serial in-rush current limiter?

[justin]

I need to limit 7.5V DC motor current to approx 4 A during startup.
I'm thinking of a single transistor with serial sense resistor, but not
exclusively. Any scematic I can "borrow" ?

Thanks,


j.

 

[Robert Baer]

I need to limit 7.5V DC motor current to approx 4 A during startup.
I'm thinking of a single transistor with serial sense resistor, but not
exclusively. Any scematic I can "borrow" ?

Thanks,

j.

Use a passive part, called a thermistor; NTC (Negative Thermal
Coefficent).

 

[Winfield Hill]

Robert Baer wrote...

Use a passive part, called a thermistor; NTC (Negative Thermal
Coefficent).

A simple two-transistor circuit with a p-channel FET can do the job.

| Q1 IRFZ34 (60V, 0.08 ohms,
| R1 $1.15 from DigiKey, qty 10)
| 0.15 3W
| in o---+--/\/\--+--- S D ----o out 4A current limit
| | | G 0.22v drop at 1A
| | B |
| '----- E C ---+---/\/\--- gnd
| Q2 2n4403 R2 10k
| pnp trannie

This circuit limits the output current to about 4A, at the expense of
a modest 0.65V drop near the 4A current limit. At normal operating
currents, like 1A, the circuit delivers nearly the full 7.5V input,
less 150mV for the sense resistor plus another 80mV for the FET.

Normally the MOSFET won't need a heat sink (it dissipates under 0.1W
at 1A), but if a 4A short-circuit current-limit condition can last
for long (more than 0.35 sec, see the IRFZ34 datasheet's Effective
Transient Thermal Impedance curves), the FET will need a hefty heat
sink. The heat-sink size can be eased if you add two resistors to
create a foldback current limit (see AoE page 316), although then a
slower motor startup may be an issue.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Winfield Hill]

Winfield Hill wrote...

Robert Baer wrote...

A simple two-transistor circuit with a p-channel FET can do the job.

| Q1 IRF9Z34 (60V, 0.08 ohms,
| R1 $1.15 from DigiKey, qty 10)
| 0.15 3W
| in o---+--/\/\--+--- S D ----o out 4A current limit
| | | G 0.22v drop at 1A
| | B |
| '----- E C ---+---/\/\--- gnd
| Q2 2n4403 R2 10k
| pnp trannie

This circuit limits the output current to about 4A, at the expense of
a modest 0.65V drop near the 4A current limit. At normal operating
currents, like 1A, the circuit delivers nearly the full 7.5V input,
less 150mV for the sense resistor plus another 80mV for the FET.

Normally the MOSFET won't need a heat sink (it dissipates under 0.1W
at 1A), but if a 4A short-circuit current-limit condition can last
for long (more than 0.35 sec, see the IRF9Z34 datasheet's Effective
Transient Thermal Impedance curves), the FET will need a hefty heat
sink. The heat-sink size can be eased if you add two resistors to
create a foldback current limit (see AoE page 316), although then a
slower motor startup may be an issue.

I have corrected the p-channel FET's part number in my post above.
The IRFZ34 is an n-channel FET; the IRF9Z34 is its p-channel mate.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[justin]

Use a passive part, called a thermistor; NTC (Negative Thermal
Coefficent).

Tried it. It is battery powered, to much of a voltage drop.
j.

 

[justin]

Winfield Hill wrote...

I have corrected the p-channel FET's part number in my post above.
The IRFZ34 is an n-channel FET; the IRF9Z34 is its p-channel mate.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

Thanks, that will work.
In case I need to change I limit: Vgate is a function of V drop over
R1 so when I need to "open" the FET I need .65 V over R1 (Vbe) to
close Q2, correct?
What is the purpose of R2, to pull the gate to the ground? If I leave
R2 floating, could I use it as an off/on switch?

j.

 

[Bob Wilson]

Use a passive part, called a thermistor; NTC (Negative Thermal
Coefficent).

It is definitely not good practice to put a resistance in series with a DC
motor. The result will be absolutely lousing starting under load. Worse is a
thermistor, since its cold resistance will be too high to allow the motor to
start properly, and the motor will slowly struggle to get up to speed.

