Arduino Randomly Resets

[slash11]

Hi !
I was running my Arduino 328P based datalogger on the go-kart last weekend for the first time and I found it randomly resets itself when the engine is running and the higher the RPM the more it resets.
I suspected vibration to be the problem first, but after I have put as much paddng as I could, it would still happen.
Is it possible that the Electro magnetic pulse from the engine ignition (2-stroke so quite high frequency) causes the Arduino to reset itself? the board is in the plastic box at the moment and there are several sensor cables connected, that run along the engine RPM lead, which goes to another logger. ( I don't log RPM on Arduino yet).
Would I need some RC filters for each Analogue input?

Many thanks.

[adafruit_support_bill]

How is it powered? You could be getting ignition noise via that route as well as your sensor cables.

[slash11]

Oh sorry, forgot to mention, its powered from 7.2V 1600mA/h battery connected directly to Arduino board ( soldered to Vin and GRD pins) the lead is about 20cm long.
Hope it helps.

[adafruit_support_bill]

If you have access to a scope, you could poke around and see what kind of noise you might be picking up. Other than that it's trial and error.

Every piece of wire is an antenna. You can try using shielded wire. And try to route your wires as far away and non-parallel to any ignition system wiring as possible. Car-audio shops will have an assortment of noise suppression filters you could try.

[Alan Chatham]

Try wrapping it in tinfoil? That's a cheap way to potentially reduce external EMI issues. Also, there may be issues caused by the engine inducing currents in the wires you've got going to your sensors. Test this by taking off the long wires to the sensors, then setting up a LED that's going to be visible on your Arduino inside the box (outside if you have to, but keep the wires short as possible) and have it run a long-cycle blink. You should be able to have a visual guide to if it's resetting or not without the wires. If it is a problem with the wires, there are ways to try and reduce problems with the currents, but I'm not really an expert on these things (Econ major here...). One way you can try is to add a small resistor in series to the sensors, limiting the current coming into the Arduino caused by EMI spikes. Another way is to attach small capacitors (like 100pF?) between the sensor input lines and ground. It'll be trial and error to figure out what values are big enough to absorb the EMI spikes, but small enough to not mess with the data you're collecting.

Hope it helps!

[StandardRockets]

disconnect the sensor to see if there is a difference, then do the paranoid tin-foil jobby.

[slash11]

Thanks to all who replied.
Good a few things to try. First I will disconnect all the sensors and only leave power and GPS and wrap it in the foil and ground it. I have also searched for a simple low pass filter diagram and its shown below. I will place on every analogue input to protect it from high frequency spikes. This circuit, I hope shouldn't affect genuine sensor input, as it is low freq. anyway (throttle, brake steering).
Will keep you posted.

Low-pass filter

Low-pass filter.JPG (4.92 KiB) Viewed 2458 times