BQ25890: How to understand ILIM pin for BQ25890

Garrett Kreger said:

1) The ILIM pin performs both functions at the same time. The ILIM pin is always used to set input current limit, as well as, function as a current monitor unless EN_ILIM is set to disable (REG00[6] = 0b). 

-->So if EN_ILIM is set to disable (REG00[6] = 0b), neither current limit setting nor input current monitor can be used,right? I found some description from the datasheet" The ILIM pin function can be disabled by setting EN_ILIM bit to 0. When the pin is disabled, both input current
limit function and monitoring function are not available".

Garrett Kreger said:

2)The voltage of ILIM pin is expected to change dynamically. The input current is proportional to voltage on ILIM pin and this is how you can use ILIM as an input current monitor. If you do not wish to use ILIM pin as a current monitor all you need to consider is what is your desired input current limit and select the appropriate resistor accordingly. 

-->So the input current limit formula is as same as input current monitor formula  in fact, right? That's to say, voltage on ILIM(VILIM) increases as Iin increases according to formula Iin=(KILIM*VILIM) / (RILIM*0.8), and when VILIM rise from below 0.8V to 0.8V, the input current climbs to the maximum value =(KILIM*0.8) / (RILIM*0.8)=(KILIM) / (RILIM), which is exactly the formula of input current limit, and then IinDPM activates due to the input current exceeding the ILIM current limit setting(namely VILIM exceeds or equals to 0.8V). Anyway , it's the actual input current value that determines the  voltage on ILIM(VILIM), and not vice versa..  Is this understanding correct?

Another question, as we know input current limit set by ILIM pin can be used for IinDPM, but what's the real usage of using ILIM pin as input current monitor? That's to say why should we monitor the input current before it hits the ILIM set limit value if not considering DPM function? please help give some usage scenarios using ILIM pin as input current monitor.

Garrett Kreger said:

4) V(BAT) > or < VMINSYS has no impact on ICO operation. 

-->Please refer to SLVA812C and in this document, ICO operation differs between V(BAT) < VMINSYS and  V(BAT) > VMINSYS. How should I understand this part?

Garrett Kreger said:

3) Please refer to datasheet section 9.2.4 for details on ICO operation. The ICO feature identifies maximum input current of adapter and is able to raise or lower the input current limit register setting. Please be aware the ICO will not overwrite ILIM pin setting. If the ICO feature is used the IDPM_LIM value in REG13 is the actual input current limit in effect and it may not match IINLIM value in REG00. 

What‘s the relationship between ILIM hardware pin set limit value and IINLIM register REG00 bit0～bit5 limit value？ Is there any register to restore the value of ILIM hardware pin limit value？

Garrett Kreger said:

ILIM pin voltage can be ignored for input current limit formula. Please use the following formula for determining Rilim, Input current limit = Kilim/Rilim. This formula is listed in ILIM pin description table and in datasheet section 9.2.12. For example, a 260ohm resistor will set input current limit to 1.5A given KILIM = 390 max.

There is no requirement for using ILIM pin as an input current monitor. The device will limit current accordingly on its own without using ILIM pin as a current monitor. It is intended only as an extra feature. Some customers like the ability to easily monitor input current via their MCU using ILIM pin. 

So the input current monitor is just for external use such as being connected to extra ADC or Voltage Compare device to do something else, right?

I always cannot understand the working mechanism of ICO.

1, How does ICO identify (detect ) the unknown adapter's maximun input current--the bottomline / affordability? Namely if an unknown adapter maximum current is 5A, how can the charger ICO know this adapter  maximum current is 5A.

2, ICO uses Current Limit DPM to prevent Voltage Limit DPM,but still must undergoing a Voltage Limit DPM one time to identify the bottomline / affordability of the unknow adapter ,right?

3, I still cannot understand why the ICO initial current limit setting value differ between Vbat>Vsysmin and Vbat<Vsysmin (Vbat<Vsysmin: IinDPM by host; Vbat>Vsysmin:500mA ), can you examplify this difference and let know what will happen if we don't set the ICO_DPM current limit value like this?

I cannot understand the description in  SLVA812C : “When V(BAT) < VMINSYS, the battery voltage can be too low to supplement a large system load if the charger buck converter is limited to 500 mA, then ramped up by the ICO algorithm." ---If the ICO initial limit value is set to 500mA when V(BAT) < VMINSYS, why is it related to the supplement mode of the battery? and why setting the ICO initial value to IinDPM by host can solve this issue?

Let me deep think about it again. Anyway，the ICO feature must be disabled by ICO_EN due to its complexity and safety consideration in my developing practice，and no unknown adapter can be used and we will NOT accept any risk or safety complain if the end user have used an unknown adapter leading to related issue. Unknown adapter is NOT guaranteed since the qualified adapter will be delivered as an accessory of the notebook product to the user.

Garrett Kreger said:

3) BQ25890 prioritizes maintaining SYS rail above the SYSMIN threshold (default 3.5V) when converter is on. If battery voltage is below 3.5V charger IC does not have the ability to utilize supplement mode and still maintain SYS rail above SYSMIN because VSYS must be below VBAT for supplement mode operation. Due to this ICO algorithm steps down input current limit from manual REG00 setting rather than starting at 500mA and stepping up input current limit setting.

