MOSFETs in h-bridges

[loosewire]

I need a driver for a 500w motor (14A max, 24vDC). I have used the L298N driver for smaller motors and I was hoping to find something like that for the higher current, but have been unable to find any. (Does anyone know of any?)

I have found a lot of articles about building drivers for higher currents. Of course these all use mosfet h-bridges. The only times I've ever used mosfets in the past is when I was following a parts list. Now that I have to actually figure out which mosfet(s) to use, I am getting blown away by the shear number of them on the market.

Without prior knowledge of a bunch of different mosfets, how do you go about picking one for the job? Is there a resource on the internet that compares the specs of different mosfets that are on the market? Does anyone have any recommendations for a book on mosfets that would make this task easier? Even a list of commonly used mosfets for different current and output voltage ranges would be helpful.

I am aware that, since these are driving motors, the source-drain resistance needs to be as low as possible. I will also be using PWM, so the switching speed is important. Anything else specific to h-bridges that I should be looking for in a good mosfet?

As for shoot-through: I am thinking that I should find the correct mosfets for the circuit, and then design around them for shoot-through protection. Or should that figure into the mosfet selection process?

Also, what advice do you have for using a mosfet controller chip. As far as I can tell, these are used primarily to boost voltages from logic chips to levels that are suitable for a mosfet gate. I would probably use optoisolators unless there is some great benefit to using a controller chip.

Thanks in advance for any tips.

 

[philba]

It sounds like you are thinking about the right things. One more spec item to be aware of is Rds at Vgs. Logic level mosfets will have decent Rds at 4.5 or 5V Vgs but that's not everything.

Another thing to consider is whether you will use P or N MOSFETs for the high-side drivers. In order to use N MOSFET for highside, you will need to have a voltage above the motor supply V since the source will be just a bit below the the supply V. (I think it needs to be Vgs above the supply for the given Rds you want). P MOSFETs don't have that problem but may be pricier than the equivalent N MOSFET. I think the prices have come down a bit so it may not be an issue.

Switching speed is an issue for heat dissipation. the more time you spend in the transition zone, the worse your heat will be but I'm not sure the best way to model that.

Also, look carefully at the "Maximum Safe Operating Area" graph on the DS. I am always suprised by how small the safe area is when you consider the "marketing" amp ratings on the front of the DS.

good luck
Phil

 

[Papabravo]

Without knowing anything more than what you have told me I would start with a part that can handle twice the motor load current. Then I would build one prototype to test the other concepts. In the process you will find that the part you have chosen will meet your requirements or it will not.

If it does not, then you will have a solid foundation for making a second choice. If you spend a great deal of time searching for the perfect component based on specification sheets you might consume more time than building two different prototypes.

Understand the difference between the "best result possible", and the "best possible result".

As a side note you might learn a great deal by downloading the free SPICE simulator from Linear Technology called LTSpice or perhaps SwitcherCadIII. They have an active Yahoo discussion group where experts in many fields hang out.

 

[Dr.EM]

Switching speed is mainly dependant on gate capacitance, it can be quite high in MOSFETS, sever hundered pF. To switch faster, a high gate current can be used to overcome this capacitance and turn the fet full-on in the shortest time. There are still rise and fall times though, and these are just inherent to any semiconductor. To minimise this affect, a slower PWM frequency will help as less time per second will be spent rising and falling as it were, and more spent on and off.