Simple mosfet question

A

Andrew

Jan 1, 1970
0
I have a simple circuit where I want to convert a digital input from
0-5V to an inverted signal from 0-12V. ie, 0V into the circuit = 12V
out, and 5V into the circuit = 0V out.

I don't have the means to get a schematic up, but I've tried to draw
it with text below (but it probably won't show up right), so I will
describe the simple circuit and I think it will be clear:

There is a 2N7000 N-ch mosfet transistor, with the source tied to
ground. The gate is tied directly to the input. The drain is tied to
a resistor, 10k, and the other side of the 10k resistor is tied to a
battery, +12V. The output is connected to the drain, and is used to
feed the clock inputs of 2 CMOS ICs. There is a 470pF filtering
capacitor tied from the output to ground.

My question is, can anyone see how this circuit would blow the
transistor? Is this an OK way to hook it up? Should I maybe add
another 10k resistor between drain and output (before the cap) to
better protect the transistor? I realize that doing this would bring
the output voltage down very slighlty, but that's fine if it is needed
to protect the transistor.

I'm asking because I've hooked up the circuit more than a few times,
and it always works at first / for a while, but sometimes the
transistor blows unexpectedly and I don't know why yet. I think it is
because I was damaging the transistor during install (with ESD, I was
not using protection at the time, but I learned my lesson!), but I
want to make sure it isn't actually a circuit problem too.

Sorry that I can't get a schematic up, I don't have a place to upload
files while I'm at work. I hope I have made it clear enough :)


+12V
|
|
/
\ 10k
/
|_______OUTPUT
__|
| |
INPUT _____| |__
|
|
|
GND
 
T

Tim Wescott

Jan 1, 1970
0
I have a simple circuit where I want to convert a digital input from
0-5V to an inverted signal from 0-12V. ie, 0V into the circuit = 12V
out, and 5V into the circuit = 0V out.

I don't have the means to get a schematic up, but I've tried to draw
it with text below (but it probably won't show up right), so I will
describe the simple circuit and I think it will be clear:

There is a 2N7000 N-ch mosfet transistor, with the source tied to
ground. The gate is tied directly to the input. The drain is tied to
a resistor, 10k, and the other side of the 10k resistor is tied to a
battery, +12V. The output is connected to the drain, and is used to
feed the clock inputs of 2 CMOS ICs. There is a 470pF filtering
capacitor tied from the output to ground.

My question is, can anyone see how this circuit would blow the
transistor? Is this an OK way to hook it up? Should I maybe add
another 10k resistor between drain and output (before the cap) to
better protect the transistor? I realize that doing this would bring
the output voltage down very slighlty, but that's fine if it is needed
to protect the transistor.

I'm asking because I've hooked up the circuit more than a few times,
and it always works at first / for a while, but sometimes the
transistor blows unexpectedly and I don't know why yet. I think it is
because I was damaging the transistor during install (with ESD, I was
not using protection at the time, but I learned my lesson!), but I
want to make sure it isn't actually a circuit problem too.

Sorry that I can't get a schematic up, I don't have a place to upload
files while I'm at work. I hope I have made it clear enough :)


+12V
|
|
/
\ 10k
/
|_______OUTPUT
__|
| |
INPUT _____| |__
|
|
|
GND

Make your pictures with a fixed-width font, then we'll all be able to read
them.

Putting a resistor in series with the cap will spare your transistor, but
you don't need to make it 10K. You could probably get away with 200 ohms
or less if I recall the capabilities of that transistor correctly.

If you're driving it from CMOS a small resistor from gate to input would
spare the driver from the capacitive load, but may not be necessary.

--
Tim Wescott
Control systems and communications consulting
http://www.wescottdesign.com

Need to learn how to apply control theory in your embedded system?
"Applied Control Theory for Embedded Systems" by Tim Wescott
Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html
 
F

Fred Bloggs

Jan 1, 1970
0
Andrew said:
I have a simple circuit where I want to convert a digital input from
0-5V to an inverted signal from 0-12V. ie, 0V into the circuit = 12V
out, and 5V into the circuit = 0V out.

