MOSFET behave as a switch design problems~

[Electronic Swear]

I would like to use a MOSFET to behave as a switch for
control the input voltage of the circuit.
The MOSFET i am using is IRF530.
The circuit is connected as the below:

+24V -------| |-------------------
|_____^ |
------- -----------
Q1 | | COIL |
| -----------
(+5V)PWM CONTROL |
with 30KHZ |
|---
(+5V)100HZ ----|| Q2
control signal |->-
|
|
|
GROUND

I don't know why the MOSFET of Q1 is on fully turn on when
PWM control is at 5V high mode. I have used a 74HC00 (NAND)
to behave a buffer to output the PWM signal.
Or is any problem wrong in this connection?
Can anyone give me a help on it?
Thank you very much.

 

[Robert Baer]

I would like to use a MOSFET to behave as a switch for
control the input voltage of the circuit.
The MOSFET i am using is IRF530.
The circuit is connected as the below:

+24V -------| |-------------------
|_____^ |
------- -----------
Q1 | | COIL |
| -----------
(+5V)PWM CONTROL |
with 30KHZ |
|---
(+5V)100HZ ----|| Q2
control signal |->-
|
|
|
GROUND

I don't know why the MOSFET of Q1 is on fully turn on when
PWM control is at 5V high mode. I have used a 74HC00 (NAND)
to behave a buffer to output the PWM signal.
Or is any problem wrong in this connection?
Can anyone give me a help on it?
Thank you very much.

Number one, the maximum gate to sourve rating is being exceeded at Q1.
Perhaps that has caused a defect in Q1 to keep it (at least) partly
on.
Furthermore, many FETs have the (inherent) characteristic that the
source and drain are interchangeable.
And Q1 has to be ON in order that Q2 can switch current thru that
inductor.
Then...when Q2 turns off, there is an inductive collapse making the
drain of Q2 rise to perhaps hunderds of volts, which will be clamped by
the avalanche breakdown of Q2 (this part is OK), then when the inductor
current finally reverses, the Q2 drain will go negative and be clamped
by the diode in Q2.
But that means a positive voltage on the "output" side of Q1 which
will be more positive than the +24V supply.....
You have a design problem!

 

[Mac]

I would like to use a MOSFET to behave as a switch for
control the input voltage of the circuit.
The MOSFET i am using is IRF530.
The circuit is connected as the below:

+24V -------| |-------------------
|_____^ |
------- -----------
Q1 | | COIL |
| -----------
(+5V)PWM CONTROL |
with 30KHZ |
|---
(+5V)100HZ ----|| Q2
control signal |->-
|
|
|
GROUND

I don't know why the MOSFET of Q1 is on fully turn on when
PWM control is at 5V high mode. I have used a 74HC00 (NAND)
to behave a buffer to output the PWM signal.
Or is any problem wrong in this connection?
Can anyone give me a help on it?
Thank you very much.

From the design, I wouldn't really expect Q1 to ever turn off.

A MOSFET turns on when the voltage between the gate and the source (Vgs)
passes a certain threshold voltage (Vth). In your circuit, you are
controlling Vg, but the inductor is controlling Vs.

Let's imagine that everything starts out static with currents all equal to
zero. Once the gate of Q2 is elevated to 5V, the source of Q1 will be
pulled toward GROUND. So, if its gate is at 5V, Q1 will turn on, at least
a little, letting current flow in the inductor. Now let's say the gate of
Q1 goes to GROUND. The inductor will react by lowering the voltage at Q1's
source, and raising the voltage at Q2's drain, until current once again
begins to flow through Q1.

What you may not realize is that there is no way to simply shut-off the
current in an inductor suddenly. You can allow the current to redirect
itself somewhere "safe" if you wish. For example in your case, you could
add a diode to the source of Q1, so that when the node gets below ground
potential, the diode begins to conduct. Since I don't know what you are
trying to do, I don't know if that is acceptable to you. It won't stop Q2
from avalanching, but it should allow Q1 to turn off when its gate is at
GROUND potential.

