AD9288 offset output

Hi MChante ,

Thanks for your interest in AD9288.

May I ask what are the voltage swing and the gain at the output of the AD8137? There is nothing wrong in the AD9288 if the common-mode voltage of each AIN pins is equal to 0.99 since 0.3 *VD, and the VREF is equal to 1.25V. Though, I tried to simulate ADC driver and found some errors. Kindly confirm if these errors manifest in your design.

ADI-DiffAmpCalc | Analog Devices

Regards,

Xavier

Hi Xavier,

The input system voltage is 750 mVpp at the input of R141 (for ADC A). AFAIK, amplifier's input does not induce clipping. The expected (and measured) differential voltage on AinA and /AinA actually is 440 mVpp. The error you seem to have detected is related to Rf and Rg that could be lower than the (Rf+Rg) of the 2 kOhms required (not mention in the component datasheet if I'm not wrong). Is this right ? In any case, you have noticed that there are two pairs of coupling capacitors before the ADC input, respectively C20, C25, C31and C34. This should not be an issue, should it ?
Therefore, I don't understand the "error" you are reffering to, could you explain?

Thanks in advance for your reply.

Kind regards,

M. Chante

Hi MChante ,

It seems that your analog input swing is with specs. Can you provide a scope shot on the following points? As for the Rf and Rg, I also agree that it is not stated in the component data sheet, but it has recommended values on the gain setting resistors in table 7 that somehow assumes that Rf+Rg >= 2kohms.

No problem on the AC coupling capacitors since the AIN circuitry can provide an inherent VCM.

Just another point, what is the purpose of the Coax footprint? Does it have an impedance? How did you come up with the 33-ohm terminating resistance?

“Rules of the Road” for High-Speed Differential ADC Drivers (analog.com)

I apologize if I have so many questions. I just wanted to make sure that there are no stones left unturned in understanding your query. Thank you.

Regards,

Xavier

Hello Xavier,

Tanks a lot for your quick reply, I perfectly understand your questions. Here are some answers.

1/ I understand that Rf+Rg should be such that their sum should be 2 kOhms (at least ?), as it appears on the example given in table 7 of AD9288 datasheet. These resistors values could probably be increased to 1 kOhms each instead of 820 Ohms. I'll give you the results when it is done.

2/ I am OoO today and will send you the expected scope measurements by tomorrow.

3/ The 33 Ohms resistor has been selected so that the equivalent input resistance roughly is 50 Ohms (considering the single ended input (18+32.2//1.1 kOhms). The 1.1 kOhms is the equivalent input resistance of the AD8137, see its datasheet, equation 9, page 24. Therefore the equivalent input resistance in the system is 49.3 Ohms, very close to 50 Ohms seen from connectors inputs. This 33 Ohms resistor also is such that there is some attenuation of the input signal from P200 or P2. This attenuation roughly is -3.9 dB. Also, considering the three poles differential filter on the outputs of the AD8137, the whole attenuation is -4.5 dB.

4/ Input connectors: There are two possible applications: one is getting the signal from connector P200, the other from P2. In both cases the source resistance is 50 Ohms, then matching the system input equivalent resistance.

Let me know whether my explanations are clear enough for you.

Kind regards,

M. Chante

Hello Xavier,
Here are the expected scope plots.

Top of R45

U6 outputs

ADC inputs (channel A)

You will notice that U6 outputs are centered on VCOM (approximatly 3.3 V / 2 = 1.65 V)
Also, the ADC inputs (channel A) are centered on 1 V.
Signals are clean and OK for me.

Do you think some more measurements are needed ?

Regards,

M. Chante

Hi MChante ,

I also tried to simulate your ADC driver in LTSpice and it seems there is no problem with it even at 820 ohms Rf and Rg.

Have you check the ADC output if with no ADC driver input at all? Are both Channel A and Channel B have the same issue?

Perhaps you can send the digital output waveforms and FFT spectrum so that we can check it in another angle.

Thank you.

Regards,

Xavier

Hi Xavier,

Answers to you questions:
1) I cannot  do any measurements with no signal conditionning (I mean double AOPs), sorry,
2) I confirm A and B channels behave the same,
3) I'll try to send you an FFT waveform as soon as I can.

Mathieu

Hi Xavier,

Sorry for this late reply, I've lately been quite busy.
It appears that we had a flaw in PCB manufacturing unexpectedly connecting a 33 kOhm pull down on AINA (pin 2 of AD9288). Hence the digital offset. Once fixed, the system works perfectly.

I am really sorry to have bothered you with this issue which was finally due to a wrong PCB manufacturing.

I do appreciate your help, thanks alot for your hints!

Mathieu

Hi MChante ,

No worries, always happy to help. I appreciate that you replied back to give resolution to this query. 

Regards,

Xavier