DRV8462: Overheating issues

Here is the snippet of our layout.  

Hi Chris,

What is the supply voltage used? Thanks for sharing the layout. Looking at the break in the copper - see below marked with white outline - heat is not going to spread to a larger area resulting in the driver running very hot. Our recommendation would be to maximize the top layer copper and bottom layer copper area to maximize the power dissipation. You can use our EVM top layer as a reference. Altium project can be downloaded from the tool webpage. This is definitely a PCB power dissipation issue when compared with the EVM PCB. I hope this helps. Thanks.

Regards, Murugavel

Hi Murugavel, 

Thanks for the quick reply.  The voltage is 48V.  I will agree that this is a power dissipation issue.  I was revisiting the DRV8462 datasheet and layout recommendations and saw a couple of items...but first, here is the snippet of the bottom layer.  


...and here is the layout guidelines from TI: 

I think we're following all the guidelines that are offered.  The only thing that I see that we're not following is having layers at 2oz...as ours are all 1oz copper planes. And again for reminder, we added an external heatsink to cover the driver and additional copper pads (using thermal pad).  

Also, I should add that our temperature was recorded at the microcontroller (U4, bottom of the image).  

Per your reply recommendation, I can certainly adjust the surrounding flooded plane to become ground (beyond just the area you circled).  The intent for how it is currently is to attempt to impedance match layers (power, ground, power, ground).  I don't think that altering this scheme will negatively affect our circuit performance...although we've not gotten to the CE testing specs yet. 

Thanks. 

Hi Chris,

Altering the layout will improve the thermals. Bottom layer looks okay. 2oz may be needed given the area around the driver is limited. Yes I noticed you mentioned external heatsink. Where was it placed? If on the top surface of the IC it may not be of much use. If you had a bare plated surface right underneath the device in the bottom layer and stick a heatsink on it, it would work better dissipating the heat from the chip via the thermal pad. All these should help. Thanks.

Regards, Murugavel

Ok, so you're recommending that I look into moving the driver to the bottom and adjust the (above) heatsink to contact an exposed (larger) pad directly above the driver.  I can do that.  Currently, the heatsink is made to hit 3 points: the ceramic IC body and the two exposed (smaller) areas on either side of the device.  If I'm going to be throttled by my top layer and how heat is going to be held there, I'll look at a redesign.  If I had access to convection cooling (with air) I'd go that route but that's not an option.  Would you confirm again your recommendation you posted above, does that still stand as a good route (pulling heat through vias to a heatsink on the opposite side of the driver)?  

Thanks much.

Hi Chris,

Yes that's what I meant.

You have another option as well which may be a slightly up version of this device. DRV8462 in DDV package - soon to be released. This has a thermal pad on the top of the IC for a bolt on heatsink. The IC is flipped which means the footprint will be a mirror image. Details in the DRV8462 datasheet. 100% software compatible but can support up to 10A because of the bolt on heatsink option.  

Regards, Murugavel