PWM MOSFET driving large DC motor

[Peter "Sci" Turpin]

I'm trying to get a motor-controller built for my mill. The old one went
bang.

For ease and reliability, I'm looking at driving a 220vDC 1.75HP
treadmill motor by PWM on half-wave rectified mains (170-180vDC). The
speed reduction will not be a problem.

I was considering this circuit layout for controlling by
micro-controller with opto-isolation..
http://electronics.stackexchange.com/questions/17116/how-to-drive-a-mosfet-with-an-optocoupler

Scaled up higher-rated components of course.

If I've got it right, the two motor-side resistors are holding the
optotransistor side of the coupler in the middle of a voltage divider,
so they need to selected so one side will drain the MOSFET gate and the
other side will keep the relative voltage within the +/- tolerance
relative to the source (20-30v seems common). But also so the
optotransistor runs at the correct output voltage. Since output voltage
needs to be 3-5v commonly, seems well within the relative voltage but I
am having trouble keeping track of what needs to be relative to what.

Also the page above says that circuit maxes out at about 10KHz, and I've
seen it mentioned larger motors really need higher (seen 18KHz quoted,
but seems almost personal-pref based on sound). I suspect that's just
based on those specific components though, but if not I may need to use
a dedicated gate-driver instead? Like one of these?
http://www.eeweb.com/blog/avago_technologies/gate-drive-optocoupler-basic-design-for-igbt-mosfet


I'm going to work on an assumption of a 2Kw motor so there's some
tolerance in the spec, so 12A.

Thoughts and recommendations welcome. I'd rather not blow another motor
or end up with components accidentally tack-welded to each-other.

[Adrian Godwin]

It might work. Fairly likely to go through a few FETs and PCBs trying. I'm not great on power electronics so I'd strongly recommend a 3-phase motor and an inverter drive.

You can get them quite reasonably here : http://www.homeworkshop.org.uk/ - a guy called Gavin Oseman often advertises them, sometimes with a matching motor.

--
You received this message because you are subscribed to the Google Groups "London Hackspace" group.
To unsubscribe from this group and stop receiving emails from it, send an email to london-hack-space+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Peter "Sci" Turpin]

My mill still has it's induction motor with it, but regrettably with it
fitted it makes it too large for my workshop. Using the DC motor allows
me to strip off a lot of unnecessary parts and fit it in place.

On 07/01/2014 22:16, Adrian Godwin wrote:
> It might work. Fairly likely to go through a few FETs and PCBs trying.
> I'm not great on power electronics so I'd strongly recommend a 3-phase
> motor and an inverter drive.
>
> You can get them quite reasonably here : http://www.homeworkshop.org.uk/
> - a guy called Gavin Oseman often advertises them, sometimes with a
> matching motor.
>
>
>
> On Tue, Jan 7, 2014 at 9:36 PM, Peter "Sci" Turpin <s...@sci-fi-fox.com
> <mailto:s...@sci-fi-fox.com>> wrote:
>
> I'm trying to get a motor-controller built for my mill. The old one
> went bang.
>
> For ease and reliability, I'm looking at driving a 220vDC 1.75HP
> treadmill motor by PWM on half-wave rectified mains (170-180vDC).
> The speed reduction will not be a problem.
>
> I was considering this circuit layout for controlling by
> micro-controller with opto-isolation..

> http://electronics.__stackexchange.com/questions/__17116/how-to-drive-a-mosfet-__with-an-optocoupler

> <http://electronics.stackexchange.com/questions/17116/how-to-drive-a-mosfet-with-an-optocoupler>
>
> Scaled up higher-rated components of course.
>
> If I've got it right, the two motor-side resistors are holding the
> optotransistor side of the coupler in the middle of a voltage
> divider, so they need to selected so one side will drain the MOSFET
> gate and the other side will keep the relative voltage within the
> +/- tolerance relative to the source (20-30v seems common). But also
> so the optotransistor runs at the correct output voltage. Since
> output voltage needs to be 3-5v commonly, seems well within the
> relative voltage but I am having trouble keeping track of what needs
> to be relative to what.
>
> Also the page above says that circuit maxes out at about 10KHz, and
> I've seen it mentioned larger motors really need higher (seen 18KHz
> quoted, but seems almost personal-pref based on sound). I suspect
> that's just based on those specific components though, but if not I
> may need to use a dedicated gate-driver instead? Like one of these?

