The manual hasn't helped much! Many important details are unexplained. Have a look at this snippet, for instance:




This part of the circuit makes absolutely no sense to me.
AD7548 is a 12bit DAC, transforming the 12bit data into a current, which is flowing out of pin20, if Vref is positive and pin1 is pulled to virtual ground by the help of OPamp U4A.
But this means, that the output of OPamp U4A becomes negative, and so the potential at the non-inverting input of U14 (OPA547)! At "0000 0000 0000" output is 0V and at "1111 1111 1111" output is -5V x 4095/4096, which is nearly -5V.

But, when the non-inverting input of OPA547 becomes negative, then the current through coil-/coil+ flows in the opposite direction, needing the output of OPA547 to become negative. But, why this +40V is needed at OPA547 then??
Also, a current flowing in the oppposite direction will make Q2 (2N7000) to become useless!

The high positive supply voltage of OPA547 makes me believe, that the potential at non-inverting input of OPA547 must become positive. But the circuit tells something different!

Should output of U4B be connected to OPA547 instead?
Makes no sense either, because its output is configured to produce positive and negative output voltages.

But even if OPA547 would get a positive control voltage, the circuit arround this chip makes no sense either! To drive an inductance being in the feedback loop of an OPamp is highly critical, because of the phase lag it adds to the overall phase response. Driving an inductance in this way, would need a sophisticated phase compensation, means the add of some phase lead, for instance by the paralleling of a suited capacitance to the inductance. To avoid unsane impedance maximum due to the resonance of these both parts an additional resistance must be paralleled to this combination.

But here we do not see any phase compensation. Only thing we see is a capacitance connected to ground, namely C6. Such a capacitance is a big "No, No", additonally adding phase lag. It looks like that the developer has already experienced instability and tried to cure the situation by inserting C6, but not actually knowing what he is doing.

The dangerous thing is, that C6 might cure the instability for the one reed relay inductance being connected during the test. But connecting another reed relay can make the whole circuit to oscillate again...

Also, Q2 makes no sense either, in this feedback loop. Why would someone want to destroy the feedback of a OPamp? The only result would be a clipping of output of OPA547...

So, I tend to believe, that this circuit is a faulty design. Not knowing more about it would make me to keep my fingers off.

Kai