3000W h-bridge vehicle motor control

[Beurnii]

I want to control 2 motor (separately) and they are kind big, their stall amp is about 100amps so I guess I would use around 30amps continuous with a mini-fridge sized vehicle.

I need to buy the components for the circuit and I think that I have all I want but I want to be sure to not forgot anything.

I already have the two motor : EV Warrior
They will be control with high power logic mosfet (IRF3708PbF) in an H-bridge and there will be some MBRF30100CT SCHOTTKY diode as flyback diode.

I control the mosfet with my arduino uno and some 220 Ohms resistor on the gate. Every mosfet and diode will be stock on an random peice of metal (as heatsink) with some thermal conductive pad to keep them cold. I may try to find a way to compress them on the pad if necessary.

I already have a medium sized 12v lead battery tha should do the job. I may double it up and run everything on 24v if needed.


I just want to be sure that I'm not forgeting anything because I dont want to pay twice the shipping. I'm buy from Digi-Key and I'm in Canada. I'm also buying everything 12x in case I burn some.


It may look stupid, but as a beginner, I would had love to find a post like this at the first place and I'm sure I'm not the only one in my situation.

Thanks,
Bernard

[MarkT]

Well yes, you need MOSFET drivers, and all MOSFET driver chips assume MOSFETs with 12V or so gate
drive, yet you have logic level MOSFETs with an absolute max Vgs of 12V, so they cannot work from
standard MOSFET drivers.

Look at the datasheets for chips like IRS2001, FAN7380, HIP4081 to see how its done.

At these power levels I would not trust to driving direct from a logic signal as the drain-gate
capacitance could easily overcome your logic output and blow up the logic chip on transients.

Furthermore without high-low MOSFET driver its hard to drive the top-side switches easily.

[ What you are attempting is fraught with issues, such bridges will simply pop MOSFETs if
anything goes wrong, protection circuitry is important and MOSFET drivers give a first level
of defence - high power brings high risk ]

[TomGeorge]

Hi,

 random piece of metal (as heatsink)

Not a good idea, a properly made heatsink and component layout will be needed to help with current and heat flow.

The conductors to the MOSFETS and terminals will need to be layed out efficiently and you will need to have some form of current monitoring to prevent peak load currents from damaging your circuit.

There is a reason why the electric scooter and trikes are so expensive, a lot of the cost is in the motor controller.

Good luck, Tom ......

[Beurnii]

Well yes, you need MOSFET drivers, and all MOSFET driver chips assume MOSFETs with 12V or so gate
drive, yet you have logic level MOSFETs with an absolute max Vgs of 12V, so they cannot work from
standard MOSFET drivers.

Look at the datasheets for chips like IRS2001, FAN7380, HIP4081 to see how its done.

At these power levels I would not trust to driving direct from a logic signal as the drain-gate
capacitance could easily overcome your logic output and blow up the logic chip on transients.

Furthermore without high-low MOSFET driver its hard to drive the top-side switches easily.

[ What you are attempting is fraught with issues, such bridges will simply pop MOSFETs if
anything goes wrong, protection circuitry is important and MOSFET drivers give a first level
of defence - high power brings high risk ]

The HIP4081A seem a good bet, the Application Block Diagram is pretty strait foward, but the problem is with the PWM mode, I dont understand how is it working and there is not a lot for information online about this driver. I guess the timing is important because there is only one imput.


And with a mosfet driver like this, I can use regular high power mosfet such as the IRFZ44N

[raschemmel]

It may look stupid, but as a beginner, I would had love to find a post like this at the first place and I'm sure I'm not the only one in my situation.  

Maybe if you changed your post title to "3000W Electric vehicle motor control " you might get some replies by people
who have done the same thing.

[MarkT]

It is far from straight forward as I have warned - get anything wrong and *bang* exploding
MOSFETs, burnout MOSFET drivers.

You almost certainly want to think about using fast opto-isolators into the HIP4081 and
all the timing (such as PWM) is your problem still.  Because of the bootstrapped high side
gate supply you cannot leave the low side switches off for more than a ms or so at a time,
otherwise the high side driver bootstrap capacitor will discharge.

Fortunately the HIP4081 has extra circuitry to help here, but most MOSFET drivers supporting
bootstrap don't.

If you don't have an oscilloscope and a good current-limited bench power supply you are not equiped to
tackle such a project really...

[Beurnii]

If you don't have an oscilloscope and a good current-limited bench power supply you are not equiped to
tackle such a project really...

yeah you are right, I'm not ready for the chip or any really high power stuff, I will think of a solution.

Seriously, before starting, I never though that controlling a motor would be so complicated compared to very small motor that everyone have use in high school or small rc car.

Thanks guys
Bernard

[raschemmel]

I would throw a CT current probe in that bag of things to have when messing with high power.
(obviously a Tektronics A621 wouldn't hurt either)

[zoomkat]

You might check the below discussion for info.

http://forum.arduino.cc/index.php?topic=324134.msg2239306#msg2239306

[Beurnii]

I found this online and I think that I will try it for fun. But I have a small concern:

When one side is power up (INPA, INPB) and the mosfet and turning close or open. Is there a small moment where both mosfet will be partially open and so creating a short? How much time does it take to a mosfet (p and N) to completely open or close.

[CrossRoads]

Could be a while - tens or hundreds of nanoseconds.
https://www.fairchildsemi.com/datasheets/ND/NDB6020P.pdf
Look at the switching characteristics on sheet 2 for example, and figure 11 & 12.

I would prefer separate highside & lowside control lines to ensure the on/off timing can be controlled.  The motor won't care too much, but you can blow a transistor pretty quick.

[raschemmel]

I know people with a lot more electronics experience than you would think twice before trying build their own DIY 3000W H-BRIDGE. As already mentioned, you might as well jump into a pit of snakes.

But if you're going to jump, you should learn something about "Shoot-Through" and how to prevent it.


Google Search:: Shoot-Through Protection

Shoot-Through Protection

HINT: You need some ADAPTIVE DEADTIME between the closing of the low side switching device and the opening of the high side switching device so that there is never a time when they are both on.

[MarkT]

I found this online and I think that I will try it for fun. But I have a small concern:

When one side is power up (INPA, INPB) and the mosfet and turning close or open. Is there a small moment where both mosfet will be partially open and so creating a short? How much time does it take to a mosfet (p and N) to completely open or close.

You say "partially open" when you mean "partially on" - but its worse than this, there is potentially a time
when both high and low-side devices are fully on, which is called "shoot-through" and is very bad.

The high- and low-side switches on the same arm must be switched at separate times with enough
dead-time between so one can fully turn off before the other conducts.  Most decent MOSFET driver
chips give you some dead-time and thus prevent shoot-through (as well as other comforts).  There's
little point designing a MOSFET H-bridge without using MOSFET drivers.

[raschemmel]

Shoot-Through Protection

[mauried]

These devices are pretty good for driving hi power Mosfets.
http://www.irf.com/part/_/A~IR2102
They provide both hi and lo side drive, have in built shoot thru protection, and seem to be pretty indestructable.
You will need 2 for a H bridge, but they are cheap, around $2 each.