HTH

Mac
--

 

[Ted Wilson]

I would like to use a MOSFET to behave as a switch for
control the input voltage of the circuit.
The MOSFET i am using is IRF530.
The circuit is connected as the below:

+24V -------| |-------------------
|_____^ |
------- -----------
Q1 | | COIL |
| -----------
(+5V)PWM CONTROL |
with 30KHZ |
|---
(+5V)100HZ ----|| Q2
control signal |->-
|
|
|
GROUND

I don't know why the MOSFET of Q1 is on fully turn on when
PWM control is at 5V high mode. I have used a 74HC00 (NAND)
to behave a buffer to output the PWM signal.
Or is any problem wrong in this connection?
Can anyone give me a help on it?
Thank you very much.

Don't entirely agree with responses received so far.

Problem with Q1 is that it's being used as a source-follower.

The gate threshold voltage for the IRF530 is between 2 and 4V, for a
drain/source current of only 250uA and, (ssuming for the moment that
Q2 is on), if you take Q1's gate to +5V, the source will therefore be
significantly below 5V, in order to for Q1 to be able to drive current
through the coil. The coil is never going to see anything like the
24V of the supply rail.

When you try to turn Q1 off, the negative di/dt in the coil will cause
the voltage on Q1's source to swing -ve, until it biases Q1 back on,
sufficiently to sustain coil current. Current will then decay at a
rate determined by the coil inductance and the threshold voltage of Q1
- not a very well defined state of affairs.

It's not too clear from your circuit just what you're trying to
achieve here, but I assume that you are trying to switch the coil at
100Hz using Q2, while applying PWM control to coil current by means of
Q1, in which case you are basically trying to AND the function of the
two MOSFETs. If this is what you're after, this would be much better
done by doing the AND function in logic, prior to the output stage,
doing away with Q1, and driving Q2 with the AND function.

Also, setting aside the fact that the IRF530 is not a logic level
device and therefore may well need more than the 5V available in your
circuit to saturate Q2, you can't drive power MOSFETs properly in
switching applications using the direct output of 74HC logic. You
need significantly better current drive capability to charge/discharge
the MOSFET's input capacitance fast enough to keep switching losses
down to reasonable proportions - you can achieve this either by
discrete buffering, or by using one of the many proprietary drivers
available.

Finaly, unless you plan to utilise the repetative avalanche rating of
the IRF530 to deal with inductive fly-back, (in which case you need to
study and understand the curves in the spec in terms of your
requirement, to check the device can take it), you need to add some
provision such as a reverse-recovery diode to deal with it.

Regards

Ted Wilson

 

[Fred Bloggs]

I would like to use a MOSFET to behave as a switch for
control the input voltage of the circuit.
The MOSFET i am using is IRF530.
The circuit is connected as the below:

+24V -------| |-------------------
|_____^ |
------- -----------
Q1 | | COIL |
| -----------
(+5V)PWM CONTROL |
with 30KHZ |
|---
(+5V)100HZ ----|| Q2
control signal |->-
|
|
|
GROUND

I don't know why the MOSFET of Q1 is on fully turn on when
PWM control is at 5V high mode.

There is a substrate-source diode like shown- so you have no control
over Q1:
Please view in a fixed-width font such as
Courier.



+-|>|-+

 

[Winfield Hill]

Fred Bloggs wrote...

There is a substrate-source diode like shown- so you have
no control over Q1:

+-|>|-+
| |
+24V -------| |-------------------
|_____^ |
------- -----------
Q1 | | COIL |
| -----------
(+5V)PWM CONTROL |
with 30KHZ |
|---
(+5V)100HZ ----|| Q2
control signal |->-
|
|
GROUND

You've got the diode in the wrong direction; it's not a problem.

His problem has got to be a failure to drive the Q1 n-channel
device correctly (a flying high-side drive is necessary - he's
got a source follower now), in addition to inadequate gate drive
for Q2, both as Ted Wilson mentioned.

