Electronics > Power/Renewable Energy/EV's

MPPT noise reduction and Blog

(1/10) > >>

I have noticed my MPPT code not always tracking very well, it runs on a PIC16LF1782. I think I have concluded it's getting confused by noise in the ADC readings for input current and voltage originating in the associated boost converter. I have tried many things like analogue filtering but without many poles (steep rolloff) it's difficult to get much improvement. I have an idea of using oversampling together with averaging as that's quite easy to do for binary modulo's like 8, 16 etc the theory being the signal will sum in the time domain but not the noise (being random) hence the result should have an improved SNR. Has anybody heard of this technique in this or similar situations or have any other idea's ?

Measure the ADC input voltage with a DSO and then you can be sure if noise is a problem

Schematics, Code, pcb, photos  and measurements can all help people to give better advice.

You are quite right, the scope shot shows what happens in low light conditions, the MPPT is switching the duty cycle between nothing and something small but the PWM EMC swamps the current signal, the scope probe connection is not good so some EMC is entering that way.

As the picture shows the construction is less than ideal being stripboard though efforts have been made to limit lengths and bulk up high current connections with bare copper wire soldered along pairs of strips.

In the schematics U1 is the current amp sensing across an 0R01 resistor in the PV return leg. Missing from the schematic is a 1nF/56R snubber across the mosfet and also a 15R resistor in series with the gate to soften the mosfet a little.

Yes it can be hard to probe without noise from the high impedance probes.

Short the probe tip to its ground lead and then touch the circuit to see how much is induced by the probes.

The most obvious things to check/ change IMO.

are you using Kelvin connections to the op amp?  You should and even twist them.
ground bounce, star ground, analogue ground. IOW could the ground be improved.
can you use a current sense amp? At least u have rail to rail In and out which you will need.
larger current sense resistor. @ 1 kw  I think you can afford a little more drop across the current sense resistor.
25 mV input if that is a full power you probably want more like Vref/2 whatever that is.

Hope that helps.

capt bullshot:
First, I'd recommend not to connect U1's output directly to the ADC input. Place a series resistor (range 10R to 470R) and a capacitor as close as possible from the ADC input to ADC GND (10nF should do the job). You should use the capacitor to GND on every ADC input, it helps filtering and it also helps to charge the ADC input sampling capacitor.

And yes, oversampling and averaging can do a good job in reducing noise, I've done such things (oversampling and averaging) successfully before.


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod