cooling small component in too warm environment

[morphingstar]

I have some small circuit boards 40x40mm monitoring hall effect semiconductors. Whenever environment temperature goes >28°C the logic of the hall signals goes crazy. The output result comes from 3 synchronized elements. When the temperature rises - it is no longer the right ones, or an element signal might even be added, usually not lost. The hall magnets are moved mechanically in proper order, this can be seen. When the temperature goes below 28°C the craziness disappears, system works correct.
The circuit temperature goes close to 40°, where it is measurable, might be warmer in some poorly ventilate spot. I am considering a fan,but this will just blow the 28+ (e.g.31) air towards the components, though it may still have a good influence.

The circuitry is applied in a mechanical accordion, picking up the lever movements. The chord levers are coupled (sets of 3) to make a chord. When the chord sounds wrong the circuit has gone nuts or worse. This digital circuit is behaving worse than circuits in "old analogue days" where drift was a known problem and had known solutions.
Suggestions?

 

[(*steve*)]

1) decoupling capacitors
2) show us the schematic

 

[Chris Dancer]

Digital circuits should be absolutely stable and reliable at those temperatures. 40 degrees is pretty cool for a semiconductor device. I would suspect an intermittent dry joint somewhere. When a component warms up it expands and breaks the circuit. Examine all the solder joints and resolder any that look suspect. Also look for hairline cracks on the PCB. Try poking around the board with a wooden or plastic "probe" to see if you can make the fault happen.

I had this type of problem on a mains inlet connector a couple of years ago, it took me weeks to find the cause.

 

[morphingstar]

Thanks. The manufacturer will not release the circuit and the reason why this happens. He knows the problem is happening in tropical climate only. 99% of the products are used in moderate climates.
I don't know if there are any capacitors in the circuit environment. I was wondering about cooling devices I have not heard about. Thanks.

 

[Chris Dancer]

In cool climates parts are very often installed in enclosures where the temperature can rise well above 28 degrees, so I still don't think the problem is the temperature affecting the components themselves. Humidy might be the problem, have you considered coating the boards with lacquer?

 

[morphingstar]

Thanks for opinion / sugg's
Climate with >28°C is for most people not a cool area.
Humidity is rather stable and high.
The effect is temperature-dependent.
Lacquer will reduce cooling of the component on the sm-pcb, how much I don't know.
Still looking to find (exotic?) cooling devices (I heard about fans).