1. so it's ok to completely eliminate R2? can i make R4 a 5k resistor like R3 or make them both 10k? trying to minimize parts.

Yes, R2 is not needed.
The value of R2 is not critical and 10kΩ will be fine.

2. transients will come from engine starters, pump motors, etc. with vicious load dumps. i don't know how big they are but these are what blew zeners before without TVS protection. what do you think?

Any transient to Z1 has to go through R2, thus it would take a large transient to damage Z1.
But a TVS diode wouldn't hurt.

3. i'm already using N mosfet for polarity protection on a different project. it's hooked up like i show in last schematic. it's working good so far. my question is - do i need both zeners or will Z1 protect Nfet gate from over voltage?

You need both zeners, since each MOSFET gate needs its own protection.

. i'm still confused on the effect Z1 will have on driver chip. if voltage supply is 24 volts, the driver chip compensates and used less current from supply. will Z1 always hold voltage at 12v regardless of input voltage? if so, don't think i like that.

It holds the voltage to 12V only for Vgs of Q1.
It has no effect on any other voltage.

5. on resistance is 80mohm. .08ohm x .4a =.032w. is that right? here's the parts i picked.
N- http://www.mouser.com/ds/2/115/DMN6140L-709504.pdf 140mohm/2.3a
P- http://www.mouser.com/ds/2/115/DMP4065S-537770.pdf 80mohm/3.4a
seems both should be ok for 1amp. am i right?

The dissipated power is I²R or .08 x .4² = 12.8mW.

Both of those appear to be more than adequate.

 

all these diodes bring up another problem, i'm running out of voltage. LED Vf is 10.5 to 11.1. with D5 and D6 it's up to 11.7. if customers have a poor switch or low battery, it runs out of voltage. i need to keep total Vf close to 11 volts.

Is this still a concern (from one of you previous posts)?
If so, making D5 (D2 in recent schematic) a Schottky diode will give you abut two or three tenths of a volt more headroom.

 

It holds the voltage to 12V only for Vgs of Q1.
It has no effect on any other voltage.

i'm sorry and i know i'm slow but why isn't Vin on driver chip held to 12V too?

You need both zeners, since each MOSFET gate needs its own protection.

someone says i could connect lower end of R2 between LED3 and L1 and i wouldn't need Z1. i really don't understand that but would like to eliminate unnecessary parts.

Is this still a concern (from one of you previous posts)?
If so, making D5 (D2 in recent schematic) a Schottky diode will give you abut two or three tenths of a volt more headroom.

i've been using a schottky but still need 11.7 volt supply. i want it closer to 11volts. i've changed to an inductor with lower resistance and with the Nfet maybe it will be ok. i remember you said there was a way to reverse the action of Pfet(Q1). that sounds interesting to learn.

thanks again

 

i'm sorry and i know i'm slow but why isn't Vin on driver chip held to 12V too?

No, again.
Vin has to equal the voltage from the switch, since that's what it's directly connected to.
Z1 keeps the maximum Vgs voltage to 12V. The rest of the voltage appears across R2.

someone says i could connect lower end of R2 between LED3 and L1 and i wouldn't need Z1.

That someone is wrong.
R2 needs to go to ground to pull the MOSFET gate to ground which turns the MOSFET ON when the switch is down.

i remember you said there was a way to reverse the action of Pfet(Q1). that sounds interesting to learn.

Here's the circuit with an added N-MOSFET so that the LED is off when the switch is up and off when the switch is down

 

I LIKE IT. saves me .3V. i may be able to use schottky for reverse protection instead of Nfet. only a few cents difference but i have more confidence in diodes.

is R4 necessary in your diagram with this setup?
thanks

 

is R4 necessary in your diagram with this setup?

Yes, otherwise I wouldn't have added it.
It biases M1's Vgs to 0V when M2 is off so that M1 is also off.

