## Re: [SeattleRobotics] Re: Motor PWM frequencies

Expand Messages
• Howdy Back to the question of what frequency to use with a given motor / controller combo. Our goals are to get as much of the power used into the motor and to
Message 1 of 26 , Jun 2, 2005
Howdy

Back to the question of what frequency to use with a given motor /
controller combo.

Our goals are to get as much of the power used into the motor and to provide
good control of the motor.

Let me make a broad generalisation. Higher frequencies are better than lower
frequencies. However, there is a limit to how high you should go. With low
switching frequencies the motor will have uneven torque. It will make noise
and it will be difficult to run the motor at low speeds.

For the most part, the highest usable frequency is limited by the switching
speed of the transistors used to drive the motor.. In a given circuit a
transistor will take a certain amount of time to go from full off to full
on, and likewise, it will take sometime to go from on to off. Note: turn on
time and turn off time are often different. During this switching time the
transistor will be partly on and it will be dissipating power. If you run
the switching frequency high enough the transistors will never turn
completely on or off and worse case, much of your power will only heat the
power transistors.

One can hook up a scope, grab a current probe and puzzle it out but I think
the best method would be to put the motor in as close to a real world load
situation and measure the total power used at different switching
frequencies. For example, one could attach an RC plane prop to the motor and
apply power to it with a variable supply, adjusting the supply to get a
given prop speed. Then measure the power used to get this speed. This gives
you a best case measurement.
Next connect up the PWM bits and measure the power it takes to get the prop
up to speed. Make measurments at increasing switching frequencies. In each
case noting the power used to get the prop up to the chosen speed.
Eventually the power required to get the prop up to speed will start to
climb. This will give you a good feel for the switching frequency's effect
on efficiency. Pick a switching frequency that is high enough to give smooth
torque but low enough to avoid burning up power in the switching
transistors.

Have fun
Steve

On 6/2/05, dan michaels <oric_dan@...> wrote:
>
> --- In SeattleRobotics@yahoogroups.com, Tom Capon <robot256@g...>
> wrote:
> > My motor has about a 25A stall current at 12V.
> >
> > Would the Pololu controller work for higher currents if a large
> > heatsink were installed?
> >
> > My real question was: How can I tell which frequency gets the
> highest
> > efficiency out of my motor, assuming I try several? Is it simply
> > battery life, or will the motor heat up on a bad frequency?
> >
>
>
> From the sounds of it, the part can handle the stall current,
> at least for a short time, but the broader issue is what is the
> *continuous* run current for your motor. Greater than 6A?
>
> With power control electronics, the "limiting" factor is usually
> heat build-up. Someone on the other thread came up with a
> heatsink that would fit on the chip, but it was rated at
> 19 degC/W, whereas a 5 degC/W part would be a lot better, if
> you can find one that fits. It's difficult to draw large amounts
> of heat off of SMT parts, so this usually limits their range
> of operation. Using pcb copper helps, but it's not the best either.
> If I did the calcs right, the datasheet shows that 1 sq.in <http://sq.in>.
> of
> [presumably 2-oz] copper on the pcb produces about a 30-35 degC/W
> heatsink. Not really very good.
>
> However, I think we are needing someone to try out the part,
> and see how it goes, so we can get some 3rd party feedback.
>
>
>
>
>
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
>
>
>
>
>
>

[Non-text portions of this message have been removed]
• I have been trying to find some good explanations as to defining efficiency. There are many parts to the efficiency question. Energy transfer efficiency,
Message 2 of 26 , Jun 2, 2005
I have been trying to find some good explanations as to defining efficiency.

There are many parts to the efficiency question. Energy transfer efficiency, energy lost due to resistance in the
wires, energy loss in the motor controller, energy loss in the connectors, energy loss in the motor.

Most people get lost in discussing the efficiency of the motor controller. Because it is the easiest thing to talk

But your question was about the motor. The reason nobody can answer this question, is because most people don't even
understand the true physics inside the motors. I won't even make the claim that I do, but I am studying it. Motor
efficiency is due to both mechanical, electrical, and magnetic. They are all tied together. Ignoring how the magnetic
field collapses in the motor, the motor Indutance and winding Resistance has the biggest effect in determining how PWM
frequency effects the over all efficiency of the motor. It is all about how fast the voltage and current rise and fall
within the windings that effect the efficiency. This is why companies that sell high performance motors clearly specify
the inductance and resistance of the wires. This allows the motor contol designers to optimize the controller along
with the motor for best efficiency under the nominal operating condition.

Now for the rest of the world. Most people won't really notice the difference in motor performance if you are running
the motor controller at 1000 Hz, 5000 Hz, 10,000 Hz, or higher. I know there are people that swear that they can notice
the difference, but these are the acceptions. But for the rest of us, it doesn't matter.

Pete
• Now for the rest of the world. Most people won t really notice the difference in motor performance if you are running the motor controller at 1000 Hz, 5000
Message 3 of 26 , Jun 2, 2005
"Now for the rest of the world. Most people won't really notice the
difference in motor performance if you are running
the motor controller at 1000 Hz, 5000 Hz, 10,000 Hz, or higher. I know
there are people that swear that they can notice
the difference, but these are the acceptions. But for the rest of us, it
doesn't matter."

