Maker Pro
Maker Pro

Simple 555 PWM - disappointing performance

T

Terry Pinnell

Jan 1, 1970
0
My curtain controller has been working fine, but I'd like to 'tame' it
a bit, as it hits its limit switches a little too hard for long term
reliability. So I thought I'd make a basic PWM circuit for it from a
simple 555 astable with fully-variable duty cycle, directly driving a
power NPN, with the motor (a geared, ex-screwdriver unit, designed for
2.4 V originally) in the transistor's collector. I duly breadboarded
that today, but was disappointed in the results. For example, with a
4V supply instead of its original 2.4 (it was powered by 2 C-type
Nicads), at about 50% duty cycle its torque was very low. Not
surprising when you see the output waveform as shown here:
http://www.terrypin.dial.pipex.com/Images/PWM1.gif

What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)
 
M

Mike Monett

Jan 1, 1970
0
Terry Pinnell wrote:

[...]
What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)
Terry Pinnell
Hobbyist, West Sussex, UK

Needs lots more base drive to saturate the 3055.

Mike Monett
 
P

Pooh Bear

Jan 1, 1970
0
Terry said:
My curtain controller has been working fine, but I'd like to 'tame' it
a bit, as it hits its limit switches a little too hard for long term
reliability. So I thought I'd make a basic PWM circuit for it from a
simple 555 astable with fully-variable duty cycle, directly driving a
power NPN, with the motor (a geared, ex-screwdriver unit, designed for
2.4 V originally) in the transistor's collector. I duly breadboarded
that today, but was disappointed in the results. For example, with a
4V supply instead of its original 2.4 (it was powered by 2 C-type
Nicads), at about 50% duty cycle its torque was very low. Not
surprising when you see the output waveform as shown here:
http://www.terrypin.dial.pipex.com/Images/PWM1.gif

What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)

A couple of immediate comments.

2N3055s have lousy current gain. Are you sure a 555 can provide enough
base drive ?

*Never* use 1N4148s as catch diodes. They expire. No surge capability. Use
a 1N400x.

Might also work better if you used a half bridge configuration ( Class D
PWM ).

Graham
 
J

Joerg

Jan 1, 1970
0
Hello Terry,

As Graham said, a lone 2N3055 isn't going to cut it. I don't remember
its beta but even if it was 50 at 3A the 555 would have to supply 60mA.
And the beta probably isn't even 50.

Use a decent size FET. Or at least use another smaller transistor in
front of the 2N3055 in a Darlington configuration. Also, don't run tens
of kHz. A 2N3055 isn't exactly a rocket.

Regards, Joerg
 
P

Pooh Bear

Jan 1, 1970
0
Joerg said:
Hello Terry,

As Graham said, a lone 2N3055 isn't going to cut it. I don't remember
its beta but even if it was 50 at 3A the 555 would have to supply 60mA.
And the beta probably isn't even 50.

Use a decent size FET. Or at least use another smaller transistor in
front of the 2N3055 in a Darlington configuration. Also, don't run tens
of kHz. A 2N3055 isn't exactly a rocket.

fT is 800 kHz IIRC.

If you treat it right though...... Some surprisingly nice audio amps used
them. We're spoilt with 20-40 MHz output devices now !

Graham
 
P

Pooh Bear

Jan 1, 1970
0
Joerg said:
Hello Graham


But maybe not at three amps where Terry wants to use it.

But fT is designed to look good.

It's where Ic/Ib falls to 1 in the emitter follower configuration IIRC.

Btw - RCA made rugged single diffused 2N3055s with huge SOA and low fT
whereas Motorola made faster ones ( 2 MHz ? ) with rather less
durability. Different processes.

Took JEDEC ages to make up their mind what to do. The RCA device was
renamed 2N3055H IIRC.

Similar happened with the 2N3442. Was very popular over here in the 70s
' Linsley Hood ' kit amplifier. If you got the 'wrong' one though it
made a nice oscillator !

