Integrated power management circuits from ON Semiconductor (ONS) are already well known to domestic developers. These are AC/DC converters and PWM controllers, power factor correctors, DC/DC converters and, of course, linear regulators. However, almost none portable device cannot do without a battery and, accordingly, without microcircuits to charge and protect it. The ONS company has in its product line a number of solutions for managing battery charge, which, traditionally for ONS, combine sufficient functionality with low cost and ease of use.
In modern electronics, the most common are NiCd/NiMH and Li-Ion/Li-Pol batteries. Each of them has its own advantages and disadvantages. Nickel-cadmium (NiCd) batteries are cheap and also have the most a large number of discharge/charge cycles and high load current. The main disadvantages are: high self-discharge, as well as the “memory effect”, which leads to a partial loss of capacity when frequently charging an incompletely discharged battery.
Nickel metal hydride (NiMH) batteries is an attempt to eliminate the shortcomings of NiCd, in particular the “memory effect”. These batteries are less critical to charging after incomplete discharge and are almost twice as high as NiCd in terms of specific capacity. There are also losses; NiMH batteries have a lower number of discharge/charge cycles and a higher self-discharge compared to NiCd.
Lithium-ion (Li-Ion) batteries have the highest energy density, which allows them to surpass other types of batteries in terms of capacity with the same overall dimensions. Low self-discharge and the absence of a “memory effect” make this type of battery unpretentious to use. However, to ensure safe use, lithium-ion batteries require the use of technologies and design solutions (polyolefin films to insulate the positive and negative electrodes, the presence of a thermistor and a safety valve to relieve excess pressure), which lead to an increase in the cost of lithium-based batteries compared to other power elements.
Lithium polymer (Li-Pol) batteries is an attempt to solve the safety problem of lithium-based batteries by using a solid dry electrolyte instead of the gel electrolyte in Li-Ion. This solution allows you to obtain characteristics similar to Li-Ion batteries at a lower cost. In addition to increased safety, the use of solid electrolyte allows the thickness of the battery to be reduced (up to 1.5 mm). The only drawback compared to Li-Ion batteries is the smaller operating temperature range; in particular, Li-Pol batteries are not recommended to be charged at sub-zero temperatures.
Today's portable applications require the fastest battery charging possible, avoiding overcharging, maximizing battery life, and preventing capacity loss. MC33340 And MC33342- charge controllers from ON Semiconductor, which combine everything you need to quickly charge and protect NiCd and NiMH batteries.
MC33340/42 controllers implement:
These controllers, combined with an external linear or pulse converter, form a complete battery charging system. An example of such a charging circuit using a classic stabilizer LM317 shown in Fig. 1.
Rice. 1.
LM317 in this circuit works as a stabilized current source with the charging current set by resistor R7:
I chg(fast) = (V ref + I adjR8)/R7. The trickle charging current is set by resistor R5:
I chg(trickle) = (V in - V f(D3) - V batt)/R5. The R2/R1 divider must be designed in such a way that when the battery is fully charged, the Vsen input is less than 2 V:
R2 = R1(V batt /V sen - 1).
Using pins t1, t2, t3, three-bit logic (keys in the diagram) sets either the charging time to 71...283 minutes, or the upper and lower limits of temperature detection.
Based on the presented circuit, ON Semiconductor offers development boards MC33340EVB And MC33342EVB.
Lithium batteries require high stability of the charging voltage, for example, for the LIR14500 battery from EEMB, the charging voltage must be within 4.2±0.05 V. For charging lithium-based batteries, ONS offers a fully integrated solution - NCP1835B. This is a charge chip with a linear regulator, a CCCV (constant current, constant voltage) charge profile and a charging current of 30...300 mA. Nutrition NCP1835B can be carried out either from a standard AC/DC adapter or from a USB port. A variant of the connection circuit is shown in Fig. 2.
Rice. 2.
