"just the facts, ma'am". You did the grunt work we never got round to. Very straight delivery. No puff. No spin. Nothing but the meat. Superb. Thank you.
Excellent presentation and detail. Thank you for the video. And thank for simply presenting the material (very interesting) without annoying music like so many other youtube videos. Yours are the best! :)
My mind is the sort that understands things well, but my rote memory is terrible. So after watching all of your videos initially to get the idea of what is going on I find that with each new project I work on I simply re-view the appropriate video as if it were a topic in an engineering reference book. There are so many wandering, rambling, directionless RUclips videos, but I find your videos are detailed and specific to a topic of interest, and always lead to an optimal solution to specific engineering problems, or at least a functional one. So I feel like I have a video library of nearly all pertinent topics to support my projects, with explanations and optimizations, and it is unnecessary for me to 'reinvent the wheel' for every new project. This channel and your ham channel are like a detailed video version of the Handbook for Radio Amateurs and QEX magazine combined, and they provide the technical support I need to ensure that I can tackle difficult projects and they will work when I'm done. Also, many projects online tell me to buy this and that prefabricated device and just plug it in to make a project work. But their 'this and that' are hundred dollar items! Your projects show inexpensive devices and ingenious ways to solve engineering problems. This is what good engineering is all about. Some other RUclips channels and Internet websites seem to miss that point entirely, or they are simply incapable of doing good engineering. As I am retired and have only a modest income, knowing that your channels are here to Elmer my work enables me to spend my money wisely on my projects, and your reference library of free videos continues to grow. I can't thank you enough. - John
Thank you very much for your kind words. Indeed, I sometimes go back to one of my videos for reference ;-) Concerning cost: I have many viewers from less developed countries. This is the reason I stick with the cheaper stuff. HAMs seem to spend more money on their hobby. So this channel will feature also more expensive things. Enjoy you hobby!
Nowadays there are also 1.5 volt aa/aaa lithium ion batteries (through an internal regulator). You can feed the esp32 directly with two in series. Also, there are Nickel-Zinc batteries with a typical output of 1.6v. You can also put two in series to feed the esp32, but they supply 1.8 when fully charged so it may be risky.
NiMH maintains voltage VERY close to 1.2V across the bulk of its discharge cycle! It starts off at about 1.45V but quickly drops down to 1.25V. This means that in contrast to other types of cells which have a more slanted voltage curve, you can pre-discharge them a little bit and you will only lose little capacity. To accomplish that in-device, you should be able to use a Zener across the NiMH battery pack to clamp the voltage. After the battery is discharged below Zener voltage, mostly through the Zener itself, no current will flow through the Zener and it will no longer contribute to power consumption. Perhaps a similar trick can be accomplished with Li-Ion, but of course you'd need to limit the current, and i'd be wary all around, many angry pixies in there. Perhaps the proper solution with Li-Ion is not an LDO but an active boost/buck regulator - you can discharge much deeper down to typically 1.8V, and you'd have less efficiency loss.
The buck or boost converters usually take 20-40% of the energy (depending on the voltage difference and make). And they need quite a lot of current if the ESP is in deep sleep. Your idea with a Zener should work. However, this costs some energy. Maybe it would be better to stop charging the battery at this particular voltage instead of charging it full, and then discharging it afterwards. But I am not a battery specialist.
Andreas Spiess I don't see why there should be more than a 30% loss at the IC, and can probably be optimised down to 10-ish. Perhaps a major contributor is the output choke or transformer, but those can be done with low resistance too.
Charging profile is simple (as described in the video). As I said, the technology is new for me. So, for the moment, I do not know much more than presented in the video...
Thanks. I think, this battery can be done yourself. Just use a woodstick, cut it to length, and put a wire from one to the other side (maybe with a screw on both sides).
Hi Andreas, thank you for taking the time to create discharge curves and publish them in your channel. This is very useful. As you know, some of us live in cold places and batteries do not always do well there. Most Li ion cells are optimized for temperatures above 0C and do poorly below that. I am curious to know if your LiFePo cells do well in the cold.
Hi Andreas, Dave, I live in Canada and I have been using both the Titus 3.7 V C Cell and Titus 3.7 V AA Lithium Thionyl Chloride batteries, -55C to +85C. I highly recommend them, I used them in the military for our remote sensors and GPS units, even in the Arctic. We are currently using inside transport GPS trackers for almost a year without replacing.
Great presentation. I think its better to charge 18650 to only 3.6v max then ldo will not required. If powering from fully charged single 18650 then ldo is only useful will till the voltage drop from 4.2 to 3.6 after that not needed. This needs to be tested if waste or avoiding waste is better.
LiFePo4 (lithium iron phosphate) batteries are sold for RC applications as transmitter and receiver batteries. They have a lower self-discharge rate and better temperature characteristics as LiPo, but a far higher internal resistance. Quite good for devices like outdoor temperature sensors, if you really want to bother with rechargeables in such a scenario.
Thanks for the info. You are right, I also assume in my summary, that you might use the "hassle free" alkaline technology if current consumption is small.
This is really cool. I knew about LiFePO4, but never seen them in 18650 package before. Are there some chips similar to TP4056 but for charging small LiFePO4 cells (~500mA)? Right now you need to essentially disconnect the cell and charge it somewhere else. I do have situations where I would prefer to have micro USB socket and power coming, and if it is available charge the cell (but preferably never charge with more than 3.6V), but once the USB power is gone the chip will shutdown, and ESP8266 would work just fine. TP4056 can't be used here due to different charging curves, as well too high charging voltage, which would prevent you from using it directly with ESP8266. This could serve both as a quick recharging (2-3 hours for medium size battery), or just as a backup power. Also it is nice to know about ESP32 voltage range: 2.3V - 3.6V. The internal LDO regulators regulate it down to 1.1V / 0.9V and 1.8V for various purposes. The lower operating 2.3V is guaranteed, and it can probably operate a bit below that fine too. So you can squeeze quite a lot of power out of the battery. Possibly even too much. :D Technical specification says that it can operate down to 1.8V correctly, at least internally, but ADC will behave differently (can be compensated), and one would have trouble communicating with external chips probably. Some active cell voltage monitoring by ADC, to warn about too low voltage, and or switch off to ultra deep sleep when it goes even lower, would be a good option too. I do not know how the ESP32 behaves when the VDD is below 2.3V exactly. Optimally it would switch off completely and use less than 0.1uA in that state to prevent catastrophic (or just permanent damage) discharge of the cell.
I have exactly this question; I want to power an ESP32 with some LiFePO4 batteries, but cannot find any chargers anywhere that are reasonably priced for just some minor hobby projects. I'd be very happy if I can repurpose one of my old phone chargers, but I'll need a controller of some kind...
With LIFEPO4 you will still need some undervoltage protection... And the main advantages of LIFEPO4 are great low temperature performance and possible high charge/discharge rates... But in case of no low temperatures I prefer to use more 18650 and LDO, it does make sense if you are not space constraint...
I'm currently field testing supply by 3 x AA Eneloop accu cells and one BS120 Schottky diode in series. (mostly because i've got lots of both components in my spare parts ;) works like a charm so far.
I use lithium 3.5v battery from discarded cell phones. It works like a charm, even when you're recording a new firmware on the device. I also use a small silicon diode in order to drop the Vcc + a little.
Diodes are ok to drop voltages. However, I prefer LDO regulators because they have a variable drop (at the beginning more, at the end less than a diode). And they are also not very expensive.
Thanks for the intro to LiFePo4. I haven't played with those yet. Why not reverse or modify your battery terminals to run a pair of LiFePo4 AA's in parallel instead of the traditional AA-Series connection? Wouldn't this be a 'best of both worlds' situation? Don't most LiPo cells have an internal over charge/discharge circuit? All of the battery packs I have torn into have them (I haven't taken apart any if the round cells).
I looked at your proposal concerning turning one cell. Not easy (the minus has a spring, the plus not). So, It would be easier to buy two single holders if you want two in parallel. Concerning LiPo: No, not all have protection circuits. You can buy them with or without.
The connector batteries are interesting. I have a bunch of used but not empty 3,6 V Litium AA batterie (not accumulators however) and use them in clocks/weatherstation. I soldered wires into the battery case which also works. But this connector battries would have been the more nicely and instant reversable solution.
Hello, I know this is an old video, but I wanted to say something: 1. If you use a buck converter, you get higher efficiency and the capacity lost might be only about 5 % in the worst cases. Plus, there's almost no voltage drop so you can use it until it reaches 3 V, but for stability, it's better to cut off at 3.1 V or 3.2 V, when the battery is almost empty, and also Li-ion batteries don't like being fully discharged, so it's better for the battery. 2. There are different Li-ion chemistries. The most common are LCO (Lithium Cobalt Oxide, LiCoO2), which is the most common type and with extremely high capacity, but the most dangerous, NMC (Lithium Nickel Manganese Cobalt Oxide, LiNiMnCoO2), the next more common type, used in Panasonic's NCR18650 batteries, which are safer with almost the same capacity and LFP (Lithium Iron Phosphate, LiFePO4), the safest "cheap" ones (there's also lithium titanate, LTO, tremendously expensive), but with lower capacity.
