# Oscillation problem on a circuit with capacitor.

Discussion in 'The Projects Forum' started by abhaymv, Apr 3, 2014.

1. ### abhaymv Thread Starter Active Member

Aug 6, 2011
105
5
Hello,

I am using the circuit to charge and discharge a capacitor using a solar cell.
Circuit is attached.

Circuit Working:

The capacitor emulates the behavior of a variable load. The voltage and current across the capacitor at different points of time during the charging phase of the capacitor gives the current voltage characteristics of the solar cell. When Mosfet M1 is ON, M2 will be off and the capacitor is temporarily short circuited. The current flowing through the sense resistor R12 at this point will be (almost) the short circuit current. As the capacitor charges, the voltage across the capacitor increases, and the current flowing to it decreases. The voltage and current values at these points are sampled, and they trace through the different points of the IV characteristics of a solar cell. When the capacitor is around 99.9% charged, the voltage across it will be (almost) the open circuit voltage of the solar cell. After I and V values are taken, the capacitor is discharged through the 10 ohm resistor by
turning M1 OFF and M2 ON. M1 and M2 are controlled with a microcontroller.
This is how this circuit is intended to function.

Specifications:

MOSFET: PSMN050-80PS N channel 80 V 46 mohm standard level MOSFET,
with a threshold voltage of 1 V.

Solar Cell: Voc = 0.4 V, Isc = 70 mA. (System was designed to test solar cells
with Voc upto 12 V and Isc from 40 mA to
3.2 A. This is the data of the solar cell
I have with me.)

I'm afraid I have no clue what the ESR of the capacitor is, as I do not have an ESR meter and the datasheet (if you could call it that) doesn't mention it.

Problem:

I connected the abovementioned solar cell to the drain of M1. I applied pulses from an ATMEGA328p microcontroller. Observing the voltage across the capacitor on a CRO, I saw that the peak voltage level while charging is only 0.2 V instead of 0.4 V. I had calculated the required time period of the pulses from simulation and I had the pulses on for about 250 ms. But the capacitor didn't charge to the maximum. When I amplified the pulses applied to the gate to 0-15 V, this problem was more or less overcome. However, now I am facing another issue.

The capacitor gets charged to 0.4 V, then it gets discharged to 0 V (discharge pulse fed to M2 is 2 s long). This repeats as per the program in the microcontroller. Now the problem is, once the charging and discharging cycle starts, it doesn't stop! Even after removing the solar cell from the circuit, the capacitor is still getting charged and discharged. I viewed the charge and discharge waveform in a CRO while the solar cell was illuminated. Then I covered the solar cell with my hand. Then I totally removed it,
leaving the drain of M1 floating. But still the system continues to charge and discharge the capacitor! There is no visible change in the CRO waveform, it continues as it is in the absence of the solar cell.
I rechecked this using an Aplab Digital Multimeter configured for DC voltage
measurement. As the signal is pretty slow, I could see the voltage levels flickering up and down, sometimes touching negative values. (The negative reading may be the result of some error of the multimeter, though).

Troubleshooting steps followed:

I removed the pulse input to the gate of M1. The oscillations stopped. Same was the case when I removed the pulse input for M2. They were left floating when I removed the pulse. When I reconnected the pulse, oscillations continued.

Then I shorted the capacitor bank (three parallel 4.7mF capacitors) one by one. As long as the capacitor is shorted, the oscillations subside to 0 V DC, but the moment I remove the short, they restart!

I had to short all three capacitors at the same time to get those oscillations to stop.

I want to fix this, but I'd also like to know how this is happening. Can someone please help me?
The circuit should have worked according to simulations, but I don't think I can simulate the effect of removing the solar cell while preserving the circuit state...

File size:
23.2 KB
Views:
36
2. ### Alec_t AAC Fanatic!

Sep 17, 2013
9,444
2,289
According to the datasheet for that FET, depending on applied voltages and temperature there could be up to ~1uA drain-source leakage current and 100nA gate-source leakage current. Perhaps that is the cause of the phantom charge/discharge.
As for simulating cell removal, just insert a controlled ideal switch in series with it.

abhaymv likes this.
3. ### abhaymv Thread Starter Active Member

Aug 6, 2011
105
5
Thanks for the reply. But if they are due to the leakage currents, I'd get a waveform with a much longer time period because the leakage currents are much lower than the current supplied by an illuminated solar cell (which gives around 70 mA). The waveform I observe in the CRO continues as if the circuit was never disturbed even in the absence of solar cell...

Also, thanks for the tip on simulating cell removal!

4. ### Alec_t AAC Fanatic!

Sep 17, 2013
9,444
2,289
Looking again at your circuit I think the charge/discharge is due to the output of opamp U3 pushing/pulling current via R8-R11.

abhaymv likes this.
5. ### abhaymv Thread Starter Active Member

Aug 6, 2011
105
5
I don't understand. Why wouldn't the op amp work as a normal difference amplifier? How can I rectify this problem?