Graham
 
E

ehsjr

Jan 1, 1970
0
Terry said:
My curtain controller has been working fine, but I'd like to 'tame' it
a bit, as it hits its limit switches a little too hard for long term
reliability. So I thought I'd make a basic PWM circuit for it from a
simple 555 astable with fully-variable duty cycle, directly driving a
power NPN, with the motor (a geared, ex-screwdriver unit, designed for
2.4 V originally) in the transistor's collector. I duly breadboarded
that today, but was disappointed in the results. For example, with a
4V supply instead of its original 2.4 (it was powered by 2 C-type
Nicads), at about 50% duty cycle its torque was very low. Not
surprising when you see the output waveform as shown here:
http://www.terrypin.dial.pipex.com/Images/PWM1.gif

What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)

Make sure that the supply voltage is rock solid. If
it is routed through the breadboard to the emitter
and motor, re-wire with direct leads from the power
supply. My bet is that the supply (as "seen" by the
motor/xsistor circuit) is not capable of the current
that the motor draws. I'd also bypass at at the
breadboard. After doing that, you can address
stronger base drive as necessary. I agree with others -
you may not have enough.

Ed
 
H

Herbert Blenner

Jan 1, 1970
0
Terry said:
My curtain controller has been working fine, but I'd like to 'tame' it
a bit, as it hits its limit switches a little too hard for long term
reliability. So I thought I'd make a basic PWM circuit for it from a
simple 555 astable with fully-variable duty cycle, directly driving a
power NPN, with the motor (a geared, ex-screwdriver unit, designed for
2.4 V originally) in the transistor's collector. I duly breadboarded
that today, but was disappointed in the results. For example, with a
4V supply instead of its original 2.4 (it was powered by 2 C-type
Nicads), at about 50% duty cycle its torque was very low. Not
surprising when you see the output waveform as shown here:
http://www.terrypin.dial.pipex.com/Images/PWM1.gif

What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)

Replace Q1 with a PNP transistor in common collection configuration.
Now this circuit can take advantage of the sizable current sink
capability of the 555 timer. Further since you are now switching
voltage across the motor, D1 is not needed.

Herbert
 
B

Bob Eldred

Jan 1, 1970
0
Terry Pinnell said:
My curtain controller has been working fine, but I'd like to 'tame' it
a bit, as it hits its limit switches a little too hard for long term
reliability. So I thought I'd make a basic PWM circuit for it from a
simple 555 astable with fully-variable duty cycle, directly driving a
power NPN, with the motor (a geared, ex-screwdriver unit, designed for
2.4 V originally) in the transistor's collector. I duly breadboarded
that today, but was disappointed in the results. For example, with a
4V supply instead of its original 2.4 (it was powered by 2 C-type
Nicads), at about 50% duty cycle its torque was very low. Not
surprising when you see the output waveform as shown here:
http://www.terrypin.dial.pipex.com/Images/PWM1.gif

What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)

As has been mentioned, the saturation of the 3055 isn't very good, it needs
more base drive. For three amps of collector current it needs at least 150mA
of base current with beta of 20. Use a an emmiter follower transistor to
drive the base making a darlington connection. Definitely, as was stated,
the fly off diode should be able to handle 3 amps surge, a 1N4001 would do
the job.

However, the real question is: why does the collector turn off to a voltage
of only one volt? It should go up to the supply rail, four volts. This
implies that either the transistor is not turning off properly or, more
likely, the voltage available is no where near four volts. Maybe a bad
battery?? Or, a battery that simply can't supply the current and is dragged
down during the on time but stays down. If this is the case, it's no wonder
you have low torque. It's also possible that the 3055 has been damged by not
being protected by the 4148 diode during fly off. In this case it might be
acting like a low ohm resistor to ground with little transistor action.
These are things to check. Also, bypass the power supply near the motor with
several microfarads of capacitance to supply the short term current needs of
the circuit.
Bob
 
M

Mike Monett

Jan 1, 1970
0
Bob said:
As has been mentioned, the saturation of the 3055 isn't very good,
it needs more base drive. For three amps of collector current it
needs at least 150mA of base current with beta of 20. Use a an
emmiter follower transistor to drive the base making a darlington
connection. Definitely, as was stated, the fly off diode should be
able to handle 3 amps surge, a 1N4001 would do the job.
However, the real question is: why does the collector turn off to
a voltage of only one volt? It should go up to the supply rail,
four volts. This implies that either the transistor is not turning
off properly or, more likely, the voltage available is no where
near four volts. Maybe a bad battery?? Or, a battery that simply
can't supply the current and is dragged down during the on time
but stays down. If this is the case, it's no wonder you have low
torque. It's also possible that the 3055 has been damged by not
being protected by the 4148 diode during fly off. In this case it
might be acting like a low ohm resistor to ground with little
transistor action. These are things to check. Also, bypass the
power supply near the motor with several microfarads of
capacitance to supply the short term current needs of the circuit.