Main characteristics:
One more important point in battery charging systems is protection against exceeding the permissible input voltage. ONS solutions disconnect the output from the target circuit when an unacceptable voltage is present at the input.
NCP349- a new product from ONS that protects against input overvoltage up to 28 V. The microcircuit turns off the output when the input voltage exceeds the upper threshold or if the lower threshold is not reached. A FLAG# output is also provided to indicate input overvoltage. Typical scheme application is shown in Fig. 3.
Rice. 3.
This microcircuit is available with various lower (2.95 and 3.25 V) and upper (5.68; 6.02; 6.4; 6.85 V) response thresholds, which are encoded in the name. NCP360 has the same functionality as NCP349, except for the maximum input voltage: 20 V.
ON Semiconductor, compared to its competitors, does not have a very wide range of microcircuits for charging batteries. However, the presented solutions in their segment are characterized by competitive characteristics and price, as well as ease of use.
All radio amateurs are very familiar with charge boards for one can of li-ion batteries. It is in great demand due to its low price and good output parameters.
Used to charge the previously mentioned batteries at a voltage of 5 Volts. Similar scarves are found wide application in homemade designs with an autonomous power source in the form of lithium-ion batteries.
These controllers are produced in two versions - with and without protection. Those with protection are a little expensive.
Protection performs several functions
1) Disconnects the battery during deep discharge, overcharging, overload and short circuit.
Today we will check this scarf in great detail and understand whether the parameters promised by the manufacturer correspond to the real ones, and we will also arrange other tests, let's go.
The board parameters are shown below
And these are the circuits, the top one with protection, the bottom one without
Under a microscope it is noticeable that the board is of very good quality. Double-sided glass fiber laminate, no “couples”, silk-screen printing is present, all inputs and outputs are marked, it is not possible to mix up the connection if you are careful.
The microcircuit can provide a maximum charge current of around 1 Ampere; this current can be changed by selecting the resistor Rx (highlighted in red).
And this is a plate of the output current depending on the resistance of the previously indicated resistor.
The microcircuit sets the final charging voltage (about 4.2 Volts) and limits the charging current. There are two LEDs on the board, red and blue (colors may be different). The first lights up during charging, the second when the battery is fully charged.
Available Micro USB connector where 5 volts is supplied.
First test.
Let's check output voltage, to which the battery will be charged, it should be from 4.1 to 4.2V
That's right, no complaints.
Second test
Let's check the output current, on these boards the maximum current is set by default, and this is about 1A.
We will load the output of the board until the protection is triggered, thereby simulating high consumption at the input or a discharged battery.
The maximum current is close to the declared one, let's move on.
Test 3
Connected to the battery location laboratory block power supply on which the voltage is pre-set around 4 volts. We reduce the voltage until the protection turns off the battery, the multimeter displays the output voltage.
As you can see, at 2.4-2.5 volts the output voltage disappeared, that is, the protection is working. But this voltage is below critical, I think 2.8 Volts would be just right, in general, I do not advise discharging the battery to such an extent that the protection will work.
Test 4
Checking the protection current.
For these purposes, an electronic load was used; we gradually increased the current.
The protection operates at currents of about 3.5 Amps (clearly visible in the video)
Among the shortcomings, I will only note that the microcircuit heats up ungodly and even a heat-intensive board does not help. By the way, the microcircuit itself has a substrate for effective heat transfer and this substrate is soldered to the board, the latter plays the role of a heat sink.
I don’t think there’s anything to add, we saw everything perfectly, the board is an excellent budget option when we're talking about about a charge controller for one can of small-capacity Li-Ion battery.
I think this is one of the most successful developments of Chinese engineers, which is available to everyone due to its insignificant price.
Happy Stay!
First you need to decide on the terminology.
As such there are no discharge-charge controllers. This is nonsense. There is no point in managing the discharge. The discharge current depends on the load - as much as it needs, it will take as much. The only thing you need to do when discharging is to monitor the voltage on the battery to prevent it from overdischarging. For this purpose they use .