I don't have the means to get a schematic up, but I've tried to draw
it with text below (but it probably won't show up right), so I will
describe the simple circuit and I think it will be clear:

There is a 2N7000 N-ch mosfet transistor, with the source tied to
ground. The gate is tied directly to the input. The drain is tied to
a resistor, 10k, and the other side of the 10k resistor is tied to a
battery, +12V. The output is connected to the drain, and is used to
feed the clock inputs of 2 CMOS ICs. There is a 470pF filtering
capacitor tied from the output to ground.

My question is, can anyone see how this circuit would blow the
transistor? Is this an OK way to hook it up? Should I maybe add
another 10k resistor between drain and output (before the cap) to
better protect the transistor? I realize that doing this would bring
the output voltage down very slighlty, but that's fine if it is needed
to protect the transistor.

I'm asking because I've hooked up the circuit more than a few times,
and it always works at first / for a while, but sometimes the
transistor blows unexpectedly and I don't know why yet. I think it is
because I was damaging the transistor during install (with ESD, I was
not using protection at the time, but I learned my lesson!), but I
want to make sure it isn't actually a circuit problem too.

Sorry that I can't get a schematic up, I don't have a place to upload
files while I'm at work. I hope I have made it clear enough :)


+12V
|
|
/
\ 10k
/
|_______OUTPUT
__|
| |
INPUT _____| |__
|
|
|
GND

There's nothing wrong with the circuit, something else is blowing the
FET. There must be a transient occurring on one or more of the three FET
terminals that is causing the damage. The gate terminal is most
susceptible and this is handled by placing a zener in parallel with it.
 
D

default

Jan 1, 1970
0
I have a simple circuit where I want to convert a digital input from
0-5V to an inverted signal from 0-12V. ie, 0V into the circuit = 12V
out, and 5V into the circuit = 0V out.

I don't have the means to get a schematic up, but I've tried to draw
it with text below (but it probably won't show up right), so I will
describe the simple circuit and I think it will be clear:

There is a 2N7000 N-ch mosfet transistor, with the source tied to
ground. The gate is tied directly to the input. The drain is tied to
a resistor, 10k, and the other side of the 10k resistor is tied to a
battery, +12V. The output is connected to the drain, and is used to
feed the clock inputs of 2 CMOS ICs. There is a 470pF filtering
capacitor tied from the output to ground.

My question is, can anyone see how this circuit would blow the
transistor? Is this an OK way to hook it up? Should I maybe add
another 10k resistor between drain and output (before the cap) to
better protect the transistor? I realize that doing this would bring
the output voltage down very slighlty, but that's fine if it is needed
to protect the transistor.

I'm asking because I've hooked up the circuit more than a few times,
and it always works at first / for a while, but sometimes the
transistor blows unexpectedly and I don't know why yet. I think it is
because I was damaging the transistor during install (with ESD, I was
not using protection at the time, but I learned my lesson!), but I
want to make sure it isn't actually a circuit problem too.

Sorry that I can't get a schematic up, I don't have a place to upload
files while I'm at work. I hope I have made it clear enough :)


+12V
|
|
/
\ 10k
/
|_______OUTPUT
__|
| |
INPUT _____| |__
|
|
|
GND

mosfets are very sensitive to gate drive and spurious signals / static
electricity etc.. A small transient in the gate lead can eat a mosfet
where a BJT will shrug it off. You are doing something that kills the
transistor - bread boarding? Keep the gate lead short to minimize
inductance. Protect the gate and source/drain if there is an
inductance in the load. Use a scope and watch the transients.

Look at the data sheet for the 7000 - particularly look at the safe
operating area and you see how fragile they can be.

A good source for the use of mosfets:
http://www.irf.com/technical-info/appnotes/an-937.pdf
 
A

Andrew

Jan 1, 1970
0
mosfets are very sensitive to gate drive and spurious signals / static
electricity etc.. A small transient in the gate lead can eat a mosfet
where a BJT will shrug it off. You are doing something that kills the
transistor - bread boarding? Keep the gate lead short to minimize
inductance. Protect the gate and source/drain if there is an
inductance in the load. Use a scope and watch the transients.

