Today in Russia there is an increase in manufacturers of autonomous electric vehicles of low and medium power. These include not only electric cars and urban transport. Electric traction is successfully used for the sale of loaders, warehouse and agricultural equipment, in the fishing and hunting fields for silent hunting and fishing (buggies, boats, ATVs), as well as in the sports and entertainment fields.
Manufacturers of most of these vehicles use medium power electric drive and lithium batteries as power sources. To ensure correct and safe operation of such a system, it is necessary to monitor the charge of each battery cell. Most manufacturers use ready-made control systems for this ( BMS) foreign production (PRC, USA, Germany).
The most efficient lithium power supplies, widely used in electric vehicles, by their nature produce an operating voltage of the order of 3.2...4 V. To ensure the operation of the electric drive at a higher voltage, they are connected in series. With this configuration in the battery, if the parameters of one or several cells change, an imbalance may occur - overcharging, overdischarge of cells, reaching 30% in the worst case. This mode significantly (by several times) reduces the battery life.
System BMS allows you to control and balance the charge of series and parallel-series connected battery cells of an autonomous electric vehicle.
There are 2 main types of balancing of battery cells: active and passive.
When the pore voltage is reached, the passive balancing system begins to dissipate energy on the resistor in the form of heat, and the charging process stops; then, having reached the lower threshold voltage, the system again begins charging the entire battery. The charging process stops when the voltage of all cells is within the required range.
Passive balancing is a unidirectional system; it can only absorb the charge of the cell. The active balancing system uses bidirectional DC-DC converters, thereby allowing energy from a more charged cell to be directed to a more discharged cell under microcontroller control BMS. The matrix switch provides the routing of charges into or out of the cell. The switch is connected to DC-DC to the converter that regulates the current, it can be positive when the cell needs to be charged, negative when it needs to be discharged. Instead of using a resistor and dissipating heat, the amount of current flowing during charging and discharging is controlled by a load balancing algorithm.
Analogue passive balancing systems are the most widely used. The figure shows a typical system and its characteristics.
We have developed a mathematical model of a battery consisting of 16 LiFePO 4 cells, the charge control of which was carried out through passive BMS. Mathematical model of battery LiFePO 4 cells in the system Matlab— Simulink takes into account the non-linear charging and discharging characteristics of the battery corresponding to a given cell type, internal resistance, as well as the current level of maximum capacity that changes during the life cycle of the cell.
A passive balancer was connected in parallel to each of the cells. To control the charging and balancing process, a key was connected in series, the opening and closing of which was carried out according to a command coming from BMS. The study was carried out for the final stage of charging the battery from an ideal voltage source.
Oscillograms of the charging process of a battery consisting of 16 LiFePO4 cells, one of which was “damaged” and had a lower capacity
The figure shows a case where the parameters of one of the cells were changed, in particular, the case of loss of capacity and increase in internal resistance was simulated, which can happen in real life, for example, as a result of an impact or due to overheating.
The damaged cell charges faster and is the first to reach the required voltage. However, no further charging occurs. According to the principle described above, the balancer begins to work. The remaining cells, indicated in green, retain their current capacity level when the charging process stops, and continue to charge when it resumes.
When the voltage level of all cells reaches the required range, the charging process stops
Greetings to everyone who looked at the light. The review will focus, as you probably already guessed, on two simple boards designed to monitor assemblies of Li-Ion batteries, called BMS. The review will include testing, as well as several options for converting a screwdriver for lithium based on these boards or similar ones. For anyone interested, you are welcome under cat.
General form:
Brief performance characteristics of the boards:
Note:
I would like to warn you right away - with a balancer there is only a blue board, a red one without a balancer, i.e. This is purely a protection board against overcharge/overdischarge/short circuit/high load current. And also, contrary to some beliefs, none of them has a charge controller (CC/CV), so for their operation a special board with a fixed voltage and current limitation is required a.
Board dimensions:
The dimensions of the boards are very small, only 56mm*21mm for blue and 50mm*22mm for red: 
Here is a comparison with AA and 18650 batteries: 
Appearance:
Let's start with blue protection board :
Upon closer examination, you can see the protection controller – S8254AA and balancing components for the 3S assembly: 
Unfortunately, the operating current according to the seller is only 8A, but judging by the datasheets, one AO4407A mosfet is rated at 12A (peak 60A), and we have two of them: 
I will also note that the balancing current is very small (about 40mA) and the balancing is activated as soon as all the cells/banks switch to CV mode (the second phase of charging).