> http://www.eeweb.com/blog/__avago_technologies/gate-drive-__optocoupler-basic-design-for-__igbt-mosfet

> <http://www.eeweb.com/blog/avago_technologies/gate-drive-optocoupler-basic-design-for-igbt-mosfet>
>
>
> I'm going to work on an assumption of a 2Kw motor so there's some
> tolerance in the spec, so 12A.
>
> Thoughts and recommendations welcome. I'd rather not blow another
> motor or end up with components accidentally tack-welded to each-other.
>
> --
> You received this message because you are subscribed to the Google
> Groups "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it,

> send an email to london-hack-space+unsubscribe@__googlegroups.com
> <mailto:london-hack-space%2Bunsu...@googlegroups.com>.
> For more options, visit https://groups.google.com/__groups/opt_out
> <https://groups.google.com/groups/opt_out>.

>
>
> --
> You received this message because you are subscribed to the Google
> Groups "London Hackspace" group.
> To unsubscribe from this group and stop receiving emails from it, send

> an email to london-hack-sp...@googlegroups.com.

[Steve Todd]

Don't forget a nice hefty reverse biased diode across the motor. When you switch it off you get a big lump of back EMF which will otherwise knacker your FET.

[Jim MacArthur]

The circuit there will switch the MOSFET's gate between 0 and about
9.7V which presumably is enough to switch the FET fully on (it varies
by device, but 10V is enough for any FET I've dealt with). The gate
needs to be positive wrt the source, but since this is just a low side
driver, source is always at 0V so there's nothing to worry about. It
gets more complicated if your FET is on the positive side of the
motor, as in a H-bridge.

The thing with MOSFETs is that you want to change the gate voltage as
quickly as possible - fully on is fine, fully off is fine, but in
between that there's a lot of heat dissipation. If your MOSFET gets
hot, you might want to lower the values of both the resistors in that
divider, keeping them in the same ratio, depending on how much current
your optocoupler can switch. Or you could just get a gate driver IC.

I'd aim for at least a 25A MOSFET as the starting current will be much
higher than the motor's rated power.

Take my advice with a pinch of salt, as my last attempt to make an
H-bridge went up in flames quite spectacularly. From then on I've
always bought motor controllers.

[Peter "Sci" Turpin]

Of course. I plan to use the same sort the original treadmill did. One
of those big dual-diode modules with the inputs tied together.

[Steve Todd]

Hang a Schmitt Trigger between the opto-coupler and the FET. That should minimise rise and fall times, and will also boost the current available. Using 4000 series logic (from memory a 4106 should do) will let you run at between 10-12V so you should have no level conversion problems.

Regards,

Steve

> You received this message because you are subscribed to a topic in the Google Groups "London Hackspace" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/london-hack-space/FjYQHOtkLbw/unsubscribe.
> To unsubscribe from this group and all of its topics, send an email to london-hack-sp...@googlegroups.com.

[Peter "Sci" Turpin]

If I read the descriptions correctly, using an optotransistor based
coupler, I could build the coupler in as the first stage of a Schmitt
trigger?
Current isn't a concern though, the MOSFET only seem to need a voltage
potential, no actual current flow. And the data sheets are saying about
5v maximum.
That's actually where it gets a little confusing for me though. 5v
relative to what point? If it was to source then why list the maximum
difference between source and gate as +/-30v? And relative to drain
makes no sense, because then you'd be switching with more voltage than
you're driving with.
I'm probably visualising it wrong. These sort of confusions usually mean
I've gotten two items mixed up or I'm misinterpreting a single term. I'm
guessing it's the +/-30v difference bit. Is it actually an insulation
value and not an operational limit?

[Steve Todd]

You're missing part of the equation. The gate of a power MOSFET has significant capacitance, so you tie it to ground with a resistor so that it will quickly switch off. Because of this you need to be able to draw current from your input. The Schmitt Trigger chip ensures that the rise and fall time of the signal from the opto coupler is as close to zero as is possible (you're never going to get infinite slew, but a few nanoseconds is certainly possible) and lets you use a smaller resistor between gate and ground.

You can certainly try it without, but the point was your FET heats up during the intermediate switching state (when it is moving from infinite to low resistance), so this is a cheap and simple way to minimise that.