[SOLVED] Amplifier and external filter design (accelerometers) gives oscillation

[Abraham900]

Hello.

I am doing a prototype to register vibrations created by the human body by the use of 2 accelerometers. Everything works fine and it is able to register the vibrations that is created by the heart. I have a high pass filter to filter out the DC component of the signals at 0,5Hz but no low pass filter so this introduces some noise to the system. Reason being is that I wanted to see how much noise there is on the signal. Therefore, I created an external filter consisting of some resistors and capacitors with a component from MAXIM, MAX280 that creates a fifth order low pass filter at 2kHz. For some reason though, this accelerometer design does not work with this external filter. The signal oscillates and provides very strange values when sampled to LabVIEW.

The external filter is connected to VOUT on the accelerometer schematic.

The signal is stable on the input on the filter design but becomes fully distorted on the output. More like a triangle wave when using oscilloscope to measure it. I cannot find any solution to this issue and I am kind of lost now... And I'm certainly not getting smarter by the days that goes trying to solve this issue. Any help is fully appreciated.

The filter design and MAX280 PDF:

https://datasheets.maximintegrated.com/en/ds/MAX280-MXL1062.pdf

https://www.dropbox.com/s/5xfdpyhgjl703m7/filterdesign.jpg

The accelerometer design:

https://www.dropbox.com/s/1edhq49zhls1yiw/accelerometerdesign.jpg

Thanks!

 

[prestonee]

hmm i guess one concern but im sure you connected this is that this is a +/-5 part, i see you connected the +5 but not the -5V in your drawn schematic?
they 'claim' it works with single ended supply, but if you read the notes they do this by making agnd pin !=0 but =2.5V. and increasing the output reference to 2.5V as well... essential a poor man conversion to turn a single ended supply to a diff supply using 2.5V as the new ground.
You are not doing this though.

best luck
,Pb

- - - Updated - - -

I just noticed your ref node, so you are half way there, you biased Agnd at ref which i hope is supply /2. however you didn not provide the save divide on your out. this is probably causing your issue. your total r in=10.1k, so according to doc you need a 121.2k res between agnd(2) and vout(7), and to disconnect vout(7) from the cap on vin(1). and add a new C between these two disconnected spots.

 

[Abraham900]

Thanks for your input. I solved this. Thanks!