# Not understanding Level shifting +/-3.5V analog signal for ADC with analog input of 2Vpp and common mode voltage

I am trying to design a very basic oscilloscope and my first PCB schematic and I am having trouble wrapping my head around how to level shift the front end analog signal so it can be accepted by my ADC. I am using a AD9238 ADC with a 2Vpp input and I want the scope to accept a signal of +/- 3.5 V.

I have a level shifter that feeds into a differential op amp to give me a differential input in my ADC. What I am having trouble understanding is how much gain to give at the differential op amp, and what to level shift to.

I have set the Vocm pin on my differential op amp to Vdd/2, and my voltage source is 3.3V so that gives me a common mode of 1.65V I believe. So since my ADC accepts 2Vpp and common mode is 1.65V does that mean I am looking to level shift from +/-3.5V to .65 to 2.65V? The ADC also has an internal Vref that can be set to 1V, and if I instead use that for Vocm pin on the differential op amp would I then be looking to level shift to 0 to 2V?

The op amp for level shifting I am using is the lf351, the differential op amp is the THS4121CD and the 40msps version of the AD9238 ADC.

Do you need DC coupling? If not then just put a capacitor in series with the input. If you need DC coupling then you’ll need something more complex. Meanwhile, you may find that an LF351 gives disappointing bandwidth, depending on what you want to use the scope for. I probably haven’t quite answered your question but a little more information on what you’re trying to achieve would be helpful.

EDIT: If you need to accept inputs from -3.5 to +3.5 then you have a couple of options - either use a passive (resistive) divider to shift the level before buffering it with the LF351, or run the LF351 on a split supply (+/-5v). For simplicity you might prefer the passive approach. Perhaps try this: use your LF351 as a unity-gain buffer, with the output tied directly back to the inverting input, then connect three resistors to the non-inverting input: one to the 'scope input, one to the +3.3 supply and one to ground. Work out the values so that you get 1.65v when the scope input is at 0V, 2.65v when the input is at +3.5 and 0.65v when the input is at -3.5v. The LF351 output can feed directly into the ADC input.

• Yes I believe we will be needing DC coupling in our scope design. This is for a school project and our professor recommended that we use the LF351 to create a level shifter to get the input signal to a voltage that is acceptable by our ADC. My confusion is coming from trying to figure out the ADC's acceptable analog input voltage since it seems to have an input span of 2Vpp. I am thinking the acceptable input should be 0.65V to 2.65V since our VDD is 3.3V, and common mode voltage is set to vdd/2. Nov 11, 2020 at 8:19

So to solve this problem we ended up using a passive attenuator at the front to give the +/-3.5V signal a gain of 1/3.5, and then this +/- 1v signal is sent to a level shifter to give +1.65V to move us from +/-1V to .65 to 2.65V range. This is because the ADC has a 2vpp input range, and with our common mode at 1.65V, .65v to 2.65v is the 2vpp range we need to be in.

Are you using the evaluation board rather than the specified ADC. It seems you are as the evaluation board is 2Vpp input. I can't recall ever seeing an analogue input for an ADC being specified in Vpp terms.

If you are using the evaluation board you'll want to map +/-3.5V to +/-2V you'll need a gain of 2/3.5, no DC level shift required.

• I'm pretty sure I'm using just the ADC datasheet info and not the evaluation board. I copied and pasted a small excerpt from the datasheet which mentions the ADC performs best with an input span of 2Vpp. "The internal voltage reference can be pin-strapped to fixed values of 0.5 V or 1.0 V or adjusted within the same range as discussed in the Internal Reference Connection section. Maximum SNR performance is achieved with the AD9238 set to the largest input span of 2 V p-p. The relative SNR degradation is 3 dB when changing from 2 V p-p mode to 1 V p-p mode. " Nov 11, 2020 at 8:08
• This answer is mistaken; the AD9238 absolute maximum ratings leave no doubt about the fact that it does not accept negative inputs, therefore an offset will have to be applied. Nov 13, 2020 at 2:02
• @ChrisStratton my answer relates to the use of the evaluation board which has +/-2Vpp input. Nov 13, 2020 at 2:06
• Only because it has transform coupling, which is of course a non-starter for a question explicitly about an "oscilloscope". Nov 13, 2020 at 2:08