This is the tutorial for how to repair Whatsminer M20s hash board.
Chip Soldering Skills
1. Tin on Chip
Apply a thin layer of flux to the chip with a brush;
2. Put Tin on the Chip Pins
(add solder flux to dilute the solder paste and stir evenly), use tweezers to apply a little solder paste evenly on the chip; On both sides of the pins’ two centrally located, sizable middle pads, apply a small amount of solder paste. Avoid overfilling the solder paste.
3. Melting Tin on the Heating Table on the Chip
the working temperature of the heating table is controlled within 270℃, the chip is easy to be damaged if it exceeds this temperature.
4. Disassemble the Chip
Brush flux onto the chip pins, place the pertinent hashrate components on the heating platform until the tin melts, and then remove the core with tweezers after removing the film. In order to prevent temperature damage to the board and components, the parts of the hash board that do not require welding should not be heated on the platform.
5. Pad Processing
Apply a small amount of tin to the pad with tweezers after the old chip has been eliminated. Avoid using too much pressure when applying. The tin melts, and after that, it sticks to the pad. To make the tin on the pad flat and even, move the pad back and forth using the tweezers’ tip.
6. Install the Chip
When the tin is melted, combine it with the chip and the pad and automatically straighten it to complete the soldering, clamp the soldered chip with tweezers to the precisely aligned pad. The filled soldered pins must have an even distribution of tin that is neither less than nor continuous. To protect yourself from burns from high temperatures, wear gloves when welding.
Maintenance Ideas
Maintenance experience:
1. The reset fails and the LED light is on
When the positive and negative poles of the hashrate board are turned on during the test, the LED will turn on. This is as a result of the output of the RST signal from the hashrate board. 1.8v high level, driving the LED light, the standard board RST is low level 0V, and the light is off. Start at U1 and trace the 0V level RST signal in accordance with the signal flow on which the chip has changed and Raise, locate the change point, and fix the relevant circuit.
2. With the LED light off, the test reset fails.
This is because the RST signal could not be transmitted because U1 →U105 transmission was interrupted by a problem. To find where the RST signal changed from a high level of 1.8V to a low level and fix related circuits, use the echo 1> /sys/class/gpio/gpio99/value command to set the RST signal to a high level of 1.8V.
First, to ascertain whether there is a problem or not, compare the RST signal coming into and leaving the chip where the signal is abnormal.
The associated chips are frequently changed. Due to the parallel connections of 3 chips on the same layer, it is possible for the signal to not pass. In addition, the current leakage of the capacitance of the current layer will also impair the chip’s ability to function normally. For comparison and measurement, it is necessary to switch the multimeter to the resistance file.
3. The reset is normal, and the chip reading fails completely or partially
The chip reading fails even though RST is normal. Checking to see if the CLK 0.9V signal transmission from U1 → U105 is interrupted is necessary at this point. If there is, fixing the associated chip and LDO circuit, checking the associated chip for the CLK, and changing the same layer will have an effect. The crystal oscillator may be damaged if the CLK interrupt occurs at the crystal oscillator, which prevents reading of the chips behind it.
The capacitor at the interruption point should also be checked for leaks at the same time.
In the event that the CLK signal is sound, the 1.8V RXD signal transmission from U1 and U105 will be checked to see if it is being disrupted. If there is an anomaly, determine whether the chip is defective, whether the LDO is normal, and determine whether there is electricity leakage by measuring the relevant layer capacitance.
If the RX measurement is flawless, the measurement is taken from the 1.8V TXD signal transmission of U105 →U1; if the TXD is interrupted, the interruption may also be the result of a faulty chip, a faulty LDO, or whether the pertinent layer capacitance is leaking.
4. Measurement of the signal voltage serves as the primary foundation for the entire board’s detection, and the method of measuring the capacitor’s resistance for maintenance purposes serves as a backup.
5. When some hashrate boards began testing, it was discovered that the power current speed increased quickly and the hashrate boards quickly heated up. This occurs when an abnormal signal voltage causes the chip to begin operating; to prevent this, the signal must be quickly measured during maintenance, and it must be blown cold after heating. Retesting is only possible after that.
6. unbalanced layer voltages on boards. Repairing and replacing the chip, however, will not enable the signal to be transmitted again. For instance, some layer voltages are only 0.2V, while others are close to 0.5V. Each layer’s voltage is markedly measured when the board temperature is lower.
To maintain an even voltage across the layers, the lower layer switches a chip. Swapping a few layers will balance the voltages if there are several unbalanced layers. Fix the signal later.
7. The first chip may be reported as defective if some chips fail the software test to read them. There might be an issue with the crystal oscillator if there is one next to it.
8. If the LDO output pin 5 does not have 1.8V, see if the input pin 1 voltage is sufficient. If the input is normal but the output is abnormal, replace the LDO chip. If the LDO is still abnormal after replacing it, consider whether the current layer chip he is powering is defective and pulls down the voltage. Is the power next to the LDO leaking capacity? If the input is normal but the output is abnormal.
