THS4551: Industrial Texas Instruments Based IEPE System Schematic Review (URGENT)

Hi Ignacio,

For the first question, I would recommend you to check with the Precision Data Converters team in their forum.

For general observations regarding the THS4551 circuit,

1. What is the supply voltage being applied to the THS4551? Is it single 5V? If yes, then I don't think you can connect the THS4551 Vocm to 0V because the output will be clipped towards the negative supply. For interfacing with the ADS1257, you need to connect the Vocm to 2.5V.  

2. Also, it looks like the input is GND referenced, which means you might run into input voltage headroom issues in single-ended to differential configuration on a single +5V supply. Do you know the input voltage swing that will be applied to the THS4551 input before the 1k resistor?

3. Otherwise, the THS4551 schematic should be stable with ~37kHz of low pass cut-off frequency.

If possible, I would recommend the customer to simulate the THS4551 circuit in TINA-TI before proceeding forward with the board design. The customer's TINA-TI circuit is attached: THS4551_MFB_37kHz.TSC

Best Regards,

Rohit

Dear sir, Thank You

1) VOCM is not 0V, it is 2.5V (a virtual ground called VCOM created by (opa365)

2) By selecting the correct IEPE accelerometer with the proper dynamic range and playing with the atenuation R012 / R013 and with the gain i can adjust the headroom before the 1k resistor

Please ensure you check the virtual ground and tell me if all its ok after that ...

I know the ADC should be at least 60KSamples but firmware its already devoloped for such ADC

I forgot the Noise reduction capacitor for REF5025

But i will place 1 uf there before production

I connected the ref5025 directly to the adc, (ref 5025 has no outbut buffer) but i think its not necesssary to buffer the 5025 output for just the adc ref

Kind regards

Hi Ignacio,

One thing I noticed is that the second OPA365 (U18, for example) which is configured in G = 2V/V will have its output close to 5V at all times. This is because its DC input is 2.5V coming in from the first stage OPA365 (U17), and will 'not' be attenuated into half or 1.25V because of the OPA365 high input impedance and 1uF ac coupling cap. Any AC input signal will be clipped on the positive side due to the second stage OP365 output being close to 5V. I think you would have to include a 100k resistor to GND at IN+ of second OPA365 (U18), in-order to set its input at 1.25V. The other option would be to configure the second stage OPA365 in unity gain.

If including the 100kohm resistor and it becomes too noisy, then I believe you have to reduce the values of R012 and R013 100k-ohm resistors. Reducing the value of R012 and R013 would mean that the VOCM signal or virtual ground could be modulated by the AC input signal. So, you may want to include a placeholder for an RC filter at the first OPA365 output (U17), as shown in Figure 22 or Figure 30 of the OPA365 datasheet. The R & C values could be 10-ohms and 1nF, respectively.

Best Regards,
Rohit