This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

increased A/D noise when isolated

Other Parts Discussed in Thread: SN6501, MSP430F1222, TEST, RC4580, SN74AHC1G04

Dear Forum Members,

Im working on a data acquisition circuit with USB connection and I experience high noise level despite the presence of isolation. Plese see the attached picture showing the layout.

The isolator is a capacitive SPI type (non TI part) Of course, the power also isolated and implemented with SN6501 oscillator IC and a Würth small ferrite transformer combination.

Both part of the circuit do utilize ground plances.

The data conversion is based on an MSP430F1222 device. The MCU also generates symmetrical square wave as test signal (using TIMER_A module in Contious mode, OUT unit in Toggle mode).

The high noise level still exists if the input tied to GND.

The noise can be highly reduced if I touch the isolated analog GND (Side A on the picture) or I connect the two ground planes thus elimintaing the isolation (even if using some nF capacitor).

The SPI CLK is run on 2,5MHz.

Can you pease give some advise what can be the rootcause becasue i do not have idea.

Thanks in advance!

  • Hi Joseph,

    To be able to identify which device is responsible for the issue, could you please elaborate more on the issue?

    1. Where do you see the noise? At ADC inputs? Only at the ADC output? Or in the PCB after you read through USB?
    2. Can you please share any waveform of good and bad signal?
    3. What kind of signal that you are feed in to ADC?

    Please do share as much information as possible so that the issue can be rootcaused quickly, thanks.

    Regards,
    Koteshwar Rao

  • Hi Rao,

    thanks for the response.

    1. Noise is present bot the ADC input and read-out USB data.

    2. Unfortunately I can send waveforms on next week (currently the circuit is not here where i am..).

    3. As i mentioned, the MSP has a (test) signal output (TA2) as well (generated by TIMER_A - used as NCO), but similarly noisy if input simply tied to AGND (isolated from USB's DGND)

    I would have a question about performance of capacitve isolation:

    Whether it offers same effectiveness from aspect of common mode noise as inductive or Photo isolation? Because referring to the digital signal transitions (high frequency content!) the isolating on-chip oxide capacitor (~1pF) exhihibits a limited impedance only (not infinite as an inductive/opto can).

    Regards,

    Joseph

  • Hi Joseph,

    Thanks for your inputs.

    1. If noise is observed at ADC input and at the read-out USB data then it doesn't look to be isolator issue.
      1. If you can share the isolator part number, I can recommend you equivalent or better TI part number that you can also test and check.
    2. Sure, you can share the waveform when you get them next week.
    3. Understood.

    For TI isolators, the isolation capacitors are deep inside the device with many other circuitry preceding them to inputs. The isolation capacitor existence in the device is transparent to external signal inputs. The isolation caps primarily determine the high voltage performance of an isolator along with a few other HV EMC tests. In your application, it doesn't look like you have any high voltage transient to affect isolation barrier.

    Since this issue doesn't look to be an isolator issue, I will let the MSP430 team to respond to you. Thanks.

    Regards,
    Koteshwar Rao

  • Hi Rao,

    You may have right:

    the problem is should not caused by the isolator (since no HV signal between the two grounds):  When I touch with a piece of wire the Power transformer secunder winding or one of SPI (especially SCK) signals the noise dramatically increase. First I tried to complement the SN6501 circuitry with RC snubbers (200pF+2KOhm) but there was no effect (or only neglible).

    Note, the pre-amp has high impedance (n*100KOhm) signal paths altough solid ground plane used. It seems it is an onboard EMI issue. Unfortunately, ferrite beams start to be effective above 10..30 MHz (my SPI is run on 2MHz and SN6501 similar range).

    Waveforms: GND not touched:

    GND touched:

    Kind rergards,

    Joseph

  • Hi Joseph,

    Is it possible to share your PCB design? You can either take screen shot of layers in your CAD tool or the output gerbers (or pdf version of gerbers) that are generated by your CAD tool.

  • Hi Dennis,

    thank your for your reply, I included a layout: IC9 at the upper lefet corner