Connection: 
simpler, because it does not have a balancer: 
It is also made on the basis of a protection controller – S8254AA, but is designed for a higher operating current of 15A (again, according to the manufacturer): 
Based on the datasheets for the power mosfets used, the operating current is stated to be 70A, and the peak current is 200A, even one mosfet is enough, and we have two of them: 
The connection is similar: 
So, as we see, on both boards there is a protection controller with the necessary isolation, power mosfets and shunts for controlling the flowing current, but the blue one also has a built-in lancir. I didn’t really delve into the circuit, but it seems that the power mosfets are paralleled, so the operating currents can be multiplied by two. These scarves do not know about the charging algorithm (CC/CV). To confirm that these are precisely protection boards, we can judge by the datasheet for the S8254AA controller, in which there is not a word about the charging module: 
The controller itself is designed for a 4S connection, so with some modification (judging by the datasheet) - soldering the connector and resistor, perhaps the red card will work a: 
It’s not so easy to upgrade the blue scarf to 4S; you’ll have to add additional components to the balancer.
Board testing:
So, let's move on to the most important thing, namely, how suitable they are for real use. The following devices will help us for testing:
- a prefabricated module (three triple/quad-register voltmeters and a holder for triple 18650 batteries), which flashed in my review of the charger, however, without balancing go tail: 
- two-register ampere-voltmeter for current control (lower readings of the device): 
- step-down DC/DC converter with current limiting and lithium charging capability: 
- charging and balancing device iCharger 208B for discharging the entire assembly
The stand is simple - the converter board supplies a fixed constant voltage of 12.6V and limits the charging current. We use voltmeters to see at what voltage the boards operate and how the banks are balanced.
First, let's look at the main feature of the blue board, namely balancing. In the photo there are 3 banks charged at 4.15V/4.18V/4.08V. As we can see, there is an imbalance. We apply voltage, the charging current gradually drops (lower device): 
Since the board does not have any indicators, the completion of balancing can only be assessed by eye. The ammeter was already showing zero more than an hour before the end. For those who are interested, here is a short video about how the balancer works in this board:
As a result, the banks are balanced at the level of 4.210V/4.212V/4.206V, which is quite good: 
When a voltage is applied a little more than 12.6V, as I understand it, the balancer is inactive and as soon as the voltage on one of the cans reaches 4.25V, the S8254AA protection controller turns off the charge : 
The same situation applies to the red board; the S8254AA protection controller also turns off the charge at the level of 4.25V: 
Now let's go through the load cut-off. I will discharge, as I mentioned above, with an iCharger 208B charging and balancing device in 3S mode with a current of 0.5A (for more accurate measurements). Since I don’t really want to wait for the entire battery to discharge, so I took one discharged battery (green Samson INR18650-25R in the photo).
The blue board turns off the load as soon as the voltage on one of the banks reaches 2.7V. In the photo (without load -> before turning off -> end): 
As you can see, exactly at 2.7V the board turns off the load (the seller stated 2.8V). It seems to me that this is a little high, especially if you take into account the fact that in the same screwdrivers the loads are huge, and therefore the voltage drop is large. It is still desirable to have a cutoff of 2.4-2.5V in such devices.
The red board, on the contrary, turns off the load as soon as the voltage on one of the banks reaches 2.5V. In the photo (without load -> before turning off -> end): 
Here everything is great, but there is no balancer.
Conclusion: My personal opinion is that a regular protection board without a balancer (red) is perfect for a power tool. It has high operating currents, an optimal cut-off voltage of 2.5V, and can be easily upgraded to a 4S configuration (14.4V/16.8V). I think this is the most optimal choice for converting a budget Shurik for lithium.
Now on to the blue scarf. One of the advantages is the presence of balancing, but the operating currents are still small, 12A (24A) this is a bit too little for a Shurik with a torque of 15-25 Nm, especially when the cartridge is already almost axes when tightening the screw. Yes, and the cutoff voltage is only 2.7V, which means that under heavy load, part of the battery capacity will remain unclaimed, since at high currents the voltage drop on the banks is significant Yes, and they are designed for 2.5V. It is better to use a blue scarf in some homemade projects, but again, this is my personal opinion.
Possible application schemes or how to convert Shurik’s power supply to lithium:
So, how can you change the power supply of your favorite Shura from NiCd to Li-Ion/Li-Pol? This topic is already quite hackneyed and solutions, in principle, have been found, but I will briefly repeat myself.
To begin with, I will say only one thing - in budget shuriks there is only a protection board against overcharge/overdischarge/short circuit/high load current (analogous to the reviewed red board). There is no balancing there. Moreover, even branded power tools do not have balancing. The same applies to all tools where there are proud inscriptions “Charge in 30 minutes.” Yes, they charge in half an hour, but the shutdown occurs as soon as the voltage on one of the banks reaches the nominal value or the protection board works. It is not difficult to guess that the banks will not be fully charged, but the difference is only 5-10%, so it is not so important. The main thing to remember is that the charge with balancing lasts for at least several hours. Therefore, the question arises, do you need it?