He could use an IR2109, IR2110 or many other suitable drivers.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Ban]

Winfield Hill wrote:
|| Fred Bloggs wrote...
|||
||| Electronic Swear wrote:
|||| I would like to use a MOSFET to behave as a switch for
|||| control the input voltage of the circuit. The MOSFET i am
|||| using is IRF530. The circuit is connected as the below:
||||
|||| +24V -------| |-------------------
|||| |_____^ |
|||| ------- -----------
|||| Q1 | | COIL |
|||| | -----------
|||| (+5V)PWM CONTROL |
|||| with 30KHZ |
|||| |---
|||| (+5V)100HZ ----|| Q2
|||| control signal |->-
|||| |
|||| |
|||| GROUND
||||
|||| I don't know why the MOSFET of Q1 is on fully turn on when
|||| PWM control is at 5V high mode.
|||
||
|| His problem has got to be a failure to drive the Q1 n-channel
|| device correctly (a flying high-side drive is necessary - he's
|| got a source follower now), in addition to inadequate gate drive
|| for Q2, both as Ted Wilson mentioned.
||
|| He could use an IR2109, IR2110 or many other suitable drivers.
||
|| Thanks,
|| - Win
||
|| whill_at_picovolt-dot-com

Actually Q1 can be dropped completely and the 100Hz control signal and 30kHz
PWM combined with an And-gate controlling a single low side switch(Q2). Or
is there anything preventing it?

ciao Ban

 

[Fred Bloggs]

Fred Bloggs wrote...



You've got the diode in the wrong direction; it's not a problem.

His problem has got to be a failure to drive the Q1 n-channel
device correctly (a flying high-side drive is necessary - he's
got a source follower now), in addition to inadequate gate drive
for Q2, both as Ted Wilson mentioned.

He could use an IR2109, IR2110 or many other suitable drivers.

Thanks,
- Win

whill_at_picovolt-dot-com

It looks like the schematic for a P-channel with source on the coil to
me. As Ban pointed out, the OP needs to learn about superposition and
reduce the circuit to a single "logic level" n-channel.
This stuff about playing with circuit topologies without knowing
anything about circuit theory, and playing with components without
knowing anything about their operation, reminds me of reading a book in
back to front order.

 

[Electronic Swear]

I would like to know if I change my 74 series NAND gate to
CMOS 4011 to be my logic part and those output will connect
with the MOS transistor IRF530, will the result be better,
just the figure shown as below:


+ 24V ------------------
|
-----------
| COIL |
-----------
|
|
|---
(+15V)100HZ----|| Q2
control signal |->-
(OUTPUT FROM 4011) |
|
|
GROUND


Also, if I change the using +15V instead of +5V for IC's Vcc,
will it be better for trigger MOSFET?
Thank you very much.

 

[Ban]

Electronic Swear wrote:
|| I would like to know if I change my 74 series NAND gate to
|| CMOS 4011 to be my logic part and those output will connect
|| with the MOS transistor IRF530, will the result be better,
|| just the figure shown as below:
||
||
|| + 24V -----------------+--------+
|| | |
|| ----------- |
|| | COIL | ---
|| ----------- / \1N4001
|| | ---
|| | |
|| |--+--------+
|| (+15V)100HZ----|| Q2
|| control signal |->-
|| (OUTPUT FROM 4011) |
|| |
|| |
|| GROUND
||
||
|| Also, if I change the using +15V instead of +5V for IC's Vcc,
|| will it be better for trigger MOSFET?
|| Thank you very much.
||

Yes using the +15V with CMOS logic would be better, you do not need to use a
low gate voltage Mosfet, but can use a cheaper normal one with less gate
capacitance. Do not forget the freewheeling Diode I showed in the
schematics, depending on the coil current a larger model might be needed,
preferably a schottky diode. Put also a small 10-33 Ohms resistor into the
gate line, to prevent oscillations of the driver.

ciao Ban

 

[Kevin McMurtrie]

