5
\$\begingroup\$

I'm not particularly well acquainted with MOSFET logic design and need some verification on a few things to make sure I've got them correct, and to see if there's any better ways of doing it. I need to control a 12v battery input terminal with a 3V3 logic level from an ARM microcontroller. I've put the following circuit together while digging through some other answers on stack exchange. Power Controller

Power_Control is the ARM GPIO and operates on 3V3 logic. Q1 is an N channel Mosfet with datasheet. It has V_GS(th) = 2.1, V_GS being either 0, when the GPIO is low and 3.3 when it is high. So I assume that should work, but I honestly don't know if that's how I'm meant to be reading the datasheet. According to my understanding, Enhancement mode NFET should be off when V_GS = 0 and on when V_GS>=V_GS(th) and a Enhancement PFET should be off when V_GS=0 and on when V_GS<=-V_GS(th).

Q2 is a P channel Mosfet with datasheet. It's got max V_GS of +-20V so I assume it's rated to handle my input and it has a V_GS = -3V and it should either be -12 when Q1 is allowed to conduct or 0 when it isn't allowed to conduct. Q2 datasheet doesn't specify it, but I'm assuming it's an enhancement mode. Is that a correct assumption to be making?

Practically speaking, will this circuit work for the desired outcome? Are there better ways to achieve it? Less components, cheaper price or if it would be possible to use 2 of the same channel FET's as it'll save on cost.

\$\endgroup\$
4
  • \$\begingroup\$ you probably want to swap D and S of Q2. \$\endgroup\$
    – jippie
    Commented Dec 8, 2015 at 5:28
  • \$\begingroup\$ Take a look at Fig. 5 in the datasheet for Q1. At 3V the drain current is still very small. I would personally replace it with a BJT and a base resistor. \$\endgroup\$
    – jippie
    Commented Dec 8, 2015 at 5:31
  • 2
    \$\begingroup\$ Thanks for the feedback. Drain Current through Q1 should only need to be 1.2mA shouldn't it? 12V/10000. It looks like it should be able to handle 50-100mA from Fig. 5? \$\endgroup\$
    – NathanielJPerkins
    Commented Dec 8, 2015 at 5:36
  • \$\begingroup\$ You are right, I thought the graph was in mA but it is in A. \$\endgroup\$
    – jippie
    Commented Dec 8, 2015 at 5:41

1 Answer 1

Reset to default
3
\$\begingroup\$

It will work. The drain current of 2N7002 at 3V is still around 0.1A... which is ok with the 10K resistor (needing only 1.2mA). But neither turn-on nor turn-off will be very fast. A somewhat marginal design. I'd do a Monte Carlo simulation on it if it were for series production.

You can get high-side switches all in one package these days, if that's what you're looking for. "better ways" depends what you mean by better. More protection for the switch? Cheaper? etc. This was covered to a fair extent in a related question: Switch (Relatively) High Voltage from Logic Level (It's not the same application [different voltages], so I won't call it a duplicate).

\$\endgroup\$
7
  • \$\begingroup\$ I guess I would consider it a better application if it used less components, cheaper components or if it could be done with 2 of the same channel FET's as a minimum cost for a chip is required. It's not for any sort of fast switching circuit, but used in conjunction with a simple low voltage detector circuit for the 12v input so it can turn off the main power users in a small once off board if it gets too low. Productions less than 5 total, so nothing massive. \$\endgroup\$
    – NathanielJPerkins
    Commented Dec 8, 2015 at 5:52
  • \$\begingroup\$ @Thallazar: you'd need to specify what load current you expect, what switching times. etc. For fewer parts, use single switch with built-in level shifter, e.g. vishay.com/docs/73449/si1869dh.pdf That one can't handle the same load current as the Si2343 (if you need that much). So you need to look around for others. \$\endgroup\$
    – got trolled too much this week
    Commented Dec 8, 2015 at 5:54
  • \$\begingroup\$ You can also replace the first FET with a BJT which might be marginally cheaper (unless it's for large scale production won't matter though): electronics.stackexchange.com/questions/123710/… (Note that the first schematics there is missing a base resistor for the BJT though.) \$\endgroup\$
    – got trolled too much this week
    Commented Dec 8, 2015 at 6:08
  • \$\begingroup\$ Max load current for the 12v side should be something like 2.5A. Thanks for the help, I'll take a look into switches with level shifters. \$\endgroup\$
    – NathanielJPerkins
    Commented Dec 8, 2015 at 6:09
  • \$\begingroup\$ @Thallazar: Yeah, single package ones are usually lower drain voltage (<=12V) or low current. The reason has to do with them being usually planar rather than trench; see eetimes.com/document.asp?doc_id=1272433 for a bit more. \$\endgroup\$
    – got trolled too much this week
    Commented Dec 8, 2015 at 6:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.