ADS1278 in-line resistors and accuracy

Hi Michael,

Thanks for your question. The answers to your questions have a lot of behind them, but I will try to address them at a high level before we dive into more detail.

The characterization of the ADS1278 was done at the maximum modulator rate, 6.75 MHz. We used the THS4521 to drive the input at this rate. Smaller differential capacitor values caused ringing, which could be improved with a larger series resistor value, but this introduces error as you already observed. Larger capacitor values caused the output of the driver to slew, but reducing the series resistor to provide more current to the cap load caused the amplifier to oscillate, as you suspected.

After a lot of testing, we found that a 2.2 nF capacitor was the ideal value when running the modulator at the maximum 6.75 MHz rate, given the size of the sampling capacitor inside. 

Since you are scaling down the modulator to run at 2.56 MHz, I can see why you would consider an amplifier with less bandwidth like the THS4531A. This is will work, but the RC filter pole needs to be reduced to a lower frequency. The RC output filter serves a dual purpose: to charge the internal capacitor sampling circuit, and to prevent higher-frequency signals from aliasing into the ADC passband. 50 ohms with 2.2nF in parallel will give you about a 723kHz cutoff, which will help to knock down signals at fmod.

I hope this answers some of your concerns, but please let me know if you need more clarification.

Best Regards,

According to the datasheet, running the modulator at 6.75Mhz would result in a equivalent impedance of 14kOhms, and the two 50 resistors would result in about .7% of error, which is too much for my application.  I have a separate filter to take care of aliasing; I only care that I get an accurate result from the conversion.  Is this error going to be linear with the input voltage such that it can be calibrated out with a linear calibration?  Also, with a 723kHz filter on the output, is there really any point of using such a high speed amplifier like the THS4521?  

Hi Michael,

That gain error calculation looks correct. To answer you first question, this error will be a constant gain error for a given modulator frequency and can be calibrated. Meaning, as you decrease the modulator frequency, the input impedance will increase, and the gain error from the output resistors will decrease.

I still do not see where 50 ohms is suggested as the output series resistance. You are not referring to 50 ohms inside the feedback loop, are you? In between the capacitive feedback path and the resistive feedback path? You should be able to drive the 2.2 nF with a much lower Ro impedance and still expect a flat frequency response. Please refer to Figure 18 of the THS4531A datasheet. These curves are for a 2k load and G = 1V/V, but I can try to run simulations for you that correlate to a 14k load like in the ADS1278.

Now to the big question: what bandwidth do you really need? This holy grail of data converter questions is one that our team does not yet have a complete answer to. It really depends on the capacitive sampling structure inside the ADC. The differential input cap does most of the work to recharge the sampling caps, but how quickly you need to recharge them depends on the modulator rate, the size of the caps, the common-mode voltage that they discharge to, etc.

The modulator is sampling at 6.75 MHz, but we have not yet quantified how much more bandwidth is required. Still, the THS4531A may not be as overkill as you think after all.

What are your thoughts?

In my current setup, my modulator is running at 2.56 MHz.  I don't understand how much better the THS4521 is going to perform over the THS4531, as in both cases you're charging a capacitor through two 50-ohm resistors that form an RC filter with cutoff of 723kHz, which is still an order of magnitude lower than the bandwidth of the THS4531.  Doesn't that effectively limit the bandwidth of the output such that high bandwidth of the THS4521 isn't really helping?

Thanks,

Mike