The perceptable difference is when you start running the motor at lower
rpm and the noise it makes when it draws heavy current. That whine you hear
when the motor is clsoe to stall or on spinup is the pwm Frequincy. The
higher the frequncy the smooother you perception of how the motor is
running. Once you get out of the audible range It's not nearly as
noticable. Again how much this really matters depends on the application.
Those that claim they notice are probably running r/c car esc's at 500hz.
And yeah those are noticeable.

Visit the SRS Website at http://www.seattlerobotics.org
• ... Hell, the whine is probably a safety feature, kind of like a moving beeper
Message 4 of 26 , Jun 2, 2005
At 1:29 PM -0700 6/2/05, Rob Purdy wrote:
> The perceptable difference is when you start running the motor at lower
>rpm and the noise it makes when it draws heavy current. That whine you hear
>when the motor is clsoe to stall or on spinup is the pwm Frequincy. The
>higher the frequncy the smooother you perception of how the motor is
>running. Once you get out of the audible range It's not nearly as
>noticable. Again how much this really matters depends on the application.
>Those that claim they notice are probably running r/c car esc's at 500hz.
>And yeah those are noticeable.

Hell, the whine is probably a safety feature, kind of like a "moving beeper"
• Thanks for all the replies. Looks like the thing to do is to get the system running and experiment with what works best in my application. --Tom Capon
Message 5 of 26 , Jun 2, 2005
Thanks for all the replies. Looks like the thing to do is to get the
system running and experiment with what works best in my application.
--Tom Capon

On 6/2/05, Rich Chandler <rchandler@...> wrote:
> At 1:29 PM -0700 6/2/05, Rob Purdy wrote:
> > The perceptable difference is when you start running the motor at lower
> >rpm and the noise it makes when it draws heavy current. That whine you hear
> >when the motor is clsoe to stall or on spinup is the pwm Frequincy. The
> >higher the frequncy the smooother you perception of how the motor is
> >running. Once you get out of the audible range It's not nearly as
> >noticable. Again how much this really matters depends on the application.
> >Those that claim they notice are probably running r/c car esc's at 500hz.
> >And yeah those are noticeable.
>
> Hell, the whine is probably a safety feature, kind of like a "moving beeper"
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
>
>
>
>
>
>
>
• There are some common misconceptions expressed in Alan s reply. Hopefully I can clear some of them up. The voltage across the coil of the motor (assuming
Message 6 of 26 , Jun 2, 2005
There are some common misconceptions expressed in Alan's reply. Hopefully I
can clear some of them up.

The voltage across the coil of the motor (assuming perfect inductance) will
generate a ramp current. The duration that the voltage is applied will
determine how high that current goes. When the voltage is removed, or set
to zero, the current STILL flows, but now ramps down towards zero. If the
FET is perfect (no on resistance, zero switching time) and the inductance is
perfect, you will have ZERO LOSS and a 100% efficient conversion of 12v
power to the equivalent 6v power. That is what switching power supplies are
THE MOTOR. It is either 0 or 12v. Sometimes it is the back-EMF instead of
0, see below.

Your losses and errors come in several flavors:
1. Coil resistance. Simple IR losses. Keep the ripple current low (high
frequencies) to minimize this value.

2. Switching losses: every time the FET switches a small amount of power is
lost as heat. Minimize switching frequency.

3. Iron Core losses. Fluxuating magnetic fields causes losses in the iron
cores (current induced, heat generated). Again, keep the magnetic
fluctuation to a minimum by using high switching frequencies. Or use
coreless motors. Oh, yeah, coreless have very low inductance, so you need
high frequencies to keep the I/R losses down. Crap, you get it coming and
going.

4. Back-emf. Not a loss, but a common thing ignored when trying to develop
a PID algorithm. Any professional controller will use a current loop for
the final drive: they control current which generates torque regardless of
the I/R and back-EMF values. There are ways to estimate the torque if you
can measure the angular velocity of your motor.

5. Depending upon what you are trying to achieve, it is best to short out
the motor leads during the off phase. The back-EMF will generate a current
(equal to the back-EMF divided by the coil resistance!) and tend to preserve
the current flowing through the coils and preserving the torque generated,
further increasing the efficiency of the system. But for this to work you
need high frequencies! Check out www.barello.net/Papers/H-Bridge.pdf for a
pictorial explanation of this.

In my analysis HIGHER FREQUENCIES are better, but you need to know what you
are doing or you won't realize the benefits. The list goes on. A good
motor driver circuit and algorithm is not trivial and any attempts at
trivializing it will result in incorrect assumptions about what is happening
and frustration with poor results.

Cheers!

