[SOLVED] Buck-boost PWM scheme

[red_alert]

I want to build a solar battery charger (Uin = 20 .. 90V, Uout = 48V, P = 2000W) using a synchronous buck-boost converter topology (H-bridge).

I wonder what's the best PWM scheme to drive the bridge switches (MOSFETs). Most integrated buck-boost driver ICs are using three separate schemes: buck only, boost only or buck-boost, depending on Vin-Vout. That means one of the bridge legs has the high-side switch always ON and the low-side switch always OFF during buck only or boost only operation.

What if I drive all four MOSFETs using a single synchronous switching scheme (by driving diagonal switches simultaneously, with complementary PWM signals), regardless of Vin?

Does it significantly reduce the overall efficiency?

 

[FvM]

Does it significantly reduce the overall efficiency?

Yes. It means that transferred power has to be stored temporarily in the buck-boost inductor, otherwise only a fraction of it. Switched power and switching losses as well as inductor losses are respectively increased.

 

[D.A.(Tony)Stewart]

Better to use a PMMT regulator otherwise just a Buck regulator that keeps PV at 75~80% of Voc unless there is overvoltage on battery then regulate to match chemistry charging profile.

The energy of the PV below 48V is <<10% with low solar input, so there is little power transferred by boosting it, unless that little energy is important for some reason. With moderate to high solar input, PV efficiency drops if loaded down to <=48V so no gain by using a boost regulator. Just use a half bridge buck switch IC with low RdsOn with Smart Switch

 

[red_alert]

Better to use a PMMT regulator otherwise just a Buck regulator that keeps PV at 75~80% of Voc

Yes, I'm going to use a MPPT algorithm (P&O, most probably) to keep the output power at maximum. I'm using a uC to drive the buck-boost converter thus it will be easy to implement the MPPT function.

The energy of the PV below 48V is <<10% with low solar input, so there is little power transferred by boosting it, unless that little energy is important for some reason. With moderate to high solar input, PV efficiency drops if loaded down to <=48V so no gain by using a boost regulator. Just use a half bridge buck switch IC with low RdsOn.

That's why I'm building this charger in the first place. For now, I have all the PV panels directly connected to the battery bank (Vmp = 2 x 29V) but during low lights (sunrise, sunset, clouds, fog) or hot weather, the Vmp is drastically decreasing (actually, even Voc reaches 29V threshold) hence there's almost no charging current available.

For my 48V battery bank, it would have been better to use 3 panels (Voc = 75 .. 135) or 4 panels (Voc = 100 .. 180) in series. This way, I could have build a simple & efficient interleaved synchronous buck converter.

Unfortunately, I just have 8 panels (hence no 3x option) and I didn't find any suitable MOSFET for the 4x variant neither (Vds = 400V or more and an acceptable Id current and low Rdson).

In synchronous buck/boost configurations, all the switches have to be rated at the load current (plus ripple) thus it's much harder / expensive to build a higher input voltage one.

Is there any other better option?