question

There is no simple answer to this question. It depends on many variables. All simple answers by someone who doesn't ask a lot of questions have a good chance of being wrong.

As others have said, your latitude, but also the pitch of the roof, your seasonal weather patterns, your seasonal power usage patterns, any shading issues, how close to "true north" your roof faces... probably more...

Whereabouts are you located? roughly...

This may assist as a general guide for Panel Placement. (courtesy one of Margreet's Presentations.)

You will notice here the Facing of the Panels is almost irrelevant in decent panels - you will get up to 90% capabilities even when the Panels are flat (assuming you are not Either too far South or too far North.

For years I used to consider it almost a crime when I would see panels split across roofs or Flat instead of at an angle - but you can see here it doesn't make a huge amount of difference. That said - with your 4hrs of Peak power - that is true, however you can also see that if working backwards from 100% - you would get up to 40-45 Degrees either side of perpendicular still with up to 90%, and then only as low as 80% efficiency by the time you hit 70 Degrees either side of perpendicular.

Just wondering why you are going with 2 MPPTs if split though, is it simply to avoid having NO sun from one array of panels causing the other panels to Not provide any energy via the set of panels that are in sunlight? Good thinking if that is the case.

My personal opinion would be - if you are down south - like I am in Australia - North Facing, and angled if you can. It may just come down to doing some hard maths based on:

Hours of sunlight from 7am > 12pm and 12pm > 5pm - so 10 hours of Total sunlight - and using a chart like this image to estimate the total yield against a North facing set of ALL panels.

vs

The same 10 hours split over East and West so 5 hours per set - but higher yield for each set during that 5hr period only, and none for the other set.

I'm fairly sure mathematically you will find a single array North only will work out better, particularly if your East and West are particularly different angles, i.e. East gets no direct sunlight after 1pm for instance, and West no Sunlight until 1pm.

There is a simple way to experiment with it - Get a couple of Small Hobby Cells $5-$9 from somewhere like Jaycar, Element14, RSComponents, etc.) - set 2 of them up Flat and hook to a multi-meter or something that can log small power over a time period, use an incandecent bulb and run it across them in the same arc and angle you would expect the Sun to move. Next do the same thing but angle the 2 panels on the same angle you would have them on the roof. You need to do this with some accurate scale timing to replicate the Sun's movement, take overall measurements and compare.

Next to consider is whether you Serialise or parallelise. Serial will give you More hours of Sun power, and Parallel will increase the current during Sun periods based on how the MPPTs process the input voltage of the PV (i.e. start charging at +5v over the Battery voltage.

It is my opinion that an E/W split is generally a better fit than an N orientation. (Yes, there are site-specific considerations, so adding this to the mix of other comments).

Depending on your power usage pattern. An N orientation has the potential to deliver more energy than an E/W orientation, but as a noon peak. If you cannot utilize this energy it is wasted.

The E/W orientation whilst not having the potential to deliver as much energy in a day, delivers the energy it does provide over a longer day with less midday glut. Thereby actually delivering more useful energy.

My opinion is further reinforced by the fact that solar ancillary equipment( MPPT's, cabling and even batteries to an extent etc) has to be sized to be able to cope with peak delivery. This ancillary equipment is relatively expensive compared to PV panels.

For example with an E/W orientation, I can downsize my MPPT's but have them work to full capacity longer in the day, than have more expensive MPPT's that are idling most of the day.

But nobody downsizes in reality, you either plan the system from the outset or optimize what you have.

So either I could plan to achieve my outcome with cheaper equipment or alternatively I could optimise and buy more panels to compensate for the loss of N orientation with the savings from downsizing the ancillary equipment.

There is a balance, I started with an N orientation as a small system, most people do when they can use every watt they can get.

However, as I have expanded the system I have blended other orientations so as to try to get the most value from working my ancillary equipment over a longer day.

In practical terms, using MPPT's originally sized for N orientation, I can nearly double the amount of PV and still not have clipping. (I have to add though that I originally sized my MPPT's theoretically, in practice after monitoring the system they had a 30% headroom to start with).

I now push them 95% for hours in the day, but I still never clip. My usage pattern is such that I can use every watt, whenever I produce it, and I believe I am producing more watt/hours in a day than I would be producing with an N orientation with the same ancillary equipment.

My usage pattern is not common, but I submit I could still have adapted my usage pattern, which is a common option.

For what it's worth, I am 30 degrees South, and latitude did not influence my thinking.

Edit & PS: I just want to add that more panels beat fewer panels hands-down on an overcast day when orientation is not a factor and all panels are running at 10%. This could really have an influence on battery size, which is a big expense.