At the same time, separate controllers charge not only exist, but are absolutely necessary for the implementation of the process li-ion charging batteries. They set the required current, determine the end of the charge, monitor the temperature, etc. The charge controller is an integral part of any.
Based on my experience, I can say that a charge/discharge controller actually means a circuit for protecting the battery from too deep a discharge and, conversely, overcharging.
In other words, when we talk about a charge/discharge controller, we are talking about the protection built into almost all lithium-ion batteries (PCB or PCM modules). Here she is:
And here they are too: 
It is obvious that protection boards are presented in various form factors and are assembled using different electronic components. In this article we will look at options for protection circuits for Li-ion batteries (or, if you prefer, discharge/charge controllers).
Since this name is so well established in society, we will also use it. Let's start with, perhaps, the most common version on the DW01 (Plus) chip.
Such a protective board for li-ion batteries is found in every second mobile phone battery. To get to it, you just need to tear off the self-adhesive with inscriptions that is glued to the battery.
The DW01 chip itself is six-legged, and two field-effect transistors are structurally made in one package in the form of an 8-legged assembly.
Pin 1 and 3 control the discharge protection switches (FET1) and overcharge protection switches (FET2), respectively. Threshold voltages: 2.4 and 4.25 Volts. Pin 2 is a sensor that measures the voltage drop across field-effect transistors, which provides protection against overcurrent. The transition resistance of transistors acts as a measuring shunt, so the response threshold has a very large scatter from product to product.
The whole scheme looks something like this: 
The right microcircuit marked 8205A is the field-effect transistors that act as keys in the circuit.
SEIKO has developed specialized chips to protect lithium-ion and lithium polymer batteries from overdischarge/overcharge. To protect one can, integrated circuits of the S-8241 series are used. 
Overdischarge and overcharge protection switches operate at 2.3V and 4.35V, respectively. Current protection is activated when the voltage drop across FET1-FET2 is equal to 200 mV.
A similar protection scheme for lithium single-cell batteries with protection against overdischarge, overcharge, and excess charge and discharge currents. Implemented using the LV51140T chip. 
Threshold voltages: 2.5 and 4.25 Volts. The second leg of the microcircuit is the input of the overcurrent detector (limit values: 0.2V when discharging and -0.7V when charging). Pin 4 is not used.
The circuit design is similar to the previous ones. In operating mode, the microcircuit consumes about 3 μA, in blocking mode - about 0.3 μA (letter C in the designation) and 1 μA (letter F in the designation). 
The R5421N series contains several modifications that differ in the magnitude of the response voltage during recharging. Details are given in the table:
Another version of the charge/discharge controller, only on the SA57608 chip. 
The voltages at which the microcircuit disconnects the can from external circuits depend on letter index. For details, see the table:
SA57608 consumes quite a large current in sleep mode - about 300 μA, which distinguishes it from the above-mentioned analogues in the worst side(the currents consumed there are of the order of fractions of a microampere).
And finally, we offer an interesting solution from one of the world leaders in the production of electronic components On Semiconductor - a charge-discharge controller on the LC05111CMT chip. 
The solution is interesting in that the key MOSFETs are built into the microcircuit itself, so all that remains of the add-on elements are a couple of resistors and one capacitor.
The transition resistance of the built-in transistors is ~11 milliohms (0.011 Ohms). The maximum charge/discharge current is 10A. The maximum voltage between terminals S1 and S2 is 24 Volts (this is important when combining batteries into batteries).
The microcircuit is available in the WDFN6 2.6x4.0, 0.65P, Dual Flag package.
The circuit, as expected, provides protection against overcharge/discharge, overload current, and overcharging current.
It is important to understand that the protection module and charge controllers are not the same thing. Yes, their functions overlap to some extent, but calling the protection module built into the battery a charge controller would be a mistake. Now I’ll explain what the difference is.
