Hi friends
I have reverse engineered this tester, schematic is attached and is complete except for capacitor values. These testers annoyingly have no model number but they are the ones you find on the first page of just about any website if you type in "18650 4-channel tester". There is also a 2-channel variant and a newer 8-channel variant. They are very cheap (about £15) and can measure voltage, current, internal resistance and capacity. They have a multicolour display for all these statistics. They also have terminal blocks so that if you want to use a breakout board instead of the cell holders (e.g. to test 21700 cells) you can do so. You can set the discharge voltage cut-off between 2.5V and 3.5V in increments of 0.1V.
Here is a backside pic of the board with the base removed:
Here's some info on how it works:
Power: Two USB-C ports at 5V. Only VBUS and GND are electrically connected so USB-PD chargers will not work. The device can run from either port and is protected from reverse current by a schottky on each VBUS, but if you only plug in one cable then charging will not work on the opposite side (i.e. if you only plug in the left side USB-C port then the right two channels will not charge, and vice versa).
Charging and discharging: Each channel has a PW4056HH battery charge controller; a 3R9 10W bleed resistor; a pair of 8025A dual N-channel MOSFETs; and other associated components. The first 8025A is on the negative end of the bleed resistor and is gated by the MCU to begin discharging. The second 8025A is on the positive end of the bleed resistor and serves as reverse polarity protection. The bleed current is maintained at a continuous 1 amp.
Cooling: The bleed resistors are actively cooled by a 5V 4010 fan. This is temperature controlled by the MCU based on a thermistor on the rear of the board between the bleed resistors.
Logic: A CD4051B analog multiplexer is used to sense voltage and current. IR is sensed a moment after a cell is inserted, probably by taking measurements either side of a load pulse. A 74HC595 8-bit shifter is used to control each charge enable pin, and to drive the status LED for each channel. The MCU has no inscription whatsoever but based on its pinout appears to be an 8-bit Nuvoton N76E003AT20 or a clone of it.
Regarding reliability, it has performed well - I have 3 of these devices, and have tested about 200 cells across them over a combined 300 hours or so runtime. Two faults have occurred during this, one being a noisy fan bearing and the other being a worn out tactile switch.
Regarding accuracy, it does a much better job than my XTAR VC4, which I returned because it was wildly overestimating capacity. For brand new cells it gives capacity values within defined ratings. Out of 100 salvage cells from 19 packs there were 15 identified as deviating in capacity (relative to the rest of the pack) - when retested only 3 of these 15 cells differed significantly from their first test (100mAh or more). I have compared the results of a single cell multiple times in the same channel; a single cell in multiple channels within the same system; and a single cell across the three I own and the results are reproduceable. I'm unsure how accurate IR is, it varies a bit if you re-insert the same cell multiple times, but it is good enough to discriminate a good cell from a bad cell. For voltage, it does vary +/-0.03V from my Klein MM450 multimeter sometimes but it is generally fine.
Hope this helps someone
I have reverse engineered this tester, schematic is attached and is complete except for capacitor values. These testers annoyingly have no model number but they are the ones you find on the first page of just about any website if you type in "18650 4-channel tester". There is also a 2-channel variant and a newer 8-channel variant. They are very cheap (about £15) and can measure voltage, current, internal resistance and capacity. They have a multicolour display for all these statistics. They also have terminal blocks so that if you want to use a breakout board instead of the cell holders (e.g. to test 21700 cells) you can do so. You can set the discharge voltage cut-off between 2.5V and 3.5V in increments of 0.1V.
Here is a backside pic of the board with the base removed:
Here's some info on how it works:
Power: Two USB-C ports at 5V. Only VBUS and GND are electrically connected so USB-PD chargers will not work. The device can run from either port and is protected from reverse current by a schottky on each VBUS, but if you only plug in one cable then charging will not work on the opposite side (i.e. if you only plug in the left side USB-C port then the right two channels will not charge, and vice versa).
Charging and discharging: Each channel has a PW4056HH battery charge controller; a 3R9 10W bleed resistor; a pair of 8025A dual N-channel MOSFETs; and other associated components. The first 8025A is on the negative end of the bleed resistor and is gated by the MCU to begin discharging. The second 8025A is on the positive end of the bleed resistor and serves as reverse polarity protection. The bleed current is maintained at a continuous 1 amp.
Cooling: The bleed resistors are actively cooled by a 5V 4010 fan. This is temperature controlled by the MCU based on a thermistor on the rear of the board between the bleed resistors.
Logic: A CD4051B analog multiplexer is used to sense voltage and current. IR is sensed a moment after a cell is inserted, probably by taking measurements either side of a load pulse. A 74HC595 8-bit shifter is used to control each charge enable pin, and to drive the status LED for each channel. The MCU has no inscription whatsoever but based on its pinout appears to be an 8-bit Nuvoton N76E003AT20 or a clone of it.
Regarding reliability, it has performed well - I have 3 of these devices, and have tested about 200 cells across them over a combined 300 hours or so runtime. Two faults have occurred during this, one being a noisy fan bearing and the other being a worn out tactile switch.
Regarding accuracy, it does a much better job than my XTAR VC4, which I returned because it was wildly overestimating capacity. For brand new cells it gives capacity values within defined ratings. Out of 100 salvage cells from 19 packs there were 15 identified as deviating in capacity (relative to the rest of the pack) - when retested only 3 of these 15 cells differed significantly from their first test (100mAh or more). I have compared the results of a single cell multiple times in the same channel; a single cell in multiple channels within the same system; and a single cell across the three I own and the results are reproduceable. I'm unsure how accurate IR is, it varies a bit if you re-insert the same cell multiple times, but it is good enough to discriminate a good cell from a bad cell. For voltage, it does vary +/-0.03V from my Klein MM450 multimeter sometimes but it is generally fine.
Hope this helps someone
Comment