TP5100 LiPo charging module failing

Question

Hi, I purchased 10 of the TP5100 modules trough Banggood and already six of them failed miserably. (With all of them the TP5100 IC failed)

The problem seems to be with some external components and I studied the circuit and the badly translated document so I redid the document as well as I could.

The main problem seems to be with the Schottky, the SS34’s reverse leakage current maybe just to high for the IC (or maybe sub-standard).

Im planning to replace it with MBRS540T3G with an reverse leakage of 300uA (any suggestions of a Schottky with a DO-214AB package size) as a replacement for the four that survived. I'm also going to replace the remaining three 10uF caps for just in case and also the 100nF cap. I’m going to use a low ESR 10uF cap from TDK with a 10% tolerance (C3216X7R1E106K160AB). All the components footprint is the same as the current. I’m also thinking about replacing the two 100 mOhm resistor with a 1% tolerance 50 mOhm 1W resistor (MCS1632R050FER) (Rs) that is specifically for current sense purposes just to be on the safe side (The modules takes way to long to reach me to risk the remaining four and with sensitive devices like the LiPo batteries make it high risk)

The NTC that is on the diagram is not avaiable on the module and seems to be tied to ground with RTRIC for a 10% trickle charge current. The on board LED also have some pads for an external two color LED (common anode) next to the input. The PWR_ON is left floating as in the diagram. To enable 8.4V (two cell charging) the SET pads may be linked to establish a link between CS and the reference voltage at VREG.

I have improved the translated document on the TP5100 and made some notes in PDF format If you are interested.

Could you please give comments about the suggested changes. I just hate to loose the last four modules as this seems to be and good charging module

shabaz · Answer

Hi Eugene, This seems to be a LiPo battery charger module.. if six have already failed, I'd personally stay away from the remainder. If you can't have confidence in the design, it's not a good idea to push 2A into a LiPo. The changes you are suggesting, are they recommended by the manufacturer of the IC? I ask, because I could only find a Chinese datasheet (which I can't read).

genebren · Answer

Hi Eugene, I am not sure that changing out the capacitors will be much help. They are there to help eliminate noise on the battery voltage sense and the charger output and would not likely cause the failures that you are seeing. The sense resistor ( parallel 100mOhm ) again are very likely to be fine. They are used to measure/control the charging current and would not likely cause the failures that you are seeing. The NTC shown on the diagram is actually suppose to be a part of the battery pack, usually connected via a third wire from the battery. The charger chip uses the NTC to help determine the charging current based on voltage, current and temperature. The real questions is how does it absence effect the charging algorithm. I could not find anything in the spec sheet that I found (loose translation into English). As shabaz points out, 2A charging is highly aggressive and not something that all batteries can handle. It is possible that the battery/charger mis-match might be another factor in the high rate of failure. Please use caution when charging Li-Ion batteries as very bad things can happen if you are not managing the charge correctly. Gene

zagene · Answer

Something else I failed to mention,

The failure to the IC occurred even before I added the 18350 LiPo Cell

It happened directly after I applied a 13.8V supply to the module.

The supply is a Regulated step down switch mode power supply, 220VAC input and a (adjusted) 13.8VDC output at 600W output rated.

The modules was connected 4 in parallel input with the output to 4X TP5100 modules individually connected to each of a 18350 battery holder

The modules is placed next to each other mounted with double sided heatsink tape (fully insulated) on a Aluminum plate with the battery holder.

It seems that the problem (from other Youtube videos I watched and reviews I've read just as soon as you place it in close proximity of another circuit.

So my other conclusions from what I've read and out of own observations is that the problem maybe with the input of VIN and the pre-charge capacitor placed parallel over VIN and Ground or maybe the switching circuit that may cause some unwanted induction to the IC or maybe bad capacitors at the input or output.

Unfortunately I ordered the wrong package Schottky diode for replacement (SMC in stead of SMA size package) so fitting it as a test may not be possible.

My other question, would a TVS at the input like a 1N6377 or at the single cell output like 1N6373 not work to clamp down over voltage.

I've got 4 left, I want to use it but safely and without seeing the main electronic component manufacture compound (smoke) escaping (its a bit difficult to put it back in)

Further information, the 18350 cells I want to use C rate is 3A/h with a maximum quick charge rate at 1.3C and a standard charge rate at 0.5C to 0.66C so it seems that the charge current of 2A may be in the limits of normal rate.

Thanks for the help everyone, I am however not able to discard something or just because the module may have other flaws that could be corrected. I do not think the original engineer planned for it to fail in this way, but more that the maker of the module take some shortcuts with "cheap" external components.