The most important role of any charge controller is to implement the correct charge profile (usually CC/CV - constant current/constant voltage). That is, the charge controller must be able to limit the charging current at a given level, thereby controlling the amount of energy “poured” into the battery per unit of time. Excess energy is released in the form of heat, so any charge controller gets quite hot during operation.
For this reason, charge controllers are never built into the battery (unlike protection boards). The controllers are simply part of a proper charger and nothing more.
In addition, not a single protection board (or protection module, whatever you want to call it) is capable of limiting the charge current. The board only controls the voltage on the bank itself and, if it goes beyond predetermined limits, opens the output switches, thereby disconnecting the bank from the outside world. By the way, short circuit protection also works on the same principle - when short circuit The voltage on the bank drops sharply and the deep discharge protection circuit is triggered.
Confusion between the protection circuits for lithium batteries and charge controllers arose due to the similarity of the response threshold (~4.2V). Only in the case of a protection module, the can is completely disconnected from the external terminals, and in the case of a charge controller, it switches to the voltage stabilization mode and gradually reduces the charging current.
It's no secret that Li-ion batteries do not like deep discharge. This causes them to wither and wither, and also increase internal resistance and lose capacity. Some specimens (those with protection) can even plunge into deep hibernation, from where it is quite problematic to pull them out. Therefore, when using lithium batteries, it is necessary to somehow limit their maximum discharge.
To do this, special circuits are used that disconnect the battery from the load at the right time. Sometimes such circuits are called discharge controllers.
Because The discharge controller does not control the magnitude of the discharge current; strictly speaking, it is not a controller of any kind. In fact, this is an established but incorrect name for deep discharge protection circuits.
Contrary to popular belief, the built-in batteries (PCB boards or PCM modules) are not designed to limit the charge/discharge current, or to timely disconnect the load when fully discharged, or to correctly determine the end of charging.
Firstly, Protection boards, in principle, are not capable of limiting the charge or discharge current. This should be handled by the memory department. The maximum they can do is turn off the battery when there is a short circuit in the load or when it overheats.
Secondly, Most protection modules turn off the li-ion battery at a voltage of 2.5 Volts or even less. And for the vast majority of batteries, this is a very strong discharge; this should not be allowed at all.
Third, The Chinese are riveting these modules in the millions... Do you really believe that they use high-quality precision components? Or that someone out there tests and adjusts them before installing them in batteries? Of course, this is not true. When producing Chinese motherboards, only one principle is strictly observed: the cheaper, the better. Therefore, if the protection disconnects the battery from the charger exactly at 4.2 ± 0.05 V, then this is more likely a happy accident than a pattern.
It’s good if you got a PCB module that will operate a little earlier (for example, at 4.1V). Then the battery simply won’t reach ten percent of its capacity and that’s it. It is much worse if the battery is constantly recharged, for example, to 4.3V. Then the service life is reduced and the capacity drops and, in general, may swell.
It is IMPOSSIBLE to use the protection boards built into lithium-ion batteries as discharge limiters! And as charge limiters too. These boards are intended only for emergency battery disconnection in case of emergency situations.
Therefore, separate circuits for limiting charge and/or protecting against too deep discharge are needed.
Simple charging device on discrete components and specialized integrated circuits we looked at . And today we’ll talk about the solutions that exist today to protect a lithium battery from too much discharge.
To begin with, I propose a simple and reliable Li-ion overdischarge protection circuit, consisting of only 6 elements. 
The ratings indicated in the diagram will result in the batteries being disconnected from the load when the voltage drops to ~10 Volts (I made protection for 3 series-connected 18650 batteries in my metal detector). You can set your own shutdown threshold by selecting resistor R3.
By the way, the full discharge voltage Li-ion battery is 3.0 V and no less.
A field grass (such as in the diagram or similar) can be dug out from an old motherboard from the computer, usually there are several of them at once. TL-ku, by the way, can also be taken from there.
