Motor Controller Flyback

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
I've been asked to include a motor controller on a board I am designing.  T
he motor is from Pololu

https://www.pololu.com/product/4756

The controller is a VNH5019 from ST

https://www.st.com/content/ccc/resource/technical/document/datasheet/1c/26/
60/a8/3e/70/4c/e1/CD00234623.pdf/files/CD00234623.pdf/jcr:content/translati
ons/en.CD00234623.pdf

It seems they recommend a capacitor to absorb the flyback current on switch
ing to prevent power from flowing back into the PSU.  I have no  idea how t
o size this cap.  They recommend 500 uF for each "10 amps of current".  It  
seems to me the capacitance would also depend on the inductance of the moto
r since that would determine the energy stored in the magnetic field and so
 the energy stored in the cap.  

Is there something about motors that make the inductance not a factor, like
 it approximately scales with the current???  Also, wouldn't the capacitor  
size be related to the square of the current???

The motor is 5.5 amps stalled, so I suppose I could put a 470uF part on the
 board and be happy.  I just don't like winging it when I build things.  I'
d like to show the reasoning for selecting this part.

--  

  Rick C.

  - Get 1,000 miles of free Supercharging
We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On Thursday, June 4, 2020 at 8:40:18 AM UTC+2, Ricketty C wrote:
Quoted text here. Click to load it
 The motor is from Pololu
Quoted text here. Click to load it
6/60/a8/3e/70/4c/e1/CD00234623.pdf/files/CD00234623.pdf/jcr:content/transla
tions/en.CD00234623.pdf
Quoted text here. Click to load it
ching to prevent power from flowing back into the PSU.  I have no  idea how
 to size this cap.  They recommend 500 uF for each "10 amps of current".  I
t seems to me the capacitance would also depend on the inductance of the mo
tor since that would determine the energy stored in the magnetic field and  
so the energy stored in the cap.  
Quoted text here. Click to load it
ke it approximately scales with the current???  Also, wouldn't the capacito
r size be related to the square of the current???
Quoted text here. Click to load it
he board and be happy.  I just don't like winging it when I build things.  
I'd like to show the reasoning for selecting this part.
Quoted text here. Click to load it

The energy of the winding is related to the square of the current as you wr
ite

But, the current is continuous, and during the freewheeling period, the dea
dtime, the current is constant (your PWM is much higher than the winding ti
me constant)

So the energy fed to the rails during deadtime is proportional to the windi
ng current

Also, the deadtime is normally very small, so only limited energy is freewh
eeled to the rail (deadtime is normaly less than 1% of the switching cycle)

In average half the deadtimes, the current is freewheeled to GND, so you ar
e left with 0.5% of the energy from the freewheeling to the rails. The moto
r takes energy to run, so I would not expect to see any increase of rail vo
ltage due to this. I have run motor drives at sub 100W with less than 10uF  
rail capacitance, so I do not think this is a real problem.

Maybe I am mistaken, but it seems to me this is not a real problem

Try to do the math. For your example a 10A current, freewheeled for 1us onl
y produces 20mV of ripple on the 500uF cap. Note that the winding inductanc
e is high, so the winding and thus the motor torque will be blind to a 20mV
 voltage increase for 1us

I would say you could run with a least 10 times lower capacitance with no p
roblem a all

Cheers

Klaus

Re: Motor Controller Flyback
On Friday, June 5, 2020 at 7:30:22 PM UTC-4, Klaus Kragelund wrote:
Quoted text here. Click to load it
.  The motor is from Pololu
Quoted text here. Click to load it
/26/60/a8/3e/70/4c/e1/CD00234623.pdf/files/CD00234623.pdf/jcr:content/trans
lations/en.CD00234623.pdf
Quoted text here. Click to load it
itching to prevent power from flowing back into the PSU.  I have no  idea h
ow to size this cap.  They recommend 500 uF for each "10 amps of current".  
 It seems to me the capacitance would also depend on the inductance of the  
motor since that would determine the energy stored in the magnetic field an
d so the energy stored in the cap.  
Quoted text here. Click to load it
like it approximately scales with the current???  Also, wouldn't the capaci
tor size be related to the square of the current???
Quoted text here. Click to load it
 the board and be happy.  I just don't like winging it when I build things.
  I'd like to show the reasoning for selecting this part.


Sorry, I'm not following much of what you are saying.  

Quoted text here. Click to load it
write

Later I realized there is also the energy of the rotor that has to be spun  
down if it is turning.  But if that is the case the current is a lot less t
han 10 amps.  I believe the motor is closer to 2 amps when in normal use.  
  