You're welcome. A page with further information is this one: batteryuniversity.com/learn/article/types_of_lithium_ion. Battery University is a good website to learn about batteries.
Thanks for all your great video's! Is it also an idea to use a Lipo and a simple Germanium diode for a little bit voltage drop? Thank you. Best regards Bert-Jan Roelofs
I think LiFePo4 is the most cold tolerant of rechargeable batteries. If your device needs to work in sub zero temperatures, you may need to create some kind of heating thing to keep LiPos warm, which will likely eat most of their capacity.
700mAh at 3.2V = 2.24Wh. And seems to cost ~$3. For that price you can get NCR18650B (3.25A*3.6V=11.7Wh) or similar. Even lowered by 20% with LDO it still seems that regular 18650 is 4-5 times better deal.
Dear Andreas, thanks for this video, i have been really looking for something like this to understand which battery i can chose for remote toy car i am planning to make for my kid using power hungry esp8266. i just have a very small suggestion, if you find it appropriate, generally all such videos you make about comparisons and evaluations of different concepts, equipment or things, are full of very useful information. can you at the end make one slide with conclusions - in text form. because it helps in remembering the final outcome of the study.
Good idea.But it is not easy because my videos usually are already compressed. But I will think about it. Concerning toy: motors are usually power hungry and need higher voltage. So, LiPo is probably the best solution. And a regulator for the 3.3 volt. The ESP will anyway use much less current than the motor.
I think one important point is missing. The end users are very prone to put something in the wrong direction. So in order to not to burn your IC' s you should place a low forward voltage diode (schotky maybe) in series to your power supply design. Otherwise someone will insert the batteries in wrong direction(Murphy Laws)
@@AndreasSpiess But a diode saves you from having to use a regulator (thanks to forward drop) if you don't need so much current. And everyone can make mistakes including the designer.
Nicely informative video! The only thing missing is if you made a final comparison to show just how much battery life you would get from each battery option (possibly even with a comparison of their price). From your video it sounds like a LiFePo4 would give the most battery life to an ESP8266 since it has good efficiency and no need for a regulator? But unfortunately you didn't show that in your video. Maybe suitable for a follow-up video, comparing battery life & cost for powering ESP8266 using LiPo vs LiIon vs LiFePo4?
I do not plan to take cost into consideration as it changes over time. But it should be possible to assess it yourself if you look at the actual price of the batteries you consider to buy. You should have the needed data in the video.
Good cross examination. I would be nice to design a circuit that has the MCU check the battery voltage and switch on a voltage doubler circuit when the battery voltage runs near the shutoff voltage. I suppose we need to check the balance because we introduce a new circuit that has it's own consumption, but we might be able to compensate it due to the extra charge we can use from working to lower voltages. I know the LT 1026 can work as a voltage doubler but unfortunately it's not a low voltage low power chip, I don't know of any other versions.
I am not a fan of these designs because the batteries do not have a lot of energy left at the end of their "life". It adds complexity, size, and cost and I would invest all three in a bigger battery. "This provides more bang for the buck"
Actually, LiPo does come in the AA form factor: the 14500. Also, rather than using a linear regulator, you could use a buck or boost converter with most of the chemistries to not only get the needed voltage, but it could be regulated and the battery could be protected against overload and over discharge.
Thanks for the feedback concerning the AA form factor. Buck- and boost converters are great for big voltage differences. But they usually have higher quiescent currents. So, they are not ideal for deep sleep scenarios. And for small voltage differences, for me, they are an overkill (and also can introduce RF).
Great video! What about lead acid gel cell batteries? I think they might have better cold charging tolerance than the LiPos and come in a wide variety of sizes at ~2V/cell. For non-flying applications, their density (heavy mass per cell) is not a problem, and in fact was a benefit when the wind was strong and my project box did not blow down!
Hello Andreas. Although I knew about LiFePO batteries I never considered using them with my 8266's but I will now! I was struggling with voltage regulators and buck convertors but maybe I can skip all that and use LiFePO's instead.By the way, in the video you mention that LiPo batteries are not available in AA packages but I believe that 14500 batteries are roughly the same size. Some LED flashlights can use AA batteries or 14500's.
You are right, they are also available in this package. So, you have the same "danger" as with the LiFePo4 batteries... You shouldn't struggle with linear regulators, they usually work fine, if you add a beefy capacitor between plus and minus of the ESP. Maybe you watch my videos about deep sleep. There, You find some explanations and tips.
Hi Andreas just a quick question. I am planning on using the ESP8266 with one 1.5V AA battery and a boost converter (tps61006). The module will be in sleep mode quite often in our project and only woken up for short period. Do you think this is a good solution? Love your videos. I have learnt a lot and found it very useful for our project. Thank you very much.
Thank you very much for your prompt reply and for the advice. I am using a 1000uF SMD Aluminium Electrolytic Capacitor across VCC of the ESP. I took this from one of your videos. I'll do battery test once I get my PCB done.
Am interested in how this goes for you +Venko. Have seen those inexpensive adjustable boost converters on AliExpress, and am curious. I would worry that if the ESP decides to transmit heavy, drawing 800ma, that you would need 3x as much current on the low end ... and the high end may get choppy.
Thanks Andreas, very informative video. Just a thought: Instead of using an LDO regulator for for NiMH etc, what if you use diodes in series to drop the small amount of voltage drop required? Would that be more efficient than using an LDO. Of course at the device you would have a higher voltage swing over the useful portion of the discharge curve. Cheers.
I do not like diodes for that purpose. They have a constant voltage drop of ca. 0.6 volts. With a fully loaded battery this is not enough and if the battery is at the end of its live your device will stop too early. Efficiency can be a little bigger, but if you use a good regulator (see video #58) this is not a problem. The only positive could be, that a diode is cheaper than a regulator, but today, this is only pennies.
nice, but isnt the 3xAA percentage also 80-90% then? and Lifepo4 cells are the best with 82% then? and could i also get 5volts from the board itself(5v output pin), if the battery input is only 3.0-3.2 volts? (because some sensors require 5volts input.... even if pwm output is 3.3v....)
hi, thanks for the video!! it's really informative and i will look into the lifepo4 battery for a new project with my esp32 dev kit. one thing i was wondering though: could i also connect a power bank over the micro USB port? that would mean to supply 5V to the board, but would that not be the same as to connect to the USB port of my computer when uploading a sketch? the 5 V would then be regulated don to 3 V on the board itself, at least in my understanding. thanks
The normal dev kits have voltage regulators for the 3.3 volts needed by the ESP32 and a charging IC for the battery. But these charging ICs are not compatible with LiFePo4. They have a max voltage of 4.2 volts for Li-Ion.
Hi Andreas, Your videos are very helpful and appreciate your effort. Whats the best way to power ESP8266 by a solar panel with a battery charger. Thanks.
What about a older battery tech for stationary devices like weather station etc the 1.2v nickel iron Edison batteries. Easy even make them yourself and with the discharge of the boards you could run one years on a 3s set up
18650s are Li-Ion not Li-Po batteries. Li-Ion and Li-Po batteries may be fairly similar but they're definitely not the same. Li-Ion batteries usually have a better energy density and are more efficient, Li-Po batteries however degrade in capacity less than Li-Ion and often have a lower internal resistance and can therefore charge/discharge a lot faster than Li-Ion batteries. When connected in series Li-Po batteries drift apart more easily in voltage and therefore it is very important to balance charge them. Li-Ion however if well matched don't drift as quickly and are therefore often not balanced in laptop batteries for example (the voltages of each cell is monitored though and if they drift too far the battery will stop working). Still I would always balance either type of battery if you have multiple connected in series. Also Li-Po usually cost slightly more than Li-Ion batteries. There are other differences, but I would say those are the main differences. which depending on your project can make a very big difference. You can easily find small Li-Po batteries that can start your car without a problem. You will find no Li-Ion battery however that can deliver anywhere near enough current for that, you will definitely need a few cells in parallel (additionally to the series connection you need anyway to get around 12 V). Drones usually use Li-Po batteries for the same reason, they discharge their batteries in a couple of minutes, which is no problem if you have the right type of Li-Po batteries but can be difficult to achieve with Li-Ion batteries.
@@AndreasSpiess Of course, I'm not saying that I think your video wasn't well researched or anything like that. In the case of an ESP the difference isn't very important. Just thought it's worth mentioning that they are not identical. Another thing worth pointing out would have been that LiFePO4 batteries while being great for an ESP while discharging, they are normally charged to 3.6 V. So powering an ESP with them directly while charging is pushing the limits, at least if you charge them fully. But those are relatively minor things. I don't think you were trying to make a 100 % accurate scientific video about batteries. And for a simple comparison of batteries the video is more than accurate enough. Just though maybe someone reading the comments would be interested in the difference.
