Schottky diode between charger and lithium cell?
Вставка
- Опубліковано 18 бер 2024
- An experiment to see what would happen if a 1N5817 Schottky diode was put between a TP4056 charge module and a lithium cell, to skew the apparent cell voltage.
Charging a lithium cell to a lower voltage like 3.9 or 4V extends its lifespan and number of charge cycles greatly.
If you enjoy these videos you can help support the channel with a dollar for coffee, cookies and random gadgets for disassembly at:- www.bigclive.com/coffee.htm
This also keeps the channel independent of UA-cam's algorithm quirks, allowing it to be a bit more dangerous and naughty.
#ElectronicsCreators - Наука та технологія
The measure of an experiment being a success is whether or not you learnt something, to some extent the outcome is immaterial.
As a science tech in a secondary school i approve this message.
Provided there are minimal casualties.😊
This is quote worthy.
Perhaps the real charger was the friends we made along the way!
@@tncorgi92 It does depend upon the casualties... 😈🤣
Brother, it is a strange world where we get amazing people like Clive that teach more in 15 minutes than most teachers do in 6 months in secondary school. May the Isle of Man put a statue up for this lad.
Damn right!!
The teachers are aiming for success in exams rather than success in life.
@@ncooper8438technically, they're teaching to pass exams, not comprehend what was taught. Makes sense, as their pay and the school budget depends upon the test results, creating a conflict of interest.
@@spvillano Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure"
If it's Clive, the statue better be erected!
3.9V should be around 90% of rated capacity, exchanging that last 10% for more than doubling the life (time and cycles) makes sense in standby applications.
Nice to see an actual test, thanks BC!
Tesla recommends charging their car to 80 percent. Would that quadruppel the life?
@@franzliszt3195 I don't know, probably not but maybe it's 2.3×? Specific datasheets are hard to come by so a big dose of salt and S.W.A.G. (Stupid Wild something Guess) is the best I can do.
When I first got my Y class 2 years ago Elon said charge to 90 percent; then about 6 months ago I saw on the screen of the car recommending charging to 80 percent. So Elon's statements should be taken with a grain of salt.@@matthewellisor5835
@@franzliszt3195 yes, if you keep the lower end of the discharge above 20% SOC, 60% DoD has ~4x the cycle life compared to 100% DoD usage for most lithium chemistries.
@@franzliszt3195not really, car batteries are rarely charged up from 0%
Nice, we love a "play with electronics featuring Clive" So Clive gets the mess and we dont damage anything. great idea 2x👍
It is good practice to have the smoke extractor next to 'The pie dish"
Had some very privileged BBC training back in the day and not to push anything more than 70-80% of its peak capacity was something special that changed my whole world!
Especially with power supplies. The cooler the better.
That's because most manufacturers (especially those in China) want "good numbers" while still trying to keep building costs to a minimum. So they overrate their products - what is 70-80% of rated load (sometimes even less) is what should be 100%, for reliabile operation.
For lithium batteries, it's not about greedy manufacturers, but the limitation of the battery chemistry it self
- one can choose to charge to 4.2 V, to get maxiumum capacity out of them but realtively few charge cycles and short lifetime - or only charge to 3.9-4 V and get 70-80% of the capacity, but like 3 times more charge cycles and longer life out.
The problem is when those batteries are used in products that's built to run the cells at "too high" voltage. For example cell phones often seem to charge the battery all the way to 4.35 V and shuts down (indicated 0% charge) already when the voltage has decreased to like 3.5-3.6 V. Unless they use some other chemistry than typical lithium ion (like the 18650 cells), that exaplains why they degrade relatively fast compared to in many other applications
I do this since many years ...
The Schottky diode is in my case between the TP5056-chip and the two protection chips. I cut the power trace, remove the solder stop and put a little glass 3 Ampere Schottky-diode at this place.
If the charging current is low (20mA), then the LiIon-Cell reaches 4.0 Volt. (drop voltage over the diode is 0.2V)
If the charging current is high (1A), then the charging stops at around 3.85 Volt. (diode drop voltage is around 0.35V)
I just realize that we need virtual assistant with Clive voice: "One moment please..." "Progress report for today" "Lets plug your phone to charger as its low on battery" :D
I like the Way you make electronics videos without all the fancy bells and whistles. Most electronics channels would have told us the same Story using a 20k Keysight Oscilloscope and other expensive gear, when all you really need is a basic meter. Never change Clive!
I don't mind a bell or whistle as long as that's what's being dissected.
Yeah, the accessibility of these videos is what makes them so special.
I was thinking to link it up to something to monitor and graph the voltages rather than take manual readings! 😊
an old school paper chart recorder would be classy - but maybe that's just me.@@stepheneyles2198
Because of this channel, I started using the TP4056 module with battery protection for my LI cells. Works fantastically well. Thanks Clive.
likewise !
