A/D input stage of Analog Devices pulsar module

Thread Starter

MisterNo

Joined Jun 28, 2010
26
Attached is the input stage schematic of the Analog Devices pulsar A/D module.
For the ADC driver they used ADA4841 op-amp that has maximum 12V Vcc - Vee voltage, so op amps are powered with +7.5V and -2.5V
As the module is designed in sample differential signal in the -5V to 5V range they added summing point on each op-amp input raising each input by 2.5V through the simple voltage divider + filtering stage with 10u and .1u caps
Vcm is 5V and comes from 5V reference

My question is, why do they make things complicated? Since they powered op-ams with Vee of -2.5V they had to add 5V reference, filtering stage and precise matched resistors R1 R2 R3 and R4
in addition, using 590ohm on the input resistors will only work for low impedance sources and sources that can provide almost 10mA of current

Why not to choose op-amp that can be powered to Vee <-5V, then none of the above would be necessary.

Am I missing something here?
 

Attachments

Thread Starter

MisterNo

Joined Jun 28, 2010
26
Thanks, LMH6550 is a good one, but not pin compatible with the PCB
I have replaced the two op-amp with two MAX9632 and ditched out that level shifting.
 

MrChips

Joined Oct 2, 2009
30,706
The level shifting is required for optimum performance. I cannot see the part number of the ADC in the schematic.
It is common for a differential input ADC to have a common mode reference of 2.5V. Hence the output of the differential amplifier has to be biased by 2.5V.

I do not know why they used two separate bias circuits. I would have used a single Vcm. The ADC itself ought to have a Vcm output of 2.5V which should be used as the Vcm reference.
 

Thread Starter

MisterNo

Joined Jun 28, 2010
26
The A/D is AD7984, a truly differential A/D, so no shifting is necessary.
Even in the datasheet they have a differential op-amp driver for A/D, as you mentioned above.

To me the board is just a bad design. Attaching on the inputs a voltage source, even if it comes from a reference will just add noise and A/D on the board is of 18bits at 1.33Msps. Vcm filtering caps of 10uF and .1uF are not coupled with any resistance except the PCB trace so their effectiveness in the kHz range is zero.
Also creating an input differential impedance of just 1.2 Kohms is insane.

The whole board is supposed to be a reference board for AD pulsar line of A/D converters.

Here is the figure showing Vcm buffered reference voltage
 

Attachments

Top