Yo tengo muy buenos resultados con 2 Alcalina AA sin regulador, funciona hasta los 2.17v. Ahora voy a probar con una 18650 en serie con un diodo. Aunque tambien el regulador del wemos d1 es bastante eficiente. I have very good results with 2 AA Alkaline without regulator, it works until 2.17v. Now I'm going to try a 18650 in series with a diode. Although also the regulator of wemos d1 is quite efficient.
Wouldn't it be possible to build a bypass circuit that uses the current directly if it's within the spec range and goes through the converter when it isn't? So we'd lose the 20% when outside of spec, but keep 100% when within spec?
Old, but pretty nice and informative video. Thanks for it. Did by any chance you found rectangular flat LiFePO4 battery somewhere in the internet? Something around 1000mAh and few millimeters thickness? So far, I did not succeed...
Andreas, thanks for this great comparison. I'm looking at several types of voltage regulators for use with the 3xAA solution. I want one with the input voltage as close as possible to 3.3v to take full advantage of the battery life. Do you have any particular modules or other advice on how to ensure I get the best one?
Hello Andreas, very informative video, as always. I'm wondering what you would use if you needed to power your solution using both 3.3V and 5V. Use case is some of the sensors (ultrasonic, PH, etc.) and displays (most of the LCD's) require 5V while ESP8266 requires 3.3V. Bi-directional logic level converters can be used for the data lines but we still need to power the sensors/displays using 5V. Thank you in advance :)
many sensor modules just require 5V because they need a stable (low noise) 3.3V supply and use a linear regulator with 1.25V drop. in a low current battery application, it might be worthwile figuring out if you can supply the sensor directly, especially if you are already using a linear regulator elsewhere for 3.3V.
If you need 5 volt and 3.3 volt just use a linear 3.3 volt regulator (e.g. AMS1117). You even do not need a LDO because the voltage difference is 1.7 volt. Just make sure, that the package can dissipate the heat.
i'm trying to figure out how to achieve long battery life like your #47 video and also supply 5v and 3.3v. Will you make a video to address it? Thank you.
you should not have overlooked the advantages of NiMH cells. They can handle over charge and over discharge which the Li batteries cannot. They do not explode or leak. If you need durability and rechargability, they are the clear winner.
I did not pay too much attention to this technology because I assumed, it is well known by my viewers (it is a relatively "old" technology). But you are right: They have the advantages you describe over the Lipo.
I am still very confused about how to actually get a LiPo *into* the 8266. I hear about regulators, step up, step down, lots of things I can't seem to find any good information about. There seems to be alot of assumed knowledge about batteries out there which makes it very hard for beginners to grasp. I'd love some LiPo clarification, because those nice little square batteries look so impressive, but using them is still beyond me.
To power the ESP866 you need 3.3 volts. If your battery or other power supply has more than these 3.3 volts, you need a step-down regulator or converter (two names for the same thing). If your battery has less than 3.3 volts you need a step-up converter. You can buy linear or switched step-down converters, but only switched step-up conveters. Step-down converters are also called "buck converters" and step-up conveters "boost converters"
You are right. And because the voltage drop is always 0.6-0.7v the ESP can get unstable if you go down below 3.6 volt. the LDO has only about 0.15 volt drop in this situation. This difference sometimes matter.
The diode drop depends on current, so you need to choose wisely. About margins: - ESP works on 3.0V to 3.6V (datasheet: mintbox.in/media/esp-12e.pdf ). - LiIon discharge curve, from less than 25ºC, drops at 3.6V to safety level in no time: www.ibt-power.com/Battery_packs/Li_Ion/Lithium_ion_tech.html I have no detailed curves between 0-22ºC, but at 10ºC that solution works with almost all battery charge. You can choose a 0,5V drop diode in case you need it in colder environment, but yes, on near-0 or below-0 temperatures this solution will not work. Below 3.6V you can see usual discharge curve drops quickly.
excellent video...about LiFePO4: they can be discharge down to 2.0V, but at this current wouldnt it damage our ESP8266? there isnt any need for undervoltage protection?
Not many. Because of the lower max. voltage you cannot use a lipo charger (unless you disconnect manually when the voltage is reached. I use a normal bench power supply with max current and max voltage.
I have two different Chargers: One balancing charger (IMAX B6-AC Charger/Discharger 1-6 Cells) for big batteries and a small one: www.aliexpress.com/item/Liitokala-lii-260-18650-26650-16340-Lithium-Battery-Charger-Detection-of-Battery-Capacity-internal-resistance-voltage/1619052072.html All Other batteries are charged with the bench power supply.
Can a solar panel be connected to a power bank which can be connected to an esp32 module? I tried using a solar panel but then I need to add a rechargeable battery and then a battery current monitoring circuit as well to make sure that the battery doesn’t get overcharged. All of that requires a lot of materials. So I thought connecting a power bank would be a better option since it already has comparators, resistors and all that.
I really enjoy all your video... Started on esp8266 a few months back, still learning. I am looking for a circuit for connecting a 5+v PSU through a DC-DC converter at the same time as charging a 3.3v battery and use this battery as backup in case power fails. Any such circuit around? I don`t like to re-invent the wheel! Thanks, Simon from Québec.
The first question is: Do you only need 3.3V? then, you can avoid 5V and use a 3.3V regulator plus a device to load the battery (depends on the technology). If you need 5V then I would go for a higher voltage battery (6V or above) and use 2 linear regulators, one for 5V and one for 3.3 V. If you take a big 5V regulator you can put the 3.3 v in series. And a device to load the battery. This gives you stable voltages at both rails. Or you go for a 3.3 V battery and a boost converter to boost the voltage to 5V. You need a 3.3 volt battery (which in reality does not exist because no battery is stable). And also here, you need a loader for the battery. Here, only the 5v rail is stable, the 3.3V goes with the battery
Thank you, Andreas, for a fine video! You said that you had ESP8266 examples which did not run stable at voltages higher than 4 V. Do you have a video (or some other documentation) of these overpowered unstable examples too? In my experiments, if an ESP8266 system with a 3.3 V power supply isn't stable, I usually try it with 4.5 V ... 6.0 V and quite often I get a system that is more stable with that higher voltage. Haven't managed to destroy a single ESP8266 chp yet.
5 volt tolerant usually does not mean Vcc. It means, that, even if Vcc is only 3.3 v, the pins support 5 volts because they have protection circuits. The data sheet clearly says, 3 - 3.6 volt, and if the chip survives higher voltages, that is nice, but not very important for me, because I try to keep things inside the specs. I went up to 4.2 volt and the current started to increase considerably after about 3.7 volt. It is well possible that you get more output performance with a higher Vcc. That is to be expected, as long as the chip survives...
At 5:20 on the picture the two middle batteries say "LiFePo4" but the two batteries on the edges say "FePo4". They all seem to be the same battery, but its just weird.
Dear Mr Spiess Voltage question I am new in electronics I want to connect the ESP8266 to one 8x8 Led MAX 7219 The ESP needs to work around 3.3 Volts and the Max works with 4.5 Volts to 5Volts How can I use only one power source to use both without burning the ESP ? Thank you
Maybe you search for a project proposal which uses these components. That would be the safest. Otherwise: You find 3.3V regulators which work on 5 volts. These can supply your ESP32. Most ESP8255 development boards have such a regulator on board.
Hello, there are 3,7 LiPo rechargeble batteries in AA and AAA format. Eaven with protection curcuit like the big well known 18650. I use them in many flashlights, than can run on higher voltages. There are also lots of new formats, like 18350 (slightly bigger than the 123A Lipos 16350) or 16850 (size equivalent to two 123A). Search for 10440 (AAAS) and 14500 (AA).
I think, this is a big advantage of the Lipos: You get them in many different sizes. Good to know how to find tehm. Thanks! For me, the square sizes are also quite handy
You can find the square ones in all shapes at hobbyking.com If you buy "Turnigy" or "NanoTech" you get the capacity, that is printed on. For excample: I've super tiny NanoTech 300mA Square Lipos that support discharge rates up to 90C!!! The IMR LiPos are interesting to, if you need high currents. The capacity/size factor is lower but a 18350 IMR can deliver up to 8A and the IMR are a little robuster than the normal LiPos.
Hi Andreas, I have only just watched this video as I'm new to physical computing. Does your recommendation to use LiFePo4 batteries still hold true as it's been a couple of years since you made this video. Thanks.
What about a cr2 camera battery? They are lithium manganese batteries, rated at 3V and have a relatively good discharge profile. They can also delivery high currents when necessary?
Hi Andreas. Although you link to your blog, I can't find the supporting article to this video. Where exactly do you connect the AA batteries? The 5V pin (via the regulator) or to the 3.3V?