I as well, they are very dependable and get the job done cheap and well. Heck I even tore down a cheap solar light recently, and the charge control chip was in fact a TP4056!
If you're designing a charge module from scratch and not relying on a ready made module, you can pull lots of stunts with the Chip Enable pin. Get yourself an MIC842 (comparator with reference) and set it up up so that it's output goes low when its input goes over 3.9V. Power it from the battery. Connect the output of the MIC842 to the CE pin of the TP4056. The battery hits 3.9V and the TP4056 turns off. You can also add an LED driven by the output of the MIC842 and that will stand in stead of the "fully charged" output of the TP4056. The MIC842 costs a buck and consumes a couple of micro amps.
While that would kind of work, the cell is not charged to its final voltage without the current first having been tapered down to zero. The simple explanation would be that during charging you see a combination of actual cell EMF and voltage drop across series resistances. This means that a comparator tripping at 3.9 would not result in a cell EOCV of 3.9, but rather something considerably lower
@@Kuremyr is it worth it to apply a test load every (x) minutes, idk maybe modulated by some chip, to see the voltage drop? probably not worth it if you want to make money selling an 80% charger tho lol
You can actually get some of the charger chips with different end voltages. That should be the easy way to stop charge at the desired voltage.
@@Kuremyr You end up with an iterative process. The cell voltage drops back, the charger restarts, etc.
You could always resort to a supervisory processor that acts to gradually reduce the current setpoint of the charger.
I also had an op amp controlled pass transistor acting as a constant voltage drop. Worked, but I never had a use for it.
@@erlendse The ones I'm familiar with are the Microchip parts. They have variants for different voltages, but all are 4.2+. If you know a part number with 3.9 or 4 I'd be interested.
No smoke? No BANG!?
Yes, it was a success. Learning from any experiment is a success
Thank you for the continued education on Li-ion cells and the proper recharging of them. Honestly, I have learned more on this channel over the years than I ever did back in my schooldays.
It's a shame these charge boards don't have more control over them. For a standby usage it would be great if it could have a selectable 'charged' voltage, then allowed the cell to relax to it's nominal voltage for a period of time before charging again, a bit like the more expensive smart chargers do...
Great to see the electronics assembly and circuit experiments back , 👍. Cheers
wow how cool. I honestly never heard of this way of soft charging like this. MORE!
Your experiment proved something I've wondered about myself.
Hmmmmmm??, is that you should only ever operate at 80% capacity ?????😏
@@brucepickess8097 maybe.. but i like to think of it of having a little something extra up my sleeve .. same thing. different perspectives 😁
This is so important! The 'last' bit from ~3.9 to 4.2V is not where the bulk of the energy sits anyway. Lower charge termination voltaqe keeps the cell safe and happy. And prevents spectacular thermal runaway..or whatever fires are called!
Good video, thanks Clive!
Nice idea.
I think it would work better if you put a 1-10K resistor parallel to the diode and a bit bigger over the battery. Then the chip can monitor the voltage better to a certain level and probably not going into 'shaking mode'. But, 'd have to try it out for the best values.
I was one of them messing with a Schottky.
I think the best are circuits that use a TL431 to cut-off the voltage.
If you Google "tl431 li-ion battery charger circuit" you'll see many circuits. Just pick a resistor so that the TL431 triggers at 3.9V instead of their 4.2V
is there a 12v version?
@@Palmit_Sorry I don't know.
@@Palmit_ Tlv431s all are set for 1.24 volts. You create a resistive divider such that the target voltage you want divides down to 1.24 volts. So you could rig it for 12V.
However you need to look at the top resistor too. As when the current goes up above the target voltage the tlv431 starts to divert the current through it such that the top resistor burns off the extra voltage. This means that the top resistor can exceed its power rating and cause a fire or burn up. So you must size it correctly.
Also you must set that resistor such that the minimum current through the tlv431 is 100uA and that the maximum current doesn't exceed whatever your particular TLV431 maxes out at. This varies from maker to maker. TI for example maxes out at 15mA through the TLV431 and has the TLVH431 which allows up to 80mA.
Really for higher voltages you would maybe want to use a TL431 instead. It is internally set for 2.49 volts so your resistive divider changes. But then you can get to even higher currents.
Asking for 12V you sound like of like you are charging a lead-acid cell. If you are, know that lead-acid batteries charge differently than Li-Ions and so the circuit may not be appropriate.