Electronic Swear wrote:
|| I would like to know if I change my 74 series NAND gate to
|| CMOS 4011 to be my logic part and those output will connect
|| with the MOS transistor IRF530, will the result be better,
|| just the figure shown as below:
||
||
|| + 24V -----------------+--------+
|| | |
|| ----------- |
|| | COIL | ---
|| ----------- / \1N4001
|| | ---
|| | |
|| |--+--------+
|| (+15V)100HZ----|| Q2
|| control signal |->-
|| (OUTPUT FROM 4011) |
|| |
|| |
|| GROUND
||
||
|| Also, if I change the using +15V instead of +5V for IC's Vcc,
|| will it be better for trigger MOSFET?
|| Thank you very much.
||

Yes using the +15V with CMOS logic would be better, you do not need to use a
low gate voltage Mosfet, but can use a cheaper normal one with less gate
capacitance. Do not forget the freewheeling Diode I showed in the
schematics, depending on the coil current a larger model might be needed,
preferably a schottky diode. Put also a small 10-33 Ohms resistor into the
gate line, to prevent oscillations of the driver.

ciao Ban

A diode between the gate and ground is good too. It keeps the gate from
going too far negative, and blowing the driver, when the coil's voltage
springs down.

 

[Mac]

Electronic Swear wrote:
|| I would like to know if I change my 74 series NAND gate to
|| CMOS 4011 to be my logic part and those output will connect
|| with the MOS transistor IRF530, will the result be better,
|| just the figure shown as below:
||
||
|| + 24V -----------------+--------+
|| | |
|| ----------- |
|| | COIL | ---
|| ----------- / \1N4001
|| | ---
|| | |
|| |--+--------+
|| (+15V)100HZ----|| Q2
|| control signal |->-
|| (OUTPUT FROM 4011) |
|| |
|| |
|| GROUND
||
||
|| Also, if I change the using +15V instead of +5V for IC's Vcc,
|| will it be better for trigger MOSFET?
|| Thank you very much.
||

Yes using the +15V with CMOS logic would be better, you do not need to use a
low gate voltage Mosfet, but can use a cheaper normal one with less gate
capacitance. Do not forget the freewheeling Diode I showed in the
schematics, depending on the coil current a larger model might be needed,
preferably a schottky diode. Put also a small 10-33 Ohms resistor into the
gate line, to prevent oscillations of the driver.

ciao Ban

Ban,

I'm not sure, but I think the OP might be trying to deliberately cause Q2
to go into avalanche. If this is the case, he/she might not want the
free-wheeling diode. I say this because that particular FET is rated for
repeated avalanche.

Mac
--

 

[Mac]

A diode between the gate and ground is good too. It keeps the gate from
going too far negative, and blowing the driver, when the coil's voltage
springs down.

You must be looking at the circuit funny. Or I am. I don't see any need
whatsoever for a diode from ground to gate. You seem to thing that the
drain will have a tendency to go below ground under some set of
circumstances. I don't see this. I see the opposite, in fact.

On top of that, there is already a diode (not shown) in the MOSFET from
source to drain. So the CMOS logic will hold the gate somewhere between 0
and 15 volts, the source is connected to 0 volts, and the drain, if the
free-wheeler is connected, will always be somewhere between 24V + 1diode
and 0 volts.

Without the free-wheeling diode, when the FET is off, the drain voltage
will rise rapidly until the transistor enters avalanche mode. I think this
is what the OP is trying for.

If I have it wrong, please set me straight.

Mac
--

 

[Kevin McMurtrie]

You must be looking at the circuit funny. Or I am. I don't see any need
whatsoever for a diode from ground to gate. You seem to thing that the
drain will have a tendency to go below ground under some set of
circumstances. I don't see this. I see the opposite, in fact.

On top of that, there is already a diode (not shown) in the MOSFET from
source to drain. So the CMOS logic will hold the gate somewhere between 0
and 15 volts, the source is connected to 0 volts, and the drain, if the
free-wheeler is connected, will always be somewhere between 24V + 1diode
and 0 volts.

Without the free-wheeling diode, when the FET is off, the drain voltage
will rise rapidly until the transistor enters avalanche mode. I think this
is what the OP is trying for.

If I have it wrong, please set me straight.