-----Original Message-----
From: SeattleRobotics@yahoogroups.com
[mailto:SeattleRobotics@yahoogroups.com]On Behalf Of Alan King
Sent: Thursday, June 02, 2005 10:06 AM
To: SeattleRobotics@yahoogroups.com
Subject: Re: [SeattleRobotics] Re: Motor PWM frequencies

Tom Capon wrote:
> My motor has about a 25A stall current at 12V.
>
> Would the Pololu controller work for higher currents if a large
> heatsink were installed?
>
> My real question was: How can I tell which frequency gets the highest
> efficiency out of my motor, assuming I try several? Is it simply
> battery life, or will the motor heat up on a bad frequency?
>

Torque ripple improves and efficiency goes down with higher PWM. There
are
no double gains without any disadvantages for changing a variable in
physics.

Do the math. !2V motor with 50% and infinite PWM frequency to satisfy
the
high people. That's 6V across the motor, and 6V across the FET. 50% or
less
efficiency of the total 12V applied, and that's just to start. Motor
efficiency
of 80% of whats applied to it's terminals means that the motor will only
actually output 40% of the total 100% applied. 10% wasted in the motor, and
50%
wasted in the FET. Again, do your own math to see how things really work.

Same 50% duty, but say 12 second period. Motor gets full voltage for 6
seconds. 12V on terminals, and 80% efficient output. Motor still has only
40%
total output over that whole 12 seconds. But wonder of wonders, the FET is
full
on in the on time for almost no loss, and no power is used at all in the off
times for again no loss. In other words you stay 80% efficient instead of
wasting half of your power in a fast switching element. Just like every
other
case, fast switching increases losses and lowers efficiency not raises it,

Other cases lie between these, it is difficult to escape endpoints like
this
wiht real physics.

If you're actually wanting efficiency then the lowest PWM you can stand
the
torque ripple is the way to go. Never read a single article with good math
that
suggests otherwise, or worked a single problem that came out with a
different

But you can make efficiency be anything you want if you drop the
switching
element losses and only calculate E from the variables you want to use. But
the
right ones to use for a motor system are total output torque delivered vs
total
input power used, and that always goes down for higher PWM frequencies.

And that's before you consider that a real world 12V motor doesn't stay
80%
efficient at only 6V applied, the full answer will be even less than 40% of
total input. Fast PWM goes towards variable voltage drive, slow PWM goes
towards duty cycle drive. Variable voltage drive always has higher losses.

In other words do your own homework if you want to pass the test..

Alan

Visit the SRS Website at http://www.seattlerobotics.org
• Just one refinement on Larry s list ... [...] ... Core losses comprise hysteresis and eddy-current losses. Hysteresis is a type of magnetic nonlinearity in the
Message 7 of 26 , Jun 2, 2005
Just one refinement on Larry's list

Larry Barello wrote:
>
[...]
>
> 3. Iron Core losses. Fluxuating magnetic fields causes losses in the iron
> cores (current induced, heat generated). Again, keep the magnetic
> fluctuation to a minimum by using high switching frequencies.

Core losses comprise hysteresis and eddy-current losses. Hysteresis
is a type of magnetic nonlinearity in the core material and is
most important at the high end of the torque range. Larry's comment
about reducing losses by reducing "magnetic fluctuation", i.e.
HIGHER pwm frequency, applies to hysteresis.

Eddy-current losses are from induced electric currents which flow
within the magnetic material (steel laminations, iron powder, etc).
This is a linear phenomenon and depends on the derivative of the
flux (i.e. the voltage across the winding drives current in the
"shorted secondaries" of the iron much like a transformer). In
order to reduce eddy-current losses you want to reduce the RMS
high-frequency content of the VOLTAGE waveform, which is achieved
by using LOWER pwm frequency.

> Or use coreless motors.

Not sure about "coreless motors" but the ironless-rotor
motors I have used -- both pancake and basket-wound styles --
still have plenty of iron in the magnetic circuit; they are
just constructed so that the iron does not rotate. This
is a BIG win if you need low inertia in order to do instant
start-stop (e.g. tape drive capstans, in the good old days!)
Hysteresis and eddy-current losses still exist in the
non-moving iron parts.

Peace and motion,
Chuck
• I thought your ANT100 switched at 1k? Most of the hobbyist motor controllers (eg: for RC cars and for combat robotics) switch well under 10k. The sozbots
Message 8 of 26 , Jun 2, 2005
I thought your ANT100 switched at 1k? Most of the hobbyist motor
controllers (eg: for RC cars and for combat robotics) switch well
under 10k. The sozbots controller is at 1k, for instance. IF you're

Or maybe I'm missing something.

On 6/1/05, Larry Barello <yahoo@...> wrote:
> 10k is kind of low, depending upon your motor. Even a fairly large coreless
> motor will need much higher frequencies to be efficient.
>
• My ant controllers are around 3khz. I don t think being hyper efficient is a requirement for combat robots! ... From: Jay Prince I thought your ANT100
Message 9 of 26 , Jun 2, 2005
My ant controllers are around 3khz. I don't think being hyper efficient is
a requirement for combat robots!

-----Original Message-----
From: Jay Prince

I thought your ANT100 switched at 1k? Most of the hobbyist motor
controllers (eg: for RC cars and for combat robotics) switch well
under 10k. The sozbots controller is at 1k, for instance. IF you're

Or maybe I'm missing something.

