feynmanism said:
we drive the Hall probe with ac current and measure the voltage with lock-in amplifier but the problem is that after some time the probe does not function(no current passes through it). Hall probe is driven with the alternative current from the sine output of the lock-in and i max is approximately 0.1 - 0.2 mA.
What can be the problem although the current is so low? Is it related with the connections from the lock-in to the probe or something else related with the alternative current?
I'm not totally sure what you're asking but have some limited experience of Hall probes and SQUIDs.
It sounds like you're reading the voltage using a phase locked loop, that's the only 'lock-in' amplifier I can think of.
A Hall sensor will have two currents, one that flows through the sense channel and a second that flows out across it (the one that represents the magnetic field and could be termed the output of the sensor).
As I understand it, Hall probes are only driven with AC, or chopped up current, to help with stabilisation. The chopping can inject quite a bit of noise into the output - although a PLL could get rid of that I guess.
There would seem to be a whole load of things that could be going wrong here since your probe setup is quite complex. It could be something mechanically wrong with your connections, the probe material it's self could be heating up (and so increasing it's resistance) or the relationship between the drive current and the PLL could be drifting over time.
First thing you want to do is verify all your connections as best as possible. You could take the circuit apart and try rebuilding it. If you have a few spare parts, try replacing the currents ones with the spare to check that none of the leads have been messed up or the silicon overheated or otherwise damaged in some way. Check the ratings of your components and make sure you're not exceeding them by missing a resistor or something out prior to powering them back up. At mA of current I can't imagine a heatsink is going to be needed, so the current is either too high or it's okay.
Next thing, I'd verify that the current through the probe is actually falling, and it's not that the drive / measuring circuit relationship is somehow drifting and just registering it as dropping. Maybe you could put a sense resistor on the input to the probe and see if you can measure the drive current with a multimeter - which may be tricky if you're driving the probe with a high frequency.
If the second current measurement agrees with the one from the PLL, I would then suspect that there is something wrong with my drive circuit, especially if it's involved in some moderately complex relationship with the sense circuit. I would suspect that either the drive, the sense loop or both are drifting in terms of their phase lock and so may no longer representative of what's actually happening across the sensor. If your drive and sense circuits are able to modulate each other in anyway, not just synchronised with each other, that would allow for the current to fall.
If the drive and sense circuits are just synchronised and totally separate otherwise, and the current still falls according to a multimeter reading, then it points to the drive circuit. If the current is present in the form that it should be from the drive, then it would suggest the sense circuit is failing in someway (perhaps it's not synchronised with the drive for example, so it's no longer phase locked).
If you're looking to increase the sensitivity of your Hall probe, you should consider replacing it with a quantum Hall sensor.
Quantum versions of Hall sensors have a conduction layer that's grown using molecular beam epitaxy to be a single atom thick, so it's essentially a 2D surface. That creates a massive improvement in the sensitivity of the sensor since you no longer have electrons being wasted as they drift around in the third dimension (the Hall effect is a 2D effect, so measuring it with a 3D structure produces a distorted result).
You can buy quantum Hall sensors from http://www.ahsltd.com/" in the UK, although they seem to be quite slow in replying (they're a university spin off, not a big global company).
I've tried some of their sensors and they're pretty impressive! Neither do they cost a fortune or require cryogenic cooling, as SQUIDs do.