1-->So if VBAT < VMINSYS, the IinDPM register by host must be set high enough in case the charger bumps into the current DPM(IinDPM)  before coming across the voltage DPM(VinDPM) which will possibly lead to battery entering into Supplement mode(DPPM), right?

But the question is that the VinDPM can also result in  battery entering into Supplement mode(DPPM) where Vsys drops below Vbat<Vsysmin thus leading to system crash, and what can charger do if the first VinDPM make the Vsys drop below Vsysmin with system crash happening before the ICO takes effect?

Even if the first VinDPM didn't cause Vsys drop below Vsysmin leading to system crash when the system load is not heavy enough, how high should the limit value in IinDPM  register set by host (the max value 3.25A?) to prevent current DPM(IinDPM)  taking place and leading to Vsys dropping below Vsysmin with system crash occuring before coming across the voltage DPM(VinDPM)  when VBAT < VMINSYS?

Garrett Kreger said:

Input current limit should be set based on the current rating of your input adapter. That is the only factor that should influence this setting. The IINDPM and VINDPM settings should always be set in a way to not overload the input adapter.

1,Sorry, I must iterate the background of ICO operation where I talk about how high IinDPM register should be set when VBAT<VSYSMIN, instead of IinDPM and VinDPM alone(I know if not considering ICO, IinDPM and VinDPM register settings should be based on the actual input adapter ratings and charger input ratings.).

Namely, during the context of ICO operation when VBAT < VMINSYS, how high should the limit value in IinDPM  register set by host (the max value 3.25A?) to prevent current DPM(IinDPM)  taking place and leading to Vsys dropping below Vsysmin with system crash occuring before coming across the voltage? DPM(VinDPM)  ?

2,Refer to SLVA812C,why don't we set the input current limit register directly to the max value(3.25A for BQ25890) instead of 500mA and then step it down to the current value on VinDPM  when  V(BAT) > VMINSYS during ICO?

3, As we know, ICO operation is based on first VinDPM in advance, but what will happen if the IinDPM happens before VinDPM during ICO?

4,Will IinDPM happen simutaneously when the charger is exactly in VinDPM if we set the input current limit register directly to a lowest value of 500mA which may triger IinDPM under the heavy load during VinDPM ?

Garrett Kreger said:

Initial IinDPM setting will be based off D+/D- detection results. You are free to raise the input current limit higher via I2C write from MCU host. I do not recommend setting it higher than the current rating of your input adapter.

-->Sorry,I must remind you that we are talking about an current limit unknown adapter, and we don't know the exact rating of this unknown adapter I'm talking about. So " I do not recommend setting it higher than the current rating of your input adapter." makes nonsense for the topic of our ICO for unknown adapter.

Garrett Kreger said:

The typical adapter will not be rated for the max possible current value. The device steps up input current limit from 500mA to minimize overloading of the adapter. If ICO algorithm started at 3.25A most adapters would be overloaded and there is risk of them crashing. 

-->The reason why I set the input current limit register directly to the max value(3.25A for BQ25890) instead of 500mA and then step it down to the current value at the point when VinDPM  reaches when  V(BAT) > VMINSYS during ICO is that we just want the charger encounter VinDPM before IinDPM,not vice versa. So the system crash should not happen due to the VinDPM even supplement mode even if I set the input current limit register directly to the max value(3.25A for BQ25890) ,isn't it right?

Another evidence for this setting methold when VBAT>VSYSMIN can be supported by ICO operation description in BQ25790 datasheet.

If I set the input current limit register too low,such as 500mA, the charger will easily happen to IinDPM just by a light load, which will lead to frequent IinDPM event before VinDPM, is this the ICO's intention---as I understand, ICO's intention is using VinDPM to replace IinDPM to protect an unknown adapter,so no IinDPM should happen before VinDPM, right?

My understanding of ICO when  V(BAT) > VMINSYS is as below:

(1), Let me assume the actual max current value the unknown adapter can provide is 1A ,but remember that we don't know this value=1A in our mind.

(2), When  V(BAT) > VMINSYS, I set the input current limit register directly to the max value(3.25A for BQ25890) to avoid IinDPM happens before VinDPM.（This setting methold when VBAT>VSYSMIN can be supported by ICO operation description in BQ25790 datasheet.）

I assuem that the system crash will not happen due to VinDPM or supplement mode to keep Vsys stably above Vsysmin.

(3), As the system load increases up to 1A and then exceeds 1A, the charger will not bump to IinDPM because unknown adapter's actual current limit is just 1A but the IinDPM threshold is high enough to 3.25A  and the Vin starts to drop due to the load is over 1A and later the charger happens to VinDPM at actual input current =1.5A before IinDPM(3.25A).

If the load continue going up and the VinDPM and battery supplement mode can keep Vsys stably above Vsysmin from crash.

(4), Charger records the actual input current at the point  VinDPM starts and then reduce this current value slightly to a value to act as the ICO_DPM register value and by now the ICO-detected max current limit value is finished and the ICO operation is completed successfully.

Please help point out the risk and vicious consequence if ICO operates like these steps(especially step 2--set the input current limit register directly to the max value(3.25A for BQ25890) when  V(BAT) > VMINSYS ) ,comparing to the right procedures of ICO opertion when V(BAT) > VMINSYS stated in SLVA812C.