ADC12DJ5200-EP: Crosstalk in single channel mode

Hi Danilo,

Just to be clear, the customer wants to use the 5200RF device in single channel mode, or JMODE 1 for example? And they may still have an analog input signal applied to Channel B of the device?

If you have any other information this will help clarify.

Regards,

Rob

Hi Rob,

Sorry for my late answer.

I would like to use the 5200RF in single channel mode (@10GS/s) on input B. Unfortunately, there would still be some undesired signal present at input A.

What would be the crosstalk in that case?

BR,

N.

Hi Nicolas,

If you are using JMODE1 in single channel mode with either CHA or CHB, this is not crosstalk, this is an artifact of interleaving the two ADC cores of the devices.

If you want to send me an FFT plot, I can verify this for you. I am sure you are seeing a spur (interleaving spur or IL spur) in the middle of the spectrum. Is that the case?

Thanks,

Rob

Hi Rob,

Thank you for your answer.

I don’t have any hardware with the ADC12DJ5200 in my possession. I am making a component pre-selection for a new design.

I would like to use the ADC12DJ5200 in single channel mode Dual Edge Sampling (DES), i.e.at 10GS/s.

The analog source is either SRCA or SRCB. (see figure below)

For example, when SRCA is used I can attenuate SRCB but not completely switch it off and vice versa.

Therefore I would like to know if the crosstalk can be a problem for my application.

I have read as well that for single channel mode DES, input A is preferred. If using input B instead, what kind of performance loss can I expect?

BR,

N.   

Hi Nicolas,

Are you planning to move the signal source from one channel to the other when using JMODE1? 

Otherwise, there won't be a crosstalk issue if you use one channel and terminate the other unused channel.

Internally the mux is optimized for CHA, I would have to check with design on what performance would change if using CHB.

Regards,

Rob

Hi Rob,

Yes, we are planning to move the signal source from INA to INB when using JMODE1 according to the picture of my last thread.

We would like to use the ADC12DJ5200 as an analog mux. Therefore my concern about the crosstalk !

We would appreciate if you could check what performance would change if using CHB.

Regards,

N.

Hi Nicolas,

I checked with design and characterization. This type of measurement has not been done on the bench.

Please provide the following we can perform the measurement in our lab.

sample rate: 10.4GSPS? when in JMODE1?

Signal source for both CHA and CHB, need frequency and amplitude level. Especially for the opposite channel when not in use.

Once we have this information. I can provide a timeline on when we can get this measurement completed for you.

Regards,

Rob

Hi Rob,

Thank you for your answer.

I would like to know the same measurement as datasheet figure 6.2 (dual channel mode - Jmode3) for single channel mode Jmode1 at 10.4GS/s.

The crosstalk in dB is supposed to be linear (not dependent on the aggressor signal amplitude). For the convenience of the measurement, I would use an aggressor amplitude corresponding to 100% FS.

Regards,

Nicolas

Hi Nicolas,

Okay, we will replicate the frequency sweep as in Figure 6-2 in JMODE 1.

Regards,

Rob

Hi Rob,

Great, thank you.

Regards,

N.

Hi Camilo and Rob,

Thanks a lot for the measurement. Impressive result !

This solves my issue for this point. However there was another open point:

Internally the mux is optimized for CHA, I would have to check with design on what performance would change if using CHB.

Thanks again for your help.

Regards,

Nicolas

Hi Camilo,

Thank you for your answer.

It answers my question.

l have another question regarding the S-parameters you provide for the ADC12DJ5200 input and response.

It basically concerns the input, the measured S11 which I am simulating with.

The fact that the S11 termination is purely differential doesn't correspond to the reality where the ADC input internal 50 Ohm terminations are referred to AGND (see datasheet figure 7-1).

It is a real simulation problem for the amplifier driving the purely differential ADC input model. If the amplifier has some common mode, it will be completely reflected whereas it is not the case in the real world. This leads to erroneous simulation.

Would you have a 2 port S2P file that is a full differential measurement of INA+ (Port1) & INA- (Port2)? (see second figure below)

Thank you and regards,

N.

Hi Nicolas,

The S parameters that we currently have for the ADC12DJ5200RF are the ones in the website. Included with them is this PowerPoint presentation that contains an image in slide 3 on how to add an ideal 1:1 transformer to make port 1 into a differential. Is that enough? Or do you need the parameters for each input to be separated? I can also check with my team to make sure the parameters for the ADC12DJ5200RF are the same for the ADC12DJ5200-EP. 

Best,

Camilo

ADC12DJ5200RF S-Parameters.pptx

Hi Camilo,

Thank you for your answer.

I would really need the S-parameters corresponding to the picture below where the Vector Network Analyzer (VNA) measures the ADC inputs in full differential configuration yielding a S2P file. BTW, it is most probably the way you have measured the ADC input !

Best regards,

N.

Hi Nicolas,

Yes, this is the way we measured the ADC, but our VNA doesn't have those capabilities to do a true SDD11 measurement and capture all the common mode Sparameters.

Regards,

Rob

Hi Rob,

Since you have measured the S-parameter according to the picture below, would it be possible for you to share with me the resulting S2P S-parameter file ?

Thanks and regards,

N.