I had a PCB made recently which uses a microcontroller to stop a device from charging when the current begins to taper off. This allows you to place it in line with the power supply rather than messing with the charge circuit. I just used a current sense resistor and an FET with a USB-C connector on each end of the board. This way you can just put it in line with any USB-C device to limit the max charge. The microcontroller is programed to collect an average current at the start of charging, then switch off the FET when it sees the current drop below 10% of that average for several seconds.
My comment/guess of 4.05V during the live stream was based on a datasheet showing the termination voltage at 4.26V. I now realise that's the maximum variation with the minimum being 4.13V.
I wonder what the actual termination voltage would have been without the diode, if your sample of the chip terminates in the low range it might explain why the test ended at 3.9V.
I have a couple of lithium battery chargers which always end at 4.14V to 4.16V but I haven't looked to see if they are based on TP4056.
Interesting - thanks, Clive! Using diodes (other than zener diodes) for their forward voltage drop can be quite useful. I struggled last xmas to find an effective way to drive strings of addressable 5V LEDs with long cable runs from a Beaglebone which output 3.3V data; the thing that finally worked was to use a single WS2812B powered from 5V via a diode. It provided both level shifting on the data and cleaned up the data from the controller. Excellent blinkenlights ensued!
I do exactly this for my expensive RC model gliders. I have two batteries connected to the receiver. Each one has a Schottky diode to stop one battery charging the other. Two batteries are there in case I'm dumb and flatten one battery and that means the spare battery will at least allow me to land the glider safely. I charge the batteries on a normal charger without the diodes. The diodes are only connected to the Rx. OK, it's a little more complex than that, but that's the essence of it.
Nice idea Clive!. It would have been interesting to short out the diode when you were finished to see how the charge circuit behaved, although it would have probably jumped back into charge mode to top up the cell.
"I'll come back when something interesting happens". I was disappointed when you came back and the bench wasn't on fire. Just kidding. Very interesting results!
While the high-volume chips don't do this in order to keep down cost, I've often wished that they had a resistor to set the target voltage as they do the charging current.
While the data shows that 3.9V dramatically increases the lifespan of LMC, even 4.15V will result in a nice improvement without sacrificing relevant capacity. Unfortunately, common charging chips don't offer fine adjustments; even those that do both LMC and LFP tend to just have 2 modes and a solder bridge to switch between 3.6V and 4.2V.
Quite interesting! It does provide some option for people using these little charging modules a way of adjusting the top voltage level for the battery. No custom circuit or chips needed.
It may not be ideal, but cool to know this is an option. :)
I had the idea of using a Schottky 1n5817 as a lower drop charging diode for a 3xaa NiMh battery outdoor light with a nominal 5 volt solar panel. The problem was that the leakage current on the 1n5817 allowed the cells to discharge through the panel over night. It works fine with a 1n4007, even though it drops twice the forward voltage compared to the Schottky. Thank you for sharing!
Ouch. A 1N5817 has 1-10mA of reverse leakage current. Not the best choice. Schottky's always have fairly high leakage currents (in both directions).
You can make a perfect linear regulator out of a normal diode and a zener to ground. Solar panel -> normal-diode -> battery. And then zener from the diode/battery junction to ground (reverse biased of course for the zener). The zener must be able to dissipate the entire output of the panel, obviously.
The reverse leakage of the zener should be more around 1-5 uA.
The only other issue with using a zener is that the voltage regulation gets squirrelly below around 4.3V. So as long as the intent is to regulate to one of the standard voltages at or above that, its fine. But below 4.3V it isn't. Standard voltages: 4.3V, 4.7V, 5.1V, 5.6V, etc.
-Matt
@@junkerzn7312 I learned this lesson the hard way, too focused on voltage drop & trying to get as much from the solar panel as possible on dark winter days. For more sophisticated stuff I tend to use low drop out regulators that happily work down to a few volts, either fixed or adjustable ones with reasonable output currents at least for what I need. Thank you for your comments.
@@springwoodcottage4248 It's interesting that there is this intersection between wanting to charge little lithium batteries like these cells, and real solar + battery systems that use much larger (usually LiFePO4) batteries. Charging small packs verses large ones.
For any battery 12V or higher, a proper charge controller is really the only good solution. They are incredibly efficient too... 98% to 99% efficient, because they current-regulated buck converters taking a higher solar input voltage and bucking it down to charge the battery. Plus they do the full 3-stage or 4-stage charging algoirhtn, including float and load-support.
Because they are current regulated, the voltage regulation is fairly poor, but the charge controller still zones in on the exact desired voltage(s) when pushing against a battery while charging it. That's why they can be so efficient.