Regards

zagene · Answer

I agree and it make sense. I will definitely take it in consideration. It just seemed as a.good DC alternative than to charge sells with an RC charger, one cell at a time. TY very much for the reply

zagene · Answer

Update on the TP5100 It seems that the biggest problem here may be the input caps. On the original diagram indicate two capacitors on the input line, a pre-charge cap (from the circuit explanation) and a supposedly decoupling capacitor both between the V+ line and ground. The interesting fact seems that the pre-charge cap on the diagram is an electrolytic of some sorts (+ on the one side may indicate an electrolytic) and not a non-polarized cap (most probably a multi layer ceramic). The explanation of the circuit (as far as I could translate) the pre-charge cap could be anything between 1uF and 10uF. In the module however there is only one capacitor, a non-polarized 10uF cap. This omission may be to save costs and circuit area as I found that a 10uF SMD tant is quite big, especially the 35V (which I presume should be the minimum work voltage cap) and it is quite expensive. I made this alteration by adding a 10uF 63V radial electrolytic (the only one I had available for the experiment) by soldering it at the bottom of the circuit between the V+ and ground. The 10uF non-polarized cap between GND and Vin at the top I replaced with a 100nF 1206 multi-layer ceramic cap. (I will use a 10uF 50V SMD Panasonic 5 x 5.8 SMD electrolytic soldered to the bottom of the circuit between Vin and ground, for the remaining two modules I have left, part no EEEFK1H100UR) Furthermore, according to the circuit diagram there should be a 10uF and a 100nF non-polarized cap between both the BAT and ground and VS and ground (reason unknown) but this is not available on the module where it only seems to be a 10uF non-polarized cap and not the 100nF non-polarized cap. To be on the safe side I also added the two 100nF caps, one 0805 100nF where the one of the 100mOhm resistor is between the Vs and the ground of the 10uF cap, and one 1206 100nF between the B+ and the ground at the bottom of the PCB. At this stage it seems to work very well as the module did not fail (so far). I agree that the 2A charge rate may be a bit to high as the 18350 get a little bit of a temperature rise (still not alarming hot but just a little bit warmer than the ambient) but to save the battery life I will replace the current sense resistor to either a 1W 68mOhm (Panasonic ERJ8BWFR068V) or a 1W 100mOhm (Panasonic ERJ8BWFR100V ERJ8BWFR100V ) 1206 1% resistor specifically designed for current sense applications. I did not replaced the Schottky diode (SS34) at this stage as the one that I ordered (SMC package) is quite too large. I will be ordering a DO-214AA (SMB) package with a lower forward voltage drop than the existing one ( item number B340B-13-F) just to be on the safe side. At this stage adding a NTC may be just possible by redesigning the whole module as it seems that pin 11 and 12 is tied to the ground at the bottom of the IC. But so far so good, I hope it help a bit to sort out this modules inadequacies. Regards

zagene · Answer

Hi, Thank you for the answer I am attaching the PDF document for your review. It seems that (from other posts) the module fail as soon as it is near another circuit. Cannot confirm this, maybe lower grade caps, I just do not know. The device (IC) seems to be very solid from the document but as you have mentioned, lots of things can go wrong. I will firstly test it with older 18350 cells I have

shabaz · Answer

Hi Eugene, I can't see any attachment, but I did find a translation here: http://www.microlab.info/wp-content/uploads/sites/5/2018/06/TP5100-EN-autotranslated.pdf It didn't help much. Anyway, the design is a minimal implementation, it may work but is not rugged. Results could be disastrous if a failure causes uncontrolled power to be dumped into the LiPo. The minimal implementation may work on a bench, but not in real life. How are you powering it? How long are the wires to the power supply? What is the power source? The design uses ceramic capacitors on the input, which can (depending on the type) cause issues if you've got long leads. You'd need an oscilloscope to check, in case high voltage spikes are blowing out the chip. Without measuring, it's guesswork though. The design should have used an electrolytic or a very overrated tantalum, not a ceramic at the input. Anyway, this circuit/module is just asking for trouble, it's high-risk (personal opinion) to trust banggood for a 2A LiPo charger, when it could cause ten thousand times its cost (or worse) in damage to your home.

zagene · Answer

Hi Everyone Had quite a battle to create a shared document regarding the TP5100 but here is the shared location https://drive.google.com/open?id=1Bp_qHS-PE0gSYjCOdWoBhxl5j-tA5JqQ Regards Eugene

TP5100 LiPo charging module failing

Top Replies