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.

  • TI Thinks Resolved

DAC8718: Device getting heated after switching on the power

Prodigy 250 points

Replies: 8

Views: 79

Part Number: DAC8718

Hi,

I am Using DAC8718 (U11 in image below) for one of the projects. Output of dac is connected to Opamp "OPA1664" in voltage follower configuration followed by protection circuit.

During testing, when power is on, U11 is getting little warm and U10 is very hot. Also out2, 3 and 7 read -11.3V regardless on any request received from micro controller.

Not understanding the actual problem. Please help in debugging the issue.

Thanks in advance,

Dinesh KS

  • Dinesh,


    I don't see any problems in the schematic, but could you please show a larger portion of it? It would be good to see all of the DAC8718 and all of the output connections from the device. I looked through the pin numbering in the schematic and didn't see any mis-numbered pins, but you should double check it as well.

    AGND and DGND should be connected in the schematic. If you look at the Absolute Maximum Ratings table on page 2, there cannot be a voltage greater than ±0.3V between the two (generally, this should be 0V).

    To help debug this, I would check all the physical connections to the device. First, do a visual inspect of the part connections to make sure that the device is properly soldered down (for both the DAC and the OPA). Then measure the voltages for all the pins and other parts of the schematic. See if there are voltages that are far off from expected or grounded when they shouldn't be.

    You mentioned that OUT2, 3, and 8 is consistently at -11.3V. Are the other outputs where you expect them? Are you able to program the outputs to the correct values? I would also try the readback operation to makes sure the registers are where you think you've programmed them. This can be useful to verify that the device is responding to commands.

    Some heating may be expected but it depends on any loading that you have on the device. There is plenty of current running throuhg the device. The positive supply current AIDD can be at 6mA max and the negative supply current can be as much as 4mA. With the supplies at ±15V, the power dissipation can be 150mW. Looking at the junction to ambient thermal resistance of the QFN device, this means that you might see 4°C of temperature rise, which could be significantly higher if there's loading on the DAC output. However, in the schematic, there doesn't appear to be any loading on the DAC outputs.

    Read through my comments carefully and respond back to each of the points. It can be difficult to debug large boards, but I think this should get you started.

    Joseph Wu

  • In reply to Joseph Wu:

    Hi Joseph Wu,

    Thank you for the response.

    Here is the schematic:

    SCHEMATIC1.pdf

    Total i have 3 DAC in my schematic all have the same circuit. This is the DAC (U14 in attached schematic page) i am facing problem as i mentioned. All other DACs performing as intended.

    Also in U14, other outputs are working well except Out2, Out3, and Out7. All connected to U16 (opamp).

  • In reply to dinesh gupta:

    Dinesh,


    So on your board, you have three of these DACs and you only have problems with this one device, U14 on the schematic. In the previous post, you showed a schematic with U11 that was having problems. Is this the same schematic? Are the two DAC schematics the same or similar? If they are the same or similar, then you should compare the schematic from one DAC to another. This would be just to make sure there isn't a problem in how you've put together the circuit.

    Did you check all the things I suggested in my last post? Are AGND and DGND connected directly? Are all the connections correct on the board and did you check voltages around the entire circuit to see that they are where you expect them?

    However, because the problems you have are isolated to very specific channels of one device, I would think this is a problem is on the board itself. Perhaps this is a problem in manufacture, where there is a solder bridge across pins. Is it possible that the U16 dragging down the outputs for OUT2, 3, and 7? Have you done a physical inspect of U16 to make sure that there isn't some solder bridge on that device? Because all problems seem to be isolated to U16, it might be best to remove this device and see if the DAC outputs go to the correct voltages. Alternatively, you could remove U16 and replace it with a new device to see if that's a problem.


    Joseph Wu

  • In reply to Joseph Wu:

    Hi Joseph,

    In schematic, both AGND and DGND are directly short. Also there is one more gnd "GND_EARTH" (eg: one at D98 in schematic) which is short to main gnd using 0 Ohms.

    Yes, all three DACS are same, also circuit is same. 3x the schematic page i sent.

    Sorry for the refdes confusion, my mistake. Problem is with U14 and U16.

    Physically, U16 is fine. I dont see any problem in manufacturing.

    Testing is under progress. I will update the status.

