LM66200: Reverse current blocking and "activation current"

Hi Jim,

Please see this thread - the IRCB specification only applies to if there is a voltage source on VOUT higher than the VINs which can cause a large current to flow from VOUT to VIN. In this case, current would flow from VOUT to VIN until the 1.4A threshold is reached, in which case both inputs would be disconnected from the output and the output would go Hi-Z. This is where the figures you posted are applicable. However, during normal operation when VOUT is not greater than the active VIN, the device compares the input voltages and outputs the highest one, and the inactive input is disconnected from the output and other input. The inactive input has 50nA typical leakage into VINx.

During switchover, the active source is first disconnected from the output, then tSW time (on the order of 10us) after the other source is connected. This prevents any cross conduction between the channels from occurring.

Similarly, with LM66100, if CE > VIN, the output goes Hi-Z. However, if you need this device to start blocking reverse current only once reverse current starts flowing (this is where the 0.5A is relevant), then CE must be connected to VOUT - this is the only way that LM66100 can detect reverse current. This feature is integrated in LM66200.

In essence - the 1.4A and 0.5A reverse current thresholds are only applicable if your VOUT is being driven at a higher voltage than your inputs. There will be no current flow between inputs.

Also, thanks for pointing out the lack of room temp trace on the second figure! I will see if we can get that updated in the next data sheet revision.

Thanks,

Patrick

Hi Jim,

Yes, that's correct - the figures are taken when there is a voltage source on VOUT, VIN1 is pulled to GND, and VIN2 is floating, so RCB is active. Thanks for your feedback, I'll see if we can make this more clear in the next data sheet revision as well.

If you have a voltage source on VOUT that you'll need to block <10mA reverse currents from/to your batteries, I have two recommendations:

1. Use a blocking diode on VOUT - but that somewhat defeats the purpose of using a dual ideal diode.

2. Use a precision comparator(s) to compare VINs to VOUT with the output connected to the ON pin, and when VOUT > VIN drive the ON pin high, disabling the device. This should disable the device at reverse currents below 10mA.

Do you think any of the above could work?

Thanks,

Patrick

Hi Jim,

Thanks for the explanation - in this case, provided that the batteries aren't directly shorted together then LM66200 (or TPS2116/2117, which are very similar but feature more control options) will certainly work here. There is no risk of a new/higher voltage battery discharging into the old/lower voltage battery, either internally or through VOUT. But unfortunately, this device does not have reverse polarity protection. Either discrete diodes or ideal diodes will work here if you need reverse polarity protection - but ideal diodes usually have higher Iq and a higher minimum VIN.

Thanks,

Patrick

Hi Jim,

Sure - what the IRCB spec describes is that, if there is a voltage source that is only on VOUT (not through the inputs) from which a source flows back into the active input (the inactive input automatically blocks reverse current), then when 1.4A typ, 4A max flows from the VOUT source to the active VIN, then the device will disconnect both inputs from the outputs and make the output Hi-Z. So it only applies if you have a third power source on VOUT, and it doesn't apply if you are only powering your circuit from the sources on IN1 and IN2.

The device determines which output to use by comparing VIN1 and VIN2 - it connects whichever source is higher to the output. When the device switches from one input to the other, it goes as per the following example where the device is switching from VIN1 to VIN2:

-VIN1 active, VIN2 inactive (so VIN2 is blocking all current in and out of the IN2 pin except 50nA leakage into the pin)

-VIN1 disconnects from VOUT (so that it blocks all current except 50nA leakage into the pin), VIN2 is still inactive

-After tSW (around 10us), VIN2 connects to VOUT

As you can see here, there is no possible situation where VIN1 and VIN2 are both connected to VOUT at the same time, or connected to each other. (It's also worth mentioning here that this device cannot have both inputs enabled at the same time like you would see in a real dual diode when both inputs are the same voltage.)

So, in essence - the device selects which input to connect to the output based on the voltages of the inputs, not based on how much reverse current goes through a channel. As per the device design, only one channel can be active at a time. If there is a voltage source on the output that is greater than both inputs, current will flow from the output to the active (highest) input - if 4A max flows from the output to the input in this condition, then the device will disable the active input so that both inputs are disconnected from the output.

I can see from the data sheet how it can be confusing - does that make sense? Do you have a third voltage source on your output - one that is not the IN1 or IN2 supply?

Another perspective to look at this from is by looking at Table 7-1 of TPS2117 data sheet, in the row titled "Diode Mode." This is a very, very similar device that expands the control modes of LM66200, but the Diode Mode is the same operation as LM66200's operation.

Thanks,

Patrick

Hi Jim,

While I'm not familiar with the internals of the simulation model as I did not make it, it is accurate in portraying the no leakage from VOUT to VIN.

Yes, the table you posted is exactly correct.

1. Yes, looks great!

2. There's no hysteresis or threshold on the input comparator, so it is very sensitive.

3. Yes, this can happen and in severe cases it can cause the IC to go into soft start and/or a system reset because the output drops to far. If this turns out to be an issue (or if you want to proactively avoid this possibility), this can be mitigated by adding capacitance on the inputs so they droop less when they're connected or it can be prevented by using the previously mentioned parts, TPS2116/7, which can be programmed to use Battery 1 until it reaches a threshold of your programming, after which it switches and uses Battery 2 until Battery 1's voltage is high enough (both the falling and rising thresholds can be programmed). But most of the time we don't see this "fluttering" issue to be bad, especially at lower load currents.

4. Rather, the leakage is from VIN2 to GND, not VOUT to VIN2 - but yes, <=50nA typ (up to 2.1uA max at 105C) leakage into the inactive input. I believe there is only leakage coming out of VINx when it is below 1.6V, as otherwise it will be sourcing the 50nA - but I'm more confident in the concept of this than the exact voltage the input would stop sourcing and start sinking, though.

Best in the rest of your design! Please don't hesitate to reply here or start a new thread if you have any more questions.

Thanks,

Patrick

Hi Jim,

That's correct.

Thanks,

Patrick