On 6/1/05, Larry Barello <yahoo@...> wrote:
> 10k is kind of low, depending upon your motor. Even a fairly large
coreless
> motor will need much higher frequencies to be efficient.
>
• ... You must be referring to either me or Dan, since we both tried to clarify the difference between rating theoretical maximums on components and the
Message 10 of 26 , Jun 2, 2005
On Wed, Jun 01, 2005 at 06:20:11PM -0700, Jay Prince wrote:

> I just read the thread on PARTS. The pololu is a 30A motor
> controller. That some guy with a competing product can't
> differentiate between rated and theoretical maximum, and used that
> as an excuse to bash pololu, does not mean they are misrepresneting
> their motor controller.

You must be referring to either me or Dan, since we both tried to
clarify the difference between rating theoretical maximums on
"components" and the disconnect with rating end products intended to
operate as advertised out of the box. As such, I'll just say that I
respectfully disagree. I don't think either of us was really trying
to bash anyone, only that the advertising on that one product was

And very much to their credit, Jan of Pololu did respond to the list
clarifying the position of their product. Hopefully I'm not
oversimplifying, and please correct me if I'm wrong, Jan indicated
that the board was intended primarily as a "carrier" for the ST Micro
chip as an aid to folks not comfortable with SMT soldering, i.e.,
something along the lines of an experimenters board as opposed to a
complete h-bridge solution. It is left as an exercise to the buyer to
figure out and supply appropriate heat sinking, etc.

I do now "get it" after his explanation and I fully accept that. But
way, at least not to me. Specifically I am referring to the
description of the product here:

http://www.pololu.com/products/pololu/

See about 1/3rd of the way down where it says:

"NEW! 30-Amp motor driver board
Need more current? If you've got a larger motor waiting to be
used, this compact motor driver is just what you need. Connect
three digital lines to your microcontroller (five if you want
error condition feedback), and you're ready to go!"

When I read that I'm thinking it is a complete solution, not a carrier
board that requires additional work and components. At which point
you can add it to your cart and purchase without ever seeing the more

Further, while the above description would lead one to think so, I
don't think this h-bridge is very comparable to something along the
lines of the Devantech 20A h-bridge driver, even putting asside the
control interface of the Devantech and talking only about the current
rating:

http://www.robot-electronics.co.uk/shop/Motor_Controllers2008.htm

So that's my last word on that ... I have no intention of debating
this further, I've said all that I need to say. For some additional
information and perspective, you might also want to take a look at the
note Larry posted from the OSMC (Open Source Motor Controller) group.

So ... with that out of the way, if you are interested in
experimenting with this board or chip and you are expecting to do
currents in the 8A - 10A or higher range, you will want to look at
some heat sinking. Also, be sure and grab a copy of the datasheet for
the chip - you can find that on Pololu's site. As with most ICs like
this, heat sinking from the top is not going to be very effective.
Heat sinking this chip really needs to be done from the bottom where
the junction tabs are exposed - see page 20 of the datasheet for
possible circuit board layouts to facilitate heat sinking. So it
would be best to figure out some way to attach heat sink on the bottom
of the board assuming the PCB provided copper layout as shown on page
20 of the datasheet. There's no photo of the bare PCB from Pololu to
verify this, but I'm assuming they laid the board out as described in
the datasheet to contact the junctions and the chip is soldered on at
those points from underneath.

A nice feature of this chip is that it does appear to shut itself down
when it gets too hot, so it should not be likely that you can damage
it by not supplying adequate heat sinking.

-Brian
--
Brian Dean
BDMICRO - ATmega128 Based MAVRIC Controllers
http://www.bdmicro.com/
• dont mess with it. you read the email...6A look at oune of these. http://www.superdroidrobots.com/shop/category.asp?catid=5 MD03 Devantech 20A 50V Motor Driver
Message 11 of 26 , Jun 2, 2005
dont mess with it.

look at oune of these.
http://www.superdroidrobots.com/shop/category.asp?catid=5

MD03 Devantech 20A 50V Motor Driver
<http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
The MD03 is a medium power motor driver, designed to supply power beyond
that of any of the low power single chip H-Bridges that exist. Main
features are ease of use and flexibility.
More Detail
<http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
TE-036-000

Tom Capon wrote:

>My motor has about a 25A stall current at 12V.
>
>Would the Pololu controller work for higher currents if a large
>heatsink were installed?
>
>My real question was: How can I tell which frequency gets the highest
>efficiency out of my motor, assuming I try several? Is it simply
>battery life, or will the motor heat up on a bad frequency?
>
>On 6/2/05, dan michaels <oric_dan@...> wrote:
>
>
>>--- In SeattleRobotics@yahoogroups.com, Jay Prince <jayprince@g...>
>>wrote:
>>
>>
>>>I just read the thread on PARTS. The pololu is a 30A motor
>>>controller. That some guy with a competing product can't
>>>differentiate between rated and theoretical maximum, and used that
>>>
>>>
>>as
>>
>>
>>>an excuse to bash pololu, does not mean they are misrepresneting
>>>
>>>
>>their
>>
>>
>>>motor controller.
>>>
>>>
>>>
>>I don't think the issue was about bashing, rather perceived
>>impressions. In the end, it should be noted that the part is
>>probably good for about 6 Amp continuous, maybe a little more,
>>and I think all sides pretty much agreed on this.
>>
>>
>>
>>
>>
>>>The theoretical maximum rating is an objective value that can be
>>>calculated-- while actual performance is something that can vary by
>>>application conditions, and thus is difficult to compare across
>>>product lines. Pololu is pretty clear on what their tests showed
>>>
>>>
>>for
>>
>>
>>>this controller. (And its a darn good price if you want to run at
>>>
>>>
>>12V
>>
>>
>>>or more...)
>>>
>>>As to the original posters question-- how large is large? One rule
>>>
>>>
>>of
>>
>>
>>>thumb for current is to make the motor controllers theoretical max
>>>current equal to the motors stall current-- assuming the application
>>>is one where the motor will not be stalled much at all, or close to
>>>it.
>>>
>>>I think 10khz is very high for robotics use, and probably a very
>>>
>>>
>>good
>>
>>
>>>frequency. My robots us a motor controller in the 1khz range.
>>>
>>>
>>>On 6/1/05, dan michaels <oric_dan@y...> wrote:
>>>
>>>
>>>>--- In SeattleRobotics@yahoogroups.com, Tom Capon <robot256@g...>
>>>>wrote:
>>>>
>>>>
>>>>>Hi all,
>>>>>I have a large DC motor that I want to control via PWM. I am
>>>>>
>>>>>
>>>>planning
>>>>
>>>>
>>>>>on getting a 30A controller from Pololu.com that has 10khz
>>>>>
>>>>>
>>frequency
>>
>>
>>>>>response. When I get it up and running, how can I tell what PWM
>>>>>frequency is best for the motor? I've heard about small DC
>>>>>
>>>>>
>>motors
>>
>>
>>>>>wanting high frequencies, but I'm wondering how this can be
>>>>>
>>>>>
>>>>determined
>>>>
>>>>
>>>>>experimentally for a given motor. thanks! --Tom Capon
>>>>>
>>>>>
>>>>Hi Tom. It's questionable if that's a 30A controller. Take a
>>>>look at the long thread on PARTS that's had to be quashed by
>>>>the moderators twice now.
>>>>
>>>>http://groups.yahoo.com/group/PARTS/messages
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>Visit the SRS Website at Ihttp://www.seattlerobotics.org
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>
>>
>>
>>Visit the SRS Website at http://www.seattlerobotics.org
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>Visit the SRS Website at http://www.seattlerobotics.org
>
>
>
>
>
>
>
>
>

[Non-text portions of this message have been removed]
• ... And if you know where to look you can find the schematic of the Devantech. See page 150 (PDF page 168) of this document:
Message 12 of 26 , Jun 2, 2005
> MD03 Devantech 20A 50V Motor Driver
> <http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
> The MD03 is a medium power motor driver, designed to supply power beyond
> that of any of the low power single chip H-Bridges that exist. Main
> features are ease of use and flexibility.

And if you know where to look you can find the schematic of the Devantech.

See page 150 (PDF page 168) of this document:
http://phys.waikato.ac.nz/research/andrew_payne_2004.pdf

P.S. It's a bit cheaper (\$102) from junun.org.

--
Dave Hylands
http://www.DaveHylands.com/
• I used one of these Devantech units in a recent project, and it worked very well. It was well worth the \$100. -Ted ... From: SeattleRobotics@yahoogroups.com
Message 13 of 26 , Jun 3, 2005
I used one of these Devantech units in a recent project, and it worked very
well. It was well worth the \$100.

-Ted

-----Original Message-----
From: SeattleRobotics@yahoogroups.com
[mailto:SeattleRobotics@yahoogroups.com] On Behalf Of tbrenke@...
Sent: Thursday, June 02, 2005 8:16 PM
To: SeattleRobotics@yahoogroups.com
Subject: Re: [SeattleRobotics] Re: Motor PWM frequencies

dont mess with it.

look at oune of these.
http://www.superdroidrobots.com/shop/category.asp?catid=5

MD03 Devantech 20A 50V Motor Driver
<http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
The MD03 is a medium power motor driver, designed to supply power beyond
that of any of the low power single chip H-Bridges that exist. Main
features are ease of use and flexibility.
More Detail
<http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
TE-036-000

Tom Capon wrote:

>My motor has about a 25A stall current at 12V.
>
>Would the Pololu controller work for higher currents if a large
>heatsink were installed?
>
>My real question was: How can I tell which frequency gets the highest
>efficiency out of my motor, assuming I try several? Is it simply
>battery life, or will the motor heat up on a bad frequency?
>
>On 6/2/05, dan michaels <oric_dan@...> wrote:
>
>
>>--- In SeattleRobotics@yahoogroups.com, Jay Prince <jayprince@g...>
>>wrote:
>>
>>
>>>I just read the thread on PARTS. The pololu is a 30A motor
>>>controller. That some guy with a competing product can't
>>>differentiate between rated and theoretical maximum, and used that
>>>
>>>
>>as
>>
>>
>>>an excuse to bash pololu, does not mean they are misrepresneting
>>>
>>>
>>their
>>
>>
>>>motor controller.
>>>
>>>
>>>
>>I don't think the issue was about bashing, rather perceived
>>impressions. In the end, it should be noted that the part is
>>probably good for about 6 Amp continuous, maybe a little more,
>>and I think all sides pretty much agreed on this.
>>
>>
>>
>>
>>
>>>The theoretical maximum rating is an objective value that can be
>>>calculated-- while actual performance is something that can vary by
>>>application conditions, and thus is difficult to compare across
>>>product lines. Pololu is pretty clear on what their tests showed
>>>
>>>
>>for
>>
>>
>>>this controller. (And its a darn good price if you want to run at
>>>
>>>
>>12V
>>
>>
>>>or more...)
>>>
>>>As to the original posters question-- how large is large? One rule
>>>
>>>
>>of
>>
>>
>>>thumb for current is to make the motor controllers theoretical max
>>>current equal to the motors stall current-- assuming the application
>>>is one where the motor will not be stalled much at all, or close to
>>>it.
>>>
>>>I think 10khz is very high for robotics use, and probably a very
>>>
>>>
>>good
>>
>>
>>>frequency. My robots us a motor controller in the 1khz range.
>>>
>>>
>>>On 6/1/05, dan michaels <oric_dan@y...> wrote:
>>>
>>>
>>>>--- In SeattleRobotics@yahoogroups.com, Tom Capon <robot256@g...>
>>>>wrote:
>>>>
>>>>
>>>>>Hi all,
>>>>>I have a large DC motor that I want to control via PWM. I am
>>>>>
>>>>>
>>>>planning
>>>>
>>>>
>>>>>on getting a 30A controller from Pololu.com that has 10khz
>>>>>
>>>>>
>>frequency
>>
>>
>>>>>response. When I get it up and running, how can I tell what PWM
>>>>>frequency is best for the motor? I've heard about small DC
>>>>>
>>>>>
>>motors
>>
>>
>>>>>wanting high frequencies, but I'm wondering how this can be
>>>>>
>>>>>
>>>>determined
>>>>
>>>>
>>>>>experimentally for a given motor. thanks! --Tom Capon
>>>>>
>>>>>
>>>>Hi Tom. It's questionable if that's a 30A controller. Take a
>>>>look at the long thread on PARTS that's had to be quashed by
>>>>the moderators twice now.
>>>>
>>>>http://groups.yahoo.com/group/PARTS/messages
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>Visit the SRS Website at Ihttp://www.seattlerobotics.org
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>
>>
>>
>>Visit the SRS Website at http://www.seattlerobotics.org
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>Visit the SRS Website at http://www.seattlerobotics.org
>
>
>
>
>
>
>
>
>

[Non-text portions of this message have been removed]

Visit the SRS Website at http://www.seattlerobotics.org
• You haven t said all you need to say. You owe pololu an apology. Repeatedly you accused them of deception or being misleading. You may not think thats
Message 14 of 26 , Jun 3, 2005
You haven't said all you need to say. You owe pololu an apology.
Repeatedly you accused them of deception or being misleading. You may
not think thats bashing, but you did it again in your most recent
"last word" on the issue.

I don't know what got your hackles up in that other thread... but you
can be assured that I for one will continue to buy pololu products,
but will not consider anything from BD Micro. I don't have an
interest in this particular motor controller, but I can't abide people
misrepresenting other people's products to bash them.