Mac

The problem is when the gate's voltage is brought to zero volts. The
coil's voltage can ring negative fast enough that capacitive coupling
pushes the gate below 0V. The chip driving the MOSFET might not like
that. It damages some CMOS 555 timer variants.

Another solution is to use a large gate resistor so there's not enough
switching speed to induce ringing in the coil.

Ban mentioned using a schottky diode. Isn't a molasses-slow rectifier
better for dampening?

 

[Mac]

The problem is when the gate's voltage is brought to zero volts. The
coil's voltage can ring negative fast enough that capacitive coupling
pushes the gate below 0V. The chip driving the MOSFET might not like
that. It damages some CMOS 555 timer variants.

What do you mean by ring negative? When the gate goes to zero, Q2 will try
to shut off, and the inductor will react by raising the voltage at the
drain of Q2 relative to its other terminal. It will do this quite quickly
until current begins to flow again, either through Ban's free-wheeling
diode, or by avalanche breakdown in Q2.

Another solution is to use a large gate resistor so there's not enough
switching speed to induce ringing in the coil.

The coil is not going to ring. I don't know where you are getting this
idea. However, you are correct that the gate will probably have overshoot
and undershoot if it is driven agressively. A gate resistor can keep this
overshoot and undershoot to a safe level.

Ban mentioned using a schottky diode. Isn't a molasses-slow rectifier
better for dampening?

Dampening what? The only thing that needs dampening is the gate of Q2, and
a resistor is the way to go for that.

Also, and as I already said, there is already a diode (not shown) from the
source of Q2 to the drain. So even if the drain node did try to go
negative, it wouldn't get very far.

Mac
--

 

[Mac]

I would like to know if I change my 74 series NAND gate to
CMOS 4011 to be my logic part and those output will connect
with the MOS transistor IRF530, will the result be better,
just the figure shown as below:


+ 24V ------------------
|
-----------
| COIL |
-----------
|
|
|---
(+15V)100HZ----|| Q2
control signal |->-
(OUTPUT FROM 4011) |
|
|
GROUND


Also, if I change the using +15V instead of +5V for IC's Vcc,
will it be better for trigger MOSFET?
Thank you very much.

[older message snipped]

I don't think 4011 logic is what you want. I've never heard of it before,
but it looks like it is some kind of high voltage CMOS. The problem with
it is that if you power it from 15 volts, you will need to drive the
inputs much higher than 5 Volts to gaurantee that the 4011 logic will see
it as high. This will force you to design your entire circuit in 4011,
including any inputs coming from off-board.

Instead, why don't you use some kind of normal logic followed by one of
the chips which Win recommended, the IR2109 or IR2110? These chips are
specifically designed to drive N-channel MOSFET's. They have two driver
outputs, a high-side and a low side, but you can just use one.

Also, are you trying to get the transistor into avalanche mode, or not? If
you are, the circuit looks pretty good once you follow my suggestions
above. If you don't want the transistor to avalanche, then put a
free-wheeling diode from the drain of Q2 to the 24 volt power supply, as
Ban suggested. You will probably need a resistor in series with the gate
of Q2, but read the datasheet for the MOSFET driver.

good luck.

Mac
--

 

[HermanMunster]

For anyone that doesn't know what a 4011 is, the 4011 is THE most common
cmos chip ever made. It is fine for driving this mosfet. put the free
wheeling diode back across the coil like it was three or four posts ago. Use
a simple 1n4001. It will be fine. You can add a 100 ohm resistor between the
gate and the 4011 if you want, but I wouldn't worry about it becasue the
4011 isn't a overlly strong driver, and any oscillations will settle out far
before the coil ever actuates. There is no need to get any fancier than what
you have here.

If the 4011 input voltages are a problem, and you currently have a TTL
circuit inplace, than use a logic level mosfet, and be done with it.

good luck.

 

[claudio lorini]

I would like to know if I change my 74 series NAND gate to

CMOS 4011 to be my logic part and those output will connect
with the MOS transistor IRF530, will the result be better,
just the figure shown as below:

Look for Appl.Note AN937 and AN936 in the IR website www.irf.com.
Ciao, Claudio.