DC motor pwm, mosfet barbecue, etc.

Hi,

I built a pwm circuit (200 Hz switching frequency) using 12V as a supply, also using around 12V to the gate of an IRF530 mosfet. I tested it with an automotive turn signal light (about 1 amp) and it worked just fine (without a heat sink on the mosfet).

I then tried using it to control a 12V tire inflator/compressor (about 4 amps). The inflator operated fine, but then when I felt the mosfet I was surprised it was so hot it burned my finger.

Yes I know I should use a heat sink, but I've also heard I should use a protection diode. How do I size this? Won't the power be dissipated (wasted) in the diode anyway, transferring the heat from the motor to the diode?

Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?

If I increase the frequency, won't the voltage spikes from the motor coils be even higher?

Thanks,

Michael

 

[Tim Wescott]

On Thu, 14 Feb 2013 09:16:05 -0800, mrdarrett wrote:

On Thursday, February 14, 2013 9:14:44 AM UTC-8, mrda...@gmail.com
wrote:
Hi,



I built a pwm circuit (200 Hz switching frequency) using 12V as a
supply, also using around 12V to the gate of an IRF530 mosfet. I
tested it with an automotive turn signal light (about 1 amp) and it
worked just fine (without a heat sink on the mosfet).



I then tried using it to control a 12V tire inflator/compressor (about
4 amps). The inflator operated fine, but then when I felt the mosfet I
was surprised it was so hot it burned my finger.



Yes I know I should use a heat sink, but I've also heard I should use a
protection diode. How do I size this? Won't the power be dissipated
(wasted) in the diode anyway, transferring the heat from the motor to
the diode?



Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?



If I increase the frequency, won't the voltage spikes from the motor
coils be even higher?



Thanks,



Michael


Correction: "Won't the power be dissipated (wasted) in the diode
anyway, transferring the heat from the MOSFET to the diode?"

What is happening is that when the MOSFET turns off, the motor coils
cause the motor voltage to go high enough to force the MOSFET into
breakdown; the MOSFET conducts the full (but rapidly decaying) motor
current at this elevated voltage, which burns a lot of power.

If you use a diode as pictured below, then it will shunt the motor
current away from the MOSFET while only dropping a volt or so (more or
less, depending on the diode you choose). Not only does the MOSFET not
have its ratings exceeded in every cycle, but the overall power
dissipation is less because the 1V across the diode is much less than the
40V or whatever you'll see across the MOSFET.

200Hz is possibly _a_ correct frequency to drive a 4A, 12VDC motor. If
anything it may be low: unless you're specifically trying for it, you
want to choose a switching frequency that's high enough that the motor
ripple current is a moderate fraction (5-20%? someone will disagree with
me) of the motor's rated current. The motor itself certainly allows a
higher frequency drive, but depending on what you're using to drive the
MOSFET gate your circuit may limit things.

I'm currently putting the finishing touches on a motor control board with
a 24V, 10A motor that switches at 10kHz, if that gives you a notion of
switching speeds.

+12V
====
|
o----------------o
|
-
diode ----> ^ to motor
|
|
o----------------o
|
|
gate drive ||-+
||<- MOSFET
o----------------||-+
|
|
===
GND
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

 

On Thursday, February 14, 2013 9:14:44 AM UTC-8, mrda...@gmail.com wrote:

Hi,



I built a pwm circuit (200 Hz switching frequency) using 12V as a supply, also using around 12V to the gate of an IRF530 mosfet. I tested it with an automotive turn signal light (about 1 amp) and it worked just fine (without a heat sink on the mosfet).



I then tried using it to control a 12V tire inflator/compressor (about 4 amps). The inflator operated fine, but then when I felt the mosfet I was surprised it was so hot it burned my finger.



Yes I know I should use a heat sink, but I've also heard I should use a protection diode. How do I size this? Won't the power be dissipated (wasted) in the diode anyway, transferring the heat from the motor to the diode?



Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?



If I increase the frequency, won't the voltage spikes from the motor coils be even higher?



Thanks,



Michael

Correction: "Won't the power be dissipated (wasted) in the diode anyway, transferring the heat from the MOSFET to the diode?"

 

[Tim Wescott]

On Thu, 14 Feb 2013 10:59:40 -0800, mrdarrett wrote:

On Thursday, February 14, 2013 10:00:29 AM UTC-8, Tim Wescott wrote:


(snip)

What is happening is that when the MOSFET turns off, the motor coils

cause the motor voltage to go high enough to force the MOSFET into

breakdown; the MOSFET conducts the full (but rapidly decaying) motor

current at this elevated voltage, which burns a lot of power.



