To deliver a current of several pA on chip?

[Octago]

when a cable delivers a current of the order of several pA(or even fA) in the integrated circuit, can this small current be delivered precisely?

due to this extreme small amount of current, is this signal very sensitive to the interference when the cable is relatively long in the chip? or are there any other problems?

I am using ultrasim to simulate a large circuit, and I found that this small current can not be simulated accurately; In the simulation results, there is a large fluctuation of this small current signal.

thanks

 

[FvM]

Simulation problems may come from default simulator absolute tolerances that are changable if necessary. In real world, small currents of course can be difficult to measure. Cable capacitance may be a serious problem, also mechanic interferencs causing currents from intrinsic isolator charges or piezoelectric effects. Electrometer amplifiers sometimes use triaxial cables with an active driven guard potential to reduce cable capacitance influence, but the best measure is to move the amplifier to the signal rather than using long cables.

 

[Octago]

thanks,
actually, now I don't need to measure this current in reality, I am thinking what will happen when this current is flowing in a metal line in the integrated circuit, could it also be very critical?

Simulation problems may come from default simulator absolute tolerances that are changable if necessary. In real world, small currents of course can be difficult to measure. Cable capacitance may be a serious problem, also mechanic interferencs causing currents from intrinsic isolator charges or piezoelectric effects. Electrometer amplifiers sometimes use triaxial cables with an active driven guard potential to reduce cable capacitance influence, but the best measure is to move the amplifier to the signal rather than using long cables.

 

[FvM]

Only few amplifiers have leakage currents low enough to measure pA or fA accurately. So I think, this isn't easy to achieve in chip design. As a circuit designer, I'm glad to have some chips with suitable specification for electrometer applications.

 

[leohart]

I had use BB IVC102 to measure a photodiode's dark current. Care should be taken on the pcb layout.Although the datasheet calm that the bias current of the internal opamp's is on the order of pA(less),but through double sampling, dark current of several pA can be readily measured.

For a monolith onchip design, I think you should make sure the insulator are reliable enough(better use a negative feedback to keep the pA(or fA) carrying wire at nearly zero potential),the amplier's bias current,and also the piezo effect as mentioned by FvM...