• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 3 answers
  • Subscribers 62 subscribers
  • Views 1122 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.

bq24630 fails at higher currents

Ben Stelle
Intellectual 620 points
Other Parts Discussed in Thread: BQ24630

I have a bq24630 design charging a 4S pack to 14.6V using a 24V DC supply.  I have configured the ISET resistor network for the charge current to use a dip switch to allow for 4 different charge rates (2.0, 2.5, 5.0, and 10.0 amps).  My problem is that in 50% of my boards when I select the 5 amp setting or higher the bq24630 will fail (sometimes letting out a small wisp of smoke) and not recover.  On those boards that fail, the IC runs very hot indicating some sort of EOS.  The only waveform that seems to offend the datasheet is the common node between the switching FETs connected to PH.  It exhibits a very high ringing voltage when the high side FET switches on (sometimes on the order of 60 to 70V) for about 20ns before settling down to the 24 volt input voltage.  I have checked that the amplitude of this ringing seems to correlate to the amount of charge current selected.  So, I placed a 26V TVS diode across the lo side FET to protect that node a bit.  This reduced the amplitude of the ringing on all the targets that I tried, but I still see voltages up around the 42V range which is the maximum clamping voltage of the diode under higher current conditions.  Anyway, this change also seems to affect the 10amp setting on some targets only allowing them to achieve 7-9amps.  Also, this change doesn't seem to stop some targets from failing in exactly the same way as without the change.  It looks to me like this same ringing is present on the hi side FET driver voltage , so this probably doesn't help things.  I currently am using only a 2 layer board with a generally continuous plane across the bottom.  I will be making a 4 layer version next in an attempt to finally get this right, but need to be absolutely sure that I don't need some additional parts to protect the IC.  I did notice that the EVM seems to have several additional parts on its schematic compared to the schematic depicted in the data sheet.  Are any of those parts intended to help with inductive ringing or loop characteristics, or are my problems likely PCB related?  For reference, please see my earlier post  This post was from prior to having a better ground plane on the board, and I no longer am having those problems.  The biggest problem right now is blowing up the IC at the higher current settings.  Again, some of the boards work fine, but still have the inductive ripple that I mentioned.

Please let me know what you think

  • 0
    Intellectual 620 points

    I hope that this request has been seen, but I have not seen any responses at all thus far.  Is this the correct place to request help from?

    Regards,

    Ben

  • 0
    TI__Expert 5565 points
    Ben,

    My guess is that when your flip your dip switches, it's not a clean connection (connection bounce, slow changing impedance, etc). Therefore, as your are switching in a resistor to go to higher charging current, there is a momentary period of time where your charging current is much higher than your setting. When your ISET voltage changes, the IC tries to react by changing the duty cycle to allow more current through the SRP/SRN sense resistor. Coupled with the possibility that your layout still has some shortcomings in keeping the AC loop small (which is very hard to do with a 2-layer board), this excessive current is causing excessive ringing on the PH pin causing the charger IC to be damaged.

    I suggest putting a scope probe on the ISET pin and capturing the change in voltage when you flip the dip switch. Also, monitor the PH node, and the inductor current (or you can just measure the differential voltage across SRP/SRN to get a representation of the inductor current) and see how it changes when the voltage on ISET pin changes.
  • 0 in reply to Smriti B
    Intellectual 620 points
    Thanks for your reply. I have seen this fail even with a clean power up without changing the DIP switch. The ringing is quite large and seems to exceed the ratings not only of the PH pin, but also the low side VDS max rating of the FETs. Most recently I have had the low side FETs fail as well creating a short through the low side FET. This problem seems to have gotten better by adding the clamping diodes across that FET, but still the IC rating is exceeded by just a little bit.

    Does it make sense that if the PH pin is overstressed in this way that the VCC pin might draw too much current? I have seen this in the ICs that have failed occasionally. So much current is drawn into the IC that the resistor (1206 size) actually blows up!

