1

First off, let me state that I am pretty new to FETs, so I'm still wrapping my head around the different characteristics.

The circuit below is a switch that will trigger a 12v (VIN=12v) signal to be constantly on, based on input from an MCU (MCU_TEL_ON=3.3v logic level).

Circuit

The DMP1045UQ P-FET I originally used is wrong for the application because its Vgss is +/-8V.

As such, I've had multiple failures of this part.

I'm trying to source a P-FET suitable for the job (~5mA load, acceptable voltage drop of 0.5v and rise time is really not a factor either) and I'm between the DMP3098LQ and the PMV65XP.

My questions are:

  • Either one of these should work, correct?
  • The Vgs given my circuit is -12v, right?
  • What characteristics matter most when selecting a part for this particular job?
t3ddftw
  • 515
  • 4
  • 13

2 Answers2

1

What characteristics matter most when selecting a part for this particular job?

Well, you haven't said what the load is and what volt-drop can be tolerated in the PMOS transistor. If I were to assume that \$R_{DS(ON)}\$ was the most important critieria I could make a comparison: -

The DMP1045UQ has around 25 milli ohms on-resistance but the DMP3098LQ is 83 milli ohms on-resistance at a load of 0.5 amps. I just chose 0.5 amps as a guess for what your load might be. So, if that criteria is important, then the DMP3098LQ may not be good enough.

The PMV65XP is about 55 milli ohm on-resistance at 0.5 amps so neither of the alternatives are significantly close to the original (failing) part to be considered.

But, you might not need such a high-spec on-resistance - only you can tell us.

Either one of these should work, correct?

Impossible to say without knowledge of your load and the acceptable volt-drop that can be tolerated in the PMOS transistor when driving it.

The Vgs given my circuit is -12v, right?

Correct on the face of it but, can that 12 volts rise any higher or lower when active - this might make a difference and, also, is it subject to noise or spikes superimposed on the 12 volts - this could make a massive difference to the choice of PMOSFET.

The characteristics of your load could ruin a few little MOSFETs if not taken properly into account. I'm thinking back-emf, inductive load etc..

Andy aka
  • 456,226
  • 28
  • 367
  • 807
  • Thanks, Andy! I had edited my OP to add that the load is about 5mA and the acceptable voltage-drop can be ~0.5v. Rise time is really not a factor either -- this is a logic line that I am simply driving high at power on. – t3ddftw Jun 16 '20 at 16:52
  • 1
    @t3ddftw OK that will work for the alternatives but what about the leakage current when off? How much leakage current through the MOSFET can you tolerate? Also how much noise on your 12 volt supply? – Andy aka Jun 16 '20 at 16:57
  • The max leakage current of 100nA for the DMP3098LQ should be fine. I have to measure ripple of the power supply as I'm directly using a 12v battery input in a vehicle. Thanks again! – t3ddftw Jun 16 '20 at 17:18
  • 1
    That's the gate-source leakage - drain source leakage is 800 nA. Vehicle battery supplies are prone to massive surges - sometimes up to 100 volts positive and negative - be very sure you know what you are designing here. – Andy aka Jun 16 '20 at 17:29
  • I have a Varistor on the input prior to the gate of this p-MOS, which should help with the transients, no? I have also been considering a common mode choke to further suppress noise, but that's a bit out of scope this question :D – t3ddftw Jun 16 '20 at 17:32
  • 1
    Only if it is massively power rated or there is another current limiting device. This is not a clean easy task to get right so be aware of this. Personally I think you should be considering more than just a varistor because on high surges they can be ineffective and turn into a short circuit and melt. – Andy aka Jun 16 '20 at 18:27
  • Though it's out of scope of this question, would you recommend something other than a common mode choke, or is that the right approach? – t3ddftw Jun 16 '20 at 19:31
  • 1
    Before designing any protection circuit you need to know what surge specification you have to make your MOSFET switch resilient to. This is not trivial and I can’t conjure up a fit-all protection method. Speak to someone who you are designing this for and clearly establish what surge current rating you need Based on some form of end-user requirements. Anyone who tells you to put a varistor or TVS here or there without specific knowledge of the application and it’s environment are not worth talking to. You have to do this yourself and I cannot guess it for you. – Andy aka Jun 16 '20 at 21:02
  • Thanks Andy, that makes sense! – t3ddftw Jun 16 '20 at 21:49
  • 1
    Good luck and when you have the info, come back and ask a new question targeted at that specific protection method. I’ve designed a few but always, it has to be designed against a specific threat, usually embodied in an IEC or EN standard or something equivalent for auto environments. – Andy aka Jun 16 '20 at 23:21
1

For such a low current I would use a BSS84 because they're general-purpose, dirt cheap and I always have a reel of them around. They're 10\$\Omega\$-ish so drop about 50mV @5mA (maybe 75mV if you get them really hot).

If your 12V is a vehicle automotive electrical system you would do well to add a ~15V zener diode and series resistor to protect the p-channel gate from potential transients in the 100V range.

Spehro Pefhany
  • 397,265
  • 22
  • 337
  • 893
  • Thanks! It does indeed seem that this will work. I'll buy some for testing. Also, thanks a bunch for the zener / resistor tip -- this is going into a vehicle. – t3ddftw Jun 16 '20 at 17:54