SSM2166 with differential input

Hello GaborNL,

I am wondering why you would use an input pad of a 2K and 1K resistors? The input impedance of the buffer amp is high, 180k, so I would use a 10K feedback like the datasheet suggests and if you need the pad then a 20K input resistor. But since you are looking for gain of course, why pad it down on the input? Just use 10K ohms all around. Then you can match them to get the best CMRR. 

Also, you could use a lower input resistor to have the buffer supply some gain taking the load off of your preamp circuit. Have a 5K series input resistor and a 10K feedback to give you 6dB of gain. The 1K feedback just does not seem good from my viewpoint but I might be missing some detail that is important to your design goals. 

I agree, I wish the CMRR specs would be published however, I understand why they might not because it relies so much on the matching of external components and PCB layout of the external circuit. 

Regarding the VCA input. I would personally do this like the datasheet suggests:

The buffer output is single ended so I would use the optional 10uf cap to ground near the ground of the buffer and use the 1K resistors as suggested. 

The input buffer as you drew up should work fine with the exception of some higher value resistors. However, I do not know why you would feed the input buffer AND the VCA with the same signal out of the pre amp? Setup the input buffer to do what it is designed to do which is drive the low impedance of the VCA input. Then you can also obtain some gain as well. 

I would not be able to easily find out about the matching of the 1K internal resistors. I would be surprised if they are laser trimmed. I would not expect that. They are probably just setup to be close to each other and the same size etc., then the normal variation of the silicon would affect both similarly and so not be too far off from each other. But, I know about matching for good CMRR often involves the matching of 1%, or better, resistors!! I think if you do this with your circuit driving the buffer you will get good CMRR and then drive the VCA from the buffer you should be good. There is not much opportunity for external noise to get into the system where CMRR would matter. 

I hope this helps...

Dave T 

Dear Dave,

Thank you for your reply!

I'm using a very low noise capsule + differential preamp for this project. My goal is to minimize noise. Also, I need a -6dB scaling to fit within the max vca input voltage range. In this position, the 17nV/rtHz of the SSM2166 is too much. An unknown part of this is coming from the buffer that I wish I could completely omit. But I can't, because this is where the IC derives the rms. At least I can omit it from the signal chain though.

I intend to use the external series 1ks for the vca inputs to obtain the -6dB there. I imagine the Vca inputs as the inputs of an inverting opamp stage, the input current into this virtual ac GND will be thereby halved.

The signal for the buffer also needs the same scaling, otherwise it leads to 6dB higher RMS reading. This can be done with 10k and 20k resistors (instead of the 1k and 2k in the drawing) since the buffer input is not noise sensitive, and it eases the load of the previous stage.

As you can see, I'm avoiding increasing the number of stages and the source impedance of any kind to keep the noise very low.