What are the benefits of isolating the ADC rather than the digital signal? Or why not use an isolation amplifier? You kinda answered the first question with another post. As an EE working in high voltage battery systems that require fast protective switching -- understanding CMTI and the benefits of certain isolated topologies would be a great topic. Especially on how to properly design a PCBA taking such topics into consideration. Most of the industry is moving to higher and higher voltages for efficiency reasons.
All good questions. About isolation, there are of course multiple ways to do it, each with their benefits and noise risks/consequences. I like the ADC approach, especially this AMC component series by TI, for low-level precision measurements so I tend to advocate for them as I have seen lack of awareness on this topic. About isolating the data lines, I would say it provides less isolation between the digital and analog signals, which is one of the main reasons we would use an isolated ADCs. I like the isolation amplifier idea, I normally don't use it in the precision measurement application only because I ultimately need to get to an ADC anyways, so I just put it there. But if there was a lot of complex signal conditioning circuitry on the analog side, then I think an isolation amplifier makes more sense because you can protect some of those components on the analog side.
@@Zachariah-Peterson Great points. I wasn't even aware that isolated ADCs were a thing. From my experience, isolation amplifier are great but increased cost. I've had issues trying to isolate the data lines as transients seem to really cause headaches. Appreciate your time in answering the question.
Hey Zach , I watched the video you did about different grounds and how to tie them by using a safety capacitor in between. But what value should I use ? I am using a DIN rail isolated 12 V to 5V DC-DC converter. I am powering it from a 12V battery , but I also need to use that 12V net in other parts of the circuits, that’s why I think I need to tie the grounds. Thanks in advance.
This is a great question especially for this type of design. The point of the capacitor in this case would be to allow high frequency currents to pass across the ground regions. If you are having challenges with high frequency radiation, from the region between the two grounds, you could connect them with a cap as it will equalize the ground potential at high frequencies. In this case, the capacitor you would use would be based on the cutoff you want to set using the RC time constant, the R value is the resistance in the return current path (typically
Could another option could be using a normal ADC with an isolator for the communication interface SPI, I2C ?? It might not be great because of the high frequency clock signal in the analog plain 🤔
Yes you can enforce isolation that way. That's another option when you have a much higher analog signal level anyways, and maybe you're still exposing the analog side to an ESD-prone environment. At that point you're only using the isolation for the digital side, which could be fine if you're just routing that interface to a connector. I wouldn't use this if I needed to measure at very low levels only because I would have to amplify that analog signal so that it takes up most of the ADC's dynamic range, or I would have to use a very low ratiometric reference, which are both susceptible to noise. If amplifying the analog signal, you might then have to do oversampling w/ anti-aliasing, or just filtering the analog line directly, to remove some of that amplified noise.