If you use a diode as pictured below, then it will shunt the motor

current away from the MOSFET while only dropping a volt or so (more or

less, depending on the diode you choose). Not only does the MOSFET not

have its ratings exceeded in every cycle, but the overall power

dissipation is less because the 1V across the diode is much less than
the

40V or whatever you'll see across the MOSFET.



200Hz is possibly _a_ correct frequency to drive a 4A, 12VDC motor. If

anything it may be low: unless you're specifically trying for it, you

want to choose a switching frequency that's high enough that the motor

ripple current is a moderate fraction (5-20%? someone will disagree
with

me) of the motor's rated current. The motor itself certainly allows a

higher frequency drive, but depending on what you're using to drive the

MOSFET gate your circuit may limit things.



I'm currently putting the finishing touches on a motor control board
with

a 24V, 10A motor that switches at 10kHz, if that gives you a notion of

switching speeds.



+12V

====

|

o----------------o

|

-

diode ----> ^ to motor

|

|

o----------------o

|

|

gate drive ||-+

||<- MOSFET

o----------------||-+

|

|

===

GND

(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)



--

My liberal friends think I'm a conservative kook.

My conservative friends think I'm a liberal kook.

Why am I not happy that they have found common ground?



Tim Wescott, Communications, Control, Circuits & Software

http://www.wescottdesign.com


Ok, thanks, Tim.

Should I seek a 4-amp diode, since I'm driving a 4-amp motor? What
about the voltage? Would a 1N4001 do?

That's complicated. If you were doing more than a one-off you'd want to
calculate the average current through the diode, which is going to be
equal to the actual motor current (not rated) times the diode's duty
cycle (which is, at worst, the compliment of the MOSFET duty cycle; i.e.
for a 30% MOSFET duty cycle the diode duty cycle would be 70%).

The easy answer is, yes, find a 4A diode with a voltage rating better
than 12V. Find a Schottky, because it'll have a lower forward drop, and
it'll be guaranteed to be fast enough.

The 1N4001 may be enough even though it's not a 4A diode, because at low
MOSFET duty cycles your pump will (I suspect) not be drawing full current
and thus not driving lots of current through the diode, while at high
duty cycles the diode won't be seeing current for a very big proportion
of total time.

If you want to PWM significantly faster than 200Hz the 4001 is probably
too slow -- you'd want to find a Schottky diode or a fast recovery
diode. But I don't have the 4001 data sheet imprinted on my brain, so I
can't guarantee the truth of that.

It's all for a chemistry experiment, really. Just trying to grow some
algae to make biodiesel, and shove some air (carbon dioxide) into the
algae tank.

Well, now it's an electronics experiment, too!

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
Why am I not happy that they have found common ground?

Tim Wescott, Communications, Control, Circuits & Software
http://www.wescottdesign.com

 

On Thursday, February 14, 2013 10:00:29 AM UTC-8, Tim Wescott wrote:


(snip)

What is happening is that when the MOSFET turns off, the motor coils

cause the motor voltage to go high enough to force the MOSFET into

breakdown; the MOSFET conducts the full (but rapidly decaying) motor

current at this elevated voltage, which burns a lot of power.



If you use a diode as pictured below, then it will shunt the motor

current away from the MOSFET while only dropping a volt or so (more or

less, depending on the diode you choose). Not only does the MOSFET not

have its ratings exceeded in every cycle, but the overall power

dissipation is less because the 1V across the diode is much less than the

40V or whatever you'll see across the MOSFET.



200Hz is possibly _a_ correct frequency to drive a 4A, 12VDC motor. If

anything it may be low: unless you're specifically trying for it, you

want to choose a switching frequency that's high enough that the motor

ripple current is a moderate fraction (5-20%? someone will disagree with

me) of the motor's rated current. The motor itself certainly allows a

higher frequency drive, but depending on what you're using to drive the

MOSFET gate your circuit may limit things.



I'm currently putting the finishing touches on a motor control board with

a 24V, 10A motor that switches at 10kHz, if that gives you a notion of

switching speeds.



+12V

====

|

o----------------o

|

-

diode ----> ^ to motor

|

|

o----------------o

|

|

gate drive ||-+

||<- MOSFET

o----------------||-+

|

|

===

GND

(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)



--

My liberal friends think I'm a conservative kook.

My conservative friends think I'm a liberal kook.