@@AndreasSpiess apologies. I'm trying to understand where to connect NiMH batteries to the Wemos D1 which has a LDO regulator on the 5V pin as well as standard USB interface for power.
Do you ever use buck converters instead of linear voltage regulators to bring the voltage of your batteries down to ~3.3v? I've got a few buck converters here but never bought voltage regulators because I'd heard they were inefficient.
Hey Andreas, I must say that I really loved your videos, watched 20 of them today! I want to use a 3,7v LiPo battery for my ESP8266 and for my ESP32, Could you please provide a link to a proper regulator that I need to use? Thanks.
This one could fit? I'm asking because the Lipo is 3.7V, and here it says input 4.1V to 7V... www.aliexpress.com/item/HOT-5pcs-lot-AMS1117-3-3V-power-supply-module-AMS1117-3-3-power-module-AMS1117-3/32581622965.html?ws_ab_test=searchweb0_0,searchweb201602_3_10152_10065_10151_10068_10307_10137_10060_10155_10154_10333_10056_10334_10335_10055_10336_10054_10059_10332_100031_10099_10103_10102_10052_10053_10142_10107_10050_10051_10172_10326_10084_10083_10080_10082_10081_10110_10111_10112_10113_10114_10312_10313_10314_10316_10078_10079_10073-10333,searchweb201603_21,ppcSwitch_3&btsid=1defd6ab-974e-4fd7-abf8-e059f0310905&algo_expid=47aad1e2-0652-49aa-9abb-1f9cae4d9bc2-9&algo_pvid=47aad1e2-0652-49aa-9abb-1f9cae4d9bc2&transAbTest=ae803_2
This one has a dropout voltage of about 1.1 volts. So, it needs a minimum of 3.0 + 1.1 = 4.1 volt at the input to get 3 volts (minimum for ESP). A better choice would be a HT7333 which has a dropout voltage of only 0.15 volt.
If you don´t need a portable application you can use a simple 220v to 3.3 buck converter. The advantage is a hassle free long life (perhaps year) application.
I love your video. I am a little confused though. I am building a control panel and I want utilize the 8266 but I want a steady yet reliable and long lasting power to the module. I am thinking of using a 5V from my SMPS. Can I achieve this?
What is your opinion about LiSOCl2-batteries by the way? They are also available in AA, ½AA and other common sizes, run at 3.6V, have a high capacity and a very slow discharge profile. They are not rechargeable, but maybe more appropriate than LiFePo4 for situations where recharging is not desired (or not possible)?
Hello Andreas, thank you very much for this video. Meanwhile almost 6 years have passed after this video. What is your experience with it today? I am looking for a solution to run an ESP8266 with a protected Soshine 5500mAh rechargeable Li-ion battery. What is your recommendation to get maximum uptime with these components? I would also like to monitor the voltage. What is your solution to get this running? Thank you and best regards, Chris
Currently, LiFePo is my favorite. and usually I create a voltage divider to reduce the battery voltage to an acceptable maximum for the ADC of the MCU. You can use quite high value resistors to save energy. And then I measure the real voltage with a multimeter and make the calculation to get the same value in the sketch.
What do you think of using a LoPo cell with a simple diode in series? I have some ancient 1N4001 diodes here and they drop the voltage by .6v. I am guessing that there are better choices for a diode, but I have no experience with them. The 4.2v would be dropped to 3.6 which is within the 8266 specs. The voltage to the 8266 would be too low at the end of the discharge curve, but most of the LiPo capacity should be available.
You can use a diode as you describe. LDOs usually have only 0.15 volt difference at the end. So, you would gain a little. But the ESP usually works below 3 volt, even if it is out of specs. I usually use a LDO because they are not expensive and I have always exact 3.3 volt (also for other parts).
What about NiZn Rechargeable Batteries (1.6V)? If you search on ebay you get plenty to choose from in AA, and two NiZn AA batteries cost about the same as one LiFePo4 AA size battery. These also might be interchangable with NiCd batteries. For some marketing reasons for NiZn batteries capacity is given in mWh instead of mAh, that gives 1.6 times grater number.
Is there a tp4056 equivalent for LiFePO4 batteries? I want to switch to using LiFePO4 batteries, but need easier solution for charging and overdischarge protection like tp4056 is for 18650 batteries.
Hi Andreas, thanks for this amazing video... Do you think I can use similar batteries to power the Ikea Motion Sensors? (They use two cr2032 3v batteries but they die every week :( )
@@AndreasSpiess Hi Andreas, thanks for answering... I've changed the product 3 (three!) times and it behaved always in the same way... The last time, fortunately, it worked. Maybe they had a broken batch, I don't know.
hi sir my sekhar from india just suggest what are the best batteries for esp8266 door sensor and esp8266 with motion sensor we are going to do with sleepmode option in it
Hi Andreas, I was just thinking, when using a Lipo cell, would it not be possible to only charge the Lipo up to 3.6V, so that you don't need an LDO when using it in the circuit? Assuming you can control what voltage the charger stops at of course. How would a lipo capacity compare to lifepo4 in this case?
@@AndreasSpiess When I was looking at charts earlier, it looked like a LiPo would be 90% charged at 3.6V, but now that I'm looking at more charts, they seem to be all over the place, most being at 50% - 80% charge at 3.6V, depending on how many amps you're drawing I think... So it definitely doesn't look like a good option if you have LiFePo4 batteries.
"just the facts, ma'am". You did the grunt work we never got round to. Very straight delivery. No puff. No spin. Nothing but the meat. Superb. Thank you.
Thanks!
prefer 14 or 131 #RufDochDeineMaMa!!!
A Victorinox/Swiss army knife is the perfect pointing device for a guy with Your accent :)
:-)
Excellent presentation and detail. Thank you for the video. And thank for simply presenting the material (very interesting) without annoying music like so many other youtube videos. Yours are the best! :)
Thanks! I also do not like music videos! But of course, they are easier to produce ;-)
My mind is the sort that understands things well, but my rote memory is terrible. So after watching all of your videos initially to get the idea of what is going on I find that with each new project I work on I simply re-view the appropriate video as if it were a topic in an engineering reference book.
There are so many wandering, rambling, directionless RUclips videos, but I find your videos are detailed and specific to a topic of interest, and always lead to an optimal solution to specific engineering problems, or at least a functional one. So I feel like I have a video library of nearly all pertinent topics to support my projects, with explanations and optimizations, and it is unnecessary for me to 'reinvent the wheel' for every new project.
This channel and your ham channel are like a detailed video version of the Handbook for Radio Amateurs and QEX magazine combined, and they provide the technical support I need to ensure that I can tackle difficult projects and they will work when I'm done.
Also, many projects online tell me to buy this and that prefabricated device and just plug it in to make a project work. But their 'this and that' are hundred dollar items! Your projects show inexpensive devices and ingenious ways to solve engineering problems. This is what good engineering is all about. Some other RUclips channels and Internet websites seem to miss that point entirely, or they are simply incapable of doing good engineering. As I am retired and have only a modest income, knowing that your channels are here to Elmer my work enables me to spend my money wisely on my projects, and your reference library of free videos continues to grow. I can't thank you enough. - John
Thank you very much for your kind words. Indeed, I sometimes go back to one of my videos for reference ;-)
Concerning cost: I have many viewers from less developed countries. This is the reason I stick with the cheaper stuff. HAMs seem to spend more money on their hobby. So this channel will feature also more expensive things.
Enjoy you hobby!
Andreas, love your work and effort! I’m starting out, and you help.
Thank you!
@@AndreasSpiessyears later im on the same path. Thanks man quality content
Mr. Spiess, I am watching this video now. Thank you.
Terry Thomas
PC Tech Support
Atlanta, Georgia USA
You are welcome!
Hey Andreas, another great video as always.
I am beginning to take notes of your videos because they are so informative
Thank you for your nice words!
Thanks Andreas - great analysis and summary as usual. Very helpful.
:-)
Dear Swiss-German guy, very interesting, thx a lot!
Bravo!
Sehr gut!
Danke!
Nowadays there are also 1.5 volt aa/aaa lithium ion batteries (through an internal regulator). You can feed the esp32 directly with two in series.
Also, there are Nickel-Zinc batteries with a typical output of 1.6v. You can also put two in series to feed the esp32, but they supply 1.8 when fully charged so it may be risky.
Thank you for the additional info!
Informative, well researched, concise and unpretentious. Would that so many other videos were like that. Thanks
Thank you for your kind words!
Thank you, just the video I was looking for.
You are welcome!
NiMH maintains voltage VERY close to 1.2V across the bulk of its discharge cycle! It starts off at about 1.45V but quickly drops down to 1.25V. This means that in contrast to other types of cells which have a more slanted voltage curve, you can pre-discharge them a little bit and you will only lose little capacity.
