Actual Solar Panel Watt Capability

[Sunking]

I get it - it's a current source. But it still has to obey Ohms law, no?

Sure it does. First look at the IR curves published for solar panels. The portion you want to pay attention too is the vertical line that runs from MPPT Knee down to Zero Volts at Zero Ohms. The current is almost linear. For your panel I am guessing at MPPT voltage the Imp current is 2.9 amps, and at Zero Volts or Isc is 3.1 amps. So what is the Power? It is undefined as it depends on what resistance you are pushing the current through. The highest current possible assuming 100% irradiance input is 2.9 amps so the power is 50 watts @ Vmp of 17 volts.

But what happens if say we take a 6 volt battery and connect it to the panel. Imp has not changed as it is still 2.9 amps. But we pulled down the voltage from 17 volts to 6 volts. What is the power now? Could it be 17.4 watts?

What if there is haze in the area, or the panel is not facing directly into the sun so the irradiance has dropped to 50% input. Well since a panel is a current source and the output current is directly proportional to light input our current falls off to 50% or 1.45 amps. Vmp does not change and is still 17 volts, but now Pmax is 25 watts.

So answer some simple Ohm's Law equations using a constant current source. What is the voltage and power with 2.9 amps throwing through

0 Ohm
1 Ohm
2 Ohms
3 Ohms
4 Ohms
5 Ohms
5.94 Ohms

Now here is where magic happens with a solar panel When you go above MPP voltage, the panel now turns into a voltage source up to 22 volts at Open Circuit infinite ohms. So when at 21 volts what is the power and current at 20 ohms?

Run through those equations may shine some light in a place you have never seen before.

[inetdog]

But it still has to obey Ohms law, no? I am listening.

Ohm's law is really only useful when the resistance, R, of a circuit element is close to constant. The problem with PV and with any semiconductor device is that R is not a constant over any wide range of current and voltage.
In the I versus V plot that Dereck mentioned, the slope of the curve at any point (or its inverse, depending on how you draw the curve) is the value of R at that point. Not even close to constant. And even then, that curve is only for one specified light input. There is actually a family of curves, one for each amount of incoming light.
And, of course, the panel can modeled as the equivalent circuit of a light-dependent current source in parallel with a light-dependent loss resistance. This comes close to modeling the maximum voltage and maximum current limitations that you see, but is messy to work with.
Or you can model it as a constant voltage source in series with a light-dependent series resistance.

[Naptown]

Ohm's law is really only useful when the resistance, R, of a circuit element is close to constant. The problem with PV and with any semiconductor device is that R is not a constant over any wide range of current and voltage.
In the I versus V plot that Dereck mentioned, the slope of the curve at any point (or its inverse, depending on how you draw the curve) is the value of R at that point. Not even close to constant. And even then, that curve is only for one specified light input. There is actually a family of curves, one for each amount of incoming light.
And, of course, the panel can modeled as the equivalent circuit of a light-dependent current source in parallel with a light-dependent loss resistance. This comes close to modeling the maximum voltage and maximum current limitations that you see, but is messy to work with.
Or you can model it as a constant voltage source in series with a light-dependent series resistance.

Huh?
I'm sure the OP digested and understands that completely

[inetdog]

Huh?
I'm sure the OP digested and understands that completely

My hope was that the OP could at least handle the first paragraph; the rest was just for fun.

[PNjunction]

Was that a PowerWerx inline meter? Nice and handy. I dig mine, although it was a Turnigy. Not as accurate as my Fluke 87V however. As a 130 amp meter, it may not be totally accurate with the calculations at the very low ranges like 2 or 3 amps. A calibrated shunt will tell you. Still a good instrument though.

What you may also be experiencing is common for batteries that aren't in a moderately discharged state. They charge up rapidly (the battery terminal voltage rises during charging and tries to meet the charge voltage source.) and when it gets close to the panel voltage output (or the voltage limitation of the charge controller), the battery naturally tapers down the current in the "absorb mode". Could THIS be the 35 watts you are experiencing? In absorb, the battery itself is doing the current-regulating (albeit voltage limited by the charge controller and eventually very little current flows.

There have been a few that have hooked everything up this way with freshly charged batteries and seem to think that their solar setup is bad when they observe a miniscule amount of current flowing - unknowingly because they are immediately into the absorb stage, and not the bulk charge stage.

We have no way of knowing how that nicad is reacting, and isn't really meant for a cv solar charge controller anyway. I'd perform the test again, but this time with a *FRESH* lead-acid battery that has some amount of discharge to it. An 18 to 22ah sealed agm should be easy to find. Of course this would have to be at your solar-insolation maximum period (somewhere around noon-ish), no clouds/haze, aligned towards the sun etc.

Going from 35 watts at 18v to the max of 50 watts at 18v is about a calculated 830 milliamp difference. Let's say your panels are on the low end of their stated 3% manufacturing tolerance. Maybe a bit more. Toss in some variance in the accuracy of the powerwerx itself. A bit of loss in the controller....some non-optimal solar-insolation as compared to the manufacturer's STC flash test ....

I think you are probably doing fine.