BQ25750: BQ25750

Hi Ethan,

Thank you for your support.

since I already manufactured the PCB, I prefer to try the Auto Revers Mode.

My main concern is that according my calculations, because of the heavy load and other parameters, the response time of the reverse regulation should be very fast - only 100uSec. May i will need to add some capacitance to the load to better handle the system voltage after input voltage drop, but anyway I am going to try that and will let you know the results .

Hi Ethan,

Even before I try the auto revers mode, In second thought the chances that it will solve my problem are very slim - because a normal SMPS regulation time is about 2 mSec, but my load voltage will decay in only 100 uSec.

So, I would like to check with you again the possibility of using the BQ25756 instead of BQ25750. Since this chip solves this issue inherently because it doesn't use those BATFET and ACFET switching (I don't really need the direct power path).

I already considered the BQ25756 before BQ25750, but I thought that it can't distinguish between the battery current and the load (system) current - it is important to me also that the chip will to be able to reduce automatically the battery charge current when the load current is too high.

But i read now again its datasheet and found surprisingly that it can do that somehow - see paragraph 8.3.5.1.1 .

I would like to verify the following:

1. Can BQ25756 indeed do the above?

2. How it actually do that if the battery current and the load current are flowing through the same shunt resistor?

BTW, I know that it is possible to modify the BQ25750 to do the same, but i don't see the benefit.

B.R

Hello Shony,

Thanks for being patient. I'll get back to you later this week on these questions.

Best Regards,
Ethan Galloway

Hello Shoney,

1 and 2: The BQ25756 can't automatically reduce the battery charge current to account for the system load.

For your system, how much difference between good VAC input and bad VAC input? If the charger can detect VAC dropping quickly, capacitors should be able to keep the system powered while the BATFETs turn-on or reverse mode turns-on.

Do you think that will work for you application?

Best Regards,
Ethan Galloway

Hi Ethan,

I already made the calculations and experiments and found that for my application the system can't stand hot swap switching longer than 100uSec.

This very bad hot swap is obviously and chip design fault - The BQ25750 designers should have taken into account the following two considerations that could be solve this problem:

1. The BATFET/ACFET transistor can stand very high inrush current - up to 300A spike. So they shouldn't have to limit it to only its 8A DC rate.

2. They should not have using N-CH transistor but P-CH. Because the N-CH in this configuration needs charge pump for the BATDRV to drive its gate 10V above BATSRC - this ramp-up takes long time.

I though maybe I will change my circuit accordingly, but it is a headache now.

Now I wonder if any of the two solutions bellow could work:

a) To omit (cancel) the ACFET and BATFET transistors by shorting them and also disconnecting the wiring between them (I don't need the direct power path). Maybe this way the charger can differentiate between the load current and the battery current (Load current = Input current - Battery current). See bellow:

     b) Using the above, but also disconnecting the SYS from the up side of the battery shunt resistor and connecting it to tit low side. And now when the system powers-on, my CPU can sense the input current continuously and change the charging current as function of the load current. 

B.R

Hello Shoney,

Thanks for working with this.

1. Please keep in mind that the AONS9917 can only withstand a 320A surge current for 10us and only with VDS=1. You can look at the Safe Operating Area, or SOA, graph for more information about the surge currents the AONS66917 can withstand at different voltages and times.

2. I'll pass this information to the designers.

For your other questions:

a. The charger needs to be able to sense the current through the battery sense resistor to work. This configuration won't work.

b. This will work and I think this may be your best option. You might be able to use the BQ25756 as an alternative part as well.

Let me know if you have any more questions about this.

Best Regards,
Ethan Galloway

Hi Ethan,

1. Obviously the charger user designer can't assume 320A inrush for granted. He should calculate the expected inrush current amplitude and time and also the selected transistor limitations in order to guarantee that the SOA is not break. Also the  charger designer could add some hardware safety limitation circuit.

In my case the max Vds voltage is 30V during 30uSec. Means I could allow 100A inrush for the above transistor.

a) Seams that you are right - my first suggestion will not work for BQ25750, because when there is valid input voltage, the charger think that all the current sensed by the BAT shunt resistor is going only to the battery. So, its H-bridge will limit this current to what defined by the ICHG resistor.

b) I still have some doubt abought this solution - As we already agreed, the BQ25756 charger can't differentiate between the charging current and the system current. So, in vice versa to the configuration in paragraph a. If I will limit the CHG current, the whole battery+system current will be limited.

Hi Ethan,

Reading the current rate directly from the battery is interesting idea. I will check that too.

But after few measurements of the BQ27750 control signals, I believe that I have a way to replace the BATFET N-CH with a P-CH + N-CH drive that will hopefully work OK. I am going to test it soon.

Thanks.