Short circuit protection of PLC's digital outputs

[JesperMP]

But still, protection is needed and Ohm's law still cannot be beaten. With a current limit of 500mA, we will have 12 (24 * 0.5) watts of power on the key element, in which case each of the discrete outputs will have to be equipped with a boxed processor heatsink)

That is not how it works.
When the current limiting electronics kicks in, it reduces the voltage not by increasing the resistance but by quickly chopping input voltage, and probably with some dampening to not generate a lot of noise. In this way there is no excessively high heat dissipation in the module.

 

[kallileo]

We have developed our own Modbus remote I/O board with multiple types of analog/digital/relay outputs and inputs.
There is a shunt resistor on each 24V output which is used to measure the voltage using the analog input of the main MCU. This way in case short-circuit or overcurrent event happen on a particular output the firmware is able to detect it within 5-10ms and turn off the FET, set the appropriate Modbus coil as well as inform the user through the built-in Web server.
So we have two very useful functions on each output:

1. We now know how much current the load on the output draws
2. Overcurrent/short-circuit protection

We also have one PTC supplying each FET which drives two 24V outputs as extra protection.


On another board we use some automotive 24V, 25A FET drivers with built-in overcurrent and overtemperature protection which are pretty much indestructible.

 

[RisingEdge]

JesperMP, you mean PWM signal? So, good idea, thank you!

 

[RisingEdge]

kallileo, i want to implement exactly such a software reaction when a short circuit is detected - turning off the output transistor) And what type of output do you use, source or sink?

 

[kallileo]

Outputs are N-channel FET so It's sinking.

I'm not sure but I believe the shunt is a 1Ω 2W resistor so the output is practically limited to 1A in order to avoid damaging the shunt resistor.
Then you need a one analog and one digital i/o on the MCU for each output.

At first we didn't implement the overcurrent shutdown of the output in firmware but we noticed that eveytime there a short-circuit one one of the output the 3.3A power supply would shutdown resetting the whole control system as it's internal short-circuit protection would kick in. So we had to make the overcurrent protection in firmware to be activated faster than the short-circuit protection of the PSU as we didn't want to use external 5-20mm fuses.

 

[RisingEdge]

kallileo, 1 ohm at 24 volts will limit to 24 amperes, but this is not the point) The method of measuring and responding to a short circuit event is absolutely clear to me, and yes, the protection should work faster than the power supply) But this is not a problem, everyone's favorite ARM CORTEX M can respond to interrupt and put a bit in the output register in less than 1 microsecond, which is quite enough. Transistor turn-on and turn-off times can be ignored, since today these figures are measured in nanoseconds. And thank you for the information!

 

[kallileo]

P=I^2 * R = 1^2 * 1 = 1W
The current sensing resistor is 2W.

 

[RisingEdge]

You said that the current shunt could act as a current limiter during a short circuit, but in this case this resistor would be the load, hence the 24 amps into the load (I = U/R)