So, the most common option looks like this:
Network charger with stabilized output 12.6V and current limitation (1-2A) -> protection board ->
In summary: cheap, fast, acceptable, reliable. Balancing varies depending on the state of the cans (capacity and internal resistance). This is a completely working option, but after some time the imbalance will make itself known by the operating time.
A more correct option:
Network charger with stabilized output 12.6V, current limitation (1-2A) -> protection board with balancing -> 3 series-connected batteries
In summary: expensive, fast/slow, high quality, reliable. Balancing is normal, battery capacity is maximum
So, we will try to do something similar to the second option, here’s how you can do it:
1) Li-Ion/Li-Pol batteries, protection boards and a specialized charging and balancing device (iCharger, iMax). Additionally, you will have to remove the balancing connector. There are only two disadvantages - model chargers are not cheap, and they are not very convenient to maintain. Pros – high charging current, high can balancing current
2) Li-Ion/Li-Pol batteries, protection board with balancing, DC converter with current limiting, power supply
3) Li-Ion/Li-Pol batteries, protection board without balancing (red), DC converter with current limiting, power supply. The only downside is that over time the cans will become unbalanced. To minimize imbalance, before remaking the shurik, it is necessary to adjust the voltage to the same level and it is advisable to take cans from the same batch
The first option will only work for those who have a model memory, but it seems to me that if they needed it, then they would have remade their Shurik a long time ago. The second and third options are practically the same and have the right to life. You just need to choose what is more important – speed or capacity. I believe that the most optimal option is the latter, but only once every few months do you need to balance the banks.
So, enough chatter, let's move on to the rework. Since I do not have a shurik on NiCd batteries, therefore, about the alteration only in words. We will need:
1) Power supply:
First option. Power supply (PSU), at least 14V or more. The output current is desirable to be at least 1A (ideally about 2-3A). Нaм пoдoйдeт блoк питaния oт нoутбукoв/нeтбукoв, oт зaрядныx уcтрoйcтв (выxoд бoлee 14V), блoки для питaния cвeтoдиoдныx лeнт, видeoзaпиcывaющeй aппaрaтуры (DIY БП), нaпримeр или : 
- Step-down DC/DC converter with current limiting and lithium charging capability, for example or: 
- Second option. Ready-made power supplies for Shuriks with current limiting and 12.6V output. They are not cheap, as an example from my review of the MNT screwdriver -: 
- Third option. : 
2) Protection board with or without balancer. It is advisable to take the current with reserve: 
If you use the option without a balancer, then you need to solder the balancer connector. This is necessary to control the voltage on the banks, i.e. to assess imbalance. And as you understand, you will need to periodically charge the battery little by little with a simple TP4056 charging module if imbalance begins. That is Once every few months, we take the TP4056 card and charge one by one all the banks that, at the end of the charge, have a voltage below 4.18V. This module correctly cuts off the charge at a fixed voltage of 4.2V. This procedure will take an hour and a half, but the banks will be more or less balanced.
It’s written a little chaotically, but for those in the tank:
After a couple of months, we charge the screwdriver battery. At the end of the charge, we take out the balancing tail and measure the voltage on the banks. If you get something like this - 4.20V/4.18V/4.19V, then balancing is basically not needed. But if the picture is as follows - 4.20V/4.06V/4.14V, then we take the TP4056 module and charge two banks in turn to 4.2V. I don’t see any other option other than specialized charger-balancers.
3) High-current batteries: 
I have previously written a couple of small reviews about some of them - and. Here are the main models of high-current 18650 Li-Ion batteries:
- Sanyo UR18650W2 1500mah (20A max.)
- Sanyo UR18650RX 2000mah (20A max.)
- Sanyo UR18650NSX 2500mah (20A max.)
- Samsung INR18650-15L 1500mah (18A max.)
- Samsung INR18650-20R 2000mah (22A max.)
- Samsung INR18650-25R 2500mah (20A max.)
- Samsung INR18650-30Q 3000mah (15A max.)
- LG INR18650HB6 1500mah (30A max.)
- LG INR18650HD2 2000mah (25A max.)
- LG INR18650HD2C 2100mah (20A max.)
- LG INR18650HE2 2500mah (20A max.)
- LG INR18650HE4 2500mah (20A max.)
- LG INR18650HG2 3000mah (20A max.)
- SONY US18650VTC3 1600mah (30A max.)
- SONY US18650VTC4 2100mah (30A max.)
- SONY US18650VTC5 2600mah (30A max.)