Quoted text here. Click to load it
eadtime, the current is constant (your PWM is much higher than the winding  
time constant)

I think you are referring the times between pulses when they motor is undri
ven electrically.  I am thinking of when the motor is being stopped.  Then  
the H-bridge just opens up and the motor has to come to a halt.  


Quoted text here. Click to load it
ding current

Yes, if the voltage does not change much and in particular if the RPM does  
not change much, then the energy is proportional to the product of current  
and time... so constant power.  


Quoted text here. Click to load it
wheeled to the rail (deadtime is normaly less than 1% of the switching cycl
e)

Yes, but this design is intended to reverse direction periodically.  That w
ould seem to be worse case.  The fly back voltage on the motor is the rever
se polarity of the applied voltage and the MOSFETs are then turned on in re
verse so the diode drop isn't even there.  They likely ramp down the speed  
in all cases, so maybe this is not an issue... other than in failure modes.
  

I've been asked to provide a dead switch in the power path to be able to sh
ut down all current to the motor.  I guess they are  considering a failure  
in the motor controller chip.  Then the current to the motor is just shut o
ff from before the H-bridge.  I was thinking the 500 uF cap would protect t
he fail safe MOSFET.  


Quoted text here. Click to load it
are left with 0.5% of the energy from the freewheeling to the rails.  

I don't follow this.  When the switches are cut off, the intrinsic diodes a
ct as freewheeling diodes and conduct the energy back out the way it came i
n.  This is a brushed motor circuit where the H-bridge is used to drive for
ward or reverse.  Are you thinking of something else?  


Quoted text here. Click to load it
of rail voltage due to this. I have run motor drives at sub 100W with less  
than 10uF rail capacitance, so I do not think this is a real problem.
Quoted text here. Click to load it
nly produces 20mV of ripple on the 500uF cap. Note that the winding inducta
nce is high, so the winding and thus the motor torque will be blind to a 20
mV voltage increase for 1us
Quoted text here. Click to load it
 problem a all
  
What about when the motor is stopped suddenly?  

--  

  Rick C.

  + Get 1,000 miles of free Supercharging
We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On Saturday, June 6, 2020 at 4:55:08 AM UTC+2, Ricketty C wrote:
Quoted text here. Click to load it
ng.  The motor is from Pololu
Quoted text here. Click to load it
1c/26/60/a8/3e/70/4c/e1/CD00234623.pdf/files/CD00234623.pdf/jcr:content/tra
nslations/en.CD00234623.pdf
Quoted text here. Click to load it
switching to prevent power from flowing back into the PSU.  I have no  idea
 how to size this cap.  They recommend 500 uF for each "10 amps of current"
.  It seems to me the capacitance would also depend on the inductance of th
e motor since that would determine the energy stored in the magnetic field  
and so the energy stored in the cap.  
Quoted text here. Click to load it
, like it approximately scales with the current???  Also, wouldn't the capa
citor size be related to the square of the current???
Quoted text here. Click to load it
on the board and be happy.  I just don't like winging it when I build thing
s.  I'd like to show the reasoning for selecting this part.
Quoted text here. Click to load it
u write
n down if it is turning.  But if that is the case the current is a lot less
 than 10 amps.  I believe the motor is closer to 2 amps when in normal use.
  
Quoted text here. Click to load it
 deadtime, the current is constant (your PWM is much higher than the windin
g time constant)
Quoted text here. Click to load it
riven electrically.  I am thinking of when the motor is being stopped.  The
n the H-bridge just opens up and the motor has to come to a halt.  
Quoted text here. Click to load it
inding current
s not change much, then the energy is proportional to the product of curren
t and time... so constant power.  
Quoted text here. Click to load it
eewheeled to the rail (deadtime is normaly less than 1% of the switching cy
cle)
Quoted text here. Click to load it
 would seem to be worse case.  The fly back voltage on the motor is the rev
erse polarity of the applied voltage and the MOSFETs are then turned on in  
reverse so the diode drop isn't even there.  They likely ramp down the spee
d in all cases, so maybe this is not an issue... other than in failure mode
s.  
Quoted text here. Click to load it

When the motor is turned off, the BEMF creates a voltage equal to the suppl
y voltage, if you were running at nominal speed just before turning off.

