Best method for reverse polarity protection and without power loss?

flat5

Joined Nov 13, 2008
403
In the shunt circuit could you use a diode and solid state relay (spst nc or double it up, dpst) and not need the fuse?
Relay contacts in series with the hot lead.
Diode activates relay when voltage is reversed. Opens relay contacts.
That would have a low series resistance and hopefully be as fast as a fuse.
 

LDC3

Joined Apr 27, 2013
924
In the shunt circuit could you use a diode and solid state relay (spst nc or double it up, dpst) and not need the fuse?
Relay contacts in series with the hot lead.
Diode activates relay when voltage is reversed. Opens relay contacts.
That would have a low series resistance and hopefully be as fast as a fuse.
Yes, that would work. :)
 

crutschow

Joined Mar 14, 2008
34,284
In the shunt circuit could you use a diode and solid state relay (spst nc or double it up, dpst) and not need the fuse?
Relay contacts in series with the hot lead.
Diode activates relay when voltage is reversed. Opens relay contacts.
That would have a low series resistance and hopefully be as fast as a fuse.
A DC solid state relay has a small forward drop from its internal MOSFET switch. I see no advantage of that over just using a MOSFET, and the SSR will be much more expensive.
 

MikeML

Joined Oct 2, 2009
5,444
I simulated the PFET circuit suggested by Cruts in post #4 and in the video in post #8. The FQP47P06 mentioned in the Video

FQP.gif

seems like a good candidate, but I dont have a Spice model for it, so I am using a similar PFET, the Siliconix SUP90P06:
SUP.gif


First, here is Crut's sim repeated with the SUP90P06 at currents required by my transceiver. I am using only a DC sweep to show the behavior. I plot the voltage across the transceiver V(source) for input voltages V(drain) ranging from -1V to +15V (also shown on X axis). Note that, as expected, the transceiver is protected against negative input voltages.

106c.gif

This PFET has about the lowest Rds of any I could find at DigiKey. I was interested at the power loss at near the transceiver's peak current of ~25A. Look at the red trace, which is the power dissipation in the PFET at the input voltage increases. The bump is created as the gate voltage begins to turn on the PFET. The R2/R3 voltage divider on the gate makes this worse than it otherwise would be if the gate were grounded.

More troubling is the power dissipated by the PFET at input voltage>12V. Note that the PFET would have to be mounted on a heatsink. This is with a PFET with one of the lowest Rds's I could find...

Still more troubling results are illustrated by the following time-domain simulation. I am showing what happens if the transceiver has been powered normally, and then there is sudden reversal of input polarity, say due to some inductive spike that couples in from the car's electrical system. I model the transceiver as a >25A resistive load, with an internal bypass capacitor (taken off a actual schematic) , and its internal reversed rectifier (which is built-into the transceiver).

106c1.gif

Note that as the input reverses polarity at 3us in the simulation (where you would hope that it would shut off to protect the transceiver), the PFET remains turned on so current flows from source to drain. It is only the action of the reversed rectifier D1 inside the transceiver that keeps the voltage at the transceiver V(source) from going more negative than ~2V (green trace). Note the current I(D1) blue trace, which is what normally vaporize the fuse that would normally be in the positive wire...

The reason that the PFET remains turned on as the input polarity reverses is due to the charge stored on its gate. To turn off the PFET, the charge must flow off the gate through the Thev. equivalent of R3 in parallel with R2. If it were not for D1, the voltage across the transceiver would have gone to -14V, and damage would occur except for the clamping action of D1, long before the gate charge is dissipated.

To see it this situation could be improved by effectively lowering the impedance from the gate to ground to get rid of the stored charge quicker. Here is resim of the modified circuit. I left R2=1Ω to show the current I(R2) out of the gate which removes the stored gate charge in order to turn off the PFET. This nicely clamps V(source) red trace to 0V when the input is reversed. The current through D1 (not plotted) is now only a few mA...

106d.gif

While this seems to improve the ability of the PFET to turn off rapidly to protect the transceiver, I would be real nervous with this circuit. Note that you are now totally relying on the PFET's absolute maximum gate-to-source breakdown voltage (only +-20V for the SUP90P06). If the input spike ever exceeds -20V, then the PFET would be destroyed, and reverse protection would be lost. Even worse, a +6V positive spike riding on the nominal battery voltage of 14V would exceed the +20V limit.

I will stick with the fuse-diode method. Any body using the Video or Crut's circuit should be aware of the limits of this method.
 

crutschow

Joined Mar 14, 2008
34,284
Mike, I don't understand why you have two resistors to the gate. That's only required if the supply voltage is more than Vgs max. You just need one resistor to ground which will increase the Vgs and reduce the ON resistance.

If concerned about transients you can connect a 10V zener from the source to the gate.

Zapper
 

MikeML

Joined Oct 2, 2009
5,444
Mike, I don't understand why you have two resistors to the gate. That's only required if the supply voltage is more than Vgs max. You just need one resistor to ground which will increase the Vgs and reduce the ON resistance.

If concerned about transients you can connect a 10V zener from the source to the gate.

Zapper
I'm using the circuit you posted back in #4 ;)

The resistor to ground (with or without a Zener) will have the same problem turning off the PFET that I already explained in post #46.
 

ronv

Joined Nov 12, 2008
3,770
I think somehow we spun off to transient protection from reverse power protection for a little power supply. :D
 

ronv

Joined Nov 12, 2008
3,770
No, I don't think so. He's looking for something to keep him from burning out his 1 amp boost converter if he hooks it up backwards to his 9 volt power supply. Your trying to protect something in a car or plane with all different types of possible transients and instantaneous reversals.
 

ian field

Joined Oct 27, 2012
6,536
No, I don't think so. He's looking for something to keep him from burning out his 1 amp boost converter if he hooks it up backwards to his 9 volt power supply. Your trying to protect something in a car or plane with all different types of possible transients and instantaneous reversals.
I was too lazy to point that out.
 

crutschow

Joined Mar 14, 2008
34,284
I'm using the circuit you posted back in #4 ;)

The resistor to ground (with or without a Zener) will have the same problem turning off the PFET that I already explained in post #46.
You may have used my circuit but you didn't read my description of the circuit. :rolleyes: It stated that for voltages less then Vgs max. you didn't need the resistor between source and gate.

I believe a reverse polarity spike is a rarity in an automotive or aircraft system. Even if you have an inductive load, the spike will be isolated by the switch that initiates the spike. The spike will occur on the output side of the switch, not the battery side.
 

ErnieM

Joined Apr 24, 2011
8,377
I think somehow we spun off to transient protection from reverse power protection for a little power supply. :D
Exactly.
...He's looking for something to keep him from burning out his 1 amp boost converter if he hooks it up backwards to his 9 volt power supply. Your trying to protect something in a car or plane with all different types of possible transients and instantaneous reversals.
You miss the point: if the circuit cannot protect from any and all possible (however unlikely) power faults for some other very expensive high current unit connected to a different power source then it is completely inadequate. :eek:

This thread has been pushed so far off the rails I do hope that Fugi (the OP here, remember him?) has wandered off to another forum where his posts and questions are taken seriously.
 

Wendy

Joined Mar 24, 2008
23,415
Keep it civil. This is not an option.

Keep it simple. Whatever happened to KISS, a quick blow fuse and a diode. For added protection you could always add a transistor, since they tend to beat the fuses when blowing out anyhow.
 

ian field

Joined Oct 27, 2012
6,536
Keep it civil. This is not an option.

Keep it simple. Whatever happened to KISS, a quick blow fuse and a diode. For added protection you could always add a transistor, since they tend to beat the fuses when blowing out anyhow.
A Shottky diode is much faster than a regular silicon one, and clamps any reverse potential to a lower value as well.

Not sure - but I think an SB diode failing SC has more "thud" value as well. That is to say, if the protection diode is sacrificed to save the circuit - it does so with a bit more enthusiasm.
 

ronv

Joined Nov 12, 2008
3,770
I know I should leave this alone, but I just can't. :rolleyes:
Best method for reverse polarity protection and without power loss?
For 9v power, is a 1A diode or Schottky diode (I assume is better for lower voltage drop) good enough?
From this I take away minimum power loss and reverse protection. Maybe battery operated.

1.25 amp fuse .09 ohms $.21
fuse clips $.30
Diode .10
Spare fuse (priceless)

PFET .03 ohms $.23
Resistor $.10

You could use a higher rated fuse with less resistance but if it is a nine volt battery or a small supply it may not blow the fuse or it may take some time. Then the diode needs to be larger.
 
Top