Capacitor C1 is needed for the initial startup of the circuit when the switch is turned on (it briefly pulls the gate T1 to minus, which opens the transistor and powers the voltage divider R3, R2). Further, after charging C1, the voltage required to unlock the transistor is maintained by the TL431 microcircuit.
Attention! The IRF4905 transistor indicated in the diagram will perfectly protect three lithium-ion batteries connected in series, but is completely unsuitable for protecting one 3.7 Volt bank. It is said how to determine for yourself whether a field-effect transistor is suitable or not.
The downside of this circuit: in the event of a short circuit in the load (or too much current consumed), the field-effect transistor will not close immediately. The reaction time will depend on the capacitance of capacitor C1. And it is quite possible that during this time something will have time to burn out properly. A circuit that instantly responds to a short load under load is presented below: 
Switch SA1 is needed to “restart” the circuit after the protection has tripped. If the design of your device provides for removing the battery to charge it (in a separate charger), then this switch is not needed.
The resistance of resistor R1 must be such that the TL431 stabilizer reaches operating mode at a minimum battery voltage - it is selected in such a way that the anode-cathode current is at least 0.4 mA. This gives rise to another drawback of this circuit - after the protection is triggered, the circuit continues to consume energy from the battery. The current, although small, is quite enough to completely drain a small battery in just a couple of months.
The diagram below for self-made monitoring of the discharge of lithium batteries is free from this drawback. When the protection is triggered, the current consumed by the device is so small that my tester does not even detect it. 
Below is a more modern version of the discharge limiter lithium battery using stabilizer TL431. This, firstly, allows you to easily and simply set the desired response threshold, and secondly, the circuit has high temperature stability and clear shutdown. Clap and that's it! 
Getting TL-ku today is not a problem at all, they are sold for 5 kopecks per bunch. Resistor R1 does not need to be installed (in some cases it is even harmful). Trimmer R6, which sets the response voltage, can be replaced with a chain of constant resistors with selected resistances.
To exit the blocking mode, you need to charge the battery above the protection threshold, and then press the S1 “Reset” button.
The inconvenience of all the above schemes is that to resume operation of the schemes after going into protection, operator intervention is required (turn SA1 on and off or press a button). This is the price to pay for simplicity and low power consumption in lock mode.
The simplest li-ion overdischarge protection circuit, devoid of all disadvantages (well, almost all) is shown below: 
The principle of operation of this circuit is very similar to the first two (at the very beginning of the article), but there is no TL431 microcircuit, and therefore its own current consumption can be reduced to very small values - about ten microamps. A switch or reset button is also not needed; the circuit will automatically connect the battery to the load as soon as the voltage across it exceeds a preset threshold value.
Capacitor C1 suppresses false alarms when operating on a pulsed load. Any low-power diodes will do; it is their characteristics and quantity that determine the operating voltage of the circuit (you will have to select it locally).
Any suitable n-channel field effect transistor can be used. The main thing is that it can withstand the load current without straining and be able to open at low gate-source voltage. For example, P60N03LDG, IRLML6401 or similar (see).
The above circuit is good for everyone, but there is one unpleasant moment - the smooth closing of the field-effect transistor. This occurs due to the flatness of the initial section of the current-voltage characteristic of the diodes.
This drawback can be eliminated using modern element base, namely, using micro-power voltage detectors (power monitors with extreme low power consumption). The next circuit for protecting lithium from deep discharge is presented below: 
MCP100 microcircuits are available in both DIP packages and planar versions. For our needs, a 3-volt option is suitable - MCP100T-300i/TT. Typical current consumption in blocking mode is 45 µA. The cost for small wholesale is about 16 rubles/piece.
It’s even better to use a BD4730 monitor instead of the MCP100, because it has a direct output and, therefore, it will be necessary to exclude transistor Q1 from the circuit (connect the output of the microcircuit directly to the gate of Q2 and resistor R2, while increasing R2 to 47 kOhm).
