Zener to clamp Mosfet Q

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
hi all,
i am using n mosfet to control a solenoid valve. fet by itself can handle flyback, but i usually add my own zener across drain-source to keep the flyback V down on the fet. i see typical use is one component, just the zener (D1) across from drain to source. but my question is, why is there no use of a series resistor with the zener? w/o a series resistor what limits the current through zener when flyback V reaches zener threshold?

what does zener current of 49ma mean for the Fairchild 1N4733A?
example pic (not my circuit)
 

DickCappels

Joined Aug 21, 2008
10,153
Let me turn the question around: Why would anybody put a resistor in series with the zener? (The flyback puls is current not voltage.)

The 49 ma shown on the datasheet for the 1N4733 (I am looking at the Vishay datasheet) is the amount of current that is applied to the zener while measuring the zener voltage. This needs to be specified zener voltage is a mild function of current.
 

Picbuster

Joined Dec 2, 2013
1,047
I do not see reason to install zener.
A zener will draw always current. http://www.bing.com/images/search?q...+diode+curve&qpvt=zener+diode+curve&FORM=IGRE
The current goes non linear up till threshold(zener) voltage is reached and will blow, without a resistor, the zener.
This will happen at the point that the zener power threshold is reached.
A serial resistor will 'consume' this current resulting in a voltage over that resistor from Vconnected - Vzener hence producing a stable voltage over the zener.
My question why do you want always a current flow in your motor circuit?
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
hi all,
i am responding to the few replies.

in the circuit i posted having zener D1 is very common.

isnt the flyback voltage made by the collapsing mag field? the mag field has energy associated with it. thus when the fet goes hard off that field falls fast and the voltage will climb on fet drain, possibly to hundreds of volts. w/o a way to channel that energy it will eventually leak out. ignoring the fet for now, the zener would reverse bias at specified voltage allowing the coil energy to find a good path to zero. but if the zener is in full reverse bias what limits the current through the zener?


as for power, if the zener is rated 1w then it can handle 1J/s, but what limits this rate? a 33v zener rated 1w does not want to see more than 33mA.

as for reverse leakage, its on order of nA, and is basically part of the fet too, . i am not worried about nA leakage for my application.

i have used zeners like D1 in the past on some fairly big inductors and never had the zener die. but i am also leaning to MOV since they can take more hits.
 
Last edited:

AnalogKid

Joined Aug 1, 2013
10,986
I do not see reason to install zener.
A zener will draw always current.
No, it won't.

I've used this method of device protection when placing a diode across the load is not possible. For example, in the above circuit, if the motor voltage is 12 V and the FET Vds is 30 V or 40 V, then a 15 V or 18 V zener across the FET will not draw any static current, and will clamp both positive and negative motor transients. Many power MOSFETs have this diode built-in, but that zener voltage is equal to or greater than the transistor Vds, offering little protection. I used this technique with darlington transistor relay drivers in ambulances. That transistor type does not have a parasitic zener diode built in.

ak
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
The current in the motor at the instant of shut-off is the maximum current that will then diminish as the energy in the magnetic field dissipates.
agree, but since zener is rated above Vcc (Vcc gives us the max Amps through coil) the mag field has to collapse some before building up voltage to zener threshold, thus some energy is already gone.

are you saying the zener should be rated to handle rated coil Amps as part of the design? zeners are rated in Power (watts), 1v x 10amp is way different than 10v x 1amp for silicon, etc
 
Last edited:

crutschow

Joined Mar 14, 2008
34,282
The energy that's used to built up the voltage to the zener value is quite small, just whatever is needed to charge the stray capacitance.

The zener needs to be rated to handle the energy stored in the inductance (its pulse rating).
It also has to handle the total power dissipated (energy time frequency) if the operating solenoid frequency is very high.

Why are you not using a snubber diode across the coil?
Is it that you are concerned about the turn-off time?
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
The energy that's used to built up the voltage to the zener value is quite small, just whatever is needed to charge the stray capacitance.

The zener needs to be rated to handle the energy stored in the inductance (its pulse rating).
It also has to handle the total power dissipated (energy time frequency) if the operating solenoid frequency is very high.

Why are you not using a snubber diode across the coil?
Is it that you are concerned about the turn-off time?
diode to freewheel the energy still falls into same Q. but yes, snubber diode across coil can be done as i have no coil delay issues, in my app the diode (be it zener or other) is there as a fail safe item (my fet turns off slowly by design, but diode needed in rare cases where fet turns off fast).

silicon is rated many ways, continuous Amps (I), and then transient which can be 10x that of I for say 20us. if my coil is 10A do i need a 10A diode? in the case of freewheel diode it will fwd conduct at a very low voltage, therefore most of the stored mag energy will pass through the diode, but with zener shunt the mag field has collapsed some to reach zener threshold, which means less energy in the coil at the time zener goes into full reverse bias. do i have this right ?

maybe i am thinking about this wrong. are we calculating freewheel current using the diode voltage of ~0.6v (or zener reverse voltage) and coil ohms?

so for a 12v 12ohm coil the diode would fwd conduct 0.6v/12ohm = 50mA = 30mW ?
and for a 33v zener its 33v/12ohm = 2.75A = 90W

so i would need a 100W zener vs a 100mW std diode?

so zener allows coil energy to die fast, freewheel not so fast.

i think i need a deeper explanation on the diffs of using zener shunt across the fet vs freewheel diode across the coil.

and here's a hybrid freewheel, this would die a tad faster than diode alone. but as i have seen on the scope, this still requires RC snubber to dampen the ring that comes from the tank circuit that is formed from the components.
 
