The three 12s in parallel should still work even though the 190 has less voltage than the two 140s, and he is getting some amps, so it all does work, sort of.
The MPPT will be having trouble finding Vmp, but it should be at some sort of voltage near that. Just not ideal, but still should get somewhere near proper amps.
With the long thin wire from battery to controller, the voltage drop at those amps will make the controller "think" the battery is higher in voltage than it really is. Depending on the actual voltage seen wrt the settings, it could be in absorption based on its programming doing lower amps instead of bulk. Don't know.
Measure battery voltage at the batts and voltage at the controller's output terminals for voltage drop when there are some decent amps flowing.
This is on top of smk's point of getting the batts low enough so they will accept bulk amps, and doing it at noon so the panels will be at max amps.
Before you can decide if the amps are right, you need the actual Isc of the array at the time of testing --not when clouds keep going by and insolation is going up and down faster than you can do your tests--you need a steady sky for an hour while you poke around. Isc Must be taken with the array disconnected from the controller but you can do it at the controller with the wire ends out of the input terminals.
If Isc is less than rated from lower insolation at the time, then you won't get full amps but you can still compare what you do get with what the battery is getting to check the system.
With MPPT you need a meter to measure voltage at the controller's input terminals when connected and operating (should show what the controller is doing for a Vmp, which can be a changing value.) You should have an ammeter there too so you can get the Imp and calculate the input watts. (unless the controller will display inputs, not just outputs like many only do.)
From that you can see if the output watts (battery voltage times the output amps as seen on the controller's display) is about 95% of input watts (typical controller efficiency)
If something is wrong it should show up from poking around with a meter like that.
You might try measuring input from each panel at the combiner box inputs if the total at the controller input is out of whack. Might find a "guilty panel" in the bunch which has maybe poor wiring connection to its panel terminals.