Look at the data sheet for the 7000 - particularly look at the safe
operating area and you see how fragile they can be.

A good source for the use of mosfets:http://www.irf.com/technical-info/appnotes/an-937.pdf
--

- Hide quoted text -

- Show quoted text -

I hope I don't get a double post here, but it has been a bit and I'm
not seeing my message pop up, so I'm thinking I forgot to hit reply!
If it double posts, please forgive me.

Anyway...

I was thinking the same thing, if it isn't ESD, it must be some kind
of noise or transients on the transistor leads, which is why I was
looking for a way to protect the transistor.

I should mention that the input is being driven by a servo motor
driver which I did not design or have the schematic for, or really
know much about the design. I am just intercepting the signal and
using it to drive my circuit.

Is it common practice to use a zener as you mentioned, Fred? I'm not
sure exactly how you meant to hook it up? Did you mean with the
cathode at the gate and anode at ground? Or cathode at the input and
anode at the gate? Is doing it this way superior to using a small
resistor in between the input and gate as Tim suggested?

After protecting the gate with a zener or resistor, I should still
probably look at protecting the drain further with either a small
resistor in line with the cap or a small resistor between the cap and
10k resistor, right?

I think you are right Tim, I could use a small resistor such as 200
ohms. The Ids current capacity of the transistor is 200mA, and even a
200 ohm resistor with the levels of voltage I'm expecting wouldn't be
close to this limit.

Default - the design is actually on a PCB right now. There is a bit
of wire coming from the servo driver to the input of the board (maybe
6 inches of 24awg wire (standard servo wire)) and due to space
constraints, there is about 3 inches of 12mil trace on the PCB from
the input to the gate of the transistor. I hadn't considered the
inductance problem that longer traces could pose.
 
A

Andrew

Jan 1, 1970
0
Sorry, I made a mistake. On the PCB the trace from input to gate is
12mil wide and only about 0.5 inches total length. I was thinking of
a different portion of the circuit, sorry for the confusion.
 
D

default

Jan 1, 1970
0
I should mention that the input is being driven by a servo motor
driver which I did not design or have the schematic for, or really
know much about the design. I am just intercepting the signal and
using it to drive my circuit.

I'd want to know more about that . . . Servo motor or something like
an RC servo? This thing has a motor connected to it? Intercepting
makes it sound like it does. If that's the case, you need look no
further, that is probably the cause.

Motors are notorious for inductive transients - even the brushless
ones can be pretty bad. If the driver is applying PWM or switching
from nothing to all out, the wonder is it works at all.

Gate resistors help but it all depends on the energy and persistence
of the transient. The mosfet gate is a small capacitor (with an
equally fragile dielectric). The resistor in series with the signal
to the gate uses the gate capacity to make a low pass filter.
Hopefully, the transient will develop across the resistor and not
punch through the gate insulator.

In the case of the 2N7000 it is only 20 picofarads - so that isn't
much of a low pass filter.

Talk of motors and intercepting signals (one presumes that the signal
you are intercepting also has wires running about) would lead me to
think you will need something a little more robust than a series gate
resistor.

If you need speed (2N7000 has to follow the signal with minimum phase
shift) something like a bipolar might be better suited or if speed
isn't too important a simple RC low pass filter might be better.
Zeners are good but they have to be fast too - There are
circuit/transient protectors that specify switching times and those
may be the better choice.
 
A

Andrew

Jan 1, 1970
0
I'd want to know more about that . . . Servo motor or something like
an RC servo? This thing has a motor connected to it? Intercepting
makes it sound like it does. If that's the case, you need look no
further, that is probably the cause.

Motors are notorious for inductive transients - even the brushless
ones can be pretty bad. If the driver is applying PWM or switching
from nothing to all out, the wonder is it works at all.

