh82faouri]

At first thanks for the great job you did I think i have idea ti solve the going below 3v problem I think arduino must read the battery voltage using analog input Then when the battery voltage gets below 3v the arduino will trigger transistor to make short circuit between the output terminals of the charging board of the battery Thus the protection will work and cut the power When the main is available again the charge board will work and arduino will work and the battery will charge

[@AndreasSpiess]

Monitoring the battery voltage with the Arduino is a good idea.

[@GRBtutorials]

Why do all protection circuits cut off at 2.7 V/cell? At 3 V/cell, the battery has pretty much no charge remaining, and it's also bad to discharge Li-ion batteries that much. That would solve the problem of the trickle charge.

[@GRBtutorials]

@@dd0356 Li-ion batteries have extremely low self-discharge. If it was Ni-MH, I'd agree, but it's Li-ion. And what do you mean with "charing a 3.0V lipo"? If you mean "charging a LiPo battery which has been discharged to 3 V/cell", well, they're usually discharged to 2.7 V/cell and have no problems charging at all. Unless you're referring to batteries consisting of multiple cells in series, in which case, yes, that's horrible, and I should have specified that I mean the voltage per cell. I edited my original comment in order to add that information, thanks.

[@GRBtutorials]

@@pirateman1966 Yeah, it's not that difficult to implement. I'll end up making my own circuit because pretty much all the circuits on the market are like that.

[@MultiPowerLoRd]

Excelent found. thank you for explanation. I will buy and test for my raspberry project

[@stefanu4359]

Thanks for this video. More than I expected

[@AndreasSpiess]

Thank you!

[@thegittubaba]

I also noticed (the few power banks i dissembled) that power bank pcb contains only one inductor (and i'm pretty sure only one oscillator) , it is used as part of buck converter when charging and as part of boost converter when discharging. So it can't do both simultaneously :P

[@AndreasSpiess]

Maybe. This board also has only one inductor. It uses the TP4056 from 5V. I assume many power banks will do the same.

[@thegittubaba]

@@AndreasSpiess last one i opened was a xiaomi 20000mah one, it supports QC 2.0, i wanted to mod it to run at 12v to power my network stuffs (router, switch, ap). I could mimic the QC handshake with RC+BJTs but had to abandon the project due to not having simultaneous charge-discharge.

[@send2gl]

I have never thought about this before but have innocently charged a portable power bank whilst using the power bank to charge a phone. Didn't notice any problems at the time.

[@AndreasSpiess]

Maybe you had one which charged and discharge at the same time? Or it was already fully charged.

[@asuagar]

Andreas, thank you so much for research and broadcast effort. In your video #155 you speak about the good performance of the TP4056 as a good solar charger board. In this device, the TP4056 will extract the maximum power of a photovoltaic panel connected to the micro USB port or it will be needed an additional piece for doing the job? Regards!

[@AndreasSpiess]

You maybe have to adjust R2 as shown in the solar charger video.

[@SmithyScotland]

I seem to remember a video (possibly adafruit) showing 6v power spikes when switching from battery to mains power with solutions using the tp4056. Did you notice how clean the supplied voltage was when switching power sources? Could be all the additional circuitry on this model fixes this.

[@AndreasSpiess]

I checked it out now. I connected the oscilloscope to 5V and unplugged an re-plugged the charging USB cable. And really, I saw big spikes (>20V). If I kept ground connected to the wall charger (as it is if you lose mains power), I only saw 3x 10nS spikes (very short). They are comparable with the 200mV spikes from ripple. The big ones most probably came from the ground problems (measuring errors). So I would say this is ok.

[@MM-hk6zw]

Hello Andreas Thank you for your great video once more! Your calculation about the battery runtime is wrong in case the 5V output is used. If you've a battery with 2400mAh at 3.7V this is about 8.9W. When boosting to 5V you can only get about 1780mAh out, in optimal conditions if the efficiency of the booster would be 100%. In this case the runtime at 400mA would then be ~4:30h, so I think it is safe to say, that a 400mA load at 5V could rund about 3:30h - 4:00h hours only.

[@AndreasSpiess]

You are right. I did simplify a little too much :-(

[@anindyamitra5091]

Can't we bypass the battery power sensing whether mains is active, keeping two power FETs, one to switch the battery off and another FET to bypass the incoming 5v to the power management circuit, instead of modifying the trickle current?

[@AndreasSpiess]

I did not try it. Maybe you try if your idea works?

[@mackgoodman]

any idea how one would attach a solar cell? Where would the inputs attach? thanks

[@AndreasSpiess]

At the mini USB. But only 6v

[@Kartoffelsnack]

How much Watt or Ampere do i need (min and if there is: max) What i understood from the video is, that i need a 6V solarpanel. I'll use this board with a LiPo 18650 2600mAh Battery, and the device it will power is an wemos d1 mini with a whole bunch of Sensors (Weatherstation). Thank you in advance!