For example, take a "Victron SmartSolar MPPT 75/15". SmartSolar version rather than the BlueSolar version. The SmartSolar version has a bluetooth/phone-app support. This little charge controller is $70 and can charge 12V and 24V batteries of any chemistry, completely programmable via the bluetooth app. The only requirement is that the SOLAR input voltage by roughly +5V above the BATTERY target voltage. Which is easy since a 12V solar panel puts out 17-22V or so. (the "15" stands for maximum charging amps... 15A in this case, also programmable in 1A increments).
Really, though, when it comes to charging batteries, its impossible to beat real charge controllers outside of the price.. And the cost has dropped precipitously over the last decade. Literally just $70 for a high-end MPPT now.
Little PWM charge controllers (the Renogy is popular) are as low as $20 and still decently efficient.
-Matt
@@junkerzn7312 We are blessed to live in the time of the renewable energy revolution where the prices continue to fall & the performance continues to rise. A tiny branch from this great tree is the use of small, often free, batteries with augmented solar to run various led sculptures & lights for much longer than the small solar cells that came with the device allow. In my cases I am running an Aldi star that had no solar as bought, for several hours a night, charged by a small (3 inch x 4 inch) approx 5 volt solar panel using just a single diode as a reverse drain preventer & a two transistor plus ldr dark detector all in a snap lid plastic box from the super market, with 3 NiMh aa cells each with a capacity of 2600 mAh. This is very efficient in the use of components & practical as it does what I want.
I've discovered that if you use a protected cell and a solar panel which is capable of more than the cell's full-charge voltage but only capable of supplying a small fraction of the normal charge current (1/20th C or less), you can charge the cell through a normal diode. With the available current being so low, the cell's terminal voltage "drags" the solar panel's voltage down until the cell's terminal voltage rises high enough to trigger the protection circuit which cuts off the charge current.
I've had a protected cell from a telephone charging from a (nominally) 6 Volt solar panel this way for years, going through an automatic daylight detector which turns on a string of around 10 LED lights at night and turns them off again in the morning. I wouldn't try this with any cell which didn't have protection, though.
The LEDs provide just enough illumination in my carport to light up the drivers' side of my car at night, just in case I forget to leave my outside light on so I can see when I get home after dark.
Dear Sir BigClive! Thank you for all the knowledge in reverse engineering and in constructing life hacks!
Years ago, I ended up with a large number of Trustfire 18650 lithium battery chargers, not wanting to use them, I measured the entire batch and found one that held a maximum charge voltage of 4.15v ( due to the fine Trustfire quality control) , this one I kept, and have used for years now, and it does fully charge a battery to 4.15v , through casual observations, I believe this might extend the service life of batteries, it certainly gives one more peace of mind when charging batteries not destined for immediate service.
And best of all the Trustfire has not lived up to its name. ( you can TRUST it will set FIRE to your home)
Been picking up discarded vapes and have accumulated lots of these 'disposable' batteries .been putting them in parallel and using them in 18650 solar lamps .
Not sure if they charge correctly how about showing how to bodge the circuit to charge them properly
is
I enjoy you doing experiments like this, awesome video, thanks!
The DD28CRTA charge controller is programmable between 3.7V and 18.5V.
I wanted to use this power option for the weather monitor. The lithium battery is in buffer mode. Two diodes are connected in reverse and in parallel. Previous use of the batteries with a constant charge caused them to swell.
A big capacitor across the output of the charger would give a buffer the charger can read voltage from.
On your side
Excellent. I do this for the 3.2v Lifepo batts.
You have the patience of a saint.
With 1000x the credibility.
Oh, I have a pair of those fantasy land 20A leads too, the silicone cable is really nice and the tip is very sharp, I like them.
great experiment with a sort of successful outcome 🙂
Just out a 470uF cap across the output of the charger module, before the diode. That should stabilise it.
Just found this channel, really enjoying your content
The charger for my ebike is programmable and allows you to charge anywhere between 100% and 80%, which translates to 4.14 to 3.96 V/cell. I'd like a charger for small cells that would give me the same choice.
I love how there is a burn mark on the work bench from the plastic welder a couple videos ago 😂
I've worked with and even designed circuits that used diode voltage drop to skew reference voltages on a regulator, drop by a specific amount an output and in this, it appears that's what is intended.
I've also saw designs that used a Schottky diode for its HF switching, which wasn't as readily apparent, due to the input frequency to the diode that was passed to a circuit or more commonly, to a smoothing capacitor, as a regular protection diode wouldn't handle the frequency as cleanly.
All, niche uses for specific cases and well, good designs for what was intended.
One learns by experimenting!
Unlike some of my own experiments, this one at least didn't emit Status Smoke. ;)
Balls, fire alarm...
LinusTechTips just did one of their "handy tech under $100" videos that featured a device intended to protect battery longevity. Apparently, you just plug your charger into it and program hours on/days off based on whatever the device manufacturer recommends. Basically just a timer with a very specific intended use case. I don't remember what it was called, but I remember thinking you might find it intriguing.
