short circuit detection with adjustable threshold

[OM222O]

I need to build a circuit that can detect if the resistance between the two probes is less than a certain threshold (max value would be 100ohms since it's used for short circuit detection). I've seen a few examples of short circuit detection circuits using BJTs that activate a LED and buzzer but that's not what I want. they also have next to no explanation on how the circuit actually works, so I don't know how to modify them. their output is also quite analog, but since I want to interface the output to an MCU, that's not a huge deal (MCU pins have schmitt trigger inputs). any ideas on how to do this?

[SiliconWizard]

I guess the solution would largely depend on the surrounding circuit. Unless this would be a completely stand-alone detector without anything else connected to the load between the probes?

[max_torque]

Are there any common mode voltages on the conductors you want to "short circuit detect" on, how fast a response do you need, what are the true thresholds for detection and is there any other noise or anything going on with those conductors?

[OM222O]

it's not that complicated. noise / bandwith etc are non issues, I just want to use it as a sanity check to check shorts between nets on a PCB. there is no common mode voltage and again, the threshold is something between 10 and 100 ohms. I just want a digital signal which says if there is a short or not.

Something like an in circuit tester used in production environments. I'm just not sure how active components like ICs behave when the PCB is populated. a simple voltage test also fails if there are any resistors, since very little current flows into the IO pins, the voltage drop across the resistors is almost 0.

[Gyro]

To ensure that semiconductors are not detected as short circuits, you will want to limit the maximum voltage across the probes to less than about 0.3V.

An unbalanced Wheatstone bridge circuit is a good approach, eg.   https://www.nutsvolts.com/magazine/article/build-a-continuity-tester

[OM222O]

but the schmitt triggers on the MCU detect a high when voltage is about 2.5v that means I need a comparator for every net. maybe using a mux or "switch IC" would be a better alternative?

[krayvonk]

this is a basic question.  a transistor does it easily... so easy in fact, you dont even need one.

</img>

[Gyro]

but the schmitt triggers on the MCU detect a high when voltage is about 2.5v that means I need a comparator for every net. maybe using a mux or "switch IC" would be a better alternative?

Yes, I'm afraid it does, of you want to do it properly. You could maybe try analogue muxes, but you'll still need the bridge circuit for each net - I don't think you could find muxes with sufficiently low and predictable on-resistance to share that across the nets too.

I realize that you would really like to use the 2.5V threshold on your MCU but as I indicated, that would then register semiconductor junctions as 'shorts'.

[Mechatrommer]

I've seen a few examples of short circuit detection circuits using BJTs that activate a LED and buzzer but that's not what I want. they also have next to no explanation

https://www.eevblog.com/forum/projects/semi-conductor-safe-continuity-tester/msg1725182/#msg1725182

reading from left..
1) power supply (9V battery)
2) power led indicator
3) Vtest emitter follower BJT Q2 to provide DUT voltage testing. adjustable by changing R4, R7 voltage divider. 2V in the picture is set as default, i guess it should be safe for most semiconductor (10Kohm output impedance from R5 hence maximum 200uA shorted/drain current to 0 ohm DUT)
4) this is the heart of the adjustable continuity tester (see below)
5 upper) comparator (based on continuity result at Ip and In) that will turn on Q1 to provide power to buzzer.
5 lower) simple self oscillating circuit to beep the buzzer (i got this from just googling)
6) input protection (to some degree) can be omitted if DUT is passive device such as resistor.

4) explained...
it is some form of bridge or twin voltage dividers if i may say...

we adjust Radj to what range we want to beep. in the schematic its using 200 ohm trimpot so it basically can be adjusted to beep up to 200 ohm continuity. if we put trimpot to center position (100 ohm) anything lower resistance on probe point 1 and 2 will make the device beep. this Radj trimpot with R6 forms voltage divider that will act as voltage reference to comparator negative input In.

we put our probe on the second (twin) divider. if point 1 and 2 (probe) has lower than Radj resistance, comparator will make the circuit beep because voltage at Ip is lower than voltage at In, so comparator 101, it will pull the Q1's base PNP low and turn it on and then power the oscillator circuit hence beeping.

i cant think of any simpler circuit, this should ranked in the first beginner class of this forum. just a comparator, oscillator circuit to beep, and voltage dividers bridge. you can go mcu route but it should ends up pretty much similar (or more) complexity than this, you will need voltage dividers for adjustability, and beep circuit anyway. fwiw...