• State
  • Locked Locked
  • Replies 10 replies
  • Subscribers 63 subscribers
  • Views 259 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

BQ25756: Is there a way to share one NTC sensor between BQ25756 charger TS pin and BQ34Z100-R2 fuel gauage TS pin?

BQ25756: Manufacturing line testing recommendations?

Ali Al-Sheikh
Prodigy 150 points
Part Number: BQ25756

Tool/software:

Hi team,

Background:

We had a terrible experience using a certain cheap charger IC in one of our products that changed the way we think of testing it on the manufacturing line, then extreme testing it later, just to guarantee at least an acceptable performance in the field. It still fails in the field for reasons unknown.

IC reached EOL and we gladly adopted the BQ25756 in our design and currently happy with it, at least after testing a few manufactured units that pack this IC. 

Question:

Before we re-create testing procedures, we thought of looking for some manufacturing line testing advice from your team, starting from whether you have some references to look at, examples somewhere, or some advice already on this thread, or any directions at all.

The purpose is to ensure a proper manufacturing testing procedure that would guarantee proper functioning in the field.

Charger Operation:

If it helps, here is how the charger normally operates.

  1. Only forward operation, and only in boost-conversion mode
  2. Designed with max input current of 2.2A at 18V (40W), and an output current of 1.4A at the boost converted node of 29.2V
  3. System deploys a 29.2V LFP battery with 6.7Wh that needs full charging & termination, potentially multiple times a day
  4. System has a nominal background operation load of ~100mA at 29.2V and some repetitive extreme-loads (motor) occurring only during the daytime - starting currents of well above 10A
  5. We usually make sure the charger is disabled through CE pin before running the motor to separate them and protect the charger from these transient times

Any advice, direction or documentation would be appreciated.

Regards,

Ali

  • 0
    TI__Mastermind 31035 points

    Hello Ali,

    Thanks for being patient with this. We are swamped at the moment. I will think about this and get back to you.

    Best Regards,
    Ethan Galloway

  • 0 in reply to Ethan Galloway
    TI__Mastermind 31035 points

    Hello Ali,

    Thanks for being patient with this. The ICs are already screened out before they ship.

    I don't have a lot of specific knowledge for manufacturing line testing. I know you can use a constant voltage load to test the BQ25756 on the production line.

    For your system, what tests are you currently running on the battery charger?

    By the way, this is TI's page on reliability testing if you haven't seen this page already: https://www.ti.com/support-quality/reliability/reliability-testing.html

    Best Regards,
    Ethan Galloway

  • 0 in reply to Ethan Galloway
    Prodigy 150 points

    Hi Ethan,

    Currently, we're exercising multiple parts of the circuit in our FCT tests:

    1. Testing that connecting a fully charged battery does not trigger charging (battery level feedback)
    2. Confirm the system can operate stably only using the charger - without a battery- (stable up-conversion given the system load)
    3. Testing the current limiting feature of the charger after increasing the supply current (input and output sense resistors)
    4. Testing the reported levels of input and output voltages and currents and comparing them with a reference (sensing + accuracy)
    5. Testing for an acceptable NTC sensor voltage level at room temp (TS divider)
    6. Testing input voltage range for charger IC to tease out the performance within the upper and lower operation voltage limits (e.g. 6 - 20V)

    Does that seem sufficient? Feel free to discuss that internally if you think it is necessary. We will very much appreciate your input on this.

    Regards,

    Ali

  • 0 in reply to Ali Al-Sheikh
    TI__Mastermind 31035 points

    Hello Ali,

    Thanks for the information. Let me look at this and get back to you.

    Best Regards,
    Ethan Galloway

  • 0 in reply to Ethan Galloway
    TI__Mastermind 31035 points

    Hello Ali,

    Just for reference, we do the following tests when we test the EVMs for production. All of the tests are done with a constant voltage load of 25V and a charge current of 4A. We use a ≈2000µF on the constant voltage load terminals. Sometimes the CV load will oscillate.

    1. With VIN=40V, test the EVM charging at 4A at 25V. This makes sure the buck FETs are working.

    2. With VIN=25V, test the EVM charging at 4A at 25V. This makes sure buck-boost mode works

    3. Then, test with VIN=15V. This makes sure the boost FETs work.

    I have a few thoughts and suggestions on your system tests:

    • Are you doing HALT and HASS tests for your system?

    • For test 1, does your system have a way of detecting if a charger is plugged in or not?
    • For test  4, keep in mind that the ADC is intended to be a general indicator. I recommend using a battery monitor or gauge for greater accuracy in the system.
    • Test 2 may not be needed if you don't plan on operating the BQ25756 without a battery attached.
    • Test 5 and 6 may also not be needed. These voltage and resistor set ranges are pretty consistent.

    Best Regards,
    Ethan Galloway

  • 0 in reply to Ethan Galloway
    Prodigy 150 points

    Hi Ethan,

    Thank you for your input on this. Since we are only doing boost, only item 3 applies and this is what we will do next.

    Ethan Galloway said:
    Are you doing HALT and HASS tests for your system?

