When our cuetomer did impulse Noise test,they found attached waveform.
Vout goes to GND and then Vout returns after 5us.
Is Cause of this behavior Input Protection of INA826 ?
When they use OPA197 in same condition,they did't find this behavior.
Kura,
Your customer impulse "noise" seems to have a large magnitude that drives the output voltage to the negatve rail. Once the output stage gets trioded near the ground, it takes time for it to recover from the non-linear condition - this is known as an overload recovery time.
Having said that, there is several unanswered questons about your customer's application - for one, the voltages shown on the attached waveforms are in hundreds of volts, which is difficult to understand. Also, what is the input common-mode voltage of the application? Any Vcm higher than 2.8V or lower than 0.5V would significantly limit the linear operation of the output stage - see below.
All in all, please provide a customer schematic so we may better be able to assist you.
Marek-san
Thank you for your reply.
About circuit:
Sorry,they can't send their circuit.
But additinal information is here.
They use INA826 with DRV401.
INA826 IN+ ,IN- are connected to shunt resistor.
REF is connected to DRV401 2.5V ref.
Additional waveform is also attached.
They would like to know following.
1.Why they did't find this behavior when they use OPA197 in same condition?
2.If they use INA828,is this behavir seen same as INA826 ?
Kura-san,
INA826 is an instrumentation amplifier with 10Gohm input impedance while OPA197 used as a difference amplifier has much lower input resistance, thus they are not equivalent circuits. Do they use resistors shown on your diagram with INA826? If so, they should not do that. The bandwidth of OPA197 is much higher than INA826, which may affect the difference between AC responses in both circuits. Looking at the waveforms you provided, it is clear that unlike INA826 output collapsing on its negative rail, in case of OPA197 the output stays within its linear range. Thus, I suspect that the problem your customer sees may also be related to the fact that the inputs of INA826 are overdriven beyond its linear range, which would activate input overvoltage protection circuitry (see below) causing long recovery time.
Kura-san,
In a single 12V supply, in Gain of 4 with Vref of 2.5V, INA828 has the maximum output range of 0.1V<Vout<11.9V for Vcm=6V - see below.
For this conditions, the valid input voltage range is shown in the diagram below to be from -0.6V to 2.35V with Vcm=6V.
Kura-san,
I am not sure how to understand your statement of inverting input not being found - please clarify.
Since I do not have a circuit schematic, it is not clear to me what the input current (blue trace) translates in terms of voltage and what is its common-mode voltage. Keep in mind, there is a difference in the input common-mode voltage between two inputs being single-ended vs differential. For example for -2V and +2V inputs, Vcm=0 but for 0V, 4V inputs, Vcm=2V and NOT 0V - Vcm is an average of two inputs.
Using -5V negative supply instead of ground will result in the linear input voltage range shown below. Notice, how input common-mode voltage, Vcm, is defined: for max Vout=11.9V, Vcm=(7.175VV+4.825V)/2=6V and for min Vout=-4.9V, Vcm=(5.075V+6.925V)/2=6V