What is needed is a simple constant current source set to limit at 4A. Once the
motor is up to speed, and its current draw drops below 4 A, the current source
will effectively not even be there and the motor will run essentially as if it
were directly connected to a voltage source.

Bob.

 

[Winfield Hill]

justin wrote...

Thanks, that will work.
In case I need to change I limit: Vgate is a function of V drop
over R1 so when I need to "open" the FET I need .65 V over R1 (Vbe)
to close Q2, correct?
Correct.

What is the purpose of R2, to pull the gate to the ground? If I
leave R2 floating, could I use it as an off/on switch?

If a FET gate floats it's likely to float to a voltage that turns
on the FET, so a pullup resistor has to be added to keep it off.

| Q1 IRF9Z34 (60V, 0.08 ohms,
| R1 $1.15 from DigiKey, qty 10)
| 0.15 3W
| in o---+--/\/\--+--- S D ----o out power switch with
| | | G 4A current limit,
| | B | R2 10k 0.22v drop at 1A
| +----- E C ---+---/\/\--,
| | Q2 2n4403 | on/off switch
| | pnp trannie R3 10k | /
| '----------------/\/\-----+---o/ o--- gnd

If the on/off switch is replaced with a small transistor or FET
(2n7000), the high-current power switch can be logic controlled.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Ken Smith]

It is definitely not good practice to put a resistance in series with a DC
motor.

This is exactly what a industrial "motor starter" does. On large motors
not having the resistance in series when power is applied, results in
disasterous failure perhaps involving the motor ripping its self apart and
rolling across the floor in flames.

Worse is a
thermistor, since its cold resistance will be too high to allow the motor to
start properly, and the motor will slowly struggle to get up to speed.

A correctly sized thermistor would make a good motor starter circuit. It
starts off with a higher resistance just when you need it and lowers its
resistance once the motor is running.

 

[John Woodgate]

I read in sci.electronics.design that Ken Smith <[email protected]>

On large motors
not having the resistance in series when power is applied, results in
disasterous failure perhaps involving the motor ripping its self apart and
rolling across the floor in flames.

Ooh, goody! Can I watch? Beats exploding op-amps any day. (;-)

 

[Hal Murray]

On large motors

Ooh, goody! Can I watch? Beats exploding op-amps any day. (;-)

Back in the '50s, I went to a "technical" high school. The curriculum
needed updating. We worked on motors and generators.

The speed control on DC montors was the stator current.
(I think, it's been a long time.) More current made a stronger
magnetic field so it didn't have to go as fast to match the
voltage on the rotor. The standard setup used a huge rheostat.

One day, our lab exercise was a startup sequencer for a ~10 HP DC motor.
It was big. Probably leftover from WW II. It was mounted on a
platform with castors.

We got it all wired up and pushed the "Start" button. Klunk for
the first relay. Motor starts spinning. After a few seconds, Klunk
for the second relay (shorts out some of the starting resistance).
Motor goes faster. ...

The last stage shorted out all of the startup resistance and also
removed the short on the speed control rheostat.

The problem was that we had missed a wire. The rheostat was open.
That's the "full speed" mode. It was pretty exciting - amazing
how fast a big chunk of iron can spin. We got the stop button
before anything bad happened but there was no question that we
had screwed up big. I still have a mental picture of that big
motor slowly walking across the floor with everybody in the lab
looking in our direction in terror. (inlcuding the instructor)

 

[N. Thornton]

A correctly sized thermistor would make a good motor starter circuit. It
starts off with a higher resistance just when you need it and lowers its
resistance once the motor is running.

Trouble with thermistor surge limiting is that if you turn power off
then on, you got no current limit. So your system needs to cope
anyway, as the thermistor isnt reliable in that respect. Only if you
can somehow guarantee no short off times can a thermistor be
effective.

Why does i need to be limited to 4A? What are the present start and
run currents? What kind of power supply?


Regards, NT

 

[Winfield Hill]

N. Thornton wrote...

Trouble with thermistor surge limiting is that if you turn power
off then on, you got no current limit. So your system needs to cope
anyway, as the thermistor isnt reliable in that respect. Only if
you can somehow guarantee no short off times can a thermistor be
effective.