Thanks for pointing out the two voltage scales - I didn't even
see the collector has it's own scale on the right side in red.

Now that there are two separate scales, you can see that the blue
base signal goes well over 1 Volt, and has considerable ripple. This
indicates something is not quite right with the measurement. With 5
volts driving 100 ohms in the base, the base voltage should not get
much above 0.6V.

If the measurement were true, you'd think the base current would be
many amps, which means the 3055 would be heavily saturated, and
there should be no ripple voltage on either the base or the
collector.

Next, the collector voltage in red dips below 0.4V eight times - and
even touches zero over on the right side. I don't quite know how a
power transistor can do this.

As you point out, it is very strange the collector clamps at 1 Volt.
Even stranger, it is exactly 1 Volt. I suspect something is amiss in
the measurement, which needs to be sorted out before we can make any
sense of what the circuit is actually doing.

But everyone agrees it will need more base drive:)

Mike Monett
 
B

Bob Monsen

Jan 1, 1970
0
Mike said:
Thanks for pointing out the two voltage scales - I didn't even
see the collector has it's own scale on the right side in red.

Now that there are two separate scales, you can see that the blue
base signal goes well over 1 Volt, and has considerable ripple. This
indicates something is not quite right with the measurement. With 5
volts driving 100 ohms in the base, the base voltage should not get
much above 0.6V.

If the measurement were true, you'd think the base current would be
many amps, which means the 3055 would be heavily saturated, and
there should be no ripple voltage on either the base or the
collector.

Next, the collector voltage in red dips below 0.4V eight times - and
even touches zero over on the right side. I don't quite know how a
power transistor can do this.

As you point out, it is very strange the collector clamps at 1 Volt.
Even stranger, it is exactly 1 Volt. I suspect something is amiss in
the measurement, which needs to be sorted out before we can make any
sense of what the circuit is actually doing.

But everyone agrees it will need more base drive:)

If you notice, on the gif, his input voltage is 15V. Thus, with a 100
ohm resistor, he is supplying 150mA, which should be sufficient for the
motor. I'm not sure the circuit is doing this, but his gif sure implies
it (See the 'simulation' waveforms).

I think he has the diode connected incorrectly... It's probably
backwards. When the motor starts drawing current, once it gets to be
more than 0.6V below the rail, the current flows through the diode,
causing the motor to sputter.
 
T

Terry Pinnell

Jan 1, 1970
0
Mike Monett said:
Terry Pinnell wrote:

[...]
What is the main reason for its poor performance please? I didn't get
around to trying it, but would a cap (100nf?) across the motor help?
What other changes would get me closer to the idealised result? (BTW,
I do get that when I swap the motor for a purely resistive load.)
Terry Pinnell
Hobbyist, West Sussex, UK

Needs lots more base drive to saturate the 3055.

Thanks, but I'm not sure why I need more than the 200 mA max output
current capability of the 555? With 3A collector current, the 2N3055's
gain only has to be 15 or more to drive the motor fully from that,
yes?
 
T

Terry Pinnell

Jan 1, 1970
0
Pooh Bear said:
A couple of immediate comments.

2N3055s have lousy current gain. Are you sure a 555 can provide enough
base drive ?

I reckoned so, but both you and Mike have raised the question so far,
so I'm beginning to wonder. But 200 mA should surely be enough? (Min
gain 15, collector max 3A.) FWIW, reducing that 100 ohm base R to 50
made no difference.
*Never* use 1N4148s as catch diodes. They expire. No surge capability. Use
a 1N400x.

OK, thanks, I'll change them.
Might also work better if you used a half bridge configuration ( Class D
PWM ).

Understood. But for this undemanding application I'd like to use a
very basic single-ended approach if possible. When driven directly
from my bench power supply, the motor operates reliably from under 2V
upwards through at least 6V. So there should be plenty of tolerance
for some CE drop if I use say a 4 - 5V supply. All I want to do is
slow it down, maybe to 1/3 or 1/2 speed, while maintaining close to
full torque. There's also the issue of integrating this 'add-on' with
my existing circuit, which uses two relays, respectively to apply
power and reverse current direction. I suspect a half bridge
configuration might add complications in that respect.

