Strange Mosfet issue

T Chap

I was experimenting with using a Mosfet as a reverse polarity protection method. Using a logic level n channel 75amp mosfet, the off state resistance was around 4meg, yet the encoder in the test still was putting out a perfect signal. This was with Source connected to VSS on the encoder, Drain to the main board. No change to the encoder output whether the gate was at 0 or 5V. I flipped the source and drain just to see what happened, no change. I replaced with an IRF510, same result. Changed to IRF540, and finally the mosfet behaved as expected and stopped the encoder from operating with the gate floating or at 0. In all tests, the mosfet turned on at 5V and stayed on afterwards if floating, at least for a while, but when to megohm resistance with the gate at 0. I wanted to find out what resistance the encoder would stop working at with a fixed resistor in the VSS path to the main board, and it would stop above 100ohms. I verified numerous times that in fact the power to the encoder was in fact passing the VSS through the mosfet, and the off state was at least 4meg on both the first mosfets, maybe 8m on the IRF540. There was no other way there could be a path to VSS. And removing the mosfet definitely kills the encoder output. Any ideas how this can be?

Beau Schwabe

Could be AC coupling ... IOW the mosfet is behaving like a capacitor. Do you have a schematic?

Phil Pilgrim (PhiPi)

Ditto the schematic. Please don't ask questions like that without one. I got lost haflway through that mongo paragraph.

-Phil

Duane C. Johnson

Hi T;

While I didn't quite understand what you were asking, I think I know what you want.

The first two circuits isolate one side of the power source and the third isolates both sides.
These are only circuit snippets. You need to choose suitable MOSFETs for your application.
Keep in mind:
1. Sufficiently low ON resistance to handle the running current.
2. Sufficient Gate voltage to turn On the MOSFET.
3. Not to much voltage to breakdown the gate.

How they work:
MOSFETs have an intrinsic body diode as part of their structure. Current flows through this diode in a direction opposite to the "Normal" direction through the FET.
Think of the MOSFET as a conventional diode in series with the power source, complete with the usual .7V drop.
However:
When the MOSFET is turned ON it looks like a low value resistor. Current can flow through this resistor in either direction. This effectively shunts across the body diode.

I suspect you were attempting to pass current through your MOSFET in the "FORWARD" direction and neglecting to account for the "REVERSE" current that flows through the body diode.

Duane J

Reach

Great example Duane as I have never heard it described so efficiently before.

T Chap

I am not at my computer this week to create the schematic properly, but I cut out the N channel version of the circuit Duane posted. My schematic is exactly that circuit, except add 3 signals from a hall sensor and 2 signals from an encoder. Shown in black is my schematic (with the mosfet drain and source cropped in), the gate is not shown. I was testing the gate to 5V or VSS. The schematic shows VCC going through a resistor, this is not accurate, there is a jumper in place of the resistor.

Phil Pilgrim (PhiPi)

Duane,

In your schematics, you have sources and drains reversed:



Connecting the source to the load (i.e. "source follower"), as you have done, only works if the gate voltage is higher yet than the load supply.

-Phil

Duane C. Johnson

Hi Phil;

The D,G,S symbols are correctly placed. See:
Circuit Symbols

Duane J

T Chap

Duane, to clarify what I was asking: using the n channel mosfet configuration, I tried 3 different mosfets. IRF540, IRF510, and MTP75N03HDL. All three will switch on and off from megohm to a few ohms as they should by connecting the gate to 5V or VSS. In my schematic, the mosfet sits in the path from VSS from the main board to the sensor modules. However, in my circuit, only the IRF540 will turn off the encoder and hall sensors when the gate is at 0. The other two mentioned, even though they read megohm resistance, the encoder and hall sensors still are sending out a signal. The encoder and hall sensors will fail to operate at all with anything larger than 100ohms in the VSS path, so how do they operate with the mosfet at megohms?

Duane C. Johnson

Hi T;

We really can't see the second schematic. It is just mostly black.

