Well, 22A can use 4mm² wire, while 44A requires 10mm².
It very quickly becomes expensive to route high currents. Which is why higher power applications always tend to use higher voltages.
A good example is Tesla switching from 12V to 48V for ancillary power - dramatically reduces copper weight and cost.
I have loads of cable from previous installs. About 20 meters of 6mm, 30 meters of 8mm and 26 meters of 10mm.
Cable expenses taken care of but yes, that part I understand about volt drop being less the higher the volts and thus thinner cables being sufficient.
Cable (and expense) sorted.
Why is 22 amps with 100 Volts is better than 44 amps with 50 Volts to get my bank full in the shortest time?
If it is purely a cost consideration, then it seems that parallel is the way to go? I need to get as much out of those panels as I can. No room for further expansion.
Theoretical watts will be the same, but practice is different.
Four in parallel won’t work for you unless you have a 24V or 12 V battery.
5V higher than the battery voltage is required by the MPPT, which a 50 Voc panel won’t provide for a 48V battery.
(Remember you are dealing with Vmpp when the panel is working).
Edit: I would see how much spare capacity you have on the existing MPPTs. You may be able to rejig things.
Technically, there is very little difference, if you size the wires to have the same I²R losses (which will probably need more than 10mm² - remember - power loss is I * V so you need 1/2 the voltage drop to have the same loss at 44A).
You will probably also find that the MPPT has similar I²R issues, and will be marginally (unlikely to be as high as 5%) better at lower current.
So you will likely charge your batteries slightly faster at 100V/22A - but the difference will be marginal.
In non technical terms, are you saying that the “speed” of the 22 amps setup will be faster than the 44 amps setup?
In other words, will the rate of flow of the 22A line be faster than the 44A line because of the higher volts? Because this is the only way it makes sense to me - that 22 A can outperform 44 A
Now THIS is something I didn’t think of and a very valid and crucial point! Thank you very much for this!
Sigh. Seried parallel it is then.
The buck stops here… 
Sorry, couldn’t resist…
True, but please, if the buck was still running (and it is for me) I really need to understand.
Was my attempt at an analogy above that bad? Higher volts equals a higher amp flow rate?
Losses are proportional to the square of the current. So if you double the current, you need 4 times the copper for the same losses. You will also have much higher losses in your MPPT.
Other than that, 10A at 10V is the same as 20A at 5V.
It’s more like “a Watt is a watt and… etc etc”. An MPPT can swap volts for amps, so that you get (with 95% efficiency or thereabouts) the same number of watts regardless of input and output volts. Given that you spent money on an MPPT already, you might as well go for the highest voltage allowed, which will be 2 x series on the 150V controller. And as Phil said, going lower wouldn’t work anyway.
There was a bit of a joke involved in that, but a highly technical one. The MPPT employs a “buck converter” to drop the voltage. When @justinschoeman said “the buck stops here”, he was quite correct. The buck (converter) does stop when you don’t have enough volts on the top to make it “run” 
Believe it or not, I actually got this lol! I read some of your comments about buck converters.
Just trying to make a joke of my own.
Have not spent it yet, but are very close to doing so. Would you suggest going for an MPPT with a higher volt rating then?
Looking for the best bang for my “buck” without breaking the bank.
(was this joke better?)
If money was no object, I’d go 4-series, one string, into a 250V MPPT.
A 150/45, which is what you’d need for 2kW or so, is around 6k. A 250/60 is 10k. For 4k, I can buy a LOT of cable.
I know 22A is no biggie on 6mm^2 cable, but it makes debugging easier if the strings are separable on the ground. So I would pull in two strings, and parallel them into a 2s2p on the ground.
Thanks, definitely the way to go if I was planning on further expansion. But I have now run out of practical roof space so this is my “last” expansion. And as mentioned, I already have surplus cable.
And this “last” expansion will be easily reachable for any fault finding purposes.
Lastly, on the comment of swapping volts for amps, does this actually mean that if I have sufficient voltage, then I could get more than the 22 amps that the panels are rated for?
Geez, all these posts and I said …
… in non-technical terms.
Ps. I have 4 strings from the roof … as Plonk said, it makes for easier “debugging”.
Yes but I need to understand why.
In my little head I need to grasp the basics.
Oh absolutely. That’s MPPTs 101. Or Buck converters 101. Or Transformers 101… for AC.
You put in 100V at 22A… you get out 50V at 43A (I took 2% for conversion). You see, conservation of energy has to hold, watts in (100V times 22A) must equal watts out (50V times 43A plus 50W of fuzzy warmness).
Please tell me you are not planning on paralleling the panels on the roof. If you are, I should maybe pop in for a coffee an convince you otherwise.
Living in Pretoria now it’s less than 30 min drive
Thank you, this is all I need to know and feeling warm and fuzzy myself now!