Edit to add: The "Charge-O-Matic." Nice straightforward name.
setting the voltage would be nice, but this is cool and I approve. I think that was a success!
This experiment seems like a success in terms of battery life extension at the expense of a lengthy recharge cycle. Lowering final charge current and terminal voltage both reduce the tendency to plate metallic lithium which would ruin the cell. Clive went where I was thinking- a great use case for something like power fail protection for an emergency exit light - hopefully not a need for a very long run time and long periods between requirement to discharge the cell.
This method is great for storage charging as well, a diode with slightly higher voltage drop will leave you to around the storage level at 3,7-3,8v and that’s how you should store the lithium battery over time.
You’ve inspired me to try this
very interesiting :) Schottky barrier!
I wish 4V per cell were standard for all devices using li-ion cells. I repair a lot of stuff (also for other people), and I can't even count how many devices got thrown away because their battery packs were dead. Handheld vacuums, lawn mower robots, bluetooth speakers, power tools, e-scooters and -bikes, laptops, tablets, phones and many more. All for at most 20% extra running time.
What happens if you also add a 1000uF or so capacitor on the TP4056 side?
That’s the video i was hoping for. For a looong time.
Sadly this solution is not optimal. So still hoping for 4.0V or 4.1V variant of this charge chip which will never happen - at least not for this ultra low price like TP4056
I like the way you make these components flexible using connectors. I however recommend using xt-30 connectors as can be seen in my latest video. I have them on so many things that many operatoins/experiments go without soldering for me. It annoys me that I cannot reduce the charge voltage for these chargers.
Maybe put a couple of schottkys in parallel. Or a bigger one.
a bigger one would probably minimize those oscillations, as it typically have more thermal mass and less Vdrop variations.
You could also use a forward biased transistor in saturation. The C-E voltage drop is only about 0.2 Volrs. Too bad germanium transistors are hard to come by these days. The voltage drops is a little less and as current heats the junction, the voltage drop is even less. With the TP4056 limiting current, thermal run-away would not be an issue, but the heating of the germanium would allow the battery to get more current when it is discharged and then tail current off more as the junction cools when the battery nears full charge.
The problem is that if the current nears zero, the diode drop also nears zero. I was thinking about a 3K9 resistor over the cell (1mA) so the voltage drop never comes under .2V.
You could add a charge doctor that turns off the input to the TP4056 when the current drops below 900 mA. It would interrupt the charging to avoid the top-off CV portion of the charge. You would also get a report of how much capacity was added to the cell. Side note, have you tried using the IP2312? I don't think it would solve your issue because VSET only allows raising the voltage termination, not lowering it.
Luvvit .. a "one moment pleeeaaaze" feast! ❤
This scheme works well as a poor-man's LiFePo4 charger using a standard silicon diode. The TP4056 enters the oscillating mode at about 3.55V on the cell. All charging ceases at about 3.65V after a few hours. The best "diode" to use for this is a power transistor with the base shorted to the collector. Some experimentation may be needed to find the best "diodes" to use. No, LiFePo4 cells do not turn into volcanos at 3.65V.
Now that you've put a diode in series with the battery will it actually power something from the output of the 4056? My first thought would be to put two or more opposite facing diodes in parallel with the first to make it actually usable. Put a load on it and see where the cutoff is as well?
I forget how soothing your voice is lol
Excellent test, thank you!
That way of changing seems like a bad idea. As the battery approaches the charging voltage the current drops, so the Schottkey diode's voltage drop will drop.
Some diagrams show 0v drop at 0a, and if those are correct, it will still charge to the supply voltage, if you leave it long enough.
other diagrams stop before current stops, so they tell you nothing.
I don't like the sound of this either, I don't think it works like people believe, must have a play one day !
One moment please…
or one more 🍻 please?
I think I would be tempted to have a few waiting for that and I have been dry for over 7 months now 👍
Interesting video and the end result is what I expected but the journey there was unexpected. Thank you
I use low drop schottky diodes to limit maximum charge voltage to about 4.0VDC. Two diodes back to back allow for the charger to not oscillate as the charger sees the voltage return through the reverse diode. UNDERSTAND Li Ion batteries have a useable voltage range from 2.8VDC to 4.3VDC. What kills LI ion batteries other than over current? Deep discharge below 3VDC and charging to over 4DC both of which cause repid electrode deterioration. The fake "enhanced capacity" from charging over 4VDC and discharging to under 3VDC will fairly soon disappear in capacity loss. Four years on my batteries perform as almost new. If you need more capacity, use a larger battery!
Thanks Clive
Any thoughts on this idea???