The circuit uses a micro-ohm p-channel MOSFET IRF7210, which easily switches currents of 10-12 A. The field switch is fully open already at a gate voltage of about 1.5 V, and in the open state it has negligible resistance (less than 0.01 Ohm)! In short, a very cool transistor. And, most importantly, not too expensive.
In my opinion, the last scheme is the closest to the ideal. If I had unlimited access to radio components, I would choose this one.
A small change in the circuit allows you to use an N-channel transistor (then it is connected to the negative load circuit): 
BD47xx power supply monitors (supervisors, detectors) are a whole line of microcircuits with response voltages from 1.9 to 4.6 V in steps of 100 mV, so you can always choose them to suit your purposes.
Any of the above circuits can be connected to a battery of several batteries (after some adjustment, of course). However, if the banks have different capacities, then the weakest of the batteries will constantly go into a deep discharge long before the circuit operates. Therefore, in such cases, it is always recommended to use batteries not only of the same capacity, but preferably from the same batch.
And although this protection has been working flawlessly in my metal detector for two years now, it would still be much more correct to monitor the voltage on each battery personally.
Always use your personal Li-ion battery discharge controller for each jar. Then any of your batteries will serve you happily ever after.
In all of the above schemes for protecting lithium-ion batteries from deep discharge, MOSFETs operating in switching mode are used. The same transistors are usually used in overcharge protection circuits, short-circuit protection circuits, and in other cases where load control is required.
Of course, in order for the circuit to work as it should, the field-effect transistor must meet certain requirements. First, we will decide on these requirements, and then we will take a couple of transistors and, according to their datasheets (according to technical specifications) let's determine whether they are suitable for us or not.
Attention! We will not consider the dynamic characteristics of FETs, such as switching speed, gate capacitance and maximum pulsed drain current. These parameters become critically important when the transistor operates at high frequencies (inverters, generators, PWM modulators, etc.), however, discussion of this topic is beyond the scope of this article.
So, we must immediately decide on the circuit that we want to assemble. Hence the first requirement for a field-effect transistor - it must be the right type(either N- or P-channel). This is the first one.
Let's assume that the maximum current (load current or charge current - it doesn't matter) will not exceed 3A. This leads to the second requirement - a field worker must withstand such current for a long time.
Third. Let's say our circuit will protect the 18650 battery from deep discharge (one bank). Therefore, we can immediately decide on the operating voltages: from 3.0 to 4.3 Volts. Means, maximum permissible drain-source voltage U ds should be more than 4.3 Volts.
However, the last statement is true only if only one lithium battery bank is used (or several connected in parallel). If, to power your load, a battery of several batteries connected in series is used, then the maximum drain-source voltage of the transistor must exceed the total voltage of the entire battery.
Here is a picture explaining this point: 
As can be seen from the diagram, for a battery of 3 18650 batteries connected in series, in the protection circuits of each bank it is necessary to use field devices with a drain-to-source voltage U ds > 12.6V (in practice, you need to take it with some margin, for example, 10%).
At the same time, this means that the field-effect transistor must be able to open completely (or at least strongly enough) already at a gate-source voltage U gs of less than 3 Volts. In fact, it is better to focus on a lower voltage, for example, 2.5 Volts, so that there is a margin.
For a rough (initial) estimate, you can look in the datasheet at the “Cut-off voltage” indicator ( Gate Threshold Voltage) is the voltage at which the transistor is on the threshold of opening. This voltage is typically measured when the drain current reaches 250 µA.
It is clear that the transistor cannot be operated in this mode, because its output impedance is still too high, and it will simply burn out due to excess power. That's why The transistor cut-off voltage must be less than the operating voltage of the protection circuit. And the smaller it is, the better.
In practice, to protect one can of a lithium-ion battery, you should select a field-effect transistor with a cutoff voltage of no more than 1.5 - 2 Volts.