When it slows down, the BEMF just falls of proportional to the RPM

If you break the motor, this can be done with breaking resistors, which is  
done if you need to slow the rotor down fast. If you reverse the direction,
 then you will see up to double the voltage on the rail


Quoted text here. Click to load it
shut down all current to the motor.  I guess they are  considering a failur
e in the motor controller chip.  Then the current to the motor is just shut
 off from before the H-bridge.  I was thinking the 500 uF cap would protect
 the fail safe MOSFET.  
Quoted text here. Click to load it
u are left with 0.5% of the energy from the freewheeling to the rails.  
Quoted text here. Click to load it
 act as freewheeling diodes and conduct the energy back out the way it came
 in.  This is a brushed motor circuit where the H-bridge is used to drive f
orward or reverse.  Are you thinking of something else?  
Quoted text here. Click to load it

You have a motor that is not connected to anything. it will conduct in the  
appropiate freewheeling diode to the rail.

Quoted text here. Click to load it
e of rail voltage due to this. I have run motor drives at sub 100W with les
s than 10uF rail capacitance, so I do not think this is a real problem.
Quoted text here. Click to load it
 only produces 20mV of ripple on the 500uF cap. Note that the winding induc
tance is high, so the winding and thus the motor torque will be blind to a  
20mV voltage increase for 1us
Quoted text here. Click to load it
no problem a all
Quoted text here. Click to load it

See above

Cheers

Klaus

Re: Motor Controller Flyback
On Sun, 7 Jun 2020 11:53:46 -0700 (PDT), Klaus Kragelund

Quoted text here. Click to load it

If an efficient driver, like a switching h-bridge, drives a motor, and
the demand is to decelerate a real spinning load, energy is extracted
from the motor and load, and dumped into the power supply. Some
systems may blow up of there's nothing but supply capacitors to absorb
that energy.



--  

John Larkin         Highland Technology, Inc

Science teaches us to doubt.

We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On 8/6/20 6:26 am, snipped-for-privacy@highlandsniptechnology.com wrote:
Quoted text here. Click to load it

That's not quite the whole picture though. A disconnect a steadily  
running DC motor from its supply and it will generate about the supply  
voltage, not more. It will spin a long time like that because there's no  
current flowing. The voltage doesn't suddenly become enough to arc over  
and force current to flow.

However, while current is flowing, there is energy stored in the  
magnetism of the motor's inductance. When you disconnect it, that energy  
*will* go somewhere; current in an inductor doesn't stop flowing  
instantaneously. So you'll get a voltage spike that could arc over the  
switch and pulse a high voltage into the power supply.

The energy that causes damage is not the rotational energy of the motor,  
as sort-of implied by JL's post.

CH

Re: Motor Controller Flyback
wrote:

Quoted text here. Click to load it

Sort of implied? A handful of pm motor with a reasonable load can be
spun up in a few seconds to store hundreds of joules of rotational
energy. The energy stored in inductances is trivial in comparison.

If you want to decelerate that spinning mass in less time than
windage, that energy has to go somewhere. With a synchronous switcher
or an h-bridge, it has to go back into the power supply. Conservation
of energy always works.

Take a look at the size of the caps needed to store a few hundred
joules.



--  

John Larkin         Highland Technology, Inc

Science teaches us to doubt.

We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On 8/6/20 9:10 pm, snipped-for-privacy@highlandsniptechnology.com wrote:
Quoted text here. Click to load it

Absolutely correct. I thought you were referring to blowing up the  
driver by turning off the motor. That can happen, but not from  
rotational energy.

Re: Motor Controller Flyback
On 8/6/20 9:10 pm, snipped-for-privacy@highlandsniptechnology.com wrote:
Quoted text here. Click to load it

Also, to get the energy to go back into the supply (given that the  
free-wheel voltage will be less than what was being supplied) requires  
alternating between a brief short-circuit while the current builds in  
the winding inductance, followed by switching that current back into the  
supply. Any system that can accurately manage the trick damn well ought  
to be smart enough not to blow up the supply.

Re: Motor Controller Flyback
wrote:

Quoted text here. Click to load it

No, just tweak the duty cycle of the h-bridge. A proper h-bridge (or
its unidirectional version, a half-bridge synchronous buck) works like
a pair of gears, bidirectionally, boost/buck. It can take 50 volts of
back EMF from the motor and charge/explode the caps of a 100 volt
supply.

Again, conservation of energy explains what must happen. Given that,
we just have to work out the details.



--  

John Larkin         Highland Technology, Inc

Science teaches us to doubt.