A large motor takes a while to slow to near zero speed, and
inrush limiters cool off pretty rapidly. These properties
help reduce the small window of serious malfunction.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Bob Wilson]

This is exactly what a industrial "motor starter" does. On large motors
not having the resistance in series when power is applied, results in
disasterous failure perhaps involving the motor ripping its self apart and
rolling across the floor in flames.

I meant a perminently connected resistor. An example of this is that
companies that make small cell phone vibrator motors specifically state that
they do not recommned putting a resistor in series for speed control.

I agree that some "effective" resistance can be a good thing with larger
motors, particularly perminent magnet or shunt types, which had massive
starting curent. But this resistance should be shorted out once the motor
comes up to speed and the back EMF cuts the current to reasonable values.
Having a significant value of resistance in series with these types of
motors once they are up to speed, results is lousy speed regulation under
load.

A correctly sized thermistor would make a good motor starter circuit. It
starts off with a higher resistance just when you need it and lowers its
resistance once the motor is running.

Unfortunately, a thermistor is useless if the motor has to be stopped and
restarted quickly. The thermistor will have no time to cool (and revert to a
higher resistance) before the motor is restarted. Also, it can be hard to
find a value that has a high enough resistance when starting, a low enough
one under running conditions, and a fast enough thermal time constant that
it reduces its resistance fast enough so it doesn't cause the motor to
labour when starting.

Bob.

 

[Bob Wilson]

I read in sci.electronics.design that Ken Smith <[email protected]>



Ooh, goody! Can I watch? Beats exploding op-amps any day. (;-)

Years ago we had a young engineer fresh out of school who had a habit of
installing electrolytics in backwards. He earned the name "Blaster", which
stuck for years.

Bob.

 

[Ken Smith]

Trouble with thermistor surge limiting is that if you turn power off
then on, you got no current limit.

It depends a lot on how the time constants match etc. If you turn the
motor off and right back on, its no problem because it is still spinning
near full speed. You have to wait for the motor to spin down before you
switch on to create the surge. Somehow you have to keep the thermistor
from cooling back off in that amount of time or you have to stop the motor
more quickly than it naturally would.

Why does i need to be limited to 4A? What are the present start and
run currents? What kind of power supply?

I really don't remember. I was just responding to someones bogus claims
about a resistance being bad.

 

[Ken Smith]

N. Thornton wrote...

A large motor takes a while to slow to near zero speed, and
inrush limiters cool off pretty rapidly. These properties
help reduce the small window of serious malfunction.

Or perhaps remove it all together. A DC motor is usually run up to speed
at more than its rated continuous current. This means that the run up
"time constant" will be shorter than the stopping "time
constant". Although the curves are not exactly exp(X) shaped, I suspect
that with some design time, it wouldn't be too hard to make the surge
limiter cool fast enough.

 

[Ken Smith]

I meant a perminently connected resistor. An example of this is that
companies that make small cell phone vibrator motors specifically state that
they do not recommned putting a resistor in series for speed control.

If this is what you mean, then we agree. A resistor in series with a
perminent magnet motor hardly reduces the running speed at all unless the
resistor has a large value or the motor is loaded.


[...]

Having a significant value of resistance in series with these types of
motors once they are up to speed, results is lousy speed regulation under
load.

Unless "lousy speed regulation" was part of the goal. At one time it was
common to have printer imbedded in products to record what happened over
time. The paper started on one roll and ended up printed and on the
other. In the take up motor on a printer like this, a series resistor
limits the stalled torque and the supply voltage limits the top
speed. Typically the motor would be one intended for 12V operation and
instead be run from 5 or 6V. This gave it very long life too.

Unfortunately, a thermistor is useless if the motor has to be stopped and
restarted quickly.

If the motor is stopped in less than its natural stopping time then yes
this is a problem.

The thermistor will have no time to cool (and revert to a
higher resistance) before the motor is restarted.

They cool quite quickly though if you heat sink them so I don't see a huge
problem here.

[...]

it reduces its resistance fast enough so it doesn't cause the motor to
labour when starting.

What do you mean about labor?