Why do the thermistor inputs break sometimes?

AETEKAETEK Posts: 106🌟 Super Member 🌟
Hello,
I have heard several times that the MKS GEN-L mainboard has problems with the inputs of the temperature sensors in combination with the A5. I am developing circuits myself around the AVR microcontrollers and I am interested in what the cause of the problems is. The MKS board is RAMPS 1.4 compatible. This means that the Ramps 1.4 circuit diagram is largely consistent with the MKS Board. If you look at the wiring of the temperature sensor inputs, then you realize that there are only two components per input. One part is a resistor that forms a voltage divider in combination with the thermistor. The second component is an electrolytic capacitor, which is used to filter interference signals. This corresponds to a low frequency filter. The input resistors on the MKS board (R25 ... R27) have a value of 4700 ohms. Due to a short circuit of the thermistor, the resistors can not burn because the maximum possible current through the resistor is limited to 1.06mA @ 5V DC.
A defective electrolytic capacitor can corrupt the measured value. Therefore I would first connect a normal 100k ohm resistor to the input to be tested and then solder out the corresponding 10uF capacitor. If the measured value then shows approx. 20C, the capacitor is defective and should be replaced.
The analog inputs of the microcontroller can handle both + 5Volt VCC and ground (0V) potential without harming it. It has also been heard that on the case of the A5 under certain circumstances + 24V DC may be present. If the 24V gets in contact with any of the Thermistor (Analog ADC) inputs, then the AVR Controller (MEGA2560) will be damaged. It does contain internal clamping diodes; but these are not designed for such a burden. Maybe we should secure the thermistor inputs externally. There are clamp circuits or in the simplest case a 5.1V Zener diode and a resistor for current limiting.

It would be interesting for me to find out if users who have problems with the thermistor inputs also have problems with the 24V DC on the chasis.






Thanked by 1Samuel Pinches
Post edited by AETEK on

Comments

  • Samuel PinchesSamuel Pinches Posts: 2,296Administrator
    edited June 2018
    The USB type-b port on the A5 is connected to the motherboard, which means that the ground from the connected device will be connected through to the motherboard with low resistance. Many computers and devices are not directly grounded either, but may have very high floating voltages - I can measure 100V on my iMac, and I can get small sparks touching grounded USB cables to my iMac chassis. It is easy to get high stray voltages on the gnd, and I think this is how these inputs are getting damaged.
    Post edited by Samuel Pinches on
  • Samuel PinchesSamuel Pinches Posts: 2,296Administrator
    edited June 2018
    I do not consider these analog ports sufficiently protected - and the method you have suggested is the same as used on the "ruggeduino" which is an Arduino that has been specifically engineered to be robust against damage from excessive voltage, and other circuit faults.


     (from rugged-circuits.com)



    Post edited by Samuel Pinches on
  • AETEKAETEK Posts: 106🌟 Super Member 🌟
    edited June 2018
    Hello Samuel,
    I had to open my A5 again, because my motherboard fan is mounted too high and the Y-axis can not drive the whole distance.

    On this occasion, I then carried out measurements.
    For me, USB GND is not connected to the chasis, but to the motherboard (s) GND
    The chasis is not conneted to GND.

    I looked at the Arduino 2560 referencing circuit.
    Stray voltage in my opinion can cause no defects in the ADC (thermistor) inputs. More likely is the destruction of the Mega 2560 microcontroller and the CH340 USB chip. This is what other users have already reported.
    The topic Stray Voltage should be considered separately.
    For me, only one possibility comes into consideration.

    Because the chasis is not connected to GND it can be used as an electrically conductive connection. If the insulation hoses from the hotend and thermistor are displaced or the two components are not properly fixed, then this situation may occur. The ADC (thermistor) input is immediately destroyed as soon as A5 is switched on. The hotend does not need to be switched on because the switch mosfet switches GND. 24V is applied when the A5 is powered on. So 24V can get into the ADC (Thermistor) input and will destoy the thermistor input.
    In this case the clamping circuit (zener diode & restistor) would save the mainboard.


    This has already happened to other users. That's why I carefully checked the electrical installation of the hotend before switching on the A5 for the first time.

    What is your mains voltage?
    Post edited by AETEK on
  • Samuel PinchesSamuel Pinches Posts: 2,296Administrator
    I agree with everything you have suggested... I think your hypothesis is the most likely cause of thermistor input damage....

    Regarding the stray voltage:
    My mains is 240V (+/-10V). The stray voltage is common to many devices with ungrounded metal chassis - the y-class filter capacitor in a switch mode power supply can pass AC voltage at very low current, there are many reports of this in metal MacBooks if you search "macbook tingling". I worry that this could pass via ground USB into the motherboard, and then cause damage by reverse polarity or voltage damage.

    However,  I have not yet seen any users with a damaged Mega2560 or CH340? Have you heard of some?
  • AETEKAETEK Posts: 106🌟 Super Member 🌟
    edited June 2018
    Stray Voltage
    Please note the difference between Ground (GND) and Earth.
    GND is the DC potential (galvanically isolated from the mains).
    Earth is zero potential (N)
    In contrast, in my experience, only helps the galvanic separation between the two devices.
    Strictly speaking, the A5 switching power supply is capable of this (galvanic isolation). I have already analyzed the A5 power supply and found about 4 capacitors connected between ground (green / yellow) and neutral (blue). Here in Germany you can plug the power plug in two directions into the wall socket. Then you suddenly have a stray voltage between live wire (brown) and earth on the capacitors in the switching power supply. If the Chasis of the A5 is grounded, then this voltage is due to Chasis. If you now connect another grounded device, such. PC via USB then a current can flow over the ground contact, which can damage equipment. I know that's all pretty hard to understand. We've also written about this topic before. Well, to avoid the problem (Stray Voltage) I have earthed my A5 power supply, but mounted on isolation blocks. So that no stray voltage from the grid via earth can get into the printer. So almost made of the internal power supply an external power supply, which allows galvanic isolation.
    If you do not want that you can:
    a) Use an USB isolator (always!)
    b) Use an isolating transformer.
    c) For a notebook, only use the power supply via the battery (not mains) when connecting to A5
    d) Add a WiFi connection to the 3D printer (on my ToDo list). This is not recommended for firmware updates.

    You could temporarily break the earth contact of the A5 with a piece of electrical tape and then again measure voltage on the shield between computer and A5 and report the result.



    >> However,  I have not yet seen any users with a damaged Mega2560 or CH340? Have you heard of some?
    Yes, in the comments of DaHai*s video
    on youtube a user reported:


    Here's another one:
    >>I had my 2 days old A5 giving a voltage spike when I first connected my USB cable (at USB3.0 connector on my PC...>>


    Post edited by AETEK on
Sign In or Register to comment.