I recommend the time-tested cheap Samsung INR18650-25R 2500mah (20A max), Samsung INR18650-30Q 3000mah (15A max) or LG INR18650HG2 3000mah (20A max) acc.). I haven’t particularly come across other cans, but my personal choice is Samsung INR18650-30Q 3000mah. Skis had a slight technological defect and fakes with low current output began to appear. I can post an article on how to distinguish a fake from the original, but a little later, you need to look for it.
How to combine all this economy:
Well, a few words about connection. We use high-quality copper stranded wires of decent cross-section. These are high-quality acoustic or conventional ball-and-socket screws/PVS with a cross-section of 0.5 or 0.75 mm2 from household goods (we rip open the insulation and get high-quality wires of different colors). The length of the connecting conductors should be minimal. Batteries are preferably from the same batch. Before connecting them, it is advisable to charge them to the same voltage so that there is no imbalance for as long as possible. Soldering batteries is not difficult. The main thing is to have a powerful soldering iron (60-80W) and active flux (soldering acid, for example). It's soldering with a bang. The main thing is to wipe the soldering area with alcohol or acetone. The batteries themselves are placed in the battery compartment from old NiCd cans. It is better to have a triangle, minus to plus, or as popularly called “jack”, by analogy with this (one battery will be located in the opposite direction): 
Thus, the wires connecting the batteries will be short, therefore, the drop in precious voltage in them under load will be minimal. I do not recommend using holders for 3-4 batteries, they are not intended for such currents. Side-by-side and balancing conductors are not so important and can be of smaller cross-section. Ideally, it is better to put the batteries and the protection board in the battery compartment, and the DC step-down converter separately in the docking station. The charge/charge LED indicators can be replaced with your own and displayed on the docking station body. If you wish, you can add a minivoltmeter to the battery module, but this is extra money, because the total voltage on the battery will only indirectly indicate the residual capacity. But if there is a desire, why not. Here: 
Now let's estimate the prices:
1) BP – from 5 to 7 dollars
2) DC/DC converter – from 2 to 4 dollars
3) Protection cards - from 5 to 6 dollars
4) Batteries – from 9 to 12 dollars ($3-4 per piece)
Total, on average, $15-20 per alteration (with discounts/coupons), or $25 without them.
Advantages:
I have already mentioned the advantages of lithium power supplies (Li-Ion/Li-Pol) over nickel ones (NiCd). In our case, a head-to-head comparison – a typical Shurik battery made of NiCd batteries versus lithium:
+ high energy density. A typical 12S 14.4V 1300mah nickel battery has a stored energy of 14.4*1.3=18.72Wh, and a 4S 18650 14.4V 3000mah lithium battery has a stored energy of 10.8*3=43.2Wh
+ lack of memory effect, i.e. you can charge them at any time without waiting for full discharge
+ smaller dimensions and weight with the same parameters as NiCd
+ fast charging time (not afraid of high charge currents) and clear indication
+ low self-discharge
The only disadvantages of Li-Ion can be noted:
- low frost resistance of batteries (they are afraid of negative temperatures)
- balancing of the cans when charging and the presence of overdischarge protection is required
As we can see, the advantages of lithium are obvious, so it often makes sense to change the power supply...
Conclusion: The monitored scarves are not bad, they should be suitable for any task. If I had a Shurik on NiCd banks, I would choose a red scarf for the conversion, :-)…
The product was provided for writing a review by the store. The review was published in accordance with clause 18 of the Site Rules.
This board had been lying in storage for a long time until the chance to use it for its intended purpose came along. If you like diagrams and tools, it will be interesting.
If anyone remembers, I have a converted screwdriver
For more than 2 years it worked actively and regularly, discharging and charging it 40 times.
Until he severely overloaded it himself, making a ventilation hole in the OSB with a 102 mm crown, barely holding the tool with both hands :) 
The corded screwdriver also couldn’t cope with this job, and there wasn’t a powerful drill at hand. The result is that one of the batteries could not withstand the abuse and went into a break. At all:(
After partially disassembling the battery, it turned out that the aluminum strip contact to the roll had burnt out. I don't know how to repair batteries yet :( 
The tool was urgently needed, so the first thought was to buy the same 26650 LiMn2O4 battery and quickly restore the battery pack. But the same battery was not found in stores. Ordering from China and waiting is too long...
In addition, I decided to add a BMS protection board to the unit to prevent this from happening again. But the problem is that there is absolutely no free space in the battery pack :(
In short, I bought relatively inexpensive high-current SONY US18650VTC4 (2100mAh 30A peak 60A). They cost 750 rubles for 3 pieces - this is slightly more expensive than ordering from China, but here and now! Took
The capacity of 2100 mAh is of course significantly less than the former 3500 mAh, but I’ll survive it somehow, you still get tired faster than it discharges. During your next smoke break and snack, you can recharge it, especially since now I will charge it with a new charger with a high current :)
I checked the remaining two 26650 3500mAh batteries that were previously working for residual capacity - I got 3140mAh. A 10% drop in capacity is quite acceptable and the batteries can still be used somewhere.
