I have a confusing problem with ADA4938-2. I connected ADA with unity gain and dual supply to AD9253. Vocm is provided by ADC. Schematic and parts values seem to be correct but output of ADA is very noisy. Any suggestion??
Could you share your schematic for us to check? Thanks.
Anyway, I would assume the following conditions: AD9253 AVDD = 1.8V and ADA4938-2 is in single supply (let's say 5V). Given the condition that the Vocm is coming from AD9253's VCM which is AVDD/2 = 0.9 V or from the VREF which is 1.0 V, in these cases those values are out of range for ADA4938's Vocm.
The ADA4938 operates at Vocm of 1.3V away from the supply rails. So at single supply (VCC=5V), this range will be from 1.3 V to 3.7 V. You may use + 2.5V supply to power ADA4938 if you wish to drive the AD9253, but the Input and Output Ranges would change.
Thanks for your quick reply. I connected Vcm of AD9253 to Vocm of ADA4938. So common mode voltage is set on 0.9 volts. The main problem is strong spurs in op amp output. Here is some information.
2. FFT of ADC samples (op amp Connected to ADC ):
3. FFT of ADC samples (op amp disconnected from ADC)
Thank you for your interest in ADA4938 and AD9253. I'm sorry but I cannot see the schematic clearly. Would you please confirm what the ADA4938 supply voltages are?
I believe Jino's point is that, depending on the amplifier supply voltages, the ADA4938 might not be able to support an output VCM = 0.9V even if 0.9V is applied to the VOCM pin.
ADA4938-2 supplies are +5V and -5V
Thank you for the reply. +5V and -5V sounds good to me, though Jino can confirm (I am not the amplifier guy).
Can you tell me more about the signal you are applying? Is the upper FFT with the amplifier connected (and with the spurs) with a tone applied as well? Can you tell us more about your test signal?
Both FFTs are noise floor of ADC. First with op amp and second without it. I really confused about this. I've designed many software radio cards including ADC, DAC, FPGA and ... and this circuit is first of them that failed in analog section.
Few questions on your circuit.
What is you input voltage level? I need to know if the amplifier output is still in the specified output swing. With your current circuit, VCM=0.9V and output swing limited to + 3.8 V, the input should be between -4.7 V to +2.9 V.
Also, 26ohms could cause an unbalanced termination for the inputs. I believe 27.4ohms is a standard value, you can use it instead.
I noticed on your schematic that you are terminating your coax inputs in 62 ohms, but you've used 26 ohms to terminate the unused differential input. This mismatch may very well be the source, as it will apply ground noise to the amp inputs in differential mode instead of common mode.
Just a hunch.
Your termination scheme also unbalances the feedback resistances such that the inverting and non-inverting sides have slightly different gains at less than unity. This is problematic.
You may wish to consider terminating R7, 13, 25, and 33 directly to ground and adding unity-gain buffers to your coax inputs with the terminating resistors on the input sides of the amps. Tie the buffer amp outputs directly to R1, 10, 22, and 30. This will preserve the matched input impedances and matched unity-gain factors on the ADA4938s.
Per previous posts, you may wish to power the ADA with both + and - rails to keep the Vocm well away from the ground rail.
Retrieving data ...