Been toying with replacing some. Cr 2032 cells I use with salvaged flashing fairy lights. Ie a single white led originally used in the Christmas display signs to simulate twinkling stars.
I have modded a few to red led lights, that I attached to my dash cams to give the impression they are on 24-7 record.
I usually get 8 months between battery changes.
However I’m thinking if I use a tp4056 with a salvaged vape cell. That charges when vehicle is running I can save my self the bother of getting batteries.
So one diode on charging circuit and then 2 or 3 on the output to drop voltage to Cr 2032 voltage.
I've used a normal bridge rectifier in series with my ebike charger to prevent charge to full 42V and it did work, charged to about 40.5V. Paired with a normal switch, the limited can be bypassed if needed.
Very interesting
Maybe worth selecting a diode with a minimal forward voltage drop? 80% capacity isn't a bad thing at all, and for emergency lighting that would certainly extend the cell's useful lifespan for a few pennies spent.
Can you please explain this in more layman's terms
@@keylanoslokj1806 Essentially, Lithium battery chargers detect "end of charge" via cell voltage. Schottky diodes have a lower "conduction" voltage drop than your "normal" diodes, and this small voltage drop adds to the battery voltage, so when the charging circuitry "thinks" the cell is fully charged, it in fact isn't quite fully charged. This very slight under-charging can really extend the lifespan of lithium cells, so it's a useful (and cheap) way of keeping your cells going far longer! Hope this helps!
For convenience maybe it is good to use the tp4056 and the Schottky diode in this case but it may be better to make a circuit with a couple of transistors, leds resistors and caps and a pot to change the charging voltage to 3.9v instead, maybe not a good idea to upset the tp4056 module with oscillating current while connected to a wall charger. Lithium batteries are great but it is better to have the same care and safety level and responsibility as if dealing with a loaded gun to no have angry lithium battery issues.
Two thigs worthy of note: The 1N5817 is the higher voltage Schottky diode, maybe try the lower voltage ones? Also a 3A diode will have a much lower voltage drop, so perhaps a 20V 3A diode will do much better...
Additionally, a capacitor before the diode may help with the oscillations... I used this combination in a device I made with an aux. lithium battery that was soldered, and a removable main battery which charged to 4.2V, and fed into the aux. via the diode.
Why wouldn't you just have a current limited supply that has a voltage limit of 3.9v? At 3.9v you're still in the bulk charge region so you can charge at the bulk rate.
You would :)
@@andymouse Thats alright then... I guess when it comes to batteries people always wanted all the capacity available but for things left on charger all day except for the 5 mins you need them perhaps 80% is good enough. The other point that I didn't make is that its also better not to charge them at 1C just because you can. The lower the better but lithium don't seem to mind the charge rate too much unless they start to heat. So 0.5C will probably make quite a big different bit 0.1 not much better.
Interesting idea! I agree, it's better to use a 3.9V or 4,0V output charger.
You just reminded me that I would love to replace the 3x AAA cells in the HarborFreight $1 flashlight with some sort of rechargeable option that can recharge wirelessly.... so a) it is always a maximum brightness when fully charged, b) no more finding small screwdriver to change batteries, and c) no more cheap alkaline battery acid death.
Seems I keep having homeowner reasons to be in the crawl space / on the roof, and these are great handsfree flashlights - that suck when they go out at an inconvenient time / leak acid when not in use.
Looks like even more ideas to consider. Thanks!
Note that some cheap flashlights rely on battery impedance to limit current through the LEDs.
@@bigclivedotcom Thanks for the info. That is probably the case here, because I don't really recall seeing much of anything inside. :)
Something I had been meaning to explore, really helpful. The results were probably as expected but useful. TP4056 charging to a reduced voltage works but only slowly and to ~3.9v. I wondered if it might help to put a cap across the diode - even a small super-capacitor ?
I can imagine it has an issue when it thinks it’s full and turns off as there is then no voltage for it to sense because of the diode blocking it which would send it into the low battery recharge mode which it would rapidly jump out of as the voltage leaps up. I use a similar diode method for float charging my boat’s starter lead acid battery from the solar charged LiFePO4/lead hybrid domestics. That works well by avoiding repeatedly charging to an absorb voltage and instead goes no higher than 13.7v ish I also use a 10W halogen lamp in series so that the voltage depression on starter operation doesn’t overload the diode yet still provides a low current path when cold for the small trickle current. It added 7 years to the useful life of the starter battery.
The same diode drop concept can be used to keep a battery-less Android phone running as a webcam for example. The phone without its battery can remain plugged in indefinitely. There would be no battery maintained at full charge to its detrement.
