0

I'm using the circuit given in the example here: Understanding an 'ideal' diode made from a p-channel MOSFET and PNP transistors

[a] Circuit

Basically there are two issues.

  1. In the forward direction [Vin = 15V on IN] it appears as if the gate voltage is only just below the positive rail, it is not being pulled to ground, which may have implications for higher currents. Purple is the gate voltage, also indicated by cursor A which reads 11.6V.

[b] Forward direction 15V

  1. The ideal diode works below 13.4V in the reverse direction [VIN 10V applied to OUT] but as soon as the voltage rises above this current flows to the input side.

[c] 10V applied in reverse

[d] 15V applied in reverse

From these c and d you can see that the mosfet gate is kept at VIN and at 10V there is no voltage rise at the input, but at 15V there is 2.7V at the input. If the mosfet is off it where is the current leakage coming from? The IRLML9301 is rated to 30V Vds and 20V Vgs so it should be fine. It blocks 20V in the reverse direction just fine when the gate is pulled high.

I have tried a few resistor combinations with no change and my simulation in iCircuit does not reproduce this behaviour.

SticilFace
  • 51
  • 3
  • Are you using a matched pair of transistors? Or just two random 2N3906s? – Hearth May 20 '21 at 14:49
  • Also, just to double check, are you sure you have your pinouts right in the physical circuit? You don't have drain and source swapped or anything? – Hearth May 20 '21 at 14:54
  • they are not a matched pair - i did test them and there are 5mV between the 2N3906 (657mV and 652mV). My understanding was that this only matters when the voltage either side is close... i could be very wrong though. – SticilFace May 20 '21 at 15:04
  • I've not looked into this circuit in particular, but the BJT pair looks like a current source of some sort, which can be very sensitive to not just voltage difference at any given moment, but temperature difference as the parts heat up in the circuit. I don't know if that matters for this application, though. – Hearth May 20 '21 at 15:05
  • re pins: I have double checked the mosfet but not the BJT. I think you are right about it being the source. I'm much happier with FETs never really figures transistors out.. – SticilFace May 20 '21 at 15:08
  • Oh, BJTs are much easier to use in some applications. And you can barely get small-signal FETs anymore. – Hearth May 20 '21 at 15:09
  • *I'm much happier with FETs never really figures transistors out.* The **T** in FE**T** stands for **transistor**. You probably mean BJT: NPN / PNP. Indeed here the PNPs need to be PNPs as you will have a hard time finding a MOSFET that can deliver the amplification that is needed for this circuit to switch decisively (which is what you want). – Bimpelrekkie May 20 '21 at 15:12
  • yes of course. specifically PNP and NPN i've not used a lot – SticilFace May 20 '21 at 15:15

0 Answers0