To accomplish that in-device, you should be able to use a Zener across the NiMH battery pack to clamp the voltage. After the battery is discharged below Zener voltage, mostly through the Zener itself, no current will flow through the Zener and it will no longer contribute to power consumption. Perhaps a similar trick can be accomplished with Li-Ion, but of course you'd need to limit the current, and i'd be wary all around, many angry pixies in there. Perhaps the proper solution with Li-Ion is not an LDO but an active boost/buck regulator - you can discharge much deeper down to typically 1.8V, and you'd have less efficiency loss.
The buck or boost converters usually take 20-40% of the energy (depending on the voltage difference and make). And they need quite a lot of current if the ESP is in deep sleep.
Your idea with a Zener should work. However, this costs some energy. Maybe it would be better to stop charging the battery at this particular voltage instead of charging it full, and then discharging it afterwards. But I am not a battery specialist.
Andreas Spiess I don't see why there should be more than a 30% loss at the IC, and can probably be optimised down to 10-ish. Perhaps a major contributor is the output choke or transformer, but those can be done with low resistance too.
The idea with the Zener is interesting. That would worth a test - and a video ;)
Very useful one, Expecting more videos on LiFePo4, charging profile details etc...
Charging profile is simple (as described in the video). As I said, the technology is new for me. So, for the moment, I do not know much more than presented in the video...
Hi Andreas, nice video! Must get a connection battery, very handy.
Thanks. I think, this battery can be done yourself. Just use a woodstick, cut it to length, and put a wire from one to the other side (maybe with a screw on both sides).
Hi Andreas, thank you for taking the time to create discharge curves and publish them in your channel. This is very useful.
As you know, some of us live in cold places and batteries do not always do well there. Most Li ion cells are optimized for temperatures above 0C and do poorly below that. I am curious to know if your LiFePo cells do well in the cold.
I never tried and there is not a lot of literature about that fact.
Hi Andreas, Dave, I live in Canada and I have been using both the Titus 3.7 V C Cell and Titus 3.7 V AA Lithium Thionyl Chloride batteries, -55C to +85C. I highly recommend them, I used them in the military for our remote sensors and GPS units, even in the Arctic. We are currently using inside transport GPS trackers for almost a year without replacing.
+ExNavy But these are non-rechargeable, right?
Yes, that's correct. For rechargeable use, I just received their new High Pulse Super Capacitor HP1520 rated at 4.0Vdc (actually 3.6Vdc)
Great presentation. I think its better to charge 18650 to only 3.6v max then ldo will not required. If powering from fully charged single 18650 then ldo is only useful will till the voltage drop from 4.2 to 3.6 after that not needed. This needs to be tested if waste or avoiding waste is better.
Then I would not use a standard Li-Ion battery. I would use a LiFePo4. You lose too much capacity from 4.2 to 3.6 volts.
LiFePo4 (lithium iron phosphate) batteries are sold for RC applications as transmitter and receiver batteries. They have a lower self-discharge rate and better temperature characteristics as LiPo, but a far higher internal resistance. Quite good for devices like outdoor temperature sensors, if you really want to bother with rechargeables in such a scenario.
Thanks for the info. You are right, I also assume in my summary, that you might use the "hassle free" alkaline technology if current consumption is small.
This is really cool. I knew about LiFePO4, but never seen them in 18650 package before. Are there some chips similar to TP4056 but for charging small LiFePO4 cells (~500mA)? Right now you need to essentially disconnect the cell and charge it somewhere else. I do have situations where I would prefer to have micro USB socket and power coming, and if it is available charge the cell (but preferably never charge with more than 3.6V), but once the USB power is gone the chip will shutdown, and ESP8266 would work just fine. TP4056 can't be used here due to different charging curves, as well too high charging voltage, which would prevent you from using it directly with ESP8266. This could serve both as a quick recharging (2-3 hours for medium size battery), or just as a backup power.
Also it is nice to know about ESP32 voltage range: 2.3V - 3.6V. The internal LDO regulators regulate it down to 1.1V / 0.9V and 1.8V for various purposes. The lower operating 2.3V is guaranteed, and it can probably operate a bit below that fine too. So you can squeeze quite a lot of power out of the battery. Possibly even too much. :D Technical specification says that it can operate down to 1.8V correctly, at least internally, but ADC will behave differently (can be compensated), and one would have trouble communicating with external chips probably. Some active cell voltage monitoring by ADC, to warn about too low voltage, and or switch off to ultra deep sleep when it goes even lower, would be a good option too. I do not know how the ESP32 behaves when the VDD is below 2.3V exactly. Optimally it would switch off completely and use less than 0.1uA in that state to prevent catastrophic (or just permanent damage) discharge of the cell.
I have exactly this question; I want to power an ESP32 with some LiFePO4 batteries, but cannot find any chargers anywhere that are reasonably priced for just some minor hobby projects. I'd be very happy if I can repurpose one of my old phone chargers, but I'll need a controller of some kind...
Andreas thx for compare. its really useful
You are welcome
With LIFEPO4 you will still need some undervoltage protection... And the main advantages of LIFEPO4 are great low temperature performance and possible high charge/discharge rates... But in case of no low temperatures I prefer to use more 18650 and LDO, it does make sense if you are not space constraint...
Great Video! Love the Swiss Army Knife at the end!! lol
Thanks! You see the Army knife even more in the mailbag videos
I'm currently field testing supply by 3 x AA Eneloop accu cells and one BS120 Schottky diode in series. (mostly because i've got lots of both components in my spare parts ;) works like a charm so far.
I had the same idea, but the problem is that the voltage of diodes is dependant on the current.
Very good analysis!
Thanks!
I use lithium 3.5v battery from discarded cell phones. It works like a charm, even when you're recording a new firmware on the device. I also use a small silicon diode in order to drop the Vcc + a little.
Diodes are ok to drop voltages. However, I prefer LDO regulators because they have a variable drop (at the beginning more, at the end less than a diode). And they are also not very expensive.
Thanks for the intro to LiFePo4. I haven't played with those yet. Why not reverse or modify your battery terminals to run a pair of LiFePo4 AA's in parallel instead of the traditional AA-Series connection? Wouldn't this be a 'best of both worlds' situation? Don't most LiPo cells have an internal over charge/discharge circuit? All of the battery packs I have torn into have them (I haven't taken apart any if the round cells).
I looked at your proposal concerning turning one cell. Not easy (the minus has a spring, the plus not). So, It would be easier to buy two single holders if you want two in parallel.
Concerning LiPo: No, not all have protection circuits. You can buy them with or without.
Thank you, that was both useful and interesting.
You are welcome!
I love the lithium batteries 😍 those are awesome.
Me too!!
I often run ESP8266's directly from a Li-ion 18650,without any real problems. I use TP4056 modules for charge/discharge control.
If these modules have a voltage regulator that is fine.
The connector batteries are interesting. I have a bunch of used but not empty 3,6 V Litium AA batterie (not accumulators however) and use them in clocks/weatherstation. I soldered wires into the battery case which also works. But this connector battries would have been the more nicely and instant reversable solution.
Perfect Video and super useful!!!
Thank you!
thank you, very informative!
+Giuseppe :-)
Hello, I know this is an old video, but I wanted to say something:
1. If you use a buck converter, you get higher efficiency and the capacity lost might be only about 5 % in the worst cases. Plus, there's almost no voltage drop so you can use it until it reaches 3 V, but for stability, it's better to cut off at 3.1 V or 3.2 V, when the battery is almost empty, and also Li-ion batteries don't like being fully discharged, so it's better for the battery.
2. There are different Li-ion chemistries. The most common are LCO (Lithium Cobalt Oxide, LiCoO2), which is the most common type and with extremely high capacity, but the most dangerous, NMC (Lithium Nickel Manganese Cobalt Oxide, LiNiMnCoO2), the next more common type, used in Panasonic's NCR18650 batteries, which are safer with almost the same capacity and LFP (Lithium Iron Phosphate, LiFePO4), the safest "cheap" ones (there's also lithium titanate, LTO, tremendously expensive), but with lower capacity.
Thank you for your information. I am only a "battery user" and do not understand too much. So I learned something...
You're welcome. A page with further information is this one: batteryuniversity.com/learn/article/types_of_lithium_ion. Battery University is a good website to learn about batteries.
3:44 correct me if I'm wrong but I think the icr14500 are AA size
You are right. I was not aware. Another commenter had the same comment. I will insert a remark.
Thanks for all your great video's! Is it also an idea to use a Lipo and a simple Germanium diode for a little bit voltage drop? Thank you. Best regards Bert-Jan Roelofs
Diodes have a constant voltage drop, also at the end of the battery life. LDOs have a variable drop.
I think LiFePo4 is the most cold tolerant of rechargeable batteries. If your device needs to work in sub zero temperatures, you may need to create some kind of heating thing to keep LiPos warm, which will likely eat most of their capacity.
Sub-zero degrees are really a problem for these batteries...
That exactly what a need .thx sir 😊
You are welcome!