The phone would be powered by its charging port while at its battery connection connectors in its empty battery bay one feeds a faked sub-charged battery voltage using a diode rated above the phone’s power draw. Often these phones will not operate battery-less while powered by its charging port alone. Here the phone thinks it has an almost charged battery, so it operates.
Would Mr. BigClive consider making a definitive video on this topic? There are a lot of old Android phones sitting around waiting for a purpose suitable only when there is no phone battery to become like successful Jiffy Pop popcorn.
Shhhchkematic! U made my day. Thank You!
A Schottky diode's voltage drop disappears at very low current, and measuring that is actually rather difficult because your multi-meter also pushes a bit of current to measure the voltage. Basically, though, the problem is that you can't get a reliable output voltage if when the intent is to fully or nearly-fully charge a battery. It will happily trickle-charge at ones or tens of microamps right up to the input voltage as if no diode were there at all, if you let it.
You can force a more reliable voltage drop by forcing a minimum current through the diode by putting a resistor to the ground return after the diode. Usually a few mA will do it, so like a 1Kohm resistor does the job. But of course, if the charger stops charging entirely now the resistor will start draining the battery you just spent so much effort charging up!
This also means that it can work quite well for larger LiFePO4 battery systems that always have some sort of load on them. You can maintain a grid-supported "float" voltage in a UPS application or for low-battery load support where the voltage is lower and it doesn't matter that the output voltage is a bit squirrelly.
I do precisely this on a 48V solar + battery backup system using a Mean Well HLG-320H-54A power supply (320W, isolated, adjustable current and voltage limits, and fanless). I run the output through a set of 4 x 20SQ060's in parallel to provide low-battery load-support from the grid.
But you don't want to "float" a NMC or NCA lithium chemistry at any voltage. Not really. LiFePO4 can be floated. NMC/NCA should not be.
-Matt
The cycling, is the 4056 turning off and checking battery voltage then turning on again ?
During the check the diode isn't letting it happen, so it's seeing nothing
Maybe another diode in reverse ?
I tried that, and also a parallel high value resistor. But no difference.
Thanks Clive, I can now save some time during the Zombie Apocalypse when I'm whacking together my scrounged parts to make equipment.
What would be the alternative budget charge module where you can select cut-off voltage and charge current?
Nice experiment, a slightly lower Vf schottky might make this worthwile, you can get ones with about 150mV@100mA which would let you charge to 4.05V. But I agree a controller with an adjustable termination voltage would be better.
That behaviour at the end may be its low charge recovery which allows you to rescue a cell which had discharged out of the 'safe' range by effectively trickle charging the cell until it reaches a recognized voltage.
It good to know there are other electro curious people in the world.
"Electro Curious" hmmmmmm, is that for persons of a persuation who enjoy mutual coupling as in electronics, electric power and telecommunication, coupling is the transfer of electrical energy from one circuit to another, or between parts of a ........ Ooooo, Errrrrrrrr.😏
Talk to Clive about other things he's 'Curious' about !
When I go into the electronics store, they always put the vintage non -SMD resistors and transistors in a plain brown paper bag. So considerate of them.
Once I bought a filament bulb rather than a LED and the young lass behind the counter went bright red and didn't know where to look. So embarrassing. @@brucepickess8097
My local electronics store has a non SMD component section down the back behind a curtain. They put the resistors and transistors in a plain brown paper bag so as not to embarrass me.
Recently I bought a filament type pilot bulb instead of a SMD dual color led.
I know some people say it's not natural and extravagant. The young lass behind the counter turned bright red and didn't know where to look. @@andymouse
My local electronics store has a non SMD component section down the back behind a curtain. They put the resistors and transistors in a plain brown paper bag so as not to embarrass me. Recently I bought a filament type pilot bulb instead of a smd dual color led. I know some people say it's not natural and extravagant. The young lass behind the counter turned bright red and didn't know where to look. @@brucepickess8097
Great video !! Also - Love those connectors with the voltage probe slots !! What are they called ? Where did you purchase them ?
Check on eBay for NS25 connectors. I buy mine from a UK supplier called Rapid Electronics and tend to refer to them as miniature Molex connectors.
Thanks a ton for the response and the info Bigclive!!!
Until I realised that you meant "lithium-ion", I thought we were in for a fireworks display! My EV stops at 4.16 V as 100% which allows it to achieve a viable product life but I mostly maintain it between 3.65 (45%) and 3.8 OCV (65%).
You have to be careful with that when dealing with any lithium pack more than 3.7v (1S) because packs have a balancing system that works by bleeding charge above a certain voltage, like 4.18v per cell. If you reduce the charge voltage, it messes up the balancing.
