OPA4209: Measurement Variation with the OPA4209.

Hi Felipe,

Felipe de Andrade said:

I want to better understand why the high input impedance is causing such significant variations and why there is a greater variation when the device is powered on. According to the OPA4209 datasheet, temperature should not affect the measurement to this extent.

Could you show us the schematic of your test circuit? 

My guess is that you are monitoring the offset voltage, Vos of the op amp. With such small input signal with respect to the Vos performance, the errors or variations could be large. If you are looking for long term stability or low drift, you may consider chopper amplifier for such application. 

OPA4209 has 18MHz BW. For instance, 10MΩ may contribute to the total noise of the circuit. Your power supply's ripple voltage can affect the Vos errors. The CMRR can affect the Vos as well. Thus why I need your testing schematic and drift requirement (<0.5% of what range). Vos/1mV can contribute large errors, if you are talking about small input signals. If this is the requirements, you may need higher precision and lower drift op amp.  The total offset of op amp and over time is determined by the following equation. 

Best,

Raymond

Hello, Raymond!

The circuit I am using is shown below. At the relay inputs, I am applying 1mV, and the relay outputs go directly to the ADS131M08. I have removed the diodes for +VCC and -VCC from the circuit, and I am changing the resistors for testing purposes.

The results I obtained from the latest data collections with 1mV at the input were:

• With attenuation: Using an input impedance of 1MΩ per channel and a feedback resistor of 300kΩ, I observed a variation of 8.6%, with the measured voltage ranging from 1.987mV to 2.676mV. In this case, the equipment had just been powered on after being off for a long time.

• With unity gain: Using 10kΩ impedances, I observed a 1.225% error, with the voltage ranging from 1.214mV to 1.274mV. Similarly, the equipment had just been powered on after being off for a long time.

Measurements taken after the equipment has been powered on for some time show greater stability, staying within the 0.5% criterion I need. However, this criterion is somewhat challenging to meet over long periods because, in my calculations, I also factor in the readings from two channels. Over short periods, such as 5–10 seconds, the circuit performs very well, but when this time is exceeded, it starts to deviate significantly from what I need.

Regarding chopper amplifiers, could you suggest a model, preferably one with 4 channels?

BR,

Felipe.

Hi Felipe,

OPA4209 is bipolar precision op amp, and the Ib should be in ±1nA (typ.) to 8nA range. With the low input difference voltage of 1mV, the Ib feeback current is too low for the topology. It would be better to use op amp with CMOS construction. Of course, OPA4209 is low noise op amp, you should not use 1MΩ (resistors' thermal noise) or even 10kΩ feedback resistor for such low differential input voltage. Use 1kΩ to a few hundred ohm feedback resistors for 1mV differential signal would be better. My guess is that your input and feedback resistor values are too high for the small input signal. 

Felipe de Andrade said:

Regarding chopper amplifiers, could you suggest a model, preferably one with 4 channels?

https://www.ti.com/amplifier-circuit/op-amps/precision/products.html#480=4&3247max=1%3B10&sort=7typ;asc&

https://www.ti.com/amplifier-circuit/op-amps/precision/products.html#480=4&3247max=0.02%3B1&sort=3247max;asc&

Here is the latest article about the chopper amplifier. 

https://www.ti.com/lit/wp/sboa586/sboa586.pdf

If you can provide me the BW, voltage rails requirements, I can narrow it down further. As it is right now, I am proposing

OPA4140, OPA4197, LMP7704, TLV4333 (chopper) and may be others once I know what output ranges you are looking for your ADC. 

Best,

Raymond

Good morning, Raymond!

I know that one of the analysis parameters for sizing the input impedance is the input bias current of the operational amplifier. Regarding the sizing of the input impedances, are there any other OPAMP criteria I should consider? I also noticed that you mentioned the feedback resistance used in the circuit. What parameter should I analyze to determine the resistance values in this case?

Regarding the amplifier, I was thinking about using the OPA4187. What do you think? I use a DC voltage at the input of my circuit, so I believe BW is not an issue. The lowest voltage level I measure is 660 µV, and I apply a gain of x200.

BR,

Felipe.

Hi Felipe, 

It may be able to use OPA205 or OPA2205 or OPA4205 or even the initial OPA4209 part. The issues are your input voltage is too small and you do not have gains in the test circuit. If you have gain of 200 initially per your test configuration, the Vos and drift are likely stable for your test duration. Your test circuit's feedback current is below the Ib required current for a proper operation in OPA4209.  It will be good to repeat the test scenario with the actual gain settings. 

https://www.ti.com/lit/ds/symlink/opa2205.pdf?ts=1733515026830&ref_url=https%253A%252F%252Fsearch.yahoo.com%252F

OPA4187 is chopper amplifier, and it is not recommended to use large feedback resistors, unless you can limit the gains. 

https://www.ti.com/lit/wp/sboa586/sboa586.pdf

Felipe de Andrade said:

The lowest voltage level I measure is 660 µV, and I apply a gain of x200.

Is this current sensing application? Please let us know. 

Best,

Raymond

Hello, Raymond!

I am using the OPA4209 to measure voltage across a load and a shunt resistor. This measurement is performed using a 4-wire configuration, and I provide the applied resistance value at the measurement terminals as the output. In summary, my application is an ohmmeter.

In the current circuit, I am encountering an issue when measuring the 20kΩ shunt resistor. The response provided by the OPA4209 is not linear when my input impedance is lower. In the case I demonstrate below, my load resistor can vary between 100Ω and 1kΩ. It is noticeable that the response obtained in the TINA software is also not linear when I use a 4.7kΩ input impedance.

However, when I increase the input impedance, the response becomes more linear. In this case, I apply a gain of 21.27 and am forced to increase the input impedance to achieve a more linear response. Of course, I do not provide only 1mV at the input but rather a range from 12mV to 120mV.

In the cases I mentioned in previous posts, where I am applying 1mV, I am already using a higher impedance to try to mitigate the OPA4209's non-linear response. In a real application, I will not apply such a low value. The real case I tested on the bench is as follows:

• OPAMP1 – Input voltage: 2.9994V – Attenuation of x0.3
• OPAMP2 – Input voltage: 599.88µV – Gain of x209

In the case demonstrated above, most of the variation occurs in the channel that applies an attenuation of 0.3, causing oscillation in my resulting value. In this scenario, I am applying a voltage lower than 1mV but with a gain of x209.

BR,

Felipe.