If desired, this board can be easily converted into 2S by simply connecting B2 and B+ with a jumper, while the power switches can heat up more due to an increase in the resistance of the field switch channels.
Controllers provide protection
Without violating my principles, I copied the original circuit diagram. 
Although the scheme looks complicated, it works simply and clearly. Naturally, mistakes have not gone away - the Chinese keep their mark :)
The numbering of transistors is shown conventionally.
A level converter and a signal adder with HY2210 are assembled on p-n-n transistors Q1-Q6
A simple transistor logic for controlling power switches is assembled on n-p-n transistors Q7-Q9
Q7 unlocks when any battery is overdischarged to a voltage below 2.40V, recovery occurs when the voltage exceeds 3.0V (after removing the load or connecting to charging).
Q8 ensures that the protection latches after it is triggered until the load is completely removed. At the same time, it provides high-speed protection in the event of a load short circuit when the current jumps above 100A.
Q9 unlocks when any battery is recharged to a voltage above 4.28V; restoration occurs under load at a voltage below 4.08V. In this case, the power switches do not interfere with the flow of discharge current.
I did not check the exact thresholds of all controllers, because... this is labor-intensive, but in reality they do not differ much from those stated in the specification.
S1 and S2 are just control points and have nothing to do with thermal protection. Moreover, they cannot be connected to each other. I’ll tell you and show you below how to properly connect thermal protection.
A signal appears on S1 when any element is overdischarged.
A signal appears on S2 when any element is overcharged, as well as after the current protection is triggered.
The board's current consumption is very small (several microamps).
New batteries 
The batteries are signed and tested, the capacity corresponds to the nominal
Despite the presence of a resistance welding machine, I soldered the batteries, because... in this case this is the best solution.
Before soldering, it is necessary to tin the batteries well. 
Batteries are soldered and installed in place
The board is soldered (in the photo the board has already been redesigned)
Be careful not to short the ends of the batteries.
Power wires - in silicone insulation 1.5 sq. mm
Control wires - MGTF-0.2
The typical board connection diagram is not optimal, because There are as many as 4 power wires going to the board. I connected it using a simpler scheme, when only 2 power wires go to the board. This connection is allowed when the length of the connecting wires to the batteries is short
Under load, when you sharply press the trigger, the board protection immediately triggers:(
At first, I logically assumed that it was cut off due to current overload, but shorting the board shunt did not change anything. It became clear that it is not the current overload of the board that triggers the protection.
Next, I connected the oscilloscope in recording mode to the batteries and checked the voltage on them under load. The voltage managed to drop below 7V and the protection immediately worked :(
This is the reason the protection is triggered. Why did the voltage drop so much, since the batteries are high-current? Let's get down to measurements and calculations:
- battery voltage 11.4V (HP890CN)
- internal resistance of batteries from the datasheet at DC-IR 66 mOhm (3x22 mOhm)
- measured motor resistance 63mOhm
- resistance of connecting wires and screwdriver switch - 23 mOhm
- resistance of the protection board - shunt + MOSFET + connection wires - 10 mOhm
Total circuit resistance 66+63+23+10=162mOhm
Circuit current 11.4/0.162= 70A
A lot, however...
But the problem is not the current, but the voltage drop across the batteries.
At a current of 70A, the voltage of each battery decreases by 70*0.022=1.54V and becomes 3.8-1.54=2.26V. This is the real reason the protection is triggered!
It is not advisable to adjust or remove the protection - the safety of use is reduced, so it should simply be slowed down while the engine is starting. Add a 0.47uF capacitor to the right place and the delay is ready :)
If it is difficult for someone to solder small change onto the board, you can solder the capacitor with a surface-mounted connection between S1 and B-
It was easier for me to install an SMD capacitor :)
There is now enough time for the engine to spin up under load. When the engine is severely blocked at full throttle, the protection is activated after 0.3 seconds, and not instantly, as before.
Redesigned board 
Don’t pay attention to the 470kOhm resistor - the original 510kOhm resistor was damaged as a result of experiments and was replaced with whatever came to hand :)
The board contains high-resistance circuits, so after soldering it is necessary to thoroughly wash the board.
Scheme after rework 
Description of all improvements
1. An unnecessary 0.1 µF capacitor was soldered from pin 2 of the HY2210 to the shunt. It’s unclear why they installed it at all; it’s not in the datasheet for the HY2210. It doesn’t affect the work, but I soldered it out of harm’s way.
2. A base-emitter resistor has been added for normal recovery after the protection is triggered.
Without it, auto-recovery of protection after removing the load is extremely unstable, because The slightest interference on P- prevents the protection from being reset. Suitable resistor value is 1-3MOhm. I soldered this resistor carefully directly to the terminals of the transistor. Be careful not to overheat it!