Duane J

Phil Pilgrim (PhiPi)

Duane C. Johnson wrote:

The D,G,S symbols are correctly placed. See: Circuit Symbols

No, they're not. See this schematic from the link you posted:



Note that the source is connected to V-, not to the load. In your schematics, the source is connected to the load.

-Phil

Beau Schwabe

"The D,G,S symbols are correctly placed. See:" - Duane, what Phil is saying is that the Source typically goes to the Supply, where the Drain goes toward your load.

BTW) There is also another MOSFET configuration where two of the same channel family are placed back to back.... i.e. Both Drains and both Gates are tied together... each Source goes about it's separate way , one to the Supply, and the other to the Load

Duane C. Johnson

Hi Phil;

Phil Pilgrim (PhiPi) wrote: »

No, they're not. See this schematic from the link you posted:



Note that the source is connected to V-, not to the load. In your schematics, the source is connected to the load.

-Phil

That circuit is used to turn a load on or off, in your example an LED load.

However, T Chap clearly said:

I was experimenting with using a MOSFET as a reverse polarity protection method.

Your circuit can't do this.
In fact, if the battery is reversed the load will see the reversed battery voltage less 0.7V, whether the switch is closed or not. If the load were an integrated circuit it would have probably been damaged.

The examples I showed are useful for protecting the load from a power source that may be inadvertently "REVERSED" in polarity.

Of course, the simplest method of reverse polarity protection is a simple diode but there would still be that pesky 0.7V diode drop in voltage.
The advantage of using a MOSFET instead of a diode, as in my examples, is the voltage drop can be much lower than 0.7V, maybe only a few mV.
MOSFETs can be used in similar circuits as nearly zero voltage drop rectifiers.
In essence the MOSFET reverse polarity protection circuits are a form of half wave rectifier.
MOSFETs can be used to make full wave bridge rectifiers.
See:
Passive Full-Wave MOSFET Rectifiers for Electromagnetic Harvesting
Chapter 3 and figure 3.3
Cool huh!!!

Duane J

Duane C. Johnson

Hi Beau;

Beau Schwabe (Parallax) wrote: »

"The D,G,S symbols are correctly placed. See:" - Duane, what Phil is saying is that the Source typically goes to the Supply, where the Drain goes toward your load.

Granted, however, Phil's circuit can't protect against a reversed power supply as T Chap was looking for.

BTW) There is also another MOSFET configuration where two of the same channel family are placed back to back.... i.e. Both Drains and both Gates are tied together... each Source goes about it's separate way , one to the Supply, and the other to the Load

Sure, the pair of MOSFETs connected with both SOURCES connected together, not drains, form one of the versions of MOSFET SSRs, Solid State Relays.

Duane J

Phil Pilgrim (PhiPi)

Duane C wrote: »

While I didn't quite understand what you were asking, ...

Duane C. Johnson wrote:

However, T Chap clearly said: ...

Neither of us, apparently, saw anything in T Chap's intial post that was completely clear.

Anyway, I do now see what your circuits are trying to accomplish. However, it does look like the body diode will have to hande enough current at a low dropout to guarantee adequate gate voltage to turn the transistor on. This will require a very careful perusal of the MOSFET's datasheet to make sure it can be done that way.

-Phil

jmg

Phil Pilgrim (PhiPi) wrote: »

Anyway, I do now see what your circuits are trying to accomplish. However, it does look like the body diode will have to hande enough current at a low dropout to guarantee adequate gate voltage to turn the transistor on. This will require a very careful perusal of the MOSFET's datasheet to make sure it can be done that way.

Certainly a 5V polarity reversal design will need a Logic level Fet, and a higher voltage polarity reversal design can need a Gate clamp, so you do not exceed the G-S ratings.

In operation, the lower of the S or D acts as electrical Source, so the diode does not add to the threshold voltage.
Below the Threshold voltage, it acts as a body-diode, but most circuits are drawing reduced currents at such low voltages.