Thus, the main requirements for field-effect transistors are as follows:
Now let's look at specific examples. For example, we have at our disposal the transistors IRF4905, IRL2505 and IRLMS2002. Let's take a closer look at them.
We open the datasheet and see that this is a transistor with a p-type channel (p-channel). If we are satisfied with this, we look further.
The maximum drain current is 74A. In excess, of course, but it fits.
Drain-source voltage - 55V. According to the conditions of the problem, we have only one bank of lithium, so the voltage is even greater than required.
Next, we are interested in the question of what the drain-source resistance will be when the opening voltage at the gate is 2.5V. We look at the datasheet and don’t immediately see this information. But we see that the cutoff voltage U gs(th) lies in the range of 2...4 Volts. We are categorically not happy with this.
The last requirement is not met, so discard the transistor.
Here is his datasheet. We look and immediately see that this is a very powerful N-channel field device. Drain current - 104A, drain-source voltage - 55V. So far everything is fine.
Check the voltage V gs(th) - maximum 2.0 V. Excellent!
But let's see what resistance the transistor will have at a gate-source voltage = 2.5 volts. Let's look at the chart: 
It turns out that with a gate voltage of 2.5V and a current through the transistor of 3A, a voltage of 3V will drop across it. In accordance with Ohm's law, its resistance at this moment will be 3V/3A=1Ohm.
Thus, when the voltage on the battery bank is about 3 Volts, it simply cannot supply 3A to the load, since for this the total load resistance, together with the drain-source resistance of the transistor, must be 1 Ohm. And we only have one transistor that already has a resistance of 1 ohm.
In addition, with such an internal resistance and a given current, the transistor will release power (3 A) 2 * 3 Ohm = 9 W. Therefore, you will need to install a radiator (a TO-220 case without a radiator can dissipate somewhere around 0.5...1 W).
An additional alarm bell should be the fact that the minimum gate voltage for which the manufacturer specified the output resistance of the transistor is 4V.
This seems to hint that the operation of the field worker at a voltage U gs less than 4 V was not envisaged.
Considering all of the above, discard the transistor.
So, let's take our third candidate out of the box. And immediately look at its performance characteristics.
N-type channel, let's say everything is in order.
Maximum drain current - 6.5 A. Suitable.
The maximum permissible drain-source voltage V dss = 20V. Great.
Cut-off voltage - max. 1.2 Volts. Still alright.
To find out the output resistance of this transistor, we don’t even have to look at the graphs (as we did in the previous case) - the required resistance is immediately given in the table just for our gate voltage.
In this article we will talk about the Li-Ion charge controller on the MCP73833.
Picture 1.
Up to this point I have been using LT4054 controllers, and to be honest, I was pleased with them:
It allowed charging compact Li-Pol batteries with a capacity of up to 3000 mAh
Was ultra-compact: sot23-5
Had a battery charging indicator
It has a bunch of protections, which makes it a practically indestructible chip
Figure 2.
An additional advantage is that before I started doing anything with it, I bought 50 of them, at a very modest price.
I identified shortcomings in the work, and, frankly speaking, they put me in a partial stupor:
The maximum declared current is 1A, I thought. But already at 300 mA during charging, the chip warms up to 110 * C, even in the presence of large radiator polygons and a radiator attached to the plastic surface of the chip.
When the thermal protection is turned on, a comparator apparently triggers, which quickly resets the current. As a result, the microcircuit turns into a generator, which kills the battery. This way I killed 2 batteries until I figured out what was wrong with the oscilloscope.
In view of the above, I got a problem with the device charging time of about 10 hours. Of course, this greatly dissatisfied me and the consumers of my electronics, but what can I do: everyone wanted to increase the service life with the same parameters of the device, and sometimes they consume a lot.
In this regard, I started looking for a controller that would have much better parameters and heat dissipation capabilities, and my choice so far has settled on the MCP73833, mainly due to the fact that my friend had these controllers in stock, and I whistled a couple of pieces quickly( faster than him) soldered the prototype and carried out the tests I needed.