Sehr interessant. Vielen Danke
700mAh at 3.2V = 2.24Wh. And seems to cost ~$3. For that price you can get NCR18650B (3.25A*3.6V=11.7Wh) or similar. Even lowered by 20% with LDO it still seems that regular 18650 is 4-5 times better deal.
If you only look at the dollars, you might be right.
Dear Andreas,
thanks for this video, i have been really looking for something like this to understand which battery i can chose for remote toy car i am planning to make for my kid using power hungry esp8266.
i just have a very small suggestion, if you find it appropriate, generally all such videos you make about comparisons and evaluations of different concepts, equipment or things, are full of very useful information. can you at the end make one slide with conclusions - in text form. because it helps in remembering the final outcome of the study.
Good idea.But it is not easy because my videos usually are already compressed. But I will think about it.
Concerning toy: motors are usually power hungry and need higher voltage. So, LiPo is probably the best solution. And a regulator for the 3.3 volt. The ESP will anyway use much less current than the motor.
I think one important point is missing. The end users are very prone to put something in the wrong direction. So in order to not to burn your IC' s you should place a low forward voltage diode (schotky maybe) in series to your power supply design. Otherwise someone will insert the batteries in wrong direction(Murphy Laws)
You are right if you want to sell your product. But a diode loses a lot of power. So it is up to you to include one.
@@AndreasSpiess But a diode saves you from having to use a regulator (thanks to forward drop) if you don't need so much current. And everyone can make mistakes including the designer.
How about pre-discharging some of the cells that are a bit too high voltage before using them? That way you could skip the regulator.
A possibility. But you would loose a lot of capacity.
Thanks ,Clear as usual but why you didn't consider lithium 3.6 non rechargeable battery
I do not know this technology.
Nicely informative video! The only thing missing is if you made a final comparison to show just how much battery life you would get from each battery option (possibly even with a comparison of their price). From your video it sounds like a LiFePo4 would give the most battery life to an ESP8266 since it has good efficiency and no need for a regulator? But unfortunately you didn't show that in your video. Maybe suitable for a follow-up video, comparing battery life & cost for powering ESP8266 using LiPo vs LiIon vs LiFePo4?
I do not plan to take cost into consideration as it changes over time. But it should be possible to assess it yourself if you look at the actual price of the batteries you consider to buy. You should have the needed data in the video.
Thanks for the video Andreas! I couldnt tell from the discharge charts, do you have stats on real-world battery life for the different options?
No, I did not do these measurements in real life
Very useful!! Thanks!
You are welcome
Good cross examination.
I would be nice to design a circuit that has the MCU check the battery voltage and switch on a voltage doubler circuit when the battery voltage runs near the shutoff voltage.
I suppose we need to check the balance because we introduce a new circuit that has it's own consumption, but we might be able to compensate it due to the extra charge we can use from working to lower voltages.
I know the LT 1026 can work as a voltage doubler but unfortunately it's not a low voltage low power chip, I don't know of any other versions.
I am not a fan of these designs because the batteries do not have a lot of energy left at the end of their "life". It adds complexity, size, and cost and I would invest all three in a bigger battery. "This provides more bang for the buck"
Actually, LiPo does come in the AA form factor: the 14500. Also, rather than using a linear regulator, you could use a buck or boost converter with most of the chemistries to not only get the needed voltage, but it could be regulated and the battery could be protected against overload and over discharge.
Thanks for the feedback concerning the AA form factor.
Buck- and boost converters are great for big voltage differences. But they usually have higher quiescent currents. So, they are not ideal for deep sleep scenarios. And for small voltage differences, for me, they are an overkill (and also can introduce RF).
Great video! What about lead acid gel cell batteries? I think they might have better cold charging tolerance than the LiPos and come in a wide variety of sizes at ~2V/cell. For non-flying applications, their density (heavy mass per cell) is not a problem, and in fact was a benefit when the wind was strong and my project box did not blow down!
You are right. If you have no space constraints this is probably the best solution. And also well known over many years.
Hello Andreas. Although I knew about LiFePO batteries I never considered using them with my 8266's but I will now! I was struggling with voltage regulators and buck convertors but maybe I can skip all that and use LiFePO's instead.By the way, in the video you mention that LiPo batteries are not available in AA packages but I believe that 14500 batteries are roughly the same size. Some LED flashlights can use AA batteries or 14500's.
You are right, they are also available in this package. So, you have the same "danger" as with the LiFePo4 batteries...
You shouldn't struggle with linear regulators, they usually work fine, if you add a beefy capacitor between plus and minus of the ESP. Maybe you watch my videos about deep sleep. There, You find some explanations and tips.
Hi Andreas just a quick question. I am planning on using the ESP8266 with one 1.5V AA battery and a boost converter (tps61006). The module will be in sleep mode quite often in our project and only woken up for short period. Do you think this is a good solution?
Love your videos. I have learnt a lot and found it very useful for our project. Thank you very much.
Thank you very much for your prompt reply and for the advice. I am using a 1000uF SMD Aluminium Electrolytic Capacitor across VCC of the ESP. I took this from one of your videos. I'll do battery test once I get my PCB done.
Am interested in how this goes for you +Venko. Have seen those inexpensive adjustable boost converters on AliExpress, and am curious. I would worry that if the ESP decides to transmit heavy, drawing 800ma, that you would need 3x as much current on the low end ... and the high end may get choppy.
TCP61006 has a 50uA quiescent current. MCP1624 has a 20uA quiescent current.
Great Video!
:-)
Thanks Andreas, very informative video. Just a thought: Instead of using an LDO regulator for for NiMH etc, what if you use diodes in series to drop the small amount of voltage drop required? Would that be more efficient than using an LDO. Of course at the device you would have a higher voltage swing over the useful portion of the discharge curve. Cheers.
I do not like diodes for that purpose. They have a constant voltage drop of ca. 0.6 volts. With a fully loaded battery this is not enough and if the battery is at the end of its live your device will stop too early. Efficiency can be a little bigger, but if you use a good regulator (see video #58) this is not a problem. The only positive could be, that a diode is cheaper than a regulator, but today, this is only pennies.
Thank you for the info :)
You are welcome!
nice, but isnt the 3xAA percentage also 80-90% then? and Lifepo4 cells are the best with 82% then?
and could i also get 5volts from the board itself(5v output pin), if the battery input is only 3.0-3.2 volts? (because some sensors require 5volts input.... even if pwm output is 3.3v....)
Most boards have a linear regulator which only converts 5 volt to 3.3 volt. Not in the other direction.
Thanks for this Video.
You're welcome!
hi, thanks for the video!! it's really informative and i will look into the lifepo4 battery for a new project with my esp32 dev kit. one thing i was wondering though: could i also connect a power bank over the micro USB port? that would mean to supply 5V to the board, but would that not be the same as to connect to the USB port of my computer when uploading a sketch? the 5 V would then be regulated don to 3 V on the board itself, at least in my understanding. thanks
The normal dev kits have voltage regulators for the 3.3 volts needed by the ESP32 and a charging IC for the battery. But these charging ICs are not compatible with LiFePo4. They have a max voltage of 4.2 volts for Li-Ion.
Hi Andreas,
Your videos are very helpful and appreciate your effort.
Whats the best way to power ESP8266 by a solar panel with a battery charger.
Thanks.
I will do a video about that when we have more sun in my country...
What about a older battery tech for stationary devices like weather station etc the 1.2v nickel iron Edison batteries. Easy even make them yourself and with the discharge of the boards you could run one years on a 3s set up
You can do that if you have them available, I think.
Where do you got the connector cells which are empty cells? Any link?
18650s are Li-Ion not Li-Po batteries. Li-Ion and Li-Po batteries may be fairly similar but they're definitely not the same. Li-Ion batteries usually have a better energy density and are more efficient, Li-Po batteries however degrade in capacity less than Li-Ion and often have a lower internal resistance and can therefore charge/discharge a lot faster than Li-Ion batteries. When connected in series Li-Po batteries drift apart more easily in voltage and therefore it is very important to balance charge them. Li-Ion however if well matched don't drift as quickly and are therefore often not balanced in laptop batteries for example (the voltages of each cell is monitored though and if they drift too far the battery will stop working). Still I would always balance either type of battery if you have multiple connected in series. Also Li-Po usually cost slightly more than Li-Ion batteries. There are other differences, but I would say those are the main differences. which depending on your project can make a very big difference. You can easily find small Li-Po batteries that can start your car without a problem. You will find no Li-Ion battery however that can deliver anywhere near enough current for that, you will definitely need a few cells in parallel (additionally to the series connection you need anyway to get around 12 V). Drones usually use Li-Po batteries for the same reason, they discharge their batteries in a couple of minutes, which is no problem if you have the right type of Li-Po batteries but can be difficult to achieve with Li-Ion batteries.
There were quite a few discussions in the comments below. Many viewers wrote that all batteries are Li-Ion. But I am no chemist ;-)
@@AndreasSpiess Of course, I'm not saying that I think your video wasn't well researched or anything like that. In the case of an ESP the difference isn't very important. Just thought it's worth mentioning that they are not identical.
