7
\$\begingroup\$

I have some MOSFETs that turn on without any voltage applied to the gate. They turn on even without the gate connected to the circuit; I tested them in a simple switch circuit using an LED:

enter image description here

I also checked them with a multimeter. It gives a reading only between the source and the drain, the current flows only from the source to the drain.

These are the part numbers: k2519, k1985, k3115, 2d0n60f, and irf740. There is many others with the same problem. I have checked all of the datasheets, they say they are n-channel MOSFETs, so I don't know what the problem is.

enter image description here

Here you can see the LED is on without the gate even connected (the green wire). The orange wire is the drain, the blue wire is the source.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ you didn't apply a + voltage but you also didn't apply a - voltage or 0 voltage so what voltage do you expect there to be? \$\endgroup\$ Commented Feb 14, 2023 at 23:46
  • \$\begingroup\$ Try your experiment again, with a resistance of no greater than \$100\mathrm k \Omega\$ between gate and source. \$\endgroup\$
    – TimWescott
    Commented Feb 14, 2023 at 23:56
  • 10
    \$\begingroup\$ This was a quiz trick-question I gave in an electronics lab years ago. What voltage is the MOSFET gate? I would literally take any answer besides 0V because it's simply an unknown voltage. Newbies often think that unconnected means grounded which couldn't be further from the truth. \$\endgroup\$
    – horta
    Commented Feb 15, 2023 at 0:08
  • 1
    \$\begingroup\$ You connected your MOSFET "upside-down". It's conducting through the drain-source body diode. \$\endgroup\$
    – brhans
    Commented Feb 15, 2023 at 20:12

6 Answers 6

12
\$\begingroup\$

The gate to the MOSFET is a capacitor. Capacitors store charges. If the gate previously had a small charge, that would stay, keeping the MOSFET on (or off), unless the charge is removed ("bled" off) by a connection to the circuit through a pull-down or pull-up resistor.

However, the circuit as-is might make a useful bottle detector! If the blue wire is connected to the gate,

  • Get a plastic bottle, such as an empty medicine vial or empty, dry, soda bottle.
  • Rub the plastic container on a piece of cloth, such as an article of clothing.
  • Wave the bottle back and forth near that gate wire.

Voila, the LED, she brightens and dims... but beware of a spark from bottle to gate, lest the MOSFET, liala tamut, RIP.

\$\endgroup\$
3
  • 13
    \$\begingroup\$ Well, if a spark happens between bottle and gate there's a very good chance that it'll be a latching bottle detector. \$\endgroup\$
    – TimWescott
    Commented Feb 14, 2023 at 23:55
  • \$\begingroup\$ I know that the gate is a capacitor and I checked it by connecting the three wires together to discharge the capacitor but the problem remained, also at first I have pulled down a 10k resistor to the ground but the same result. \$\endgroup\$
    – user320546
    Commented Feb 15, 2023 at 0:55
  • 4
    \$\begingroup\$ Since you state that ALL the n-chan MOSFETS conduct, regardless of whether the gate is pulled down or not, this seems to indicate that you may be connecting the Drain & Source wrongly. Drain is centre pin. If wrongly (reverse) connected the ALL MOSFETS will conduct through their inherent "body diodes" \$\endgroup\$ Commented Feb 15, 2023 at 12:50
12
\$\begingroup\$

Rule 1 for MOS: gotta tie that gate to a known voltage to get the result you expect.

The MOS gate is essentially one side of a capacitor. Without anything tied to it, it's a floating signal that can pick up stray voltage from, well, just about anything. So even though you haven't applied a voltage explicitly, your FET's gate nevertheless has a voltage on it.

If this behavior is a concern as you do your bench work, add a pull-down resistor to the FET between gate and source to bias the gate to 'off'. The resistor can be a pretty high value: 10k ~ 100k would work.

Also, check your drain and source connection. The pinout is:

  • pin 1 - gate
  • pin 2 - drain
  • pin 3 - source

Rule 2 for MOS: The FET will conduct through its body diode if the FET is connected in 'reverse', regardless of the gate voltage. In this case for an n-FET, if Vsource > Vdrain + 0.5V or so, it behaves as a diode.

Try this:

  • disconnect your MOSFET from your system.
  • Tie gate to source.
  • Test source to drain with a voltmeter, in both directions using 'diode' test.

If the FET is good, you should see conduction only when (+) is on source (reverse direction - body diode.) If you see continuity in both directions, congratulations, your FET is toast.

\$\endgroup\$
7
  • \$\begingroup\$ I know that the gate is a capacitor and I checked it by connecting the three wires together to discharge the capacitor but the problem remained, also at first I have pulled down a 10k resistor to the ground but the same result. \$\endgroup\$
    – user320546
    Commented Feb 15, 2023 at 0:57
  • \$\begingroup\$ Your FET may be damaged. That's another problem with MOS: static discharge can easily fry the gate capacitor. \$\endgroup\$ Commented Feb 15, 2023 at 1:01
  • \$\begingroup\$ But it's not just one or two FET with this problem, a have several all with the same problem. \$\endgroup\$
    – user320546
    Commented Feb 15, 2023 at 1:37
  • 4
    \$\begingroup\$ The other possibility is that you have swapped drain and source. The FET will conduct through its body diode if source is a higher voltage than drain. \$\endgroup\$ Commented Feb 15, 2023 at 1:46
  • \$\begingroup\$ Where did you get your MOSFETs? Lots of fake parts are being sold. Mae sure you use a reputable distributor. \$\endgroup\$
    – PStechPaul
    Commented Feb 15, 2023 at 2:00
9
\$\begingroup\$

That's normal and even expected if you left the gate unconnected, so there is no problem. It just does not work the way you expect it to work.

FET gates must not be left unconnected.

The gate has extremely high impedance and any leakage currents and electric charges can push the gate voltage high enough for it to turn on and conduct.

If you want the FET to stay off, add a reasonable sized resistor like 10k from gate to ground to keep the gate voltage discharged.

\$\endgroup\$
2
  • \$\begingroup\$ at first, I have pulled down a resistor of 10K between the gate and the ground but the LED keeped on, so I disconnect the gate from the circuit to see whats happen. \$\endgroup\$
    – user320546
    Commented Feb 15, 2023 at 0:43
  • 1
    \$\begingroup\$ @Straight-Adhi there is quite a chance that the MOSFET is already damaged by static electricity. MOSFETS are vulnerable to static unless the datasheet explicitly states otherwise. \$\endgroup\$
    – fraxinus
    Commented Feb 15, 2023 at 14:46
2
\$\begingroup\$

You have wrongly connected Drain and Source. Simply swap the connection, and if everything else is wired correctly - including gate pull-down resistor - then your circuit will work properly.

\$\endgroup\$
1
\$\begingroup\$

Remember that a mosfet is controlled by the gate-source voltage. To keep an NMOS (enhancement type- the usual) off, connect a resistor between gate and source pins. Then there won’t be any confusion. No worries about what’s ground etc.

To identify which is drain and which is gate, use a multimeter in diode mode. Once you find a combination of two pins that shows typical silicon diode forward voltage (0.6-0.7V), the pin you touched with the red (+) probe is the source, and the one touched with the (-) probe is the drain. That’s true for NMOS. For PMOS, the drain and source are swapped vs. the foregoing.

\$\endgroup\$
0
\$\begingroup\$

P-FET has inner (body) diode that connects its Drain to Source, while in N-FET, body diode connects Source to Drain. That may answer your question.

\$\endgroup\$

Your Answer

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