3. A 0.47uF capacitor has been added to slow down the response of overdischarge protection from 25ms (typical for HY2210) to 300ms. I tried connecting a 0.1uF capacitor - the protection works too quickly for a hefty RS-775 motor. If the engine is absolutely brutal, you may need to install a more capacitive capacitor, for example 1 µF
Now sharply pressing the trigger under load does not trigger the protection :)
Connecting a protective thermal switch.
Both NO and NC thermal switch can be connected to this board.
I provide the diagrams below. 
I used NO thermal switch KSD 9700 5A 70ºC 
Glued it to the batteries 
At the same time, I decided to abandon charging from the power supply through current-limiting resistors and charge the batteries with a converted 3S 12.6V 3A charger 
The final scheme turned out like this 
Charging Colaier 12.6V 3A
I've already done UV on it. kirich, but as always I have something to add 
In its original form, the charger does not hold the declared current of 3A and overheats. In addition, it emits noticeable interference to a nearby radio receiver.
The charger was disassembled even before the tests :) 
Charging differs from simple power supplies by additionally installed current limiting circuit elements. 
I'll be brief with modifications :)
- Installed the missing input filter. Now the radio does not respond to charging.
- Moved the thermistor NTC1 (5D-9) and fuse LF1 (T2A) to the right places
- There is space on the board to install discharge resistors R1 + R2. They are needed to discharge the CX1 after disconnecting charging from the network. I installed a discharge resistor OMLT-0.5 620 kOhm in parallel with CX1 :)
I installed output choke L1 instead of jumpers. The operation was not affected in any way, because the output ripple for charging is not of great importance. 
Reduced the output voltage from 12.8V to 12.65V by connecting a 390kOhm resistor in parallel with resistor R29 8.2kOhm
- Reduced the output current from 3.2A to 2A by replacing the 1.6kOhm resistor R26 with a 1kOhm resistor 
The current was reduced because, firstly, this charger cannot deliver a current of 3A without overheating, and secondly, because US18650VTC4 batteries have a maximum charging current of 2A.
The PCB layout is not done correctly, resulting in poor stability of the output voltage and current. I didn’t change it because it’s not very critical.
Conclusions:
- SONY US18650VTC4 batteries have only one drawback - small capacity
- BMS 3S 25A board is able to work normally after a little modification
- Charging 3S 12.6V 3A in its original form does not work satisfactorily and requires significant improvement, I cannot recommend it, sorry
After the modification, the screwdriver has been working normally for 4 months. The decrease in power is not felt, it charges quickly, in just over an hour.
!
Now we, together with the author of the YouTube channel “Radio-Lab”, will assemble a battery for 4 banks from individual Li-ion 18650 batteries with a protection board, also known as BMS.
For the author's future projects, such a battery will be needed. On the Internet, he bought 8 of these Li-ion batteries from disassembly, like the Sanyo company.
The voltage at the input is 19.6V, and at the output of the converter it will gradually increase to the level of the charged battery and at the end the blue LED will go out, the red LED will light up and the BMS board will turn off the battery.
There hasn't been a review of converting a screwdriver to lithium for a long time :)
The review is mainly devoted to the BMS board, but there will be links to some other little things involved in converting my old screwdriver to 18650 lithium batteries.
In short, you can take this board; after a little finishing, it works quite well in a screwdriver.
PS: a lot of text, pictures without spoilers.
P.S. The review is almost an anniversary on the site - the 58000th, according to the address bar of the browser;)
All board components are placed on one side: 
The second side is empty and covered with a white mask: 
The part responsible for balancing during charging: 
This part is responsible for protecting cells from overcharge/overdischarge and it is also responsible for general protection against short circuit: 
Mosfets: 
It is assembled neatly, there are no obvious flux stains, the appearance is quite decent. The kit included a tail with a connector, which was immediately plugged into the board. The length of the wires in this connector is about 20-25 cm. Unfortunately, I didn’t take a picture of it right away.
What else did I order specifically for this alteration:
Batteries -
Nickel strips for soldering batteries: (yes, I know that you can solder with wires, but the strips will take up less space and will be more aesthetically pleasing :)) And initially I even wanted to assemble contact welding (not only for this alteration, of course), that’s why I ordered the strips, but laziness prevailed and I had to solder them.
Having chosen a free day (or rather, having blatantly sent all other matters away), I set about redoing it. To begin with, I disassembled the battery with dead Chinese batteries, threw out the batteries and carefully measured the space inside. Then I sat down to draw the battery holder and circuit board in a 3D editor. I also had to draw the board (without details) in order to try on everything assembled. It turned out something like this: 
According to the idea, the board is attached from above, one side into the grooves, the other side is clamped with an overlay, the board itself lies in the middle on a protruding plane so that when it is pressed it does not bend. The holder itself is made of such a size that it fits tightly inside the battery case and does not dangle there.