Another thing worth pointing out would have been that LiFePO4 batteries while being great for an ESP while discharging, they are normally charged to 3.6 V. So powering an ESP with them directly while charging is pushing the limits, at least if you charge them fully. But those are relatively minor things. I don't think you were trying to make a 100 % accurate scientific video about batteries. And for a simple comparison of batteries the video is more than accurate enough. Just though maybe someone reading the comments would be interested in the difference.
Yo tengo muy buenos resultados con 2 Alcalina AA sin regulador, funciona hasta los 2.17v. Ahora voy a probar con una 18650 en serie con un diodo. Aunque tambien el regulador del wemos d1 es bastante eficiente.
I have very good results with 2 AA Alkaline without regulator, it works until 2.17v. Now I'm going to try a 18650 in series with a diode. Although also the regulator of wemos d1 is quite efficient.
It is good if it works, however, 2.17 volts is out of specifications
Wouldn't it be possible to build a bypass circuit that uses the current directly if it's within the spec range and goes through the converter when it isn't? So we'd lose the 20% when outside of spec, but keep 100% when within spec?
Of course, this is possible.
Old, but pretty nice and informative video. Thanks for it.
Did by any chance you found rectangular flat LiFePO4 battery somewhere in the internet? Something around 1000mAh and few millimeters thickness?
So far, I did not succeed...
I used a flat one in my mailbox notifier. But it had a higher capacity
@@AndreasSpiess thanks a lot, could you share the reference. Maybe from it I may find a flat one. Ideally, I would love to find one of 2-4mm thick
Unfortunately the supplier does no more sell the one I have
Andreas, thanks for this great comparison. I'm looking at several types of voltage regulators for use with the 3xAA solution. I want one with the input voltage as close as possible to 3.3v to take full advantage of the battery life. Do you have any particular modules or other advice on how to ensure I get the best one?
HT7333 is a good one
Hello Andreas, very informative video, as always. I'm wondering what you would use if you needed to power your solution using both 3.3V and 5V. Use case is some of the sensors (ultrasonic, PH, etc.) and displays (most of the LCD's) require 5V while ESP8266 requires 3.3V. Bi-directional logic level converters can be used for the data lines but we still need to power the sensors/displays using 5V. Thank you in advance :)
many sensor modules just require 5V because they need a stable (low noise) 3.3V supply and use a linear regulator with 1.25V drop. in a low current battery application, it might be worthwile figuring out if you can supply the sensor directly, especially if you are already using a linear regulator elsewhere for 3.3V.
If you need 5 volt and 3.3 volt just use a linear 3.3 volt regulator (e.g. AMS1117). You even do not need a LDO because the voltage difference is 1.7 volt. Just make sure, that the package can dissipate the heat.
i'm trying to figure out how to achieve long battery life like your #47 video and also supply 5v and 3.3v. Will you make a video to address it? Thank you.
very cool videos!!!
Thanks!
you should not have overlooked the advantages of NiMH cells. They can handle over charge and over discharge which the Li batteries cannot. They do not explode or leak. If you need durability and rechargability, they are the clear winner.
I did not pay too much attention to this technology because I assumed, it is well known by my viewers (it is a relatively "old" technology). But you are right: They have the advantages you describe over the Lipo.
I am still very confused about how to actually get a LiPo *into* the 8266. I hear about regulators, step up, step down, lots of things I can't seem to find any good information about. There seems to be alot of assumed knowledge about batteries out there which makes it very hard for beginners to grasp. I'd love some LiPo clarification, because those nice little square batteries look so impressive, but using them is still beyond me.
To power the ESP866 you need 3.3 volts. If your battery or other power supply has more than these 3.3 volts, you need a step-down regulator or converter (two names for the same thing). If your battery has less than 3.3 volts you need a step-up converter.
You can buy linear or switched step-down converters, but only switched step-up conveters.
Step-down converters are also called "buck converters" and step-up conveters "boost converters"
About LiIon cells, I'm using them with one power diode to lower voltage from 4,2--3,6vnto a safe 3,6--3,0v. No capacity waste there.
You are right. And because the voltage drop is always 0.6-0.7v the ESP can get unstable if you go down below 3.6 volt. the LDO has only about 0.15 volt drop in this situation. This difference sometimes matter.
The diode drop depends on current, so you need to choose wisely. About margins:
- ESP works on 3.0V to 3.6V (datasheet: mintbox.in/media/esp-12e.pdf ).
- LiIon discharge curve, from less than 25ºC, drops at 3.6V to safety level in no time: www.ibt-power.com/Battery_packs/Li_Ion/Lithium_ion_tech.html
I have no detailed curves between 0-22ºC, but at 10ºC that solution works with almost all battery charge. You can choose a 0,5V drop diode in case you need it in colder environment, but yes, on near-0 or below-0 temperatures this solution will not work.
Below 3.6V you can see usual discharge curve drops quickly.
excellent video...about LiFePO4: they can be discharge down to 2.0V, but at this current wouldnt it damage our ESP8266? there isnt any need for undervoltage protection?
Usually, the ESP8266 stops to work at about 2.8 volts. But it will still deplete the battery. Therefore it would be good to protect the battery.
@@AndreasSpiess how to protect it? just check the voltage and deep sleep if LiFePO4 battery is < 3V? or there are other protection circuits available?
very helpful
Thanks!
Are chargers available for the LiFePo4 cells? Can you charge them with a lipo charger like the battery shield for the WeMos 8266?
Not many. Because of the lower max. voltage you cannot use a lipo charger (unless you disconnect manually when the voltage is reached. I use a normal bench power supply with max current and max voltage.
Hi Andreas, Thanks for this excellent video very helpful to me, Have you work with any kind of charger for LIPO batteries?
I have two different Chargers: One balancing charger (IMAX B6-AC Charger/Discharger 1-6 Cells) for big batteries and a small one:
www.aliexpress.com/item/Liitokala-lii-260-18650-26650-16340-Lithium-Battery-Charger-Detection-of-Battery-Capacity-internal-resistance-voltage/1619052072.html
All Other batteries are charged with the bench power supply.
Thank you Andreas!
Can a solar panel be connected to a power bank which can be connected to an esp32 module? I tried using a solar panel but then I need to add a rechargeable battery and then a battery current monitoring circuit as well to make sure that the battery doesn’t get overcharged. All of that requires a lot of materials. So I thought connecting a power bank would be a better option since it already has comparators, resistors and all that.
All these things have to fit. You have to do your research and find projects which do what you want.
I really enjoy all your video... Started on esp8266 a few months back, still learning. I am looking for a circuit for connecting a 5+v PSU through a DC-DC converter at the same time as charging a 3.3v battery and use this battery as backup in case power fails. Any such circuit around? I don`t like to re-invent the wheel! Thanks, Simon from Québec.
The first question is: Do you only need 3.3V? then, you can avoid 5V and use a 3.3V regulator plus a device to load the battery (depends on the technology).
If you need 5V then I would go for a higher voltage battery (6V or above) and use 2 linear regulators, one for 5V and one for 3.3 V. If you take a big 5V regulator you can put the 3.3 v in series. And a device to load the battery. This gives you stable voltages at both rails.
Or you go for a 3.3 V battery and a boost converter to boost the voltage to 5V. You need a 3.3 volt battery (which in reality does not exist because no battery is stable). And also here, you need a loader for the battery. Here, only the 5v rail is stable, the 3.3V goes with the battery
Thank you, Andreas, for a fine video!
You said that you had ESP8266 examples which did not run stable at voltages higher than 4 V.
Do you have a video (or some other documentation) of these overpowered unstable examples too?
In my experiments, if an ESP8266 system with a 3.3 V power supply isn't stable, I usually try it with 4.5 V ... 6.0 V and quite often I get a system that is more stable with that higher voltage. Haven't managed to destroy a single ESP8266 chp yet.
Do you have any examples ?
This is extremely strange. If I would be inpolite, I would ask you to check your multimeter ;-) I never heard of something like that
onboard regulator ? :P
Good idea. That would explain it!
5 volt tolerant usually does not mean Vcc. It means, that, even if Vcc is only 3.3 v, the pins support 5 volts because they have protection circuits.
The data sheet clearly says, 3 - 3.6 volt, and if the chip survives higher voltages, that is nice, but not very important for me, because I try to keep things inside the specs.
I went up to 4.2 volt and the current started to increase considerably after about 3.7 volt. It is well possible that you get more output performance with a higher Vcc. That is to be expected, as long as the chip survives...
At 5:20 on the picture the two middle batteries say "LiFePo4" but the two batteries on the edges say "FePo4".
They all seem to be the same battery, but its just weird.
Yes, you are right.