Let me not engage in a complete and thorough translation of the datasheet (although this is useful), but quickly and simply tell you what I looked at first of all in this controller and whether I liked it or not.
1. General scheme inclusions are what catches your eye from the start. It is easy to notice that, with the exception of the indication (which you don’t have to do), the harness consists of only 4 parts. They include two filter capacitors, a resistor for programming the battery charge current, and a 10k thermistor to control overheating of the Li-Ion battery. This scheme shown in Figure 3. This is definitely cool.
Figure 3. Connection diagram MCP73833
2. She is much better with heat. This can be seen even from the connection diagram, since identical legs are visible that can be used for heat removal. In addition to this, looking at the fact that the chip is available in msop-10 and DFN-10 packages, which are larger in surface area than sot23-5. Moreover, in the DFN-10 case there is a special polygon, which can and should be used as a heat sink to a large surface. If you don’t believe me, then look at Figure 4 for yourself. It shows the pinouts of the legs of the DFN-10 case and the routing recommended by the manufacturer printed circuit board, with heat removal using a polygon.
Figure 4.
3. The presence of a 10k thermistor. Of course, in most cases I will not use it, since I am sure that I will not overheat the battery, but: there are tasks in which I mean a full charge of the battery in just 30 minutes of operation from the power supply. In such cases, the battery itself may overheat.
4. A rather complex battery charging indication system. As I understood and tried: there is 1 LED responsible for whether power is supplied from the charging power supply. In theory, the thing is not so necessary, but: I had cases when I broke the connector and the controller simply did not receive 5V at the input. In such cases, it was immediately clear what was wrong. An extremely useful feature for developers. For consumers, it is easily replaced by simply an LED along the 5V input line, installed with a current-limiting resistor.
5. The remaining two LEDs are broken during the charging stage. This allows you to unload the MK (if you do not need, for example, to show the battery charge on the display) in terms of processing the charge on the battery during charging (indication whether it is charged or not).
6. Programming the charge current over a wide range. Personally, I tried to increase the charging current to 1A on the board shown in Figure 1, and at around 890mA the board went into thermal protection in stable mode. As people around say, with large ranges they pulled out 2A perfectly from this controller, and technical description The maximum charge current is 3A, but I have a number of doubts related to the thermal load on the chip.
7. If you believe the datasheet, then this microcircuit has: Low-Dropout Linear Regulator Mode - a mode of reduced input voltage. In these modes, using a DC-DC converter, you can carefully reduce the voltage at the input of the microcircuit during the start of charging to reduce its heat generation. Personally, I tried to reduce the voltage, and the heat logically became less, but this microcircuit must drop at least 0.3-0.4V so that it can comfortably charge the battery. Purely technically, I’m going to make a small module that does this automatically, but I don’t have the money or time for this, so I happily ask everyone who is interested to email me. If there are a few more people, we will release such a thing on our website.
8. I didn’t like that the body was very small. Soldering it without a hair dryer (DFN-10) is difficult, and it won’t work out well, no matter how you look at it. It's better with msop-10, but it takes a lot of time for beginners to learn how to solder it.
9. I didn’t like that this controller does not have a built-in BMS (battery protection from rapid charge/discharge and a number of other problems). But more expensive controllers from TI have such things.
10. I liked the price. These controllers are not expensive.
And then I’m going to implement this chip into my various device ideas. For example, it is now already produced at the factory trial version development board based on STM32F103RCT6 and 18650 batteries. I already have a development board for this controller, which has proven itself very well, and I want to complement it with a portable version so that I can take my work project with me and not think about power and searching for a socket into which to insert the power supply.
I will also use it in all solutions that require charging currents of more than 300mA.
I hope you can use this useful and simple microcircuit in your devices.
If you are at all interested in battery power, here is my personal video about battery power for devices.