Gate resistors help but it all depends on the energy and persistence
of the transient. The mosfet gate is a small capacitor (with an
equally fragile dielectric). The resistor in series with the signal
to the gate uses the gate capacity to make a low pass filter.
Hopefully, the transient will develop across the resistor and not
punch through the gate insulator.

In the case of the 2N7000 it is only 20 picofarads - so that isn't
much of a low pass filter.

Talk of motors and intercepting signals (one presumes that the signal
you are intercepting also has wires running about) would lead me to
think you will need something a little more robust than a series gate
resistor.

If you need speed (2N7000 has to follow the signal with minimum phase
shift) something like a bipolar might be better suited or if speed
isn't too important a simple RC low pass filter might be better.
Zeners are good but they have to be fast too - There are
circuit/transient protectors that specify switching times and those
may be the better choice.

Oh sorry, I guess my wording was misleading. Yes it is a servo motor
driver that is intended to drive an RC servo motor, but instead, it is
driving my circuit. There is no servo motor.
 
D

default

Jan 1, 1970
0
Oh sorry, I guess my wording was misleading. Yes it is a servo motor
driver that is intended to drive an RC servo motor, but instead, it is
driving my circuit. There is no servo motor.

Well that doesn't sound too bad. RC servo is something like a one
millisecond pulse (if my memory is working) that is pulse position
modulated. 0-4.5 or 5 volt pulse. That shouldn't be a problem and I
wouldn't think 6" of wire would be either.

If you have a scope, look at the signal on the 10X probe range. And
look at the output (drain) too for ringing.
 
N

Nobody

Jan 1, 1970
0
Well that doesn't sound too bad. RC servo is something like a one
millisecond pulse (if my memory is working) that is pulse position
modulated.

The RF modulation is PPM, but the signal between the receiver and
servo is PWM, with a 1500us pulse corresponding to neutral. The pulse
frequency isn't significant.
 
D

default

Jan 1, 1970
0
The RF modulation is PPM, but the signal between the receiver and
servo is PWM, with a 1500us pulse corresponding to neutral. The pulse
frequency isn't significant.

You are right. It has been awhile since I messed with them. Pulse
frequency is around 80 HZ


To control the servo, you command it to a certain angle, measured from
0 degrees. You send it a series of pulses. The ON time of the pulse
indicates the angle to turn to; 1ms = 0 degrees, 2ms=max degrees(about
120) and anything in between gives a proportional output angle. 1.5ms
is generally considered to be the "center". The 1~2ms limit is
manufacturers' recommendations; you can usually use a wider range
around 1.5ms for grater throw. I did not do this, but you can use
pulses of less than 1ms, and more than 2ms for an output angle of
180degrees or more. The limiting factor is the feedback pot and the
mechanical limits built in the servo.

Years ago I built a little joystick pan and tilt from a pair of servos
to aim a tiny camera. I used a quad op amp to generate the pulses for
both servos.
 
A

Andrew

Jan 1, 1970
0
You are right. It has been awhile since I messed with them. Pulse
frequency is around 80 HZ

To control the servo, you command it to a certain angle, measured from
0 degrees. You send it a series of pulses. The ON time of the pulse
indicates the angle to turn to; 1ms = 0 degrees, 2ms=max degrees(about
120) and anything in between gives a proportional output angle. 1.5ms
is generally considered to be the "center". The 1~2ms limit is
manufacturers' recommendations; you can usually use a wider range
around 1.5ms for grater throw. I did not do this, but you can use
pulses of less than 1ms, and more than 2ms for an output angle of
180degrees or more. The limiting factor is the feedback pot and the
mechanical limits built in the servo.

Years ago I built a little joystick pan and tilt from a pair of servos
to aim a tiny camera. I used a quad op amp to generate the pulses for
both servos.

Yeah, you are right about servos and they way they are controlled.
The entire circuit I have is basically taking the servo output and
creating a digital signal from it that throws a relay, so commanding
one direction throws the relay one way, and commanding the other
direction throws the relay the other way. This all works great -
until the transistor blows. I've posted the entire portion related to
the transistor, it is basically just inverting the input and changing
it to 0-12V instead of 0-4.5V.