Dear Mr Spiess Voltage question I am new in electronics I want to connect the ESP8266 to one 8x8 Led MAX 7219 The ESP needs to work around 3.3 Volts and the Max works with 4.5 Volts to 5Volts How can I use only one power source to use both without burning the ESP ? Thank you
Maybe you search for a project proposal which uses these components. That would be the safest. Otherwise: You find 3.3V regulators which work on 5 volts. These can supply your ESP32. Most ESP8255 development boards have such a regulator on board.
@@AndreasSpiess Thank you BTW you have many interesting projects
Hello,
there are 3,7 LiPo rechargeble batteries in AA and AAA format. Eaven with protection curcuit like the big well known 18650. I use them in many flashlights, than can run on higher voltages. There are also lots of new formats, like 18350 (slightly bigger than the 123A Lipos 16350) or 16850 (size equivalent to two 123A).
Search for 10440 (AAAS) and 14500 (AA).
I think, this is a big advantage of the Lipos: You get them in many different sizes. Good to know how to find tehm. Thanks!
For me, the square sizes are also quite handy
You can find the square ones in all shapes at hobbyking.com
If you buy "Turnigy" or "NanoTech" you get the capacity, that is printed on. For excample: I've super tiny NanoTech 300mA Square Lipos that support discharge rates up to 90C!!!
The IMR LiPos are interesting to, if you need high currents. The capacity/size factor is lower but a 18350 IMR can deliver up to 8A and the IMR are a little robuster than the normal LiPos.
Hi Andreas, I have only just watched this video as I'm new to physical computing. Does your recommendation to use LiFePo4 batteries still hold true as it's been a couple of years since you made this video. Thanks.
The basic chemistry did not change and the statements are still valid
@@AndreasSpiess thanks for the speedy reply.
What about a cr2 camera battery? They are lithium manganese batteries, rated at 3V and have a relatively good discharge profile. They can also delivery high currents when necessary?
3 volts max is a bit low for 3.3 volt chips. Otherwise it is ok, I thing.
@@AndreasSpiess Thank you for your response Andreas. I have decided to use 2x 18650 Li-ion batteries with a 3V3 regulator.
Hi Andreas. Although you link to your blog, I can't find the supporting article to this video. Where exactly do you connect the AA batteries? The 5V pin (via the regulator) or to the 3.3V?
Maybe you watch the video again. You should find the diagram in the video.
@@AndreasSpiess apologies. I'm trying to understand where to connect NiMH batteries to the Wemos D1 which has a LDO regulator on the 5V pin as well as standard USB interface for power.
Do you ever use buck converters instead of linear voltage regulators to bring the voltage of your batteries down to ~3.3v? I've got a few buck converters here but never bought voltage regulators because I'd heard they were inefficient.
Only if the voltage difference is high.
Hey Andreas,
I must say that I really loved your videos, watched 20 of them today!
I want to use a 3,7v LiPo battery for my ESP8266 and for my ESP32,
Could you please provide a link to a proper regulator that I need to use?
Thanks.
Just a normal 3.3v LDO like the AMS1117 or similar. I think, I even have a video about that.
This one could fit?
I'm asking because the Lipo is 3.7V, and here it says input 4.1V to 7V...
www.aliexpress.com/item/HOT-5pcs-lot-AMS1117-3-3V-power-supply-module-AMS1117-3-3-power-module-AMS1117-3/32581622965.html?ws_ab_test=searchweb0_0,searchweb201602_3_10152_10065_10151_10068_10307_10137_10060_10155_10154_10333_10056_10334_10335_10055_10336_10054_10059_10332_100031_10099_10103_10102_10052_10053_10142_10107_10050_10051_10172_10326_10084_10083_10080_10082_10081_10110_10111_10112_10113_10114_10312_10313_10314_10316_10078_10079_10073-10333,searchweb201603_21,ppcSwitch_3&btsid=1defd6ab-974e-4fd7-abf8-e059f0310905&algo_expid=47aad1e2-0652-49aa-9abb-1f9cae4d9bc2-9&algo_pvid=47aad1e2-0652-49aa-9abb-1f9cae4d9bc2&transAbTest=ae803_2
This one has a dropout voltage of about 1.1 volts. So, it needs a minimum of 3.0 + 1.1 = 4.1 volt at the input to get 3 volts (minimum for ESP). A better choice would be a HT7333 which has a dropout voltage of only 0.15 volt.
Thanks for the full answer!
If you don´t need a portable application you can use a simple 220v to 3.3 buck converter. The advantage is a hassle free long life (perhaps year) application.
You are right. But this was not the topic of this episode. I will come back to this topic in one of the next videos.
Joseph Nicholas Why “perhaps” year? You think/know an Esp8266 only last a year connected per your testing? Thanks!
I love your video. I am a little confused though. I am building a control panel and I want utilize the 8266 but I want a steady yet reliable and long lasting power to the module. I am thinking of using a 5V from my SMPS. Can I achieve this?
It runs as long as it gets 5 volts.
What is your opinion about LiSOCl2-batteries by the way? They are also available in AA, ½AA and other common sizes, run at 3.6V, have a high capacity and a very slow discharge profile. They are not rechargeable, but maybe more appropriate than LiFePo4 for situations where recharging is not desired (or not possible)?
I never used them so far. I do not know if they are able to supply the peak current of an ESP32, but for a LoRa node they would be very good, I think.
Hello,, I use NiZn Rechargeable Batteries , I was surprised they were not part of your battery comparison ?
I did not have them at this time. And I think, they are not very common. What is your experience with them?
hello sir i like your video but have a small question ....
and...
Hello Andreas, thank you very much for this video. Meanwhile almost 6 years have passed after this video. What is your experience with it today? I am looking for a solution to run an ESP8266 with a protected Soshine 5500mAh rechargeable Li-ion battery. What is your recommendation to get maximum uptime with these components? I would also like to monitor the voltage. What is your solution to get this running? Thank you and best regards, Chris
Currently, LiFePo is my favorite. and usually I create a voltage divider to reduce the battery voltage to an acceptable maximum for the ADC of the MCU. You can use quite high value resistors to save energy.
And then I measure the real voltage with a multimeter and make the calculation to get the same value in the sketch.
What do you think of using a LoPo cell with a simple diode in series? I have some ancient 1N4001 diodes here and they drop the voltage by .6v. I am guessing that there are better choices for a diode, but I have no experience with them. The 4.2v would be dropped to 3.6 which is within the 8266 specs. The voltage to the 8266 would be too low at the end of the discharge curve, but most of the LiPo capacity should be available.
You can use a diode as you describe. LDOs usually have only 0.15 volt difference at the end. So, you would gain a little. But the ESP usually works below 3 volt, even if it is out of specs.
I usually use a LDO because they are not expensive and I have always exact 3.3 volt (also for other parts).
What about NiZn Rechargeable Batteries (1.6V)? If you search on ebay you get plenty to choose from in AA, and two NiZn AA batteries cost about the same as one LiFePo4 AA size battery. These also might be interchangable with NiCd batteries. For some marketing reasons for NiZn batteries capacity is given in mWh instead of mAh, that gives 1.6 times grater number.
They are on the way to my lab. Stay tuned ;-)
Best cell to use is one that is free. 😊 Plenty free cells available in discarded phones and most of all in discarded vapes
I agree. But most of them ar of the 4.2V type.
Is there a tp4056 equivalent for LiFePO4 batteries? I want to switch to using LiFePO4 batteries, but need easier solution for charging and overdischarge protection like tp4056 is for 18650 batteries.
I did not search for it so far.
There's no ideal battery, you need to regulate the voltage, if you want your device to work properly and safely.
You are right.
Hey. What do you think about the ch123a 3v battery? Is it worth to use it as a power supply for esp12e?
You can try. They should be able to deliver the peak currents.
Hi Andreas, thanks for this amazing video...
Do you think I can use similar batteries to power the Ikea Motion Sensors? (They use two cr2032 3v batteries but they die every week :( )
If they die every week something must be wrong. Eitehr the design or the usage pattern.
@@AndreasSpiess
Hi Andreas, thanks for answering...
I've changed the product 3 (three!) times and it behaved always in the same way...
The last time, fortunately, it worked.
Maybe they had a broken batch, I don't know.
hi sir my sekhar from india just suggest what are the best batteries for esp8266 door sensor
and esp8266 with motion sensor
we are going to do with sleepmode option in it
You decide. I just provide you with the basic know-How.
Hi Andreas, I was just thinking, when using a Lipo cell, would it not be possible to only charge the Lipo up to 3.6V, so that you don't need an LDO when using it in the circuit? Assuming you can control what voltage the charger stops at of course. How would a lipo capacity compare to lifepo4 in this case?
I do not know. I think the LiPo is charged less than 50% at 6.6 volts.
@@AndreasSpiess When I was looking at charts earlier, it looked like a LiPo would be 90% charged at 3.6V, but now that I'm looking at more charts, they seem to be all over the place, most being at 50% - 80% charge at 3.6V, depending on how many amps you're drawing I think... So it definitely doesn't look like a good option if you have LiFePo4 batteries.