This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DAC7760: Trouble with chip overheating

Part Number: DAC7760
Other Parts Discussed in Thread: DAC80501, DAC60501, DAC70501, DAC8760

I've been having some trouble with overheating of the device and I need a design review. 

In a past revision the alarm pin was tied to DVDD by mistake (someone else's design before I got handed the project). The device would output stable voltage for a good amount of time, then suddenly start overheating and burning up (smoking). I've made the correction but I am not certain that it was the cause of thee problem. Is there anything else that could be causing the problem?

  • Hi Matt,

    The !ALARM pin is an open-drain output, so there is certainly some potential to create a situation with some unexpected heat if the pin were directly connected to the digital supply. Depending on the current-sourcing capability of the DVDD supply, it could be plausible for the device to be damaged from this event.

    Looking at the rest of your schematic I do not see any issues. The most common mistake I see with this device relating to any overheating is the thermal pad being connected to GND in dual-supply use-cases. This is, of course, outside of the context of EMC/EMI events. If there were any tests applied to the digital I/Os, power, or outputs of the device during these overheating events, that would be helpful to understand.

    Can you rework any of your existing boards to float the !ALARM pin to verify you proposed solution?

    I also cannot help but notice that the IOUT is not being used in your application. I am curious how you are using the part and what may have got your attention. It is possible some other road-map devices could be a better fit for you, so I'd like to understand as much as you're willing to share. I understand that is sensitive, so with your permission we can take that chat offline if you are willing to entertain it.

  • Kevin,

    I've floated the alarm pin. At this point the chip still begins to overheat when the supply voltages (+/- 15VDC) are applied. I am using a Fluke to measure the voltage output. as soon as the supply ground is connected, the voltage reading shows a negative value (-0.8**) and the IC begins to heat up. 

    Also: I have not sent any signal to the chip since power up. 

    Any thoughts on what could still be causing this? 

  • Kevin,

    I am open to the idea of a new IC if you are able to help me out in selecting one. 

  • Matt,

    Can you probe the +/-15V rails to check their stability? Similar for DVDD. If the device is heating up / smoking there should be a significant current signature on one of the rails which will maybe give us a better idea where to look.

    Regarding suggesting a potentially simpler device, what voltage output range(s) are you using?

  • Kevin,

    Here is a more detailed description of what I have observed from the device.

    When initially connected (power supplies, SPI, etc) the device functions correctly. It can remain connected to +/-15V indefinitely with no issue. I'm able to send data and see a voltage reading from the fluke. It remains stable for several minutes allowing me to change the value using test code then without warning will begin to overheat and the fluke will show a negative voltage even though the initialization set the range at 0 to +5V. Once this occurs, I must disconnect the +/-15V supply to stop the overheating. From this point, the best I can see, the IC is in some sort of failure state because connecting the supply gnd and either 15V supply rail results in rapid overheating (even after a complete power down reset). 

    The only requirements are to use SPI to output a 0-5V range on at least 1 (eventually 2) channels with a minimum 8 bit resolution. I have 3.3V, 5V, and +/-15V supplies available. 

    - Matt

    P.S. My test code was verified with the DAC7760EVM for long duration testing. While I'm not 100% certain, I am fairly confident it's operating reliably. 

  • Matt,

    I apologize for the slow reply. I was out of office last week.

    I think we have two paths to consider.

    The first - typically things do not randomly fail, especially not in such a catastrophic fashion. We can try to determine what the pattern leading to this is (i.e. code changes, register changes, changes in loading, transients on power or other terminals etc). This is certainly not normal or expected behavior from the part, so there is something for us to debug on the board level.

    The second - we look at other options since there are a wide array of devices which can accomplish your requirements and having some of the features of DAC8760 actually seem to be an overkill for your application. As a start, I would suggest looking at the most recent device family in the DAC80501, DAC70501, and DAC60501. They are essentially the same device, just with resolution variants for 16, 14, and 12 bits respectively.

  • Kevin,

    Thanks! I'm giving those a look over. 

    I think I discovered the issue. Altium was removing the PAD via net assignments from the footprint each time I would update from schematic to PCB. In other words, the final issue was a result of a floated thermal pad that did not show up on the DRC because the pad was assigned to the net but had no actual connection to the net. 

    Thanks for the help!

    -Matt