We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
mandag den 8. juni 2020 kl. 13.10.43 UTC+2 skrev snipped-for-privacy@highlandsniptechnology.com:
Quoted text here. Click to load it

if you short the motor the energy is dumped in the motor resistance and  
FETs Rdson

if you turnoff the FETs the motor voltage is rpm * K so less than supply  
and the motor coasts to a stop

if you keep switching you basically have a buck converter in reverse,  
iow a boost converter dumping the energy into the supply


Re: Motor Controller Flyback
On Mon, 8 Jun 2020 09:33:43 -0700 (PDT), Lasse Langwadt Christensen

Quoted text here. Click to load it

Sure, but that's pretty dramatic.

Quoted text here. Click to load it

Yes, eventually.


Right. That's what a straightforward speed control loop will do if you
use it to decelerate. Mechanical things can store a lot of energy.



--  

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement  

We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On Monday, June 8, 2020 at 8:17:16 PM UTC+2, John Larkin wrote:
Quoted text here. Click to load it
It doesn't need to be. Depends on the winding resistance

Quoted text here. Click to load it

A motor is used to drive something. So often this is a quite short time to standstill. For example, I am working on pumps. It takes not more than about a second from 6000 RPM.
Quoted text here. Click to load it

Cheers

Klaus

Re: Motor Controller Flyback
On 6/8/2020 6:10 AM, snipped-for-privacy@highlandsniptechnology.com wrote:
Quoted text here. Click to load it

Not just windage. There is also load and bearing friction to slow the  
motor. The link to the motor shows it to have a gearbox attached. That  
provides some friction as well. However, the load rotational inertia is  
also involved. Characteristics of his load have not been considered...  
unless I missed that post.


Re: Motor Controller Flyback

Quoted text here. Click to load it

In the electrical machinery world, windage refers to all the losses of
a motor except the external load.

I had to take two semisters of electrical machines. I learned a lot,
actually.

--  

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement  

We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On 6/9/2020 1:12 PM, John Larkin wrote:
Quoted text here. Click to load it

I learned it as windage and friction of the motor without a load  
attached. However, we are really on the same page. I accept your definition.

Quoted text here. Click to load it

Yes. A valuable course of study. I have learned much about them too. But  
that does not mean I want to challenge you.

Re: Motor Controller Flyback

Quoted text here. Click to load it

I'm fine with discussion, and not everyone has to agree. Everybody can
learn stuff.  

I did learn a lot in the electrical machinery courses. That's the
first time I really understood a transformer or a motor. They didn't
mention small stuff, like PM motors and alternators. Half my class was
destined to work for utilities, NOPSI or LP&L. The electronics guys
didn't even socialize with the power guys.

We hated the labs. Too much smoke, for one thing.

--  

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement  

We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
On Monday, June 8, 2020 at 4:00:27 AM UTC-4, Clifford Heath wrote:
Quoted text here. Click to load it
  
Quoted text here. Click to load it
  
Quoted text here. Click to load it
  
Quoted text here. Click to load it
  
Quoted text here. Click to load it

That helps a lot.  I understand the effect of the inductance which will cre
ate a voltage proportional to the di/dt.  The effect of the rotor rotation  
is more like a transformer with an open winding, no current flow so no forc
e and no slowing.  The voltage produced will be the same as the voltage app
lied to sustain the rotation.  

One of the motor controllers I looked at showed a "braking" mode where the  
two top halves of the bridge were closed creating a shorted loop to continu
e the current and I assume letting the resistance of the FETs slow the moto
r.  

Thanks

--  

  Rick C.

  -+ Get 1,000 miles of free Supercharging
We've slightly trimmed the long signature. Click to see the full one.
Re: Motor Controller Flyback
mandag den 8. juni 2020 kl. 20.22.52 UTC+2 skrev Ricketty C:
Quoted text here. Click to load it
d
b
  
Quoted text here. Click to load it
o  
Quoted text here. Click to load it
  
Quoted text here. Click to load it
y  
Quoted text here. Click to load it
  
Quoted text here. Click to load it
,  
Quoted text here. Click to load it
reate a voltage proportional to the di/dt.  The effect of the rotor rotatio
n is more like a transformer with an open winding, no current flow so no fo
rce and no slowing.  The voltage produced will be the same as the voltage a
pplied to sustain the rotation.  
Quoted text here. Click to load it
e two top halves of the bridge were closed creating a shorted loop to conti
nue the current and I assume letting the resistance of the FETs slow the mo
tor.  
Quoted text here. Click to load it