    Yes, we will be doing that. Any documentation or advice you could give on the recommended practice for testing it?

    Ethan Galloway said:
    For test 1, does your system have a way of detecting if a charger is plugged in or not?

    Not sure it does, but we are only watching the input current become some small positive value to prove that charging is happening. If 0, which we can see through our bench tests when there is no charging, then we know no charging is happening. Does that sound like a reliable test?

    Ethan Galloway said:
    For test  4, keep in mind that the ADC is intended to be a general indicator. I recommend using a battery monitor or gauge for greater accuracy in the system.

    Noted. Yes, we are comparing that to what the fuel gauge can see + the power supply feeding the system for reference.

    Ethan Galloway said:
    Test 2 may not be needed if you don't plan on operating the BQ25756 without a battery attached.

    Noted. It was done to allow operating the system without battery for some quick flashing, testing, maintenance, etc work that isn't necessarily reflective of the real-world application. We even had to add a bulk cap at the load to stabilize the output/compensation and get a relatively stable ~29.2V to keep the system running. This test covers this bulk cap change.

    Ethan Galloway said:
    Test 5 and 6 may also not be needed. These voltage and resistor set ranges are pretty consistent.

    Agreed, makes sense! This is only done to ensure all surrounding components' values are correct with no significant shift that would cause the system to behave differently. These may be removed down the road, following your advice.

  • 0 in reply to Ali Al-Sheikh
    TI__Mastermind 31035 points

    Hello Ali,

    Thanks for the clarifications.

    Ali Al-Sheikh said:
    Yes, we will be doing that. Any documentation or advice you could give on the recommended practice for testing it?

    Got it. That makes sense. We don't have any documentation for HALT or HASS testing.

    Ali Al-Sheikh said:
    Ethan Galloway said:
    For test 1, does your system have a way of detecting if a charger is plugged in or not?

    Not sure it does, but we are only watching the input current become some small positive value to prove that charging is happening. If 0, which we can see through our bench tests when there is no charging, then we know no charging is happening. Does that sound like a reliable test

    I think this would actually test the termination current limit of the BQ25756. I think the steps and results would go like this:

    1. Fully charged battery is plugged in

    2. Charging is activated with EN_CHG=0

    3. Charging current quickly goes 0A and this triggers charging to stop.

    Let me know if this makes sense. Let me know if you think this test would still be useful.

    Thanks for letting me know about test 2 and 4. These tests make sense.

    Best Regards,
    Ethan Galloway

  • 0 in reply to Ethan Galloway
    Prodigy 150 points

    Hi Ethan,

    Ethan Galloway said:

    I think this would actually test the termination current limit of the BQ25756. I think the steps and results would go like this:

    1. Fully charged battery is plugged in

    2. Charging is activated with EN_CHG=0

    3. Charging current quickly goes 0A and this triggers charging to stop.

    Sorry for not being clearer about our test. We are simply testing whether the charger can detect a fully charged battery and basically never starts charging even though it is enabled. The test is done using a power supply that generates the full-batt voltage. We plug it into the battery terminal, the system starts and EN_CHG is enabled by default, and then we watch for a few seconds to ensure that charging never starts.

    Ali Al-Sheikh said:
    keep in mind that the ADC is intended to be a general indicator. I recommend using a battery monitor or gauge for greater accuracy in the system.

    It is interesting you mentioned this. We did notice on our boards (even the evaluation board) that there is a notable shift in detected voltage and current values (both input and output), and it almost always tends to be on the lower side (underestimated), and almost by a similar amount. We use 0.5% sensing resistors on our system for both input and output.

    We are also using a fuel gauge with fairly accurate voltage and current reporting, but the issue is it's only monitoring the output voltage/current, and we need some reliable reading for the input (solar) side. We may plan on calibrating/tweaking such values using the FW to improve accuracy (probably adding some constant to compensate for the underestimation).

    Q/ Would you advise on doing that? Please comment on what's advisable to improve the accuracy of the ADC-reported values (at least the input). 

  • 0 in reply to Ali Al-Sheikh
    TI__Mastermind 31035 points

    Hello Ali,

    Thanks for the clarifications.

    I'll get back to you next week on this.

    Best Regards,
    Ethan Galloway

  • 0 in reply to Ethan Galloway
    TI__Mastermind 31035 points

    Hello Ali,

    Thanks for being patient with this. This has been a very busy week for me.

    Ali Al-Sheikh said:
    We plug it into the battery terminal, the system starts and EN_CHG is enabled by default, and then we watch for a few seconds to ensure that charging never starts.

    This makes sense. I think this would work.

    Ali Al-Sheikh said:
    Q/ Would you advise on doing that? Please comment on what's advisable to improve the accuracy of the ADC-reported values (at least the input).

    I'm not sure of many ways to improve the ADC accuracy. You can go up to a 15mΩ sense resistor on the input side to make the ADC more sensitive. I'm not sure if your system would benefit from the increased sensitivity though.

    Best Regards,
    Ethan Galloway