I'm basically just looking for reasons why the transistor is blowing
and I think you guys are right that there is some sort of transient
noise. I'm going to add a resistor between battery and the 470pF cap,
but I'm not sure if I should add a small resistor between input and
gate, or if I should hook up a zener, or both? I'm also not sure how
the zener would connect, I'm new enough at this sort of thing
(transients, noise protection, etc) that I don't know how the zener
helps in a case like this. I'm assuming it would be cathode at the
gate and anode at the ground, and would limit the input voltage to a
"max" of whatever the zener is rated to? So I would use a zener of
like 4V? (The input signal is 0-4.5V). Is that the idea?
 
A

Andrew

Jan 1, 1970
0
I was reviewing my older schematics and PCB designs, and I just
realized something.

I originally had a 2 layer PCB version of this circuit that worked and
went through rigorous testing great! Unfortunately it was not a great
(PCB) design and was noisy. I had filtering caps here and there that
made it work, but I worried that this noisy design was going to cause
issues later, so I switch to a 4 layer board which used a power and
ground plane. When I switched to this *better* board, I was able to
remove some filtering caps as it was a much better design and was
significantly less noisy.

One of the filtering caps that I removed was a 10nF cap that went from
the gate to the source, to filter the input signal.

Was this filtering capacitor also inadvertently "protecting" the
transistor and that is why that design never failed, but now, the new,
less noisy design that does not have this capacitor is more
susceptible to noise/transients/spikes that could kill the transistor?

Or does the removal of this capacitor theoretically have nothing to do
with it because it was not "protecting" the transistor in any way?

I'm guessing that it has alot to do with it, and the capacitor was
protecting the gate input from transients and/or spikes.

What do you guys think?
 
N

Nobody

Jan 1, 1970
0
Yeah, you are right about servos and they way they are controlled.
The entire circuit I have is basically taking the servo output and
creating a digital signal from it that throws a relay,

You do have a flyback diode on the relay coil, right?
 
A

Andrew

Jan 1, 1970
0
You do have a flyback diode on the relay coil, right?

No, I don't have a flyback diode on the relay, but it is in a
completely different portion of the circuit that I haven't had any
trouble with yet?

I didn't understand the idea of a flyback diode at the time I first
designed this circuit, and I do have a little better understanding of
it now, and how to use it, I just don't fully understand what the
voltage spikes can do if they are not controlled with a flyback
diode. It seems to me that the spike would just be across the relay
coil for a short period of time? I suppose that this voltage spike
could kill the transistor (another 2N7000)? It hasn't happened yet on
any of the prototype boards, but I suppose it is a real possibility.

I think I will add one at any rate. Does it have to be a schottky in
the case of a relay, or will any old diode suffice?

Any any thoughts on the 10nF capacitor from gate to source possibly
protecting the transistor?
 
N

Nobody

Jan 1, 1970
0
No, I don't have a flyback diode on the relay, but it is in a
completely different portion of the circuit that I haven't had any
trouble with yet?

The relay coil isn't connected to the transistor that's blowing?
I didn't understand the idea of a flyback diode at the time I first
designed this circuit, and I do have a little better understanding of
it now, and how to use it, I just don't fully understand what the
voltage spikes can do if they are not controlled with a flyback
diode. It seems to me that the spike would just be across the relay
coil for a short period of time? I suppose that this voltage spike
could kill the transistor (another 2N7000)?

You cannot instantaneously change the current through a coil. If the coil
is in series with a transistor, and you turn the transistor off, the
current will keep flowing through the transistor anyhow. The effect of
"turning off" the transistor is that the voltage drop across it will
increase to whatever figure is necessary to keep the current flowing.

