High speed current sensing

Your other post implies that you want to build a fuse circuit.

Low-side or high-side sensing? What accuracy is needed?

Does "100 MHz" imply that you want to trigger on a current pulse with a length of 10 ns? There is no integrated current monitor that is that fast.

Do you want to limit the current or switch off? In the first case, you need a high-speed amplifier (see page 37 of the Simplifying Current Sensing white paper); in the latter case, you probably need a high-speed comparator.

Hi Nic,

many more details are required. Best you could show a schematic.

Kai

What I am looking for looks like this:

The 100MHz/50% duty cycle signal drives a switch to convert the current to RL. The current flowing RL is a 100MHz/50% signal as well. I want to monitor and record the RL average current in real time. (Monitoring peak current seems impossible). Required accuracy is 0.1A. 

Current monitors are limited by the communication protocol (SPI or I²C), so their bandwidth is at most a few kHz.

Current sensors have an analog output (and sometimes a comparator output if you want to quickly handle overcurrents with an interrupt). You'd need an ADC to read them. It would be possible to add a peak detector circuit with another opamp.

Nic,

I'm in agreement with Clemens here. If you need bandwidth above a few hundred kilohertz, you will most likely need a high speed op amp to perform this task (given that you only want to measure average current, I don't know how much BW you need). The fastest current sense amplifier we have in our portfolio is INA293, which tops out at 1.3MHz for the A1 version. 

With your schematic as set up, you will also need to ensure that you select a device that is capable of accepting 6.5V for V+ for the device to be able to withstand the common mode voltage on the high side. There is more information regarding this here. 

I'm looping in our high speed team to get their thoughts as well. 

Yes, I only want to monitor the average current. The average current should be considered as DC, so the bandwidth should be closed to DC, I think. The current sensing amp could be regarded as the low-pass filter. 

For 6A peak current, 50% duty cycle signal, the average current is 3A.

From this standpoint, I think most of the TI parts could fit that, right?

Hi Nic,

what voltage drop would be seen across Rsensing?

Kai