Why am I not happy that they have found common ground?



Tim Wescott, Communications, Control, Circuits & Software

http://www.wescottdesign.com

Ok, thanks, Tim.

Should I seek a 4-amp diode, since I'm driving a 4-amp motor? What about the voltage? Would a 1N4001 do?

It's all for a chemistry experiment, really. Just trying to grow some algae to make biodiesel, and shove some air (carbon dioxide) into the algae tank.

Thanks,

Michael

 

On Thursday, February 14, 2013 11:34:42 AM UTC-8, Tim Wescott wrote:

(snip)

Ok, thanks, Tim.



Should I seek a 4-amp diode, since I'm driving a 4-amp motor? What

about the voltage? Would a 1N4001 do?



That's complicated. If you were doing more than a one-off you'd want to

calculate the average current through the diode, which is going to be

equal to the actual motor current (not rated) times the diode's duty

cycle (which is, at worst, the compliment of the MOSFET duty cycle; i.e.

for a 30% MOSFET duty cycle the diode duty cycle would be 70%).



The easy answer is, yes, find a 4A diode with a voltage rating better

than 12V. Find a Schottky, because it'll have a lower forward drop, and

it'll be guaranteed to be fast enough.



The 1N4001 may be enough even though it's not a 4A diode, because at low

MOSFET duty cycles your pump will (I suspect) not be drawing full current

and thus not driving lots of current through the diode, while at high

duty cycles the diode won't be seeing current for a very big proportion

of total time.



If you want to PWM significantly faster than 200Hz the 4001 is probably

too slow -- you'd want to find a Schottky diode or a fast recovery

diode. But I don't have the 4001 data sheet imprinted on my brain, so I

can't guarantee the truth of that.



It's all for a chemistry experiment, really. Just trying to grow some

algae to make biodiesel, and shove some air (carbon dioxide) into the

algae tank.



Well, now it's an electronics experiment, too!

Yup, sure is, thanks! When the chemistry gets boring that's when I start digging into the circuits.

Thanks for the tips!

Michael

 

[Jamie]

mrdarrett@gmail.com wrote:

Hi,

I built a pwm circuit (200 Hz switching frequency) using 12V as a supply, also using around 12V to the gate of an IRF530 mosfet. I tested it with an automotive turn signal light (about 1 amp) and it worked just fine (without a heat sink on the mosfet).

I then tried using it to control a 12V tire inflator/compressor (about 4 amps). The inflator operated fine, but then when I felt the mosfet I was surprised it was so hot it burned my finger.

Yes I know I should use a heat sink, but I've also heard I should use a protection diode. How do I size this? Won't the power be dissipated (wasted) in the diode anyway, transferring the heat from the motor to the diode?

Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?

If I increase the frequency, won't the voltage spikes from the motor coils be even higher?

Thanks,

Michael

At that freq, you could be saturating the coil and your apparent
speed control that you may think you have is more mechanical than
electrical.

It is common to use ~8kHz

with out knowing more about the motor you're driving, you would
need to monitor heat output of the motor's core to determine if
you are reaching too far for the PWM carrier frequency.

Too many hz will cause heating and efficiency losses.

Jamie

 

[Phil Allison]

<mrdarrett@gmail.com>

I built a pwm circuit (200 Hz switching frequency) using 12V as a supply,
also using around 12V to the gate of an IRF530 mosfet. I tested it with
an automotive turn signal light (about 1 amp) and it worked just fine
(without a heat sink on the mosfet).

I then tried using it to control a 12V tire inflator/compressor (about 4
amps).

** It's probably more than that.


The inflator operated fine, but then when I felt the mosfet I was surprised
it was so hot it burned my finger.

** Really ?

Did you estimate the power dissipation using I squared R ?

Yes I know I should use a heat sink, but I've also heard I should use a
protection diode.

** The correct name is "freewheeling diode" - it greatly increases the
efficiency of the drive.

The diode needs to be rated for at least half the max motor current.

Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?

** No.

The best frequency is about 2kHz.

Try a pair of 3A diodes ( 1N4502 ) in parallel.



..... Phil

 

[Pimpom]

<mrdarrett@gmail.com> wrote in message
news:966ee6b7-0d91-4bd4-a0ad-465d41092cf5@googlegroups.com...

On Thursday, February 14, 2013 11:34:42 AM UTC-8, Tim
Wescott wrote:

(snip)

Ok, thanks, Tim.



Should I seek a 4-amp diode, since I'm driving a 4-amp
motor? What

about the voltage? Would a 1N4001 do?



