coil drive & diodes

Hey gang, I'm designing a large board that will control 500mA coils in an H-bridge configuration. The coils are making little electromagnets to pull & repel, not spinning a motor. I've got these MOSFETs designed in right now

N-channel MOSFET http://www.digikey.com/product-detail/en/AOI514/785-1690-5-ND/3973570

P-channel MOSFET (edit - correct part posted) [http://www.digikey.com/product-detail/en/NDP6020P/NDP6020P-ND/1055922]http://www.digikey.com/product-detail/en/NDP6020P/NDP6020P-ND/1055922]http://www.digikey.com/product-detail/en/NDP6020P/NDP6020P-ND/1055922](http://www.digikey.com/product-detail/en/NDP6020P/NDP6020P-ND/1055922)

Do you think external diodes are needed across these MOSFETs as well?

http://www.modularcircuits.com/blog/articles/h-bridge-secrets/h-bridges-the-basics/

Hm.

Any current that goes through the MOSFET intrinsic body diode is also heat that must be dissipated.

Which is only a concern for high speed PWM - is PWM involved at all? Besides with an H-bridge you can use fast-decay mode and finesse the issue.

What's the supply voltage? Do you realize both apparent links are one link to n-FET only, despite saying p-FET...

Thanks, fixed the link for the P-FET. Forum makes that a mess.

Parts will not be PWM'ed, just turned on for 1/2 second or so. Will be an array of 64 H-bridges, so any one part would be on for up to 8 pulses only. Current is 500mA, 5V source.

In the meantime, I've added this 1206 size diode across each part in case other MOSFETs are used. http://www.digikey.com/product-detail/en/ACDBN160-HF/641-1616-1-ND/5013821

Have you considered placing an appropriate voltage MOV across the coil?

No need for diodes in your scenario, a few switching cycles isn’t going to dump
much heat in the body diodes anyway and the MOSFET packages are designed
to cope with 2.5W

It was asked, and answered. I left it up to the OP to determine if that is an issue.

Thanks for the replies. The final design has the diodes, the board user can include them, or not, when populating. Tricky part was finding a layout that allowed for 64 coils in 8x8 matrix with 50mil traces - they take up a lot of room. Ended up spreading things out quite a bit on 100cm x 100cm card, with P-channels for sourcing A to B, and N-channels for sink from B to A at the top of the board, and the mating N-channels for sinking A to B, and P-channels for sourcing B to A at the bottom of the board. All thru hole MOSFETs, I think routing would have been extremely difficult with SMDs and having to add vias for their pins.

RFID chessboard?