We made two boards for AD7682/AD7689. One of them work perfectly, one of them does not. The voltage reference for it is unstable from time to time. The reference does not export 4.096V any more, but randomly move from 0V to 4.9V.
We use the internal reference of AD7682 as a voltage reference for a DAC. The reference at REF pin is followed by AD4841-2 as a voltage follower. However, the voltage reference is unstable. When we solder out the AD4841, the adc reference works again.
And when we solder the AD4841 back, it continues to work. However, on one day it suddenly does not work again....And the cycle repeats twice in these two weeks...
Is it because the capacitive input for AD4841 is too large?(10uF) Or what could be the reason? What should I check?
The schematic is like following:
We are looking into this. I'll get back to you soon.
I am assuming you are using the AD7682 internal reference by setting CFG [5:3] as 000 (for a 2.5V output on REF) or 001 (for a 4.096V output).
With the internal reference enabled, the bandgap voltage is also available on the REFIN pin. Table 3 of the datasheet shows that the REFIN output voltage have typical values of 1.2V or 2.3V, depending on what internal reference has been selected. You say that this internal reference is used for a DAC. May we know where the REF output is being used? If it’s your reference for the ADC, you don’t need to buffer it with the ADA4841 as you can see in the typical application diagrams on page 18 of the datasheet.
You mention you have two boards. What is different between the working board and the one that doesn’t?
1. Yes, we set the reference to 4.096 by configuring CFG[5:3] as 001.
2. The REF pin, not the REFIN pin has been use to provide voltage reference to a DAC.
3. The ADC does not need a buffer. The buffer is for providing voltage reference of a DAC, which is AD5760.
I dont really know the difference, since we made two exactly same boards... And the abnormal board works or fails from time to time....
The ADC starts to work abnormally again.
1. The ac noises have peak-peak value of 0.4V.
2. The output reference gradually drifts from 4.3V to 5V....
3. Finally the reference will behave in the way as the third diagram: random walk at a low frequency.
The time interval as is shown in the diagram below, is 500ms.
Thank you very much.
1-2. The input voltage range is found on Table 2 on Page 4 of the ADA4841 datasheet. Increasing your supply as you suggested will work or you may opt to use a RRIO amp for a wider dynamic range. Harry will be more knowledgable on this.
3. The ADA4941 also exhibits degraded linearity beyond its specified input range (within 1 V of VCC on the high end) and may be solved as above.