    I am migrating the design to a 4 layer board in the next pass as I mentioned, which should help me to keep the switching loop as short as possible. I assume that the ringing in this case is caused by a combination of the parasitic capacitance of the inductor and the parasitic inductance created by the routing from the input capacitors through the high side switch FET and ultimately that return path to ground. Does this sound right?

    Would any resistance in series with the PH path help or hurt here? Any other damage avoidance techniques you can think of?

    Does the trace/cable length leading up to the input capacitor matter? If so I don't have much control over those items. The power supply cable is maybe 6-8 inches long and the battery cable on the far end is equally long.

    Can you help me to understand the exact path that is likely the source of my problems?

    Regards,

    Ben
  • 0 in reply to Ben Stelle
    TI__Expert 5565 points
    Can you post a waveform of your PH node?

    You shouldn't add any resistance in series with PH. However, if you are not able to improve your layout (getting the AC loop as small as possible), you might want to consider adding a snubber circuit on the PH node. Here is a link to help you pick out the R and C for a snubber circuit for a buck converter. www.ti.com/.../snubber_circuit_design.html
    Keep in mind that the snubber circuit is dissipating the extra "ringing" energy as heat, so it will lower your battery charger's efficiency.
  • 0 in reply to Smriti B
    Intellectual 620 points
    Since I did not get a reply very quickly last time, I did a bit of thinking and ultimately looked at the eval board schematic closer. As I mentioned in my original post, there are a number of extra components shown on the eval board schematic that are not shown anywhere in the datasheet.

    This is a huge mistake on the part of TI in my opinion, as it would have saved me a couple of board revisions. it turns out that by placing a little bit of resistance (100 Ohms) in series with the gate of the high side switch, and also adding the bypass diode to allow for quick turn off that I was able to get rid of the ringing all together (at least high side ringing, which was killing the chip). The soft start on that leading edge of turning on the high side is absolutely critical especially when the charge current gets to higher levels. This appears to be the dominant effect in my board at the moment. Please consider that a revision to the datasheet to suggest some of the additional circuit tricks present in the eval board especially since most of the posts I see for this chip constantly refer to the eval board anyway.

    Also, please be aware that it is critical to respond right away to posts at least to say that you need some time to consider how to respond. The E2E community works great for support but only if the follow up is timely.
  • 0 in reply to Ben Stelle
    TI__Expert 5565 points
    100 Ohm resistance is bit too high because it will severely impact your efficiency. Ideally, you want to stay <=10 Ohms for the series resistance. Alternatively, you can put a resistor in series with the BTST pin to get the same effect (this is R14 in the EVM schematic), but the added benefit is that you don't need the diode for fast turn-off. Adding the series resistor to BTST only slows down turn-on, but leaves turn-off unaffected. But even the BTST resistor should be <=10 Ohms, or else you risk severely impacting your efficiency.
  • 0 in reply to Smriti B
    Intellectual 620 points
    I tried 10 ohms thinking the same thing about efficiency, however this limited follow on ringing but did not do very much about the initial high side spike. Only at 100 ohms did I see the kind of waveforms I wanted to see. I will keep the BTST resistor trick in mind as we move forward. I had that planned into the next revision anyway for testing. Why are those components not depicted or even discussed in the datasheet? Again, this might well have saved us a couple of board revisions.
  • 0 in reply to Ben Stelle
    TI__Expert 5565 points
    Is your charge current still 10A? You will want to keep an eye on your high side FET with the 100 Ohms resistor. It's probably running hotter now since the gate drive has been slowed down with the series resistor.

    The application drawings in the datasheet usually only show the minimum necessary components to make the IC work in a stable condition. The EVM schematic usually contains more components (such as series resistors, diodes, snubbers, etc.) to give an idea of what components/circuits can be used to help with the PCB/layout design process. With that said, it's probably a good idea to discuss some of this in the datasheet and/or refer to the EVM user's guide. We appreciate this feedback and will work on incorporating them into our future datasheets.