That's complicated. If you were doing more than a
one-off you'd want to

calculate the average current through the diode, which is
going to be

equal to the actual motor current (not rated) times the
diode's duty

cycle (which is, at worst, the compliment of the MOSFET
duty cycle; i.e.

for a 30% MOSFET duty cycle the diode duty cycle would be
70%).



The easy answer is, yes, find a 4A diode with a voltage
rating better

than 12V. Find a Schottky, because it'll have a lower
forward drop, and

it'll be guaranteed to be fast enough.



The 1N4001 may be enough even though it's not a 4A diode,
because at low

MOSFET duty cycles your pump will (I suspect) not be
drawing full current

and thus not driving lots of current through the diode,
while at high

duty cycles the diode won't be seeing current for a very
big proportion

of total time.



If you want to PWM significantly faster than 200Hz the
4001 is probably

too slow -- you'd want to find a Schottky diode or a fast
recovery

diode. But I don't have the 4001 data sheet imprinted on
my brain, so I

can't guarantee the truth of that.



It's all for a chemistry experiment, really. Just
trying to grow some

algae to make biodiesel, and shove some air (carbon
dioxide) into the

algae tank.



Well, now it's an electronics experiment, too!


Yup, sure is, thanks! When the chemistry gets boring
that's when I start digging into the circuits.

Thanks for the tips!

Michael

Cheap and common diodes -
BY399 (3A 100V fast recovery)
1N5822 (3A 40V Schottky)

Some time ago, I made a speed control for a motor of unknown
characteristics except that it's rated for 24V 3A. I used an
IRF540 at ~25kHz to avoid the possibility of an audible
whine. It was part of a more complex whole and everything
worked fine. The MOSFET was mounted on a plain 1.5mm
aluminium sheet of approx 15 sq.in. and it barely got warm
at full load. I used the BY399 because I didn't have a
Schottky then.

One factor which partly determines transistor dissipation is
the gate drive current available, which becomes more
significant as you increase the frequency. With a weak gate
drive, it takes longer to charge up the gate and fully
switch the MOSFET on. At higher frequencies, that transition
period becomes an appreciable fraction of the switching
cycle. During that period, the transistor spends some time
between fully on and fully off states and that's when it
dissipates the most power. I used about 0.3A drive current
in my project. High-power high-frequency drives use amperes.

 

[Pimpom]

"Pimpom" <Pimpom@invalid.invalid> wrote in message
news:kfkmc5$pqd$1@news.albasani.net...

mrdarrett@gmail.com> wrote in message
news:966ee6b7-0d91-4bd4-a0ad-465d41092cf5@googlegroups.com...
On Thursday, February 14, 2013 11:34:42 AM UTC-8, Tim
Wescott wrote:

(snip)

Ok, thanks, Tim.



Should I seek a 4-amp diode, since I'm driving a 4-amp
motor? What

about the voltage? Would a 1N4001 do?



That's complicated. If you were doing more than a
one-off you'd want to

calculate the average current through the diode, which
is going to be

equal to the actual motor current (not rated) times the
diode's duty

cycle (which is, at worst, the compliment of the MOSFET
duty cycle; i.e.

for a 30% MOSFET duty cycle the diode duty cycle would
be 70%).



The easy answer is, yes, find a 4A diode with a voltage
rating better

than 12V. Find a Schottky, because it'll have a lower
forward drop, and

it'll be guaranteed to be fast enough.



The 1N4001 may be enough even though it's not a 4A
diode, because at low

MOSFET duty cycles your pump will (I suspect) not be
drawing full current

and thus not driving lots of current through the diode,
while at high

duty cycles the diode won't be seeing current for a very
big proportion

of total time.



If you want to PWM significantly faster than 200Hz the
4001 is probably

too slow -- you'd want to find a Schottky diode or a
fast recovery

diode. But I don't have the 4001 data sheet imprinted
on my brain, so I

can't guarantee the truth of that.



It's all for a chemistry experiment, really. Just
trying to grow some

algae to make biodiesel, and shove some air (carbon
dioxide) into the

algae tank.



Well, now it's an electronics experiment, too!


Yup, sure is, thanks! When the chemistry gets boring
that's when I start digging into the circuits.

Thanks for the tips!

