Connecting solar panel negative to battery negative for common ground?

[Red Squirrel]

I have a little portable solar system I will be redesigning to add a bit of electronic features to it, basically a LCD screen with voltage/current/watts and a very basic UI.  I want to be able to measure the voltage of the actual panels, the amps being drawn, and ditto for the batteries.   A voltage divider will be used for voltage and then connected to an analog pin of the MCU.  For current I'll use a negative side shunt and op amp and also connecting to analog pins of MCU.   The issue is that I'm not too sure how I would go about measuring the panels as I am not sure if it's safe to just bond the negative to the system "ground" so that it's all at same potential as not sure how the charge controller will handle that.   Is this generally safe to do?

What is the best way to verify with my current charge controller?  If I take my volt meter and probe between solar negative in, and the battery positive out (with battery connected) I get 12 volts, but from battery positive to solar positive in I get like +0.4v.  If I use my ohm meter I get about 3k.  So the grounds are NOT connected internally, but also not floating, which seems a bit alarming.

If I can't connect them, what is the best way to go about measuring that voltage then?  I could have a separate MCU and then use i2c or other communication with optoisolators to talk to the main MCU I guess, but that seems overkill and I might be overlooking something.

[mtdoc]

There shouldn't be any problem with bonding negative to system gound.  For stationary PV systems, bonding panel/CC/battery negative to earth ground is common practice and considered by many to be the preferred method - though it is somewhat controversial since it may defeat the USA NEC required GFP system. (For some detailed discussion of this issue see here and here)

Note - there is at least one company (Sunpower IIRC) that makes PV panels meant to have a positive ground. But those are oddballs and you are unlikely to be using one.

[Red Squirrel]

I decided to test but got some odd behaviour.  I used a flashlight just so the panels can produce a few volts, and was measuring with the meter at the panels (connected to charge controller, with charge controller connected to batteries).  I then took a crocodile clip connected from the battery ground and connected it to the solar panel ground, and voltage dropped by about 1 volt and one of the LEDs on the controller lit.  I don't think it liked it.  So it seems with this controller they are connected but not directly and by bonding them I'm shorting something out.  Here's a pic of the controller if it helps:



That big 3 pin package is a mosfet, so I'm guessing it's doing some PWM or just adjusting voltage linearly.   


Edit: Voltage drop seems to be proportional to actual output voltage.  I got the panels up to 5.2 volts, if I connect battery ground to panel ground I get 3.4v.

[alsetalokin4017]

Well, looking at the large photo you provided, you can see what's going on. The Solar panel negative is connected to the Battery negative through the diode D3. The Solar negative is connected to the mosfet (IRF1018E) Source pin and the D3 anode... and the Battery negative is connected to the D3 cathode and the mosfet Drain pin and tab. This is also connected back to the Solar negative-mosfet Source through resistor R22.
So when you connect the two negatives you are essentially shorting out the mosfet, R22 and diode D3, at least.

[mtdoc]

Yep, I misunderstood your original question. I thought your were asking about bonding your CC negative output and battery negative to earth ground (which the panel frame and CC case would also be bonded to). Direct connection of PV negative tp battery negative won't work if there's a CC involved.

Looking again at the questions in the OP:

To measure panel voltage - measure between PV + and PV -
To measure battery voltage -measure between Battery + and Battery -
To measure amps into battery from PV - put your shunt between CC  and battery or  between battery and load to measure current to load of course.

[Red Squirrel]

Ok but to measure between the two PV doesn't the negative have to be the same as the MCU negative?  Ex: one wire would go to negative, the other wire would go to the analog pin. The MCU will measure the voltage of the analog pin in reference to the system (battery) ground.     Since I can't combine the two negatives, is there another way to measure the voltage from the MCU?

Worse case scenario I might just get another charge controller (but I have no way of knowing if it's going to be floating or not so it's kinda a trial and error) or maybe just make a 13.5 volt buck converter, I could get fancy and do MPPT but I want to keep it simple for now.

[Red Squirrel]

Been thinking this over and think I might actually remove the charge controller altogether and regulate the voltage myself.  My level of electronics knowledge is not super high so want to keep it simple.  I'm thinking a buck converter controlled by the MCU. Then I can wire it so the solar negative is common through the whole system to make voltage and current measurement easier as that is always in reference to the MCU ground.  Ex with voltage divider one end of the output goes to the analog pin and one goes to ground, so with an uncommon ground like I have with this controller this becomes difficult to do.

Within the MCU I'll then ensure the output voltage stays at 13.5v or so.  Can monitor current going to battery too and raise this and basically do a smart charger too to charge the battery faster.   This won't be as efficient as a MPPT charger but will it work for the most part? Since the MCU will see the current/voltage of both sides I can add some rules in there like, don't allow panel voltage to drop within a certain point etc by adjusting output voltage.   Basically try to find the best voltage/current curve of the panel and run at that.

Failing that I can just keep the existing controller and forget about monitoring the voltage/current of the panel.  It was just a nice feature I wanted to add as it gives a good indication of the panel's performance.