We have a very large rejection rate (20%) of our production due to the INA168 that not working. Most of the time, the output of the INA168 is 0V. I don't know what is wrong. The absolute maximum rating are not exceeded.
+VOUT is a DCDC converter output (12V)
V5116 is the Vsense point for the buck converter
+5V is the supply voltage for uController and other stuff
We are still on production for this product. Do you have any suggestion for this problem?
I do not see any issues in your circuit setup.
When the INA168 output is 0V, what is the load current through the 0.01Ohm sense resistor?
If you're not sure of the load current, can you please measure the voltage drop across the shunt resistor and also measure the voltage between the input pins of the INA168 (Vsense)? These two voltages shuld be equal, but if they are not, it could begin to help us understand the issue.
We'll get this worked out.Thank you,Collin WellsPrecision Analog
Regards,Collin WellsPrecision Linear Applications
The load current can be between 0A and 10A. I've got the voltage drop according to the current.
When this problem occur, I only change the INA for an other, and the problem is "fixed". But like I told you we are in production and that's made a lot of INA to replace.
I'm sorry to have to ask more questions, but I still am not seeing a reason for the failures. All INA168 devices leave our factory in working condition as guaranteed by our final test solution.
To re-state your problem: with some initial conditions applied to your board, the INA168 output is stuck at 0V and this is how you are detect that it is failing. You then replace the INA168 and repeat the test, the board works as planned.
What is the expected output of a "working" INA168 board? The "failing" boards have an output of 0V.
Could you please help me understand what conditions the INA168 is in when you test the boards? (Vbus = 12V, Vs = 5V, load current = ?, Vsense = ?)
Does your test sweep the input current from 0-10A when the INA168 output stays at 0V, or does the test apply a single input current value to check the INA168 output? If the latter, what is the "test" input current? If you don't sweep the input current, have you tried sweeping the input to see if the INA168 reacts at all to a changing input value?
Do you ever have to replace the failing device with more than one INA168 before the failure goes away?
Thanks,Collin WellsPrecision Analog
The input voltage of the DCDC is 24V. +Vout (12V) is provided by a 130KHz synchronous buck with a soft start of typically 50ms. We need this soft start to feed 2 halogen headlamp (60W each, 120W total). The output current is 10A, the current limit of the converter (short circuit) is 27A. The sense resistor is 0R010 and the loading resistor of the INA is 75K. With 10A, VOUT of the INA is 1V5.
When the test is starting, DCDC start in "cold" lamp load which increase the nominal current but this current is <27A. When the lamp are "hot", we've got our nominal current (10A) we take a manual measurement and we expect to get 1V5. If the output voltage of the INA is 0V, tests are stopped and we replaced the INA. (100% of the time that the INa output was 0V, it's was due by the INA) The INA need to be replaced only one time.
The supply voltage of the INA (+5V) appear before +VOUT (+12V)
Also, each time that the INA wasn't working, we change this part on the PCB and our product work well after that. At the start we were thinking that it was our PCB assembler, we changed with an other one and the problem occur also.
This is starting to sound like an EOS (electrical over-stress) or ESD related failure, but before we go down that route I have a few more questions.
Have you ever tried taking a "failing" part and try it on a different board to see if it begins working in a different system or not?
Could you try monitoring the output of a "failing" INA168 during the full startup sequence? My desire in the previous questions to see if you performed any sweeping of the input current during production testing was to try to see if the INA168 output reacted to a changing input voltage or if it was just stuck at 0V for all inputs. When the DCDC begins to power-up in the "cold" lamp scenario, you've stated that the initial input current is higher than 10A but less than 27A and then settles eventually to 10A. Please try monitoring the INA168 output during this changing current to see if the output does anything other than stick to 0V.
Finally to determine if it may be damage related to EOS, do you have any failing returns from your end customers that display this same issue? If so, is it possible that there are any transient voltage spikes that you're not noticing on the Vbus line that may violate the +60V DC spec of the INA168. Spikes can be related to the turning on and off of heavy current loads through an inductive PCB trace or cable.
Thanks and Regards,Collin Wells
Do you have any updates on this issue? Has is been resolved?
Best Regards,Collin WellsPrecision Analog
Sorry for not answering you soon, I was waiting after a new production unit. This morning I looked for a bad unit, but I was unable to find a unit with a failing INA168. The next production will be the next week, I “hope” that one test will fail with a INA168. At this moment, I will be able to get wave forms & swap INA168 with another product that is working.
This morning, I took screen shot of the first power-up of an unit.
CH2: Output voltage (VOUT, 2 levels possible: 8V & 12V)
CH4: Load current (INA sense), there is a serial switch between load and VOUT; 0A is possible
CH2: INA “output voltage” via 75K (note: probe setup was not very good)
The output voltage (Vout) is protected by MOV (V22ZA2 clamping 43V at 5A). In fact all input and output are protected against voltages spikes and reverse polarity. At this point, we never had returns from our end customers. The first unit was ship 5 years ago, we have over 2K unit in field.
As soon as I will get a failing INA168, I will take screen shots and swap with another one that is working.
Thank you for the images and description of the issue. Based on the fact that none of your units get returned it doesn't seem like it is your system that is damaging the components. This leads me to believe that they may be damaged before they get assembled or during assembly due to ESD. Please contact me directly at Collin@TI.com when you get another failing unit and we can get it examined by the failure analysis lab to determine the cause of the failure.
Just to be thorough, when/ if you do get a failing one and can take screen shots, please also include the voltage directly on both sides of the shunt resistor. Depending on your layout, J17 and J18 may be convenient places to attach your probes to look at the signals. So in total I'd like to see:
Shunt + (J17)Shunt - (J18)INA168 Output (TP9)Bus Current
Thanks,Collin WellsPrecision Analog
Sorry if I've not given new earlier, since I have no more productions problem, I forget to give you some news. In fact, I have not managed to find exactly the problem, but I realized that the resistor Rense was "low quality" because the purchases changed it without my approval. Since we installed the "good ones" everything is back to normal.
All content and materials on this site are provided "as is". TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with regard to these materials, including but not limited to all implied warranties and conditions of merchantability, fitness for a particular purpose, title and non-infringement of any third party intellectual property right. TI and its respective suppliers and providers of content make no representations about the suitability of these materials for any purpose and disclaim all warranties and conditions with respect to these materials. No license, either express or implied, by estoppel or otherwise, is granted by TI. Use of the information on this site may require a license from a third party, or a license from TI.
TI is a global semiconductor design and manufacturing company. Innovate with 100,000+ analog ICs andembedded processors, along with software, tools and the industry’s largest sales/support staff.