[@NicksStuff]

8:56 So, when devices accept to charge a battery and operate at the same time, it's either because the low voltage protection is ≥3V or because there is a current-sensing resistor between the battery and the boost converter and some magic (op-amp ?) that substracts that current from the actual output of the charger, so that the charging circuit know the actual charging current? But why not leave everything as it *and* connect the power supply to LDO/Step-up (after a couple diodes, so that the battery isn't charged that way)? The charger would still trickle charge the battery at 60 mA while a much bigger current is supplied to the load

[@AndreasSpiess]

You can try it if it works. I never tried.

[@chrisz7234]

Andreas thank you for sharing. Very very helpful as always. But can you suggest a different reliable 18650 batteries provider? That shop only ships to a few countries!!

[@AndreasSpiess]

I only can recommend something I have used :-(

[@ACTlVISION]

I'd really love to see a vid about the problem you mention with the Pi battery voltage protection, I've looked into some of these ICs but not a lot of examples

[@zgryx8428]

Thank you for the wonderful video. When my 18650 battery shield is charging and you remove the charger suddenly, the shield will turn off and then turn on again for several seconds. I believe this is a protection mechanism, but I am using Arduino, and if the power source restarts, the Arduino also restarts, which is not appropriate for my project. How can we avoid this? Thank you a lot.

[@AndreasSpiess]

I do not know. Maybe it is not intended for this usage as many powerbanks (they do not allow to charge and power a device at the same time.

[@JanJeronimus]

Interesting video as usual. THX Andreas. (Don't understand the thumbs down. If someone adds thumbs down at least explain why!) In (small) electronic project the power circuit is an important also for the cost. My experience (of at least my feeling) is that if there is a (mains) power failure a period can start with several power failures. Manual intervention (for each device) to reactivate the charging is for some applications not a problem, however for most applications not an option. Also it would be nice if a MCU can have access to some (diagnostic) parameters of the power circuit. (Battery status, temperature, mains functioning, is the solar system delivering something). For 2$ you currently can not expect this. However who could imagine several years ago an MCU with Wifi for less than 2$.

[@wm6h]

Jan Jeronimus I think many thumbs downs are a “Bad Touch” on a phone or iPad. Not intentional. It would be an interesting statistics project.

[@AndreasSpiess]

Maybe I come back to the topic of shutdown and power monitoring in a later video...

[@gersonfer]

Another very good explanation as always. Tks

[@AndreasSpiess]

Thank you!

[@ats89117]

Thanks for another excellent video! At 2:00 you mention that an LDO is used to convert from 5V to 3.3V, which is pretty horrible. I guess paying a few cents more for a switching converter isn't possible? Or maybe they don't care that they are wasting 1/3rd of the power before they start because they are running from mains? The efficiency could go from 67% to 95%. Doesn't seem like a good solution to me...

[@AndreasSpiess]

Maybe you look at the diagrams and discover your thinking mistake?

[@ats89117]

You mean the fact that the power is usually wasted on the development board? That's probably even worse...

[@AndreasSpiess]

The LDO is powered by the battery. No big loss in most situations.

[@TheMvieira]

Very helpful Andreas!! Thank you from Brazil!

[@AndreasSpiess]

You are welcome!

[@kwazar6725]

Sali. Another fun sunday morning with andreas. Just a few comments.. lipo and lion min 3.0v... for longer lasting i would do lifepo4. They are much more robust....anyway.. I think ups for pi is a must because its fat sdcard corrupts easily.

[@AndreasSpiess]

I did videos about LiFePo. But this device (and many others) use Li-Po). Maybe it will change in the future. But Li-Po has a higher capacity for the same weight/space.

[@kwazar6725]

@@AndreasSpiess ur correct. What about getting early warning from ups that pi needs gentle shutdown.

[@AndreasSpiess]

I also did a video about that. Including software ;-)

[@ranger81de]

I'm currently looking for an UPS for my main RPi running my smart home system and it turns out it's not an easy thing to find. I have even tried expensive modules like the "Strom Pi 3" but was not satisfied with it. When switching from mains to battery, the voltage dropped to 4V for a brief moment. The manufacturer said, I have a faulty module but this is generally my experience. All of these UPS modules for the RPi claim to have enough output current and voltage to run the newest RPi 3B+. But in fact I'm seeing a lot of flaws and under voltage... I have now ordered a classic "real" UPS with a big lead-acid battery finally and I hope that this UPS will finally provide the reliability I need. Thanks for the video, I'll give these modules a try.

[@AndreasSpiess]

I tried it with my RPi3 and it did not crash when I unplugged USB. I did it quite a few times.

[@zacsubach]

@@AndreasSpiess before or after you changed the resistor?

[@ranger81de]

@@AndreasSpiess Would be interesting to monitor the voltage while the module switches to battery. In case of the (maybe defective) Strom Pi, it was below 4V for about 0.4 seconds which is in my opinion not acceptable. I have just ordered two of the modules, I'll check it. Thanks for your update Andreas, as always very helpful !!!

[@AndreasSpiess]

I did not check it. But since nothing switches I assume no drop. A LiPo has a very low inner resistance.

[@gibsonblogger]

Tobias, have you looked at lifepo4wered.com/lifepo4wered-pi+.html ? I have not tried it.