 

UPDATE
here is the circuit i built.
it works perfect...THANK YOU CRUTSCHOW !!!
a concern.
i measured voltage on the Pfet gate.
it goes up and down same as the supply voltage when i change it
i went up to 16 volts on supply and gate had 16 volts.
i'm using 12v zener at Z1.
have i got it wired wrong? is zener bad? maybe i don't understand how a zener works.
thanks

 

i measured voltage on the Pfet gate.
it goes up and down same as the supply voltage when i change it
i went up to 16 volts on supply and gate had 16 volts.
i'm using 12v zener at Z1.

The gate will go up and down with the supply voltage when the switch is up.
But remember, it's the gate-source voltage that's of interest, not the gate voltage to ground, and in this case the gate-source voltage stays at essentially 0V, which is fine.

But when the switch is down, the gate voltage will go towards ground until the Zener starts to conduct.
Thus the gate voltage (measured to ground) will stay at about the supply voltage minus 12V if the supply voltage goes above 12V.
So for a 16V supply, the gate voltage to ground will be about 4V, giving a gate-source voltage of -12V (MOSFET on).

 

it works exactly as you said.
i think i got it now. i keep forgetting about the gate/source 'difference' is all that matters. my mind is not wired for electronics.
i never question what you say, i just ask so i can try to understand.
i need to phrase my questions so that's clear.

is there an easy way to test my transient voltage suppression on the bench?
i'm scared to death of zeners from my previous fiasco.
are mosfets as easily damaged?
thanks

 

is there an easy way to test my transient voltage suppression on the bench?
i'm scared to death of zeners from my previous fiasco.

Not easy.
You need a pulse generator with sufficient power to inject the high voltage pulses into the power line.
You have to decide when magnitude and duration of transients you want to use for the test.

But I think there's little chance of damaging the zener in your circuit for any level of transient you are likely to see.

are mosfets as easily damaged?

Only the gate is particularly sensitive.
The gate-source voltage must never exceed the Vgs max rating in the data sheet or the gate oxide insulator can be damaged.

 

my tvs diode clamps at 42v.
if i use a 60v fet, will it be completely safe until the tvs blows open?
i'm trying to stay with sot23 package for fet.

another plus with this circuit, overhead voltage is enough that i won't need the extra Nfet to reverse Pfet action.
thanks again.

 

my tvs diode clamps at 42v.
if i use a 60v fet, will it be completely safe until the tvs blows open?

It should.
But TVS diodes normally short when they fail, not open.

 

a stator based charging system can produce over 200 volts ac.
when the engine's rectifier blows(a common problem due to loose battery cables) it puts this on the boat's supply line.
i've intentionally connected the tvs to ground with a very thin .006" wide trace. is this ok?
i'm using 1 oz. copper on aluminum core board.
if the tvs shorts, do you think the trace will burn through?
is that even possible with transients?
i need a way to prove the customer had a voltage supply problem instead of a light failure alone.
thanks

 

You could add a small resistor in series with the 12V supply to limit the transient current through the TVS, minimizing the possibility of TVS failure.
What is the transient current rating of the TVS?

What is the purpose of the thin trace to the TVS?

 

You could add a small resistor in series with the 12V supply to limit the transient current through the TVS, minimizing the possibility of TVS failure.
What is the transient current rating of the TVS?

What is the purpose of the thin trace to the TVS?

i thought it might blow the trace if the trace was small enough.
smaj26

 

i thought it might blow the trace if the trace was small enough.

Better to put a fuse in series with the power so that it blows if the current transient is too high.
Easier to replace a fuse than a trace.

 

since i don't understand transients, i'm not sure what parameters i need.
i thought transients were so short duration, a fuse couldn't protect the circuit?
can i use a resettable fuse? will they trip fast enough?
how much current does it take to blow a mosfet?
can high voltage(250) blow a fet or the driver chip at low current?
can i put fuse on the low side right before ground connection so i don't need one on both inputs?
need one that can carry 12 watts from 10 to 26 volt input.
thanks