Hi Clive, i see that you have been using that Plastic welder device on your lovely wooden bench 👍
I agree with your conclusion. A charger IC with a programmable upper voltage will be better. Also i looked at 1N5817 (1A Schottky) from Onsemi. At 100mA the forward voltage is 0.29 volt (ish). But at 500mA the same voltage is 0.37 volt ish. This may have perhaps have confused the charger IC.
I realized that the charger has to stop periodically and monitor the battery voltage at zero charging current. But with a single Schottky diode blocking the voltage in the reverse direction, all the charger sees is some very small leakage current through the diode. So what may happen is the diode's leakage current may be highly dependent on the temperature, making the maximum charge unreliable.
Adding another diode in antiparallel with the original is not a solution. It will subtract its voltage from the battery, giving the charger the voltage showing it's undercharged. It may be best to instead put a high value resistor across the diode to swamp out the variable leakage current and give a stable charge. I'm not sure what value, but I would experiment with 100k and go down to see if it makes a difference.
I tried the inverse Parallel diode and a resistor and neither made a significant difference.
It's more likely that military and other "rugged" industry applications go for a lower charge voltage because Li-tech batteries are more reliable in high-temperature locations if they aren't charged until the nuts fall off, so to speak. Charging a battery to its max voltage in an equipment bay reaching 50 deg C plus is going to kill it quite quickly and that would mean downtime to fix the failure whereas the battery is more likely to survive for longer in such conditions if it's only charged to, say, 4.0V maximum.
Next project, reverse-engineer the classic TP4506 module and redesign it/hack it to charge batteries to 4.0V without using an external diode dropper.
Sí, se sabe que cargar al 100% una batería de litio acorta su vida y descargarla por debajo del 10% también. Limitar la tensión de carga me parece una idea bastante interesante y fácil de implementar, en aplicaciones donde la vida de la batería es algo importante, sacrificando unos mA/h de almacenamientos.
De todos modos, hay muchos factores que también acortan o alargan la vida, sobre todo, la calidad de la química de la batería, las temperaturas de carga, la antigüedad de la batería, si hay picos de descargak
How easy would it be to modify the charge board to cut off at a lower voltage? or is it permanantly set by the chip?
Unfortunately thats the rub, most of the chips including this one have a regulated output set internally. You would think there would be a solution available for those who want to set both voltage and current.
Have you tried limiting the input voltage of the charging chip? Datasheet says minimum input is 4V
That's worth trying.
It would be interesting to plug it into an oscilloscope and see exactly what it happening when it is oscillating.
I have been using that trick to fully charge Li/FePO4 cells for a few years now.
The diode I use is rated for higher current and has a nominal drop of 0.445 vdc.
I’ve always wondered how these things "measure" the voltage across the cell. Does it pause the charging for a bit then measure the voltage, across a high impedance load? Technically voltage can’t be measured without a tiny current, right?
I think it goes by its own output voltage.
Right at the start you mentioned this was a disposable cell that you were charging? I thought it wasn't safe to charge non-chargeble LiIon cells. Or was this cell actually a chargeable cell?
They use rechargeable cells because they are cheaper than non rechargeable ones.
So would this be a feasible way to force a DW01 li-ion protection chip to stop charging when a voltage has been reached?
AFAIK, you can keep a li-ion cell 'floating', i.e. keep it hooked up to a constant voltage supply for a prolonged time, as long as that voltage is sefely below the fully charged voltage of 4.2V. Something like in the range of 3.7 to 4.0V should be fine. Of course the catch is - that power supply has to be current limited, otherwise you can have a bad time it you hook up a discharged cell.
But, if it is current limited, you could theoretically just hook it up overnight and come back in the morning to a cell charged up to your desired voltage.
You don't keep them hooked up for storage, thats just BS the manufacturers will give you a shelf voltage to charge (or discharge ) to for storage and its typically 3.5 ish not overly critical and of course cheap cells with high internal resistances will self discharge quite quickly.
@@andymouseI dont think hes suggesting you 'should' more that you 'can' (as long as the power supply is below the fully charge voltage, as as soon as the power supply voltage matches the battery voltage there will be no potential difference, and then no current will go into the battery, so it wont be charging anymore.
Although you shouldn't do this in my opinion, as if the cell has a high internal resistance, or gets warm when charging, has dendrites etc, it will sit there burning off power in the form of heat, which will leave it constantly charging = constantly getting hotter and hotter until it goes nuclear. Its in my opinion the common failure mode of most cheap Chinese battery packs (that lack thermal protection, cell level voltage monitoring). And why its important in my opinion to never leave Li batteries charging unattended. You never know if a cell has gone bad and if there is sufficient protection to stop it going nuclear.
It seems we agree but people shouldn't chime in with BS regardless of 'should' or 'can' :)@@ratgreen