E.g. if you have 1A flowing through the coil and transistor, and you
instantaneously increase the collector-emitter (or drain-source) impedance
to 1MOhm, you would get 1MV across the transistor. In practice, you can't
get "instantaneous" turn-off, but you can still generate very high
voltages; you could easily get 1KV from a relay coil if the switching is
fast enough.
It hasn't happened yet on
any of the prototype boards, but I suppose it is a real possibility.

I think I will add one at any rate. Does it have to be a schottky in
the case of a relay, or will any old diode suffice?

Recovery time won't matter here. It only applies if you switch the relay
off then switch it on again before the coil current has dropped to zero
(which is likely to be in the millisecond range unless you have a really
large coil).
Any any thoughts on the 10nF capacitor from gate to source possibly
protecting the transistor?

It will typically extend the turn-off time, potentially allowing the
inductor's current to drop to zero before the transistor's impedance gets
too high.
 
D

default

Jan 1, 1970
0
I've posted the entire portion related to
the transistor, it is basically just inverting the input and changing
it to 0-12V instead of 0-4.5V.

I like gate transient as the cause but . . .

Looking at the characteristic curves for a 2N7000, you might just be
blowing it because you aren't turning it on with enough voltage.
Don't know what your load resistor is, but with 4.5 volts on the gate,
your drain current has to be relatively low to avoid cooking the
transistor.

Look at the chart that shows "On-Resistance Variation with Gate
Voltage and Drain Current" then calculate the load current you are
using and compare that to the absolute maximum values. Little buggers
are only good for 200 milliamps and 400 milliwatts

The gate voltage has to be high enough to turn on the transistor fully
or you end up generating heat in the linear range. The gate threshold
for a 2N7000 (point where it begins conducting not where it is fully
saturated) is three volts - you only have four point five - that might
not be enough with a batch of marginal spec transistors.
 
A

Andrew

Jan 1, 1970
0
The relay coil isn't connected to the transistor that's blowing?

Nope. There is a relay with a 2N7000 in a different portion of the
circuit. I haven't had any trouble with it, but I'm going to add the
flyback diode for good measure.

The circuit in question is the one described in the first post. The
cliff notes, though, are that the circuit works great for a while, but
after a varying amount of time, for an unknown reason, the transistor
will blow. I'm guessing it is transients or noise of some kind. The
exact same circuit, except WITH a 10nF cap from gate to source, was
tested rigorously in four separate PCBs for many many hours in hot/
cold temperatures and had zero problems.

Since the time varies, I'm wondering if I was just lucky with those 4
boards that they worked so great for so long, or if it is possible
that the 10nF cap from gate to source was acting to block transients
that otherwise would have killed the transistor?

And thanks for the additional help on flyback diodes and the
description of what can happen!

Just one question on that matter..
The effect of "turning off" the transistor is that the voltage drop across
will increase to whatever figure is necessary to keep the current flowing.

This is only for a short period of time though? Or will current flow
through the transistor indefinitely (voltage across the transistor)?
 
N

Nobody

Jan 1, 1970
0
This is only for a short period of time though?
Yes.

Or will current flow
through the transistor indefinitely (voltage across the transistor)?

No; that would give you free energy.

The current will drop at a rate proportional to the voltage across the
coil (minus the voltage drop for the coil's resistance) and inversely
proportional to the coil's inductance. dI/dt = V/L.
 
A

Andrew

Jan 1, 1970
0
Look at the chart that shows "On-Resistance Variation with Gate
Voltage and Drain Current" then calculate the load current you are
using and compare that to the absolute maximum values. Little buggers
are only good for 200 milliamps and 400 milliwatts

My transistor is a fairchild (if it matters) and from what I can tell,
with 4.5V Vgs, I don't have to worry about on resistance until ~
400mA. The load resistor for the circuit that is blowing the
transistor is 10k when the transistor is on. With a max voltage of
around 14V, that is 1.4mA Id, way way way under the worry zone as far
as Rds on, right?

Can anyone comment one way or the other on whether a 10nF cap from
gate to source should have any real effect on protecting the
transistor? I keep thinking yes, but it seems you guys don't agree?
Wouldn't it limit potentially damaging spikes to the gate?
 
Top