I am designing board with very large dynamic range.
Therefore I am using AD8331 amplifier in Low Gain Mode. According to Datasheet it can be used for driving 12bit ADC and in some cases with 14bit ADC(output filter is needed).Bandwith of interest is 100kHz - 30 MHz.
AD8331 output voltage is 2Vpp. Unipolar Power supply 5V.
According to my calculation AD8331 limits ENOB to 11.5bits
AD9253 has VCM=0.9V and AD8331 has VCM=2.5V. Therefore some kind of translation has to be added.
At first, I was thinking of adding ADA4930 that is typically used for driving low voltage ADC, but I want to minimize component number and power consumption.
I am wondering if I can use AD8331 to directly drive ADC AD9253?
Idea is to add this type of circuit to output of VGA and AC couple it.
Will it impact ADC and all chain performance?
Thank you for your interest in AD9253 and AD8331. As you have observed there is a mismatch of the AD8331 output common mode voltage compared to the input common mode voltage requirement of the AD9253. For this reason you cannot directly drive (DC couple) to the AD9253.
Your schematic showing the AC coupled configuration looks good conceptually, though at 100kHz the impedance of the AC coupling capacitors could start affecting the loading of the amplifier and the voltage at the ADC inputs.
The AD9253 datasheet performance is for the ADC alone without an amplifier. Adding an amplifier will always add some noise and distortion. I'll ask the AD8331 engineer look at this to see if he has any additional comments.
Thank you for an answer.
I understand that any amplifier in chain contributes to noise. Actually my question was:
Is solution with ADA4930(Ac Coupling (30k-30MHz), Gain 1) better or worse regarding noise/performance than this solution when ADC is drived directly from AD8331?
If possible can you put some numbers or point to some document with example calculations?
I'll be waiting for an answer from AD8331 engineer regarding loading of VGA.
Regarding loading of the AD8331, the engineer suggested increasing R6 and R7 from 100 Ohms to 250 Ohms so you’re closer to the drive capability of the AD8332.
Regarding comparing the noise performance of ADA4930 vs. AD8331, please look at the datasheets for these amplifiers. I'll also ask the amplifier engineer his/her opinion.
Thank you for kind reply.
However, my question is not to compare noise performance between ADA4930 and AD8331.
My question is to compare noise performance between following
1. AD8331(LO mode) -> ADA4930 -> AD9253
2. AD8331(LO mode) -> passive network above -> AD9253
Bandwidth of interest is 100kHz - 30 MHz.
I am wondering if I can remove ADA4930 from board without impact on performance.
This would save me money and minimize on-board power dissipation.
I'm sorry about my misunderstanding your question!
Yes, I think your option #2 is good. You might have to experiment with changing some component values (e.g. AC coupling capacitor size, and R6 and R7 values, but that looks like a good option.
Do C4 and R5 in your passive network represent components on your board, or do they represent the ADC input loading? If they represent the ADC inputs, I think it is a good idea to add pads for a shunt capacitor and even single ended capacitors to ground, like C and C1 in Figure 58 and 59 in the AD9253 datasheet.
I hope your project goes well.
C4 and R5 represent ADC Input Loading, but schematic was just concept sketch.
Of course I have additional shunt capacitor and capacitances to the ground.
I have already connected AD8332 eval board to our board with AD9253 and this is functioning very well.
Again, thank you for support.
Great job! I'm very glad to hear your results!
Please take care.