resistance of the fets and resistance of the motor


Re: Motor Controller Flyback
On Sunday, June 7, 2020 at 2:53:51 PM UTC-4, Klaus Kragelund wrote:
Quoted text here. Click to load it
ning.  The motor is from Pololu
Quoted text here. Click to load it
t/1c/26/60/a8/3e/70/4c/e1/CD00234623.pdf/files/CD00234623.pdf/jcr:content/t
ranslations/en.CD00234623.pdf
Quoted text here. Click to load it
n switching to prevent power from flowing back into the PSU.  I have no  id
ea how to size this cap.  They recommend 500 uF for each "10 amps of curren
t".  It seems to me the capacitance would also depend on the inductance of  
the motor since that would determine the energy stored in the magnetic fiel
d and so the energy stored in the cap.  
Quoted text here. Click to load it
or, like it approximately scales with the current???  Also, wouldn't the ca
pacitor size be related to the square of the current???
Quoted text here. Click to load it
t on the board and be happy.  I just don't like winging it when I build thi
ngs.  I'd like to show the reasoning for selecting this part.
Quoted text here. Click to load it
you write
pun down if it is turning.  But if that is the case the current is a lot le
ss than 10 amps.  I believe the motor is closer to 2 amps when in normal us
e.  
Quoted text here. Click to load it
he deadtime, the current is constant (your PWM is much higher than the wind
ing time constant)
Quoted text here. Click to load it
ndriven electrically.  I am thinking of when the motor is being stopped.  T
hen the H-bridge just opens up and the motor has to come to a halt.  
Quoted text here. Click to load it
 winding current
Quoted text here. Click to load it
oes not change much, then the energy is proportional to the product of curr
ent and time... so constant power.  
Quoted text here. Click to load it
freewheeled to the rail (deadtime is normaly less than 1% of the switching  
cycle)
Quoted text here. Click to load it
at would seem to be worse case.  The fly back voltage on the motor is the r
everse polarity of the applied voltage and the MOSFETs are then turned on i
n reverse so the diode drop isn't even there.  They likely ramp down the sp
eed in all cases, so maybe this is not an issue... other than in failure mo
des.  
Quoted text here. Click to load it
ply voltage, if you were running at nominal speed just before turning off.
Quoted text here. Click to load it
s done if you need to slow the rotor down fast. If you reverse the directio
n, then you will see up to double the voltage on the rail

Ok, I assume you know what you are talking about, but I thought an inductor
 did not limit the voltage, rather it would create a voltage proportional t
o the dI/dt which can be very high if the switch is opened.  So the cap all
ows the motor to spin down and the inductance to drop current more slowly w
ithout a large voltage jump.  

I take it the BEMF does not work the same way as the inductance?  


Quoted text here. Click to load it
o shut down all current to the motor.  I guess they are  considering a fail
ure in the motor controller chip.  Then the current to the motor is just sh
ut off from before the H-bridge.  I was thinking the 500 uF cap would prote
ct the fail safe MOSFET.  
Quoted text here. Click to load it
you are left with 0.5% of the energy from the freewheeling to the rails.  
Quoted text here. Click to load it
es act as freewheeling diodes and conduct the energy back out the way it ca
me in.  This is a brushed motor circuit where the H-bridge is used to drive
 forward or reverse.  Are you thinking of something else?  
Quoted text here. Click to load it
e appropiate freewheeling diode to the rail.

Yes, and that voltage will be the same polarity as the applied power was be
fore the switch connecting the PSU is opened.  


Quoted text here. Click to load it
ase of rail voltage due to this. I have run motor drives at sub 100W with l
ess than 10uF rail capacitance, so I do not think this is a real problem.
Quoted text here. Click to load it
us only produces 20mV of ripple on the 500uF cap. Note that the winding ind
uctance is high, so the winding and thus the motor torque will be blind to  
a 20mV voltage increase for 1us
Quoted text here. Click to load it
h no problem a all
Quoted text here. Click to load it

I guess I'll have to take your word for it as I don't even have a mental im
age of what the motor is doing once the circuit is opened.  

But I don't think I'll be building this after all.  Looks like I wasn't the
 only one wondering why we were building yet another ventilator and we are  
merging with another group who is further along and have their hardware rea
dy to start testing.  They picked a very similar controller that is not 3.3
 volt compatible.  The CPU is the AVR from the Arduino Uno, so it's a 5 vol
t device anyway.  I would have tried to steer them toward an ARM based modu
le like Teensy or similar and just used the 3.3 volt compatible VNH5019 con
troller that has a bit more going for it.    

--  

  Rick C.

  -- Get 1,000 miles of free Supercharging
We've slightly trimmed the long signature. Click to see the full one.

Site Timeline