On 6/2/05, Brian Dean <bsd@...> wrote:
> On Wed, Jun 01, 2005 at 06:20:11PM -0700, Jay Prince wrote:
>
> > I just read the thread on PARTS. The pololu is a 30A motor
> > controller. That some guy with a competing product can't
> > differentiate between rated and theoretical maximum, and used that
> > as an excuse to bash pololu, does not mean they are misrepresneting
> > their motor controller.
>
> You must be referring to either me or Dan, since we both tried to
> clarify the difference between rating theoretical maximums on
> "components" and the disconnect with rating end products intended to
> operate as advertised out of the box. As such, I'll just say that I
> respectfully disagree. I don't think either of us was really trying
> to bash anyone, only that the advertising on that one product was
>
> And very much to their credit, Jan of Pololu did respond to the list
> clarifying the position of their product. Hopefully I'm not
> oversimplifying, and please correct me if I'm wrong, Jan indicated
> that the board was intended primarily as a "carrier" for the ST Micro
> chip as an aid to folks not comfortable with SMT soldering, i.e.,
> something along the lines of an experimenters board as opposed to a
> complete h-bridge solution. It is left as an exercise to the buyer to
> figure out and supply appropriate heat sinking, etc.
>
> I do now "get it" after his explanation and I fully accept that. But
> way, at least not to me. Specifically I am referring to the
> description of the product here:
>
> http://www.pololu.com/products/pololu/
>
> See about 1/3rd of the way down where it says:
>
> "NEW! 30-Amp motor driver board
> Need more current? If you've got a larger motor waiting to be
> used, this compact motor driver is just what you need. Connect
> three digital lines to your microcontroller (five if you want
> error condition feedback), and you're ready to go!"
>
> When I read that I'm thinking it is a complete solution, not a carrier
> board that requires additional work and components. At which point
> you can add it to your cart and purchase without ever seeing the more
>
> Further, while the above description would lead one to think so, I
> don't think this h-bridge is very comparable to something along the
> lines of the Devantech 20A h-bridge driver, even putting asside the
> control interface of the Devantech and talking only about the current
> rating:
>
> http://www.robot-electronics.co.uk/shop/Motor_Controllers2008.htm
>
> So that's my last word on that ... I have no intention of debating
> this further, I've said all that I need to say. For some additional
> information and perspective, you might also want to take a look at the
> note Larry posted from the OSMC (Open Source Motor Controller) group.
>
> So ... with that out of the way, if you are interested in
> experimenting with this board or chip and you are expecting to do
> currents in the 8A - 10A or higher range, you will want to look at
> some heat sinking. Also, be sure and grab a copy of the datasheet for
> the chip - you can find that on Pololu's site. As with most ICs like
> this, heat sinking from the top is not going to be very effective.
> Heat sinking this chip really needs to be done from the bottom where
> the junction tabs are exposed - see page 20 of the datasheet for
> possible circuit board layouts to facilitate heat sinking. So it
> would be best to figure out some way to attach heat sink on the bottom
> of the board assuming the PCB provided copper layout as shown on page
> 20 of the datasheet. There's no photo of the bare PCB from Pololu to
> verify this, but I'm assuming they laid the board out as described in
> the datasheet to contact the junctions and the chip is soldered on at
> those points from underneath.
>
> A nice feature of this chip is that it does appear to shut itself down
> when it gets too hot, so it should not be likely that you can damage
> it by not supplying adequate heat sinking.
>
> -Brian
> --
> Brian Dean
> BDMICRO - ATmega128 Based MAVRIC Controllers
> http://www.bdmicro.com/
>
>
> Visit the SRS Website at http://www.seattlerobotics.org
>
>
>
>
>
>
>
• ah, what bashing? what he stated was true. there is no bash in that. the surface mount can only disapate so much heat. it will run 30A but for such a short
Message 15 of 26 , Jun 3, 2005
ah, what bashing?
what he stated was true.
there is no bash in that.

the surface mount can only disapate so much heat.
it will run 30A but for such a short time to make that figure useless
for real world use.

if you can point out the "bashing part" I would be happy to agree with
you on this, I just do not see it.

Tony Brenke

Jay Prince wrote:

>You haven't said all you need to say. You owe pololu an apology.
>Repeatedly you accused them of deception or being misleading. You may
>not think thats bashing, but you did it again in your most recent
>"last word" on the issue.
>
>I don't know what got your hackles up in that other thread... but you
>can be assured that I for one will continue to buy pololu products,
>but will not consider anything from BD Micro. I don't have an
>interest in this particular motor controller, but I can't abide people
>misrepresenting other people's products to bash them.
>
• I agree with Tony and Mr. bdmicro. If a controller can t do an honest 30A continous then they owe all of us an apology. The closest thing he has said to
Message 16 of 26 , Jun 3, 2005
I agree with Tony and Mr. bdmicro. If a controller can't do an honest 30A
continous then they owe all of us an apology. The closest thing he has said
to" bashing " is that they are "missleasding". Hardly what I'd call bashing.
The simple fact that it is advetised as being an off the shelf 30A power
solution isn't true. Unfortunaltly there is not recognised standard as to
how to rate a power controller. Kind of like all those R/C car contollers
claiming 500 to 600A. Sure in theory it will peak that for a fraction of a
micro second. Hardly what I'd call a practicle rating. To me if it says
30A then I should be able to hook up a 30A load and walkaway have a pizza
and a pitcher of beer come back and fine the magic smoke is still contained
in the controller and it's still running. Some how I don't think the pizza
and beer standard will ever take I don't think too many will aregue it's
merits.

----Original Message Follows----
From: "tbrenke@..." <tbrenke@...>
To: SeattleRobotics@yahoogroups.com
Subject: Re: [SeattleRobotics] Re: Motor PWM frequencies
Date: Fri, 03 Jun 2005 21:00:04 -0700

ah, what bashing?
what he stated was true.
there is no bash in that.

the surface mount can only disapate so much heat.
it will run 30A but for such a short time to make that figure useless
for real world use.

if you can point out the "bashing part" I would be happy to agree with
you on this, I just do not see it.

