Parallel chargers (synchronous buck-boost)

[kathmandu]

Hello,

I would like to charge a battery using two current sources (solar/wind). I thought of using two buck-boost converters in the following configuration:



The two chargers are designed to run in CCM mode hence the inductor currents (i1 and i2) flows like in the diagram above.

During normal operation, i1 and i2 are (supposedly) flowing through the battery only. But what happened when one of the current sources (i1/i2) is much smaller (or almost zero)? Does the strongest current flow through the opposite inductor, "charging" it with a negative current?

Anyway, for this particular case, I could define a control scheme for (not) driving the synchronous Mosfet when the current through that particular inductor is too small (thus changing the operating mode into an asynchronous one).

But, when both i1 and i2 currents are large, is there any (bad) influence if the converters run in synchronous mode?

I hope I made myself clear: in synchronous mode, the two (charged) inductors are present in the same loop, with opposite current senses.

Thank you very much for your time.

 

[BradtheRad]

Suppose you were to use identical switching frequencies, and stagger switching of the two sources? That way you should be able to manage current flows so as to prevent conflicts of direction or polarity.

 

[kathmandu]

Unfortunately, I can't synchronize the two converters. Looks like I have to implement a current sense solution, to disable the control of the synchronous (high side) switch when the corresponding inductor current falls bellow a (small) threshold.

Of course, there is another handy solution - to put a diode at every converter output but that's not economical (and I could simply use asynchronous converters to achieve that current flow separation).

Later edit:

Actually, one can not use a synchronous converter without implementing that output current sense control. With or without an additional parallel converter, if the inductor current is null when the synchronous (high side) switch is activated, the battery itself would generate a negative inductor current.

 

[mtwieg]

During normal operation, i1 and i2 are (supposedly) flowing through the battery only. But what happened when one of the current sources (i1/i2) is much smaller (or almost zero)? Does the strongest current flow through the opposite inductor, "charging" it with a negative current?

Anyway, for this particular case, I could define a control scheme for (not) driving the synchronous Mosfet when the current through that particular inductor is too small (thus changing the operating mode into an asynchronous one).

This is sometimes done with synchronous converters. Allowing negative current when average current is low is called "forced continuous conduction mode" (FCCM). Many controllers can prevent this by either shutting off the synchronous switch when current falls to zero and keeping it off for the rest of the switching period, effectively making the synchronous switch act like an ideal diode so the converter operates in DCM. A few will operate in BCM instead, and increase the switching frequency to modulate Iavg.

FCCM doesn't damage anything, but its light load efficiency is worse than DCM or BCM.

 

[kathmandu]

Thanks for sharing this, never heard of FCCM before.

I was trying to understand where the negative current goes when the lower switch is activated. Seems like is flowing through the input capacitor only (as the input sources - solar/wind - has reverse current protections).

Anyway, I better stay away from this back flowing. I'm going to use a current sensor anyway ( for metering/short circuit protection) hence it only takes few lines of plain C language to achieve that ideal diode operation.

Once again, thank you all for your kind support!

 

[mtwieg]

I was trying to understand where the negative current goes when the lower switch is activated. Seems like is flowing through the input capacitor only (as the input sources - solar/wind - has reverse current protections).

The input capacitor should absorb all the high frequency ripple, yes. None of your components, including the source, should behave any differently if you enter FCCM. The real concern is if your average current becomes negative too long for the input capacitor to absorb it, and suddenly you are trying to push DC current back into the source. That's what you really need to worry about.

Anyway, I better stay away from this back flowing. I'm going to use a current sensor anyway ( for metering/short circuit protection) hence it only takes few lines of plain C language to achieve that ideal diode operation.

Simply disabling the synchronous fets entirely when Iavg is below a certain point is a fairly simple way to avoid the problem. But be aware that the converter's transfer function will change between DCM and FCCM. And when you change from one mode to the other, you may see transient disturbances and possibly even some oscillation between the two operating modes. Check its behavior carefully.

 

[kathmandu]

Actually, I'm talking about a 25A solar charger (CCM). If the average output current drops bellow 1/2/5A or something (clouds, sunset), I could switch to asynchronous mode.

I mean, I don't have to continuously watch for zero average current, to play with FCCM/DCM and stuff like this. I simply can afford the freewheeling diode voltage drop at 1-2-5A and bellow.

 

[mtwieg]

Actually, I'm talking about a 25A solar charger (CCM). If the average output current drops bellow 1/2/5A or something (clouds, sunset), I could switch to asynchronous mode.

I mean, I don't have to continuously watch for zero average current, to play with FCCM/DCM and stuff like this. I simply can afford the freewheeling diode voltage drop at 1-2-5A and bellow.

Switching to "asynchronous mode" will make it operate in DCM. The point at which you switch modes should as close to the boundary between the two as possible. Otherwise you will see transient disturbances to Iavg when changing modes.

 

[kathmandu]

Switching to "asynchronous mode" will make it operate in DCM.

I'm afraid I don't understand your warning. By "asynchronous mode" I meant not driving the high-side switch (as pictured in the diagram above). The converter operation will not be affected - excepting the increased conduction losses due to freewheeling diode forward voltage.

How could this affect CCM/DCM operation??