Michael

Cheap and common diodes -
BY399 (3A 100V fast recovery)
1N5822 (3A 40V Schottky)

Some time ago, I made a speed control for a motor of
unknown characteristics except that it's rated for 24V 3A.
I used an IRF540 at ~25kHz to avoid the possibility of an
audible whine. It was part of a more complex whole and
everything worked fine. The MOSFET was mounted on a plain
1.5mm aluminium sheet of approx 15 sq.in. and it barely
got warm at full load. I used the BY399 because I didn't
have a Schottky then.

One factor which partly determines transistor dissipation
is the gate drive current available, which becomes more
significant as you increase the frequency. With a weak
gate drive, it takes longer to charge up the gate and
fully switch the MOSFET on. At higher frequencies, that
transition period becomes an appreciable fraction of the
switching cycle. During that period, the transistor spends
some time between fully on and fully off states and that's
when it dissipates the most power. I used about 0.3A drive
current in my project. High-power high-frequency drives
use amperes.

Typo: BY399 is rated for 800V, not 100V.

 

[Nobody]

On Thu, 14 Feb 2013 09:14:44 -0800, mrdarrett wrote:

Yes I know I should use a heat sink, but I've also heard I should use a
protection diode. How do I size this?

While conducting, it will need to carry the same current as the motor. To
be on the safe side, it should be sized for the maximum current you will
be supplying to the motor.

If you're thinking of sizing it according to duty cycle, bear in mind that
the diode's duty cycle will rise to 100% when you switch the motor off
(even if you do so by removing the power; the energy stored in both the
motor's inertia and its inductance has to go somewhere).

Won't the power be dissipated
(wasted) in the diode anyway, transferring the heat from the motor to the
diode?

The diode will dissipate power equal to the motor current multiplied by
the diode's voltage drop (this is lower for a Schottky diode than a P-N
diode) multiplied by duty cycle. This will be much lower than the power
you're currently dissipating in the MOSFET.

Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?

There isn't a "correct" frequency. Too low a frequency may produce audible
vibrations. Too high a frequency and you have to start worrying about the
effect of load capacitance on efficiency.

If I increase the frequency, won't the voltage spikes from the motor coils
be even higher?

No. Without a flyback diode, the voltage will be whatever it takes to
force the MOSFET to conduct (i.e. its breakdown voltage). With a flyback
diode, the voltage will be whatever it takes to force the diode to conduct
(i.e. some fraction of a volt above the supply voltage, depending upon
whether you're using a P-N diode or a Schottky diode).

 

[Phil Allison]

"Nobody"

If you're thinking of sizing it according to duty cycle, bear in mind that
the diode's duty cycle will rise to 100% when you switch the motor off
(even if you do so by removing the power; the energy stored in both the
motor's inertia and its inductance has to go somewhere).

** Absurd drivel.

There isn't a "correct" frequency. Too low a frequency may produce audible
vibrations.

** So what ?

The real issue is a VERY high RMS current flow in the motor and excess
dissipation in the switch.

With a sensible choice of frequency, motor current is nearly pure DC.



..... Phil

 

[amdx]

On 2/14/2013 9:13 PM, Jamie wrote:

mrdarrett@gmail.com wrote:
Hi,

I built a pwm circuit (200 Hz switching frequency) using 12V as a
supply, also using around 12V to the gate of an IRF530 mosfet. I
tested it with an automotive turn signal light (about 1 amp) and it
worked just fine (without a heat sink on the mosfet).

I then tried using it to control a 12V tire inflator/compressor (about
4 amps). The inflator operated fine, but then when I felt the mosfet
I was surprised it was so hot it burned my finger.
Yes I know I should use a heat sink, but I've also heard I should use
a protection diode. How do I size this? Won't the power be dissipated
(wasted) in the diode anyway, transferring the heat from the motor to
the diode?

Also, is 200 Hz the correct frequency to drive a 4A, 12V DC motor?

If I increase the frequency, won't the voltage spikes from the motor
coils be even higher?

Thanks,

Michael

At that freq, you could be saturating the coil and your apparent
speed control that you may think you have is more mechanical than
electrical.

It is common to use ~8kHz

with out knowing more about the motor you're driving, you would
need to monitor heat output of the motor's core to determine if
you are reaching too far for the PWM carrier frequency.

Too many hz will cause heating and efficiency losses.

Jamie

I have noted that most golf cart motor drivers run between 18khz and

22Khz. Also seen it written that they went that high to get above the
hearing range.
Yes, I know these run 200 to 400 amps at 36 to 48 volts, and would have
more commutations per revolution than a 4 amp motor. Hmm... so maybe my
point is moot I'll let Phil tell me!
Mikek