Maker Pro
Maker Pro

Flyback diode?

jsondag

May 28, 2010
5
Joined
May 28, 2010
Messages
5
I am planning to use 2 of this MOSFET to control a motor, using an arduino. The motor is a 24V 140A motor .

The question is do I need a flyback diode, it looks like one is included in the package, and it says it's avalanche rated. Even if it is, would I be better off putting one on the motor anyway, so less power is dissipated in the MOSFET.

Also would I need a 140A schottky?

I have it all hooked up and am aching to hit the button, but I only have 2 MOSFETs on hand, and don't want to blow them.

Thanks.
 

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
It depends on what configuration the transistors are connected.
Half-bridge (bi-directional) with +/- supply needs no extra diodes.
Common drain (source follower, motor on source & to -) needs no extra diode.
Common source (motor on drain & to +) needs a diode on the motor.
The diode doesn't need to be a schottky nor to be rated at the full motor current. It's only supposed to catch the short pulses on each turn-off of the transistors.
Thus it's useful if it has a high pulse rating, but the average rating depends on the switching frequency.
Also, if you have not determined that the transistors will take the full stall current of the motor, I'd ramp up the duty cycle according to the motor RPM.
I'm confused about what the Id25 rating means, haven't seen that before.
 

jsondag

May 28, 2010
5
Joined
May 28, 2010
Messages
5
I have the drain to one side of the motor (chassis ground), the source to battery negative, and the other side of the motor to battery +.

Right now I am using only 12V not the 24V.
I think the Id25 is really Ids (current drain to source)
That's how it was shown on digikey anyway.

Not sure why common source, would be different than common drain, as the spike would still hit the reverse diode in the MOSFET, in the same way.
But that is why I ask, to be sure...

Edit: The switching speed is about 500Hz. Not sure what kind of diode would be best, I had simply heard that schottky diodes were better because of the lower voltage drop...
 
Last edited:

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
On common source the flyback voltage goes positive, beyond supply voltage, with nothing to stop it except transistor breakdown voltage.
On common drain the flyback voltage tries to go negative, below ground, but is stopped by the transistor then getting a gate voltage - thereby conducting.
The internal reverse diode is not conducting in either case. That only happens in the half-bridge configuration - and then in the transistor opposite of the one shutting off.

Schottkys are better for that reason, but it's not a property important to the flyback catch. In fact you could even be using a transient protector diode (zener) of up to 16V.
It's only purpose is to keep the flyback voltage lower than the 40V breakdown voltage of the transistor. Knowing the motor inductance would allow a current calculation.
The higher the transient voltage the shorter the flyback current spike will be. Use primarily a fast switching diode, but even an ordinary rectifier diode will do the job.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I'm confused about what the Id25 rating means, haven't seen that before.

It's not made real clear in the specs, but I assume it means Id at Vgs=25 volts.

If so, and even if not, I would suspect that you'd want to have an appropriate gate driver to ensure the mosfet gets turned on and off quickly. At those currents, being tardy in switching will very quickly destroy the device.
 

jsondag

May 28, 2010
5
Joined
May 28, 2010
Messages
5
You're completely right of course, I have no idea why I thought that the internal diode, would in any way be across the motor. Brain fart I guess.

I will see if there some way I can measure the inductance, my meter does it, but when I turn on the oscillator in the meter (needed to test inductance) the meter shuts off, and won't turn on until I turn off the oscillator.

I guess it's back to the old function generator, oscope, and an RC circuit standby.


This page claims that I will need a main capacitor. I would have thought the battery (110AH deep cycle) could hold it's voltage, and clamp the reverse spike on it's own (as long as there was a flyback diode). I guess on microsecond time scales it may not, especially with the inductance of the wires.

Not sure what kind of capacitor I would need for that. EDIT: The Open Source Motor Control (OSMC), uses a 1200uF electrolytic. It can handle 160A so....

It also shows using another MOSFET as the flyback protection. Which is a though, if I am going to have to order more parts anyway.....

EDIT: I also can't for the life of me figure out, how he's claiming that circuit does regenerative breaking. Reverse current from the motor is shorted throught the flyback MOSFET and lost as heat, and forward voltage when the drive MOSFET is off, wouldn't do anything but develop voltage across the flyback mosfet, which is always off because the gate is grounded to source.

EDIT2:I think I finnally "get" the regenerative breaking. If the Drive FET is off a voltage greater than the supply will develop on the motor, and that would put the drain of the Drive FET negative with respect to ground, meaning power would flow through the reverse diode in the Drive FET, allowing the side of the motor attached to V+ to go higher than V+ due to reverse voltage across the Drive FET, charging the battery. Is that right?
 
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The idea is that the capacitor presents a lower resistance load and/or source (for spikes) than the battery does. Remember that for these purposes "the battery" means not only the internal resistance (and possibly impedance) of the battery, but also resistances (and impedances) of battery leads and other wiring.

Alternatively, the capacitor may be present to limit the variation on the supply rail to the rest of the circuit. In this case it will almost certainly be isolated using a resistor or a relatively large value inductor.
 

jsondag

May 28, 2010
5
Joined
May 28, 2010
Messages
5
Out of these 3 caps, are any of them suitable, and which is the best option?
Or all they all insane overkill in every respect. (that's ok :D)

Also, connecting a large cap across the main supply will cause a large current draw when you first connect the battery. Will that hurt anything, apart from the nice sparks that you'll get....

http://cgi.ebay.com/Hitachi-HCG-FA-...tem&pt=LH_DefaultDomain_0&hash=item335d1a9d73

http://cgi.ebay.com/Capacitor-Elect...tem&pt=LH_DefaultDomain_0&hash=item414e824d26

http://cgi.ebay.com/Aero-M-CGS-2800...tem&pt=LH_DefaultDomain_0&hash=item22f5ec6af7
 
Last edited:

Resqueline

Jul 31, 2009
2,848
Joined
Jul 31, 2009
Messages
2,848
The page you referred to is very useful. I've come across it before but forgot about it.
Using a MOSFET instead of a flyback diode is smart. It makes the output a square wave. At 50% duty you have 50% battery voltage out. And thinking of the motor as an inductance in series with an EMF generator it becomes clear that if the EMF is greater than the duty cycle then current flows from the motor and into the supply, actually charging the battery. Ingeniously & simple.

I don't consider those capacitors overkill but I'm unable to decide which one's are best. They're old and datasheets won't be easy to find but that's where it says anything about the ripple current rating. Physical size is a factor though. A Sangamo AF size is rated at only 25A but I was unable to find data on the BF size.
Connection with sparks shouldn't hurt anything but you could use a resistor for "soft-starting" it.
Remember to keep +&- leads close together to reduce inductance like that page says.
 

jsondag

May 28, 2010
5
Joined
May 28, 2010
Messages
5
I found out that for CGS type, a 40VDC cap at 28,000uF has a RMS ripple current rating of about 10A @120hz & +85 ºC

with multipliers of

Ambient Temperature
+45 ºC +55 ºC +65 ºC +75 ºC +85 ºC
2.24 2.00 1.73 1.41 1.00

60 Hz 120 Hz 300 Hz 1000 Hz 10 kHz
0.9 1.0 1.15 1.25 1.30

I would be running it at 500Hz, and <40 ºC
so conservatively... 10A * 1.15 * 2.24 = 25A RMS ripple current.

I don't know what the RMS current would be since it only needs to smooth dips, and spikes for a small portion of each cycle, though my intuition says this is enough....
 
Top