The analogue one will be easier to fix than the digital one. Chances are it's the calibration pot needs a bit of contact cleaner and exercise as they're pretty sturdy animals or maybe the needle has stuck - mine does from time to time. Never had a digital one but i know in the analogue one there is an LM7805 voltage regulator IC to drop the voltage to 5V for the meter and deteecting circuits. Other than that and obviously the meter itself, there are no polarity conscious components in the anaogue one.
The analogue one operates on the Wheatstone Bridge principle, +ve and -ve supply to the bridge, calibration pot and reference thermistor in one leg with one side of the meter between them, sensing thermistor and a resistor in the other leg with the other side of the meter between them.
Both thermistors heat up due to current flow in them, the meter deflects to 2% in free air under calibration as (to the best of my knowledge) there is 2% CO in free air.
The reference sensor is hidden away from gas flow so it just gets up to temperature and maintains it, like wise the sensing thermistor will achieve a certain temperature and maintain it - or it will until CO passes over it.
As the CO passes over the sensing thermistor, the CO cools the thermistor, increasing the resistance (should have said before, they are both NTC thermistors) and hence the voltage developed across it. This causes the meter to deflect higher, conversely if there is less CO than 2% in the gas sample, it allows the temperature to rise on the sensing thermistor reducing the resistance in it and hence the voltage developed across it causing the meter to deflect towards zero/lower.
In the analogue onees, there is once central screw underneath if memory serves, remove it and you can separate the top and the bottom. If you have a basic understanding of electronics and can drive a soldering iron, you should be able to fix it now.