Tony Brenke

Jay Prince wrote:

>You haven't said all you need to say. You owe pololu an apology.
>Repeatedly you accused them of deception or being misleading. You may
>not think thats bashing, but you did it again in your most recent
>"last word" on the issue.
>
>I don't know what got your hackles up in that other thread... but you
>can be assured that I for one will continue to buy pololu products,
>but will not consider anything from BD Micro. I don't have an
>interest in this particular motor controller, but I can't abide people
>misrepresenting other people's products to bash them.
>

Visit the SRS Website at http://www.seattlerobotics.org
• ... http://www.superdroidrobots.com/shop/category.asp?catid=5 ... ...
Message 17 of 26 , Aug 29, 2005
--- "tbrenke@..." <tbrenke@...> wrote:

> dont mess with it.
>
> look at oune of these.
>
http://www.superdroidrobots.com/shop/category.asp?catid=5
>
> MD03 Devantech 20A 50V Motor Driver
>
<http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
> The MD03 is a medium power motor driver, designed to
> supply power beyond
> that of any of the low power single chip H-Bridges
> that exist. Main
> features are ease of use and flexibility.
> More Detail
>
<http://www.superdroidrobots.com/shop/item.asp?itemid=471&catid=5>
>
> TE-036-000
>
>
>
>
> Tom Capon wrote:
>
> >My motor has about a 25A stall current at 12V.
> >
> >Would the Pololu controller work for higher
> currents if a large
> >heatsink were installed?
> >
> >My real question was: How can I tell which
> frequency gets the highest
> >efficiency out of my motor, assuming I try several?
> Is it simply
> >battery life, or will the motor heat up on a bad
> frequency?
> >
> >On 6/2/05, dan michaels <oric_dan@...> wrote:
> >
> >
> >>--- In SeattleRobotics@yahoogroups.com, Jay Prince
> <jayprince@g...>
> >>wrote:
> >>
> >>
> >>>I just read the thread on PARTS. The pololu is a
> 30A motor
> >>>controller. That some guy with a competing
> product can't
> >>>differentiate between rated and theoretical
> maximum, and used that
> >>>
> >>>
> >>as
> >>
> >>
> >>>an excuse to bash pololu, does not mean they are
> misrepresneting
> >>>
> >>>
> >>their
> >>
> >>
> >>>motor controller.
> >>>
> >>>
> >>>
> >>I don't think the issue was about bashing, rather
> perceived
> >>impressions. In the end, it should be noted that
> the part is
> >>probably good for about 6 Amp continuous, maybe a
> little more,
> >>and I think all sides pretty much agreed on this.
> >>
> >>
> >>
> >>
> >>
> >>>The theoretical maximum rating is an objective
> value that can be
> >>>calculated-- while actual performance is
> something that can vary by
> >>>application conditions, and thus is difficult to
> compare across
> >>>product lines. Pololu is pretty clear on what
> their tests showed
> >>>
> >>>
> >>for
> >>
> >>
> >>>this controller. (And its a darn good price if
> you want to run at
> >>>
> >>>
> >>12V
> >>
> >>
> >>>or more...)
> >>>
> >>>As to the original posters question-- how large
> is large? One rule
> >>>
> >>>
> >>of
> >>
> >>
> >>>thumb for current is to make the motor
> controllers theoretical max
> >>>current equal to the motors stall current--
> assuming the application
> >>>is one where the motor will not be stalled much
> at all, or close to
> >>>it.
> >>>
> >>>I think 10khz is very high for robotics use, and
> probably a very
> >>>
> >>>
> >>good
> >>
> >>
> >>>frequency. My robots us a motor controller in
> the 1khz range.
> >>>
> >>>
> >>>On 6/1/05, dan michaels <oric_dan@y...> wrote:
> >>>
> >>>
> >>>>--- In SeattleRobotics@yahoogroups.com, Tom
> Capon <robot256@g...>
> >>>>wrote:
> >>>>
> >>>>
> >>>>>Hi all,
> >>>>>I have a large DC motor that I want to control
> via PWM. I am
> >>>>>
> >>>>>
> >>>>planning
> >>>>
> >>>>
> >>>>>on getting a 30A controller from Pololu.com
> that has 10khz
> >>>>>
> >>>>>
> >>frequency
> >>
> >>
> >>>>>response. When I get it up and running, how
> can I tell what PWM
> >>>>>frequency is best for the motor? I've heard
> >>>>>
> >>>>>
> >>motors
> >>
> >>
> >>>>>wanting high frequencies, but I'm wondering how
> this can be
> >>>>>
> >>>>>
> >>>>determined
> >>>>
> >>>>
> >>>>>experimentally for a given motor. thanks!
> --Tom Capon
> >>>>>
> >>>>>
> >>>>Hi Tom. It's questionable if that's a 30A
> controller. Take a
> be quashed by
> >>>>the moderators twice now.
> >>>>
> >>>>http://groups.yahoo.com/group/PARTS/messages
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>Visit the SRS Website at
> Ihttp://www.seattlerobotics.org
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>
> >>
> >>
> >>Visit the SRS Website at
> http://www.seattlerobotics.org
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >
> >
> >Visit the SRS Website at
> http://www.seattlerobotics.org
> >
> >
> >
> >
> >
> >
> >
> >
> >
>
>
> [Non-text portions of this message have been
> removed]
>
>
>
> Visit the SRS Website at
> http://www.seattlerobotics.org
>
>
> SeattleRobotics-unsubscribe@yahoogroups.com
>
>
>
>
>

____________________________________________________