Comparison of reverse-polarity PMOS circuits

[dknguyen]

Does anyone know what benefit is supposedly gained from the current mirror in the circuit as opposed to the two simpler circuits? It's not clear to me. About the only thing I can come up with is that it the current mirror produces it's own voltage drop across the gate-source without relying on the load itself which might make things things more predictable or reliable, but I am unsure if that actually would be the case.

 

[tomizett]

From a quick look, it seems that the first should be more sensitive to the direction of current. If you just turn the FET on by bringing the gate low through a resistor (as in the other 2 circuits) then it can remain turned on while the "load" sources current to the "supply", as long as it can keep enough current flowing to prevent the source falling below the turn-on threshold of the FET.
So, the simpler circuits would be fine for reverse-polarity protection where you're not concerned about stored charge in the load dumping into the supply, but not for active rectification.

 

[dknguyen]

Hmmm backdriving was something I had not considered. I am having trouble analyzing what the current mirror would do when backdriven though.

 

[MikeMl]

 

[dknguyen]

View attachment 111072 View attachment 111073

Thanks. Could you please graph the voltage of R1 and R2? I seem to have two contradicting ideas of what those voltages should be doing. There's some hole in my understanding of current mirrors.

 

[crutschow]

Here's an explanation of the circuit operation that may help.

 

[dknguyen]

Here's an explanation of the circuit operation that may help.

Ah, thanks. The website I ran into that circuit on was the only one its kind that came up on Google. Every else PMOS reverse polarity related is the other run-of-the-mill stuff so I had nothing else to go on.

How would you use unipolar TVS diodes on this circuit? It doesn't seem like it would survive if the PMOS was closer to the supply than than the TVS. However, if the TVS is closer to the supply than the PMOS then the TVS would try and shunt all the current in a reverse polarity situation. Is a bipolar TVs the only way to go?

 

[crutschow]

How would you use unipolar TVS diodes on this circuit?

I don't understand the question.
If you are referring to the current-mirror circuit, there are no diodes.
What would be the purpose of the TVS?

 

[dknguyen]

I don't understand the question.
If you are referring to the current-mirror circuit, there are no diodes.
What would be the purpose of the TVS?

I mean, if one wanted to combine ESD protection and reverse polarity protection on a supply input line. The rest of the circuit only needs a unipolar TVS but a bidirectional one is needed to not blow during reverse polarity.

 

[crutschow]

I mean, if one wanted to combine ESD protection and reverse polarity protection on a supply input line. The rest of the circuit only needs a unipolar TVS but a bidirectional one is needed to not blow during reverse polarity.

The current mirror circuit is not appropriate for reverse polarity protection for more than about 5V, since higher voltages will exceed the reverse-bias voltage the current mirror transistor.

 

[tomizett]

Is that your circuit Crutschow? It's rather cunning, I like it!

 

[crutschow]

Is that your circuit Crutschow? It's rather cunning, I like it!

I came up with it somewhat independently but I think others on the web designed a version before me.