We current have a AD620 part that we want to test for Quiescent Current, Gain Error, Voltage Offset, Input Current, and Dynamic Response. I wanted to know what test setup (schematic) Analog Devices uses for these specs.
Maybe it is not a Rarely Asked Question, but the recent RAQ article from James Bryant may point you in the right direction.
Remarkably easily, and absolutely not!
How we test each part is proprietary information and we don't give out those schematics. These schematics are designed for accuracy but also repeatability, test speed, and robustness. They have a lot of stuff you probably don't need. Instead I'll give you a description of how to take these measurements in the lab. This not the exact way we test, but should give you excellent results. This is similar to how we characterize the part in our lab before we release. But first some setup stuff:
All specs are checked according the parameters at the top of Table 1 in the AD620 unless otherwise noted: at room temperature, with supplies of +/-15V and with a load of 2 kohm.
In order to differentiate between the device we are testing and test circuitry (which may also include a AD620), I will refer to the device we are testing as the DUT (device under test.)
During this testing you will need to be very careful about your ground connections. If you have a wire with current going through it, it will generate a voltage, so if you calling the end of this wire "ground", it may be quite a few uV or even mV away from the other wire you call "ground".
For this testing you will also need a second in amp for amplification of the small error signals. Another AD620 will work great. I will call this amplifier the "signal conditioning amp". Place the signal conditing amp close to the DUT on the same board to minimize errors. Configure this signal conditioning amp in a gain of 1000. Use +/-15V supplies. Ground its reference. You need to know this in amp's gain and offset so you can subtract it out.
Measure gain as follows: Ground negative input. Apply +10 mV to positive input. Measure output. Apply -10 mV to positive input. Measure output. Gain = (First measurement - Second measurement) / 20 mV.
Measure its offset as follows: Ground inputs. Measure output with DMM. Divide output by gain to get input referred offset (which is what I will use in the equations below)
A DMM can sometimes have enough capacitance to cause an amplifier to go into oscillation. If you get unexpected results, check your measurement with an oscilloscope with the DMM still connected. If you see oscillation, you can normally fix the issue by placing a 100 ohm resistor between the DMM and the amplifier output.
Quiescent current: Connect a DMM in current measuring mode between the +Vs of the DUT and your +15V power supply. Ground inputs and reference pin of the DUT.
Gain Error, G=1: Connect voltage source to positive input of DUT. Ground negative input and reference pin. Connect the signal conditioning amp positive input to DUT output. Connect signal condition amp negative input to voltage source. Measure the output of the signal conditioning amp with your DMM. Set your voltage source to 10V. Measure. Set your voltage source to -10V. Measure.
Gain Error = (Measurement1 - Measurement2) / 20V / Signal Conditioning Amp Gain
Gain Error, G>1: Similar procedure as G=1, but use a high precision, low temperature drift voltage divider to divide down the voltage before going into the DUT. You will need to know the attenuation of the voltage divider very accurately, since any error in your knowledge of the voltage divider will impact your measurement.
Voltage Offset: The AD620 has two offset specs: VOSO and VOSI.
Ground DUT inputs and reference. Connect signal conditioning amp inputs to DUT output and DUT reference. Connect DMM to output of signal conditioning amp.
Place DUT in gain of 1. Measure.
Place DUT in gain of 1000. (For more accurate gain, measure as described above when we measure the signal conditioning's amp's gain.) Measure
Solve for VOSI and VOSO using the following two equations:
Measurement1/Signal Conditioning Amp Gain - Signal Conditioning Amp Gain Offset = VOSO + VOSI
Measurement2/Signal Conditioning Amp Gain - Signal Conditioning Amp Gain Offset = VOSO + (VOSI * DUT GAIN)
Input Current: Place DUT in gain of 1000. To get the exact gain, measure as described above when we measure the signal conditioning's amp's gain. To measure bias current, place a 10 Mohm resistor between one of the DUT's inputs and ground. Ground the other DUT input and reference. Connect signal conditioning amp to output of DUT.
Bias current = ((Measurement Voltage / Signal Conditioning Amp Gain - Signal Conditioning Amp Gain Offset) / DUT Gain - DUT offset at gain of 1000) / 10 Mohm
(Note: if the bias current of the DUT is too high, you may rail the signal conditioning amp, in which case you can lower the gain of the signal conditioning amp or remove it altogether).
For input offset current, place a 10 Mohm resistor between each input and ground.
Dynamic Response. You can use either a network analyzer or a waveform generator + oscilloscope for this measurement. The AD620 is not a high speed device, so you can get a reasonable number for the bandwidth without taking the precautions you would normally do for a higher speed device. Ground the reference and negative input. Send the signal into the positive input. Measure at the output. If you measurement device has 50 ohms input impedance, you will need to buffer the output of the AD620. We like to use a FET probe for this. Make sure the signal you send into the AD620 is small enough that the AD620 does not slew rate limit. Don't use the signal conditioning amp for this measurement.
If you go to this page, and select "amplifiers and comparators" for the product page and refine your search by checking "tutorials", you'll see quite a few tutorials that may be useful. MT-037 through MT-046 may be particularly helpful.
Perfect! Exactly what I needed. Thank you for taking the time on that response.
I may leave this open in case I have a few minor questions (doubtful with how thorough your response was).
Retrieving data ...