Hi everyone !

I need to measure the voltage of small solar cells, and use different resistors to make the I-V curve. Basically, I apply a resistor of 12ohms, measure the voltage, apply 47 ohms, measure...
To change the resistor quickly, I put these resistors at the emitter of parallel transistors in saturation: this way I can use the resistor(s) I want.
I make the measures and I apply the current base to transistors using an Arduino.

However, my electronics or calcul are probably wrong, because I don't get the right values, they are too high. I know it because I have a device that does exactly the same thing (but it's expensive), so I used it to compare the values.
For example, if I only apply current to the first transistor, who has a 12ohms resistor at its emitter, I get around 0.064V instead of 0.024. Also, if I apply a current to the base of the two first transistors (12 and 47 ohms, in parallel = 9.5 ohm) the voltage increases to 0.94V, instead of decreasing.
Maybe my circuit or a resistor calcul is wrong, I don't know...
I attached the part of my circuit corresponding to the problem

Thanks

Edit : I use 4x 2N2222 and 4x BC547 transistors

 

Your parallel transistors are NOT saturating. They are emitter follower, or common collector circuits, which cannot saturate. The base voltage never can be above the collector voltage (a critical requirement of bjt saturation) because both are powered by the same potential. Because of this, they are introducing a fixed DC offset to your measurements. Also, the base currents are flowing through your test resistors, another error source. To fix this, move the test resistors to the collectors, and tie the emitters to GND. There still will be a DC offset voltage in your measurements, but it will be Vcesat, not Vbe.

And make sure the solar cell negative terminal is tied to your system GND. In the drawing it is floating. Wait... it is tied to GND, it just isn't clear because of the way the schematic is drawn. Use a GND symbol to avoid confusion and mistakes.

ak

 

Maybe use FETs with low Rds(on)?

 

Use the MOS-transistor FDV303N (Rds ~ 1Ohm). Sources of the transistors to negative supply (GND). Resistors turn to the drain. More powerful transistors will have a big leak.

 

Base current might be throwing the readings off to some extent. Another reason to go with a MOSFET per Bordodynov.

(Also note: If the source resistance is >> the emitter resistor then for practical purposes the device can be considered saturated.)