Practicalities apart, I'm also curious why it doesn't give a cleaner
output.
 
T

Terry Pinnell

Jan 1, 1970
0
Joerg said:
Hello Terry,

As Graham said, a lone 2N3055 isn't going to cut it. I don't remember
its beta but even if it was 50 at 3A the 555 would have to supply 60mA.
And the beta probably isn't even 50.

See my replies to Mike and Graham. A 555 can supply 200mA.
Use a decent size FET. Or at least use another smaller transistor in
front of the 2N3055 in a Darlington configuration. Also, don't run tens
of kHz. A 2N3055 isn't exactly a rocket.

?? As the 'scope screen shot shows, it's running at about 300 Hz.
 
T

Terry Pinnell

Jan 1, 1970
0
ehsjr said:
Make sure that the supply voltage is rock solid. If
it is routed through the breadboard to the emitter
and motor, re-wire with direct leads from the power
supply. My bet is that the supply (as "seen" by the
motor/xsistor circuit) is not capable of the current
that the motor draws. I'd also bypass at at the
breadboard.

Thanks. I'll take another look at the supply, but I did try both my
home made bench power supply and a 4.8V battery set. The former should
be well able to supply 3A. The latter was 4 full-charged D-type
NiCads. Both gave similar result. But I'm wondering if my earth
returns were properly wired up; I'll check later this morning. I'll
also try another set of batteries.
After doing that, you can address
stronger base drive as necessary. I agree with others -
you may not have enough.

See my earlier replies on this.
 
T

Terry Pinnell

Jan 1, 1970
0
Herbert Blenner said:
Replace Q1 with a PNP transistor in common collection configuration.
Now this circuit can take advantage of the sizable current sink
capability of the 555 timer. Further since you are now switching
voltage across the motor, D1 is not needed.

Thanks, Herbert, but according to the LM555 datasheet, a 555 "output
can source or sink 200 mA."
 
B

Ban

Jan 1, 1970
0
Terry said:
Understood. But for this undemanding application I'd like to use a
very basic single-ended approach if possible. When driven directly
from my bench power supply, the motor operates reliably from under 2V
upwards through at least 6V. So there should be plenty of tolerance
for some CE drop if I use say a 4 - 5V supply. All I want to do is
slow it down, maybe to 1/3 or 1/2 speed, while maintaining close to
full torque. There's also the issue of integrating this 'add-on' with
my existing circuit, which uses two relays, respectively to apply
power and reverse current direction. I suspect a half bridge
configuration might add complications in that respect.

But you do not need any relays in this case, the H-bridge can do all what
you need. Also the supply voltage should be increased, otherwise you might
not be able to put enough current into the motor. Do not only monitor the
voltages, but far more important is the current.
Practicalities apart, I'm also curious why it doesn't give a cleaner
output.

Because when the switch is off, the motor keeps turning and creating the
back-EMF into basically an open circuit. This may even cause problems with
the bearings, additional heating and vibration.
 
T

Terry Pinnell

Jan 1, 1970
0
Ban said:
But you do not need any relays in this case, the H-bridge can do all what
you need. Also the supply voltage should be increased, otherwise you might
not be able to put enough current into the motor. Do not only monitor the
voltages, but far more important is the current.


Because when the switch is off, the motor keeps turning and creating the
back-EMF into basically an open circuit. This may even cause problems with
the bearings, additional heating and vibration.

The existing circuit has been working reliably for nearly a year.

A small, robust geared screwdriver motor is surely unlikely to suffer
from problems of heating and vibration?

What switch are you referring to in "...when the switch is off, the
motor keeps turning..."?
 
B

Ban

Jan 1, 1970
0
Terry said:
The existing circuit has been working reliably for nearly a year.

A small, robust geared screwdriver motor is surely unlikely to suffer
from problems of heating and vibration?

What switch are you referring to in "...when the switch is off, the
motor keeps turning..."?

Well, the switch is the 3055, isn't it? With 300Hz the motor will keep
turning when the PWM is switching the transistor on and off. The motor
generates a voltage when turned, I suppose it is a DC-motor with magnets on
the stator.
 
Top