Last edited:

#12

Joined Nov 30, 2010
18,224
are you saying the zener should be rated to handle rated coil Amps as part of the design?
No. I answered your question. That's all.
As for rating the zener, it has a time constant for heating and a maximum current, both of which I don't know.
I can tell you that energy is 1/2 L I^2 for each time you stop the motor. You might use that to arrive at how much energy you're putting into the zener by figuring stops per time (times 1/2 LI^2).
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
No. I answered your question. That's all.
As for rating the zener, it has a time constant for heating and a maximum current, both of which I don't know.
I can tell you that energy is 1/2 L I^2 for each time you stop the motor. You might use that to arrive at how much energy you're putting into the zener by figuring stops per time (times 1/2 LI^2).
E of the inductor just tells us total energy, but we need a time period to get power, and its power that the diode is rated, but this rating is typically tied to temp, thus we can usually see much higher than rated for very small transient times. if the coil is 12ohm and the zener is 24v, we'll see 2A flow through the zener for a period of time to dump a portion of E until zener reverse bias stops when flyback V drops just below 24v.

we know the relationship that volts*amps = watts
 
Last edited:

crutschow

Joined Mar 14, 2008
34,282
You need to know the inductance of the coil to determine the energy dissipated in the zener.
The resistance of the coil determines the coil current but the inductance determines the stored energy that must be dissipated.
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
You need to know the inductance of the coil to determine the energy dissipated in the zener.
The resistance of the coil determines the coil current but the inductance determines the stored energy that must be dissipated.
true, and i can measure it, but this relates to the time needed to dump that energy, which ultimately gives is a power rating, but transient is like a special zone because silicon can survive higher currents if the transient times go down. thus is why i think we see smaller rated zeners because the transient times are way shorter.


lets look at amps

1A = 1C/s
J = V*(A*t)
1V @ 1A with t=1 = 1J
1V @ 10A with t=0.1 = 1J

same energy yet we had amps being 10x
so in essence, we could dump a lot of charge very fast but for very short time
that said, are the specs based on temp and not really amps ?

could we realistically expect a 1 watt (1V*1A) zener to survive a 1kA pulse for 1ms ?
 
Last edited:

crutschow

Joined Mar 14, 2008
34,282
If the zener has a surge rating, you can likely use that as the maximum solenoid current if the solenoid is operated at a low frequency.
If you look at the 1N4733A data sheet for example, it shows a surge current rating of 890mA.
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
lets use 1A 2H as example, this get's us 1J of energy.
and coil is 10ohm dc
lets say its a 30v TVS

so when coil begins to climb in V it will reach 30v and then allow to cross the zener. 30v/10ohm = 3A
3A @ 30v = 90watt, but we only have 1J of energy to dump, so it will take just 11.1ms to dump 1J @ 33V @3amp

its 1J total but over a very short period, this energy flux is way higher than the spec'd 1watt, its 90x bigger.

can a 1watt zener handle 1J/11.1ms ?

not all that energy makes it into zener since zener will shut down when flyback V makes it below 30v, etc.
 

crutschow

Joined Mar 14, 2008
34,282
The current zener current will never be higher than the inductor current (how can it).
The 30V appears across the zener, not he 10 ohm inductor resistance.

The time to dump the energy is theu 1J / (30V * 1A) = 33.3ms.
All that energy essentially goes into the zener because when the voltage drops below 30V, the current has dropped to zero, giving zero inductive energy left.

Pick out a 1W zener and look at it's surge current rating to see whether it can tolerate it.
For example the 30V, 1W, 1N4751A zener is rated at only 150mA surge, so that's certainly not adequate for this example.
So if the inductance were that high you would need to go to a higher wattage zener or a TVS diode rated for that surge.
 

Thread Starter

DC_Kid

Joined Feb 25, 2008
1,072
so what limits the current when zener goes full reverse conducting? energy w/o resistance means big amps. why would the current be less than the coil current. if clamp voltage is low then we see low amps and longer period to dump the energy, if we use higher clamping voltage we get bigger amps with short period to dump the energy. so why is amps limited to what the coil was at just before current was cut off?

the collapsing field wont stay above 30v the whole time, there is still a short period near end of collapse where the induced voltage falls below 30v, the mag field has diminished, so from there until very end of collapse there is still induced voltage, and some energy that has nowhere to go.
 
Top