At first I thought about making spring contacts for batteries, but abandoned this idea. This is not the best option for high currents, so I left cutouts in the holder for nickel strips with which the batteries will be soldered. I also left vertical cutouts for the wires, which should extend from the inter-can connections beyond the lid.
I set it to be printed on a 3D printer from ABS and after a few hours everything was ready :) 
When screwing everything on, I decided not to trust screws and fused these M2.5 plug-in nuts into the body: 
Got it here -
Great item for this type of use! It is fused slowly with a soldering iron. To prevent the plastic from packing inside when melting into blind holes, I screwed a bolt of suitable length into this nut and heated its head with a soldering iron tip with a large drop of tin for better heat transfer. The holes in the plastic for these nuts are left slightly smaller (0.1-0.2 mm) than the diameter of the outer smooth (middle) part of the nut. They hold very tightly, you can screw in and unscrew the bolts as much as you like and don’t be too shy with the tightening force.
In order to have the possibility of cell-by-can control and, if necessary, charging with external balancing, a 5-pin connector will stick out in the back wall of the battery, for which I quickly threw on a scarf and made it on the machine: 
The holder has a platform for this scarf.
As I already wrote, I soldered the batteries with nickel strips. Alas, this method is not without its drawbacks, and one of the batteries was so outraged by this treatment that it left only 0.2 volts on its contacts. I had to desolder it and solder another one, fortunately I took them with a reserve. Otherwise there were no difficulties. Using acid, we tin the battery contacts and nickel strips cut to the required length, then thoroughly wipe everything tinned and around it with cotton wool and alcohol (but you can also use water), and solder it. The soldering iron must be powerful and either be able to react very quickly to the tip cooling, or simply have a massive tip that will not cool instantly upon contact with a massive piece of iron.
Very important: during soldering and during all subsequent operations with the soldered battery pack, you must be very careful not to short-circuit any battery contacts! In addition, as indicated in the comments ybxtuj, it is very advisable to solder them discharged, and I absolutely agree with him, this way the consequences will be easier if something does short out. A short circuit of such a battery, even a discharged one, can lead to big troubles.
I soldered wires to three intermediate connections between the batteries - they will go to the BMS board connector for monitoring the banks and to the external connector. Looking ahead, I want to say that I did a little extra work with these wires - they can not be led to the board connector, but soldered to the corresponding pins B1, B2 and B3. These pins on the board itself are connected to the connector pins. 
By the way, I used silicone insulated wires everywhere - they do not react to heat at all and are very flexible. I bought several sections on Ebay, but I don’t remember the exact link... I really like them, but there is a minus - silicone insulation is not very mechanically strong and is easily damaged by sharp objects.
I tried on the batteries and the board in the holder - everything is excellent: 
I tried on a handkerchief with a connector, used a Dremel to cut out a hole in the battery case for the connector... and missed the height and took the size from the wrong plane. The result was a decent gap like this: 
Now all that remains is to solder everything together.
I soldered the included tail onto my scarf, cutting it to the required length: 
I also soldered the wires from the inter-can connections there. Although, as I already wrote, it was possible to solder them to the corresponding contacts of the BMS board, there is also an inconvenience - in order to remove the batteries, you will need to unsolder not only the plus and minus from the BMS, but also three more wires, but now you can simply pull out the connector.
I had to tinker a little with the battery contacts: in the original version, the plastic part (holding the contacts) inside the battery leg is pressed by one battery standing directly under it, but now I had to think about how to fix this part, so as not to be tight. Here's the detail: 
In the end, I took a piece of silicone (left over from pouring some form), cut off a roughly suitable piece from it and inserted it into the leg, pressing that part. At the same time, the same piece of silicone presses the holder with the board, nothing will dangle.
Just in case, I laid Kapton insulating tape over the contacts, and grabbed the wires with a few snot drops of hot glue so that they would not get between the halves of the case when assembling it. 
And I decided to try to get to the root of the problem.
I dismissed the assumption that the overload protection was triggered during inrush currents, since even without the shunt nothing changed.
But still I looked with an oscilloscope at a homemade 0.077 ohm shunt between the batteries and the board - yes, PWM is visible, sharp consumption peaks with a frequency of approximately 4 kHz, 10-15 ms after the start of the peaks the board cuts off the load. But these peaks showed less than 15 amperes (based on the shunt resistance), so it’s definitely not a matter of current overload (as it turned out later, this is not entirely true). And the ceramic resistance of 1 Ohm did not cause a shutdown, but the current was also 15 amperes.
