[SOLVED] Solar supplied PIC unwanted overclocking

[firefarzad]

Hello!

I have designed a circuit for a specific job, I used a LM2576 for regulating input voltage to 4.3V and a PIC18F4550 for doing the work!
if I supply the circuit with a 12V DC source, it works Fine!
but if I use a 17V PV panel, it works 10times faster!!!! and all the time synchronizing will fail.
I don't know why this happen? I used a 12MHz external crystal, and 2 capacitor 22p pulling down the crystal pins as always,
I have also tried internal crystal, but no difference!
Please help me with this unwanted overclocking!

 

[betwixt]

That simply doesn't make sense. The execution speed of the PIC is entirely decided by the crystal and if it did appear to run at 120MHz, Microchip would be throwing money at you to find out how you did it!

I would suggest you post a schematic and code for us to see. Most likely, it isn't running 10 times faster, it's not running 10 times slower! Some other factor that was causing a delay is not present when the PV voltage is higher.

Brian.

 

[firefarzad]

i found the problem!
but i dont know how to solve it!
the voltage output of lm2576 is very noisy, and the PIC will work so fast, at least 5times faster! I can not measure it.
I have use a 3300uF, but no use

 

[Okada]

Use the circuit provided in LM2576 datasheet. I have used circuit provided in datasheet with out problems. It needs just 2 capacitors and an inductor. In my design I needed 2A and hence used a 100 uH 3A inductor.

 

[betwixt]

In particular, make sure the capacitors are wired to ground close to the ground pin of the LM2576. Switching regulators tend to be very noisy so it's important that ground currents are not allowed to build up voltages along the resistance and inductance of the wiring.

Also use low ESR capacitors, if you use one with high ESR, a bigger value will not improve the noise you see.

Brian.

 

[Alloy]

I'd suggest you using LDO regulator after that LM you have.
Also make sure that output and input voltages are inside the range specified in the datasheets.
Also remember to put 100nF capacitors between GND and VCC pins of MCU (for every pin!)

 

[Easyrider83]

I could suggest to use low pass filter based on inductor 1/10 less from main one, but no suggestions before I can see actual schematics and pcb layout. It is obviosly useless.

 

[firefarzad]

I found the problem!
I tried so much to filter the noisy voltage but no use!
I observed that when the lm2576 is supplying more than 1A suddenly voltage drop to 3.5V!!! and I fixed it by adding some inductance in default schematic!