2
\$\begingroup\$

Update:

A Renesas Tech pointed me to the documentation for the HIP4080, which is basically a two leg version of the HIP4086, but with much more thorough documentation. I now have three working theories for the failure:

1. My bootstrap capacitors are too small for the PWM frequency.
2. The diode that I used are not fast enough.
3. PWM frequency was too high.

All can, apparently, cause insufficient power to be available to handle the high side MOSFET gate charge, thereby causing more of the current to be drawn from the HIP4086. Also, given that (3) is known to cause heating issues with the 4080, I think a redesign is warranted.



I am at my wit's end on this one and am really hoping someone can point me in the right direction.

I have a custom designed 3 phase motor driver using the HIP4086A driver chip, the schematic is based heavily off the driver/MOSFET section of the demo board, found on page 17 of the demo board manual. Only four differences exist between that design and the one I am using, all other components are identical those used on the demo board:

  1. Substituted MBR2H200SFT1G for the ES1B diodes
  2. 3.3v logic is used instead of 5v.
  3. High and low inputs are linked so as to require only 3 inputs.
  4. Dead time is permanently enabled via 30K resistor from pin 7 to Vdd.


My problem is that when I power the board with 12v to the HIP4086A and 24v to the MOSFET bridges, everything works great. My scope shows well defined 3ph output waveforms from the bridges of the correct period, timing and amplitude. Nothing heats up (I inadvertently left it running for an hour like this and no problems once so ever).

BUT when I connect and power a motor via PWM at low RMS current (<1A), the following occurs:

1. Motor operation is sporadic at best, but this may be due to the test motor being a fairly cheap motor with high torque ripple. This lasts about 30 seconds on average before (2).
2. HIP4086A stops output to the MOSFETs. HIP4086A shows an increased constant current draw of 0.35A @ 12v. MOSFET bridges show zero current draw @ 24v.
3. Usually this is where I catch the error. The HIP4086A is dead at this point. I did not catch it as quick the last time and the HIP4086 actually melted the solder on its pads and a small bubble formed on the top center surface of the IC, adjacent to pin 19. No other components or traces show any noticable temperature increase.

Can anyone offer any advice as to why these driver ICs keep self destructing as soon as an inductive load is connected?

Schematic (Please forgive the roughness, I am not in-office right now) enter image description here

PCB Layout images, latest version. 4 layer board, first image is composite, each additional image is one layer only. Composite, 4 layers Layer 1, Top Layer 2 Layer 3 Layer 4, Bottom


Added reformatted partial schematic. I hope this is more readable.
Note: Current Sense resistor connections omitted. Current sense not enabled, currently. Reformatted schematic

\$\endgroup\$
  • \$\begingroup\$ I need to point out that for the latest failure, I was using two separate external power supplies, one at 12vdc and one at 24vdc. All grounds are linked within the schematic, but in this case, the MOSFET bridges could not have been directly pulling the Vdd to the HIP4086A low. \$\endgroup\$ – JDD May 7 at 2:44
  • \$\begingroup\$ Show your PCB --- all layers. \$\endgroup\$ – analogsystemsrf May 7 at 2:47
  • \$\begingroup\$ I hate to do this to you, especially since you don't have your EDA with you... but the schematic would be much easier to read if I didn't have to twist my head. And the bus doesn't help readability either \$\endgroup\$ – user60561 May 7 at 2:49
  • \$\begingroup\$ Sounds like Ground_A is not the same as Ground_B, and inductive spikes on Ground_A are endangering the driver IC. \$\endgroup\$ – analogsystemsrf May 7 at 3:45
  • 1
    \$\begingroup\$ PCBs added; I thankfully had these on me. Ground A is linked to Ground B through a single connection point, see "sh1" just above "R33". Admittedly they are connected through only two vias, more vias are to be added at this point in the future. I am downloading eagle on my PC, now, to work on building a more neck friendly version. \$\endgroup\$ – JDD May 7 at 3:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.