There was also the option of a short-term drawdown on the banks during startup, which triggered the overdischarge protection, and I went to see what was happening on the banks. Well, yes, horror is happening there - the peak drawdown is up to 2.3 volts on all banks, but it is very short - less than a millisecond, while the board promises to wait a hundred milliseconds before turning on the overdischarge protection. “The Chinese indicated Chinese milliseconds,” I thought and went to look at the voltage control circuit of the cans. It turned out that it contains RC filters that smooth out sudden changes (R=100 Ohm, C=3.3 uF). After these filters, already at the input of the microcircuits that control the banks, the drawdown was smaller - only up to 2.8 volts. By the way, here is the datasheet for the can control chips on this DW01B board -
According to the datasheet, the response time to overdischarge is also considerable - from 40 to 100 ms, which does not fit into the picture. But okay, there’s nothing more to assume, so I’ll change the resistance in the RC filters from 100 Ohms to 1 kOhm. This radically improved the picture at the input of the microcircuits; there were no more drawdowns of less than 3.2 volts. But it didn’t change the behavior of the screwdriver at all - a slightly sharper start - and then shut up.
“Let’s go with a simple logical move”©. Only these DW01B microcircuits, which control all discharge parameters, can cut off the load. And I looked at the control outputs of all four microcircuits with an oscilloscope. All four microcircuits do not make any attempts to disconnect the load when the screwdriver starts. And the control voltage disappears from the mosfets gates. Either mysticism or the Chinese have screwed up something in a simple circuit that should be between microcircuits and mosfets.
And I started reverse engineering this part of the board. With swearing and running from the microscope to the computer. 
Here's what we ended up with: 
In the green rectangle are the batteries themselves. In blue - the keys from the outputs of the protection chips, also nothing interesting, in a normal situation their outputs to R2, R10 are simply “hanging in the air”. The most interesting part is in the red square, which is where, as it turned out, the dog rummaged. I drew the mosfets one at a time for simplicity, the left one is responsible for discharging to the load, the right one is for the charge.
As far as I understand, the reason for the shutdown is in resistor R6. Through it, “iron” protection against current overload is organized due to the voltage drop on the mosfet itself. Moreover, this protection works like a trigger - as soon as the voltage at the base of VT1 begins to increase, it begins to reduce the voltage at the gate of VT4, from which it begins to reduce conductivity, the voltage drop across it increases, which leads to an even greater increase in the voltage at the base of VT1 and an avalanche-like a process leading to the complete opening of VT1 and, accordingly, the closing of VT4. Why does this happen when starting a screwdriver, when the current peaks do not even reach 15A, while a constant load of 15A works - I don’t know. Perhaps the capacitance of the circuit elements or the inductance of the load plays a role here.
To check, I first simulated this part of the circuit: 
And this is what I got from the results of her work: 
The X axis is time in milliseconds, the Y axis is voltage in volts.
On the bottom graph - the load is turned on (you don’t have to look at the numbers on Y, they are arbitrary, just up - the load is on, down - off). The load is a resistance of 1 ohm.
In the top graph, red is the load current, blue is the voltage at the mosfet gate. As you can see, the gate voltage (blue) decreases with each pulse of load current and eventually drops to zero, which means the load is turned off. And it is not restored even when the load stops trying to consume something (after 2 milliseconds). And although other mosfets with different parameters are used here, the picture is the same as in the BMS board - an attempt to start and shutdown in a matter of milliseconds.
Well, let’s take this as a working hypothesis and, armed with new knowledge, let’s try to chew on this piece of Chinese science :)
There are two options here:
1. Place a small capacitor in parallel with resistor R1, this is: 
The capacitor is 0.1 uF, according to the simulation it is possible even less, up to 1 nf.
The result of the simulation in this version: 
2. Remove resistor R6 altogether: 
The result of the simulation of this option: 
I tried both options - both work. In the second option, the screwdriver does not turn off under any circumstances - start, rotation is blocked - it turns (or tries with all its might). But somehow it’s not entirely peaceful to live with the protection turned off, although there is still protection against short circuits on the microcircuits.
With the first option, the screwdriver starts confidently with any pressure. I was able to achieve shutdown only when I started it at second speed (increased for drilling) with the chuck blocked. But even then it jerks quite strongly before turning off. At first speed I could not get it to turn off. I left this option for myself; I am completely satisfied with it.
There are even empty spaces for components on the board, and one of them seems to be specially designed for this capacitor. It was designed for the size of SMD 0603, so I soldered 0.1 uF here (circled it in red): 
PS: damn, it took me less time to remodel the screwdriver than it took me to write this review :)
ZZY: perhaps my comrades who are more experienced in power and analogue circuitry will correct me on something, I myself am a digital and analogue person through the roof :)