    U16 being the primary suspect (as you pointed out), i am trying to know whether by any chance U14 (DAC) is malfunctioning.

    Thanks,

    Dinesh

  • In reply to dinesh gupta:

    Hi Joseph,

    Circuits are working well. As you pointed out issue is with U16.

    Output circuit are supposed to protect from reverse vlotage protection. As recommended by TI ("https://www.ti.com/lit/ug/sbau229/sbau229.pdf"), i used protection circuit for my requirement.

    My requirement is: I need to protect opamp (Eg: U16), if any operator plugs opamp output to power supply (24 VDC) accidentally, then protection circuit should protect my opamp. I think U16 failed while testing this case.

    Is there any recommendation for this requirement?

    Thanks,

    Dinesh KS

  • In reply to dinesh gupta:

    Dinesh,


    I'm glad that you were able to get that part of the circuit working.

    If you want to protect an op-amp, I would post on the amplifier forum for a definitive answer. However, I think this is the method that you would want to use. Here is a figure:

    In this figure, your overvoltage is accidentally placed on the output terminal. Depending on the overvoltage supply's ability to source current, you may or may not need a high power resistor to reduce the current going into the circuit.

    If the output terminal is taken higher than nominal, the rectifier diode dumps current into VDD (and if VDD cannot sink substantial current, the TVS diode will limit the voltage. If you have ±15V supplies, then you probably would want to use ~15V TVS diodes. Note that at 15V, there may be some amount of conduction through the TVS diodes, and it may not clamp until a higher voltage (perhaps 18V).

    This clamp voltage, the TVS and diode current (and wattage), and the overvoltage sourcing current will determine the resistance that you need to protect the device from a continuous overvoltage event.

    Similarly if the output terminal is taken low, current is pulled out of VSS through the rectifier diode. Again, if VSS isn't able to source much current, the current will be pulled from the TVS diode.


    Joseph Wu

  • In reply to Joseph Wu:

    Hi Joseph,

    Thank you for the suggestions. 

    We tested on 2 boards. The  first board had an issue with U16 OpAmp. After removing U16 and applying power supply (24V)  to  opamp output. The other DACs are working as expected. 

    On the 2nd board, we tested with the below criteria:

    1. With a correctly powered board ( no reverse polarity), measured all correct DAC outputs from U11 (DAC1) and U17 (DAC3)

    2. Briefly touched 24V to 4 DAC1 outputs working right to left:

    DAC1 Vout 5, Vout 6, Vout 1 and then Vout 2.  

    1. First three were briefly touched on, Vout 1 maybe a second or two until I saw a puff of smoke.
    2. Subsequently, OPA1664 component U10 is now hot to the touch.
    3. DAC3 outputs are still fine.
    4. DAC1 outputs Vout3,4,7,8 are still fine.
    5. DAC1 outputs read the following:

    Vout 5 : still fine (gives expected voltage)

    Vout 6: 4.35V (expected 5.0V )

    Vout 1: 0.26V

    Vout 2:   -14.46 V       

    Is it possible the TI reference design was intended to survive 24V transients, but not a continuous 24V?   

    We use resettable fuses for the power supplies, could we consider adding a smaller ( in form factor and current rating ) resettable fuse for each of the DAC outputs as well?  How would this impact the accuracy of the voltage output?

    Please advise if you can think of a useful next step with this board.

  • In reply to dinesh gupta:

    Dinesh,


    If you're referring the protection on the DACx760EMC-EVM as shown in the User's Guide, then it will provide only some basic protection. It may provide continuous protection for a lab supply that is current limited, but for an industrial supply that can source a couple of amps, no. There's no way that the output diode could try to hold of a 24V supply that could supply amps of current. You would blow out the diode, and then you would blow out the device.

    Output protection is all about safely shunting away current from the device when there's an overvoltage event. In the example that I had shown, you still need a high wattage resistor. This resistor would provide some voltage drop so that the TVS could do it's job. You'd need to determine the TVS breakdown voltage and then calculate the current across the resistor and going through the TVS. Then you'd need to determine the wattage of both the resistor and the TVS diode to ensure they could safely handle that amount of continuous power.

    As for the devices, it's likely if you touched the output of U10, then U10 is damaged. Depending on what failed next, then it's also possible that the DAC is damaged.


    Joseph Wu

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.