BQ76920: Driving current sense inputs

Thanks WM5195;

The amps I selected are RRIO, and so I "think"(?) they both should see only positive voltage on their non-inverting inputs (I hope). I agree about the filtering location, and have made that change. If you have any other suggestions, improvements, corrections, I'd really appreciate them.

Thanks again!3414.afe.pdf

Hi Jeffrey,

If the BQ76920 VSS is referenced to the battery side of the sense resistor, discharge currents will give positive voltages.  When charging voltages from the sense resistor will be negative.

I would think you may still want some differential capacitance near the SRP/SRN pins to quiet any noise picked up by the traces.

Hello WM5295;

I've modified the power supply to include a negative reference for the CHG amplifier. I'm running this by the AMPS forum as a sanity check. My intent is to bias the LPV821's VSS to -230mV which should allow it to drive SRN to -200mV. 

I'd appreciate any suggestions/comments/corrections

Thanks!AFE and Current Sense.pdf

Hello again WM5295;

I don't think my previous circuit would work because only this AM, I managed to  accomplished the task of understanding that SRP isn't intended to be driven off Vss. Table 6.3 says SRP may range +/- 10mV. 

So, rather than 2 amps, I think this corrected design should(?) work, since the amplifier is driving SRN above (discharge) and below (charge) Vss.

Would you please sanity check .

Thanks

6607.AFE and Current Sense.pdf 

Hello again WM5295;

I don't think my previous circuit would work because only this AM, I managed to  accomplished the task of understanding that SRP isn't intended to be driven off Vss. Table 6.3 says SRP may range +/- 10mV. 

So, rather than 2 amps, I think this corrected design should(?) work, since the amplifier is driving SRN above (discharge) and below (charge) Vss.

Would you please sanity check .

Thanks

8130.AFE and Current Sense.pdf 

Hi Jeffrey,

I had assumed you had the SRP side of the sense resistor at the reference.  In the dual amp design the amp should make any correction for the reference to resistor connection, if the reference side (for SRP) moved 1 mV the amp would adjust this to 5mV.  One might wonder if the offsets of the amp and tolerance in the circuit might introduce more uncertainty than it removes. So the single amp may be a good solution.  I have not built either version of the circuit.

With the new circuit it is not clear what benefit R135, C61 and the capacitor below it might provide.  Also C14 might be best across the SRP & SRN inputs to the BQ76920.  A differential cap between SRP and SRN would be good with the dual amplifier circuit also.  Be sure to check the amplifier for recommendations on load capacitance and input resistance ranges.

Hello Willy (Matt told me your name)

Matt emailed and said he'd be out on holiday, but that you'd be 2cd best at handling my issues (he didn't say who was best ;-D).

Anyway, I'm so easily confused anymore. You said "...it is not clear what benefit R135, C61 and the capacitor below it might provide"..but I actually don't see what R115 and C60 (?) provide??? Both sides are at Vss?

Matt suggested modifying the BAT and REGSRC filters as per slua749a Figure 13. bq76920 REGSRC Power Option ..right side.

But it's unclear what the value of rd is recommended. I simply specified a 1K, but just guessing. I'll place the diff cap across SRN/SRP

Will you please write off on this change, and/or recommend anything else?

Thanks in advance!

Hi Jeffrey,

C60 provides a capacitance to the IC reference pin to help keep the voltage in operating range during transients. R115 senses the voltage at the battery side of the high current path.  In a compact system with no current they should all be the same potential.  With a dynamic system and/or high currents there will be differences. 

The Rd needed will vary with the design.  We used 10k on the BQ76920EVM and 100k on the BQ76930EVM.  The diode will block discharge of the BAT and REGSRC pins during short circuit or heavy load of the battery to maintain an operating voltage for the part and drive level for the FETs, but will also be a peak detector for noise or transients on the battery.  The load on BAT and REGSRC can be quite light, so the discharge of an accumulated voltage on the pins may be slow.  The Rd will provide a path to discharge the peak voltage back to the battery stack.  It should be small enough to discharge the filter caps sufficiently that they don't eventually push above the pin ABS MAX limit with the frequency of the system noise pulses, but large enough to prevent significant voltage drop during short circuit or heavy load.  Some designers use a zener (D3 on the EVM) to limit the voltage to which the node can be pumped.  Some designers don't like the diode so they use a larger filter capacitor to hold the voltage through the heavy load event and eliminate the diode and Rd.  Sorry it is not an easy answer, you will need some analysis and/or testing of your system.