How low can I go? Voltage?

I am designing a device similar rogowski coil, which is detecting the di/dt of a current.

We are trying to detect an extremely fast di/dt, however this means that the inductance of the coil needs to be low and therefore the voltage produced at the output of the device is low.

So my question is:

how low of a voltage can I get away with in a real analog world system amplify to an opamp and still have decent enough accuracy? 0-100mV, 0-10mV, 0-1mV?

The sensitivity of a device is dominated by the noise. The minimum amplitude you can detect depends from the signal/noise ratio of the incoming signal and the noise added by the conditioning circuits.

digi001 said:

I am designing a device similar rogowski coil, which is detecting the di/dt of a current.

We are trying to detect an extremely fast di/dt, however this means that the inductance of the coil needs to be low and therefore the voltage produced at the output of the device is low.

So my question is:

how low of a voltage can I get away with in a real analog world system amplify to an opamp and still have decent enough accuracy? 0-100mV, 0-10mV, 0-1mV?

Click to expand...

With good "drift-less" opamps, the typical low voltages are at 100 uV level. Typical offsets and drifts in good "precision" type opamps are of the order of 10 uV. Using good temperature-stable resistors is also good. Possibly, stabilizing temperature of the opamp is also good. The Rogowski coil typically has a low resistance, so noise will be rather determined by the opamp. I would advise to use two similar coils, one "active", to symmetrize the input circuit. Dissimilar metals should be avoided or compensated by symmetry, too.

---------- Post added at 22:11 ---------- Previous post was at 22:09 ----------

As you did not specify how "fast" the signal could be, the ordinary opamps are good up to < 1 MHz. Fast opamps like MAX4286 can go to > 20 MHz with ~18 dB gain. Similar stability as above.

1 mV is a usable signal.

I have not made a rogowski coil, so I can only talk theory.

To obtain a meaningful range of readings, you need a fast enough time constant. Meaning low time constant. The L/R formula suggests a high resistance will help in this regard. However a high resistance limits current.

Experimentation should arrive at a happy medium. It may involve finding the right tradeoffs. Such as Henry value versus resistance.

Avoid having excess resistance in the coil. You want to maximize purity of coil response.

Experiment with resistance in surrounding branches. The object is to maintain a certain degree of isolation so that coil emf stands out (assuming you are amplifying its emf). Then it is more readily detected and amplified.

---------- Post added at 15:33 ---------- Previous post was at 15:21 ----------

In case this helps...

A high Henry value is associated with less current.

A low Henry values is associated with higher current.

Thanks. These suggestions all help.

I understand it is a pretty open question since their are various levels of noise throughout many different types of systems, but this would be for industrial type applications.

We are looking to detect a di/dt in around 2-5uS, so maybe a 1Mhz opamp would be just enough but I think going faster could be better.

Agree with the above low-drift op amps, and instrumentation amp comments.

Analog Devices is the first place I usually start looking for op-amp devices, as they make a decent variety. You can typically get samples, so you might even be able to play with a few different devices to find a good pairing for your final coil design (+1 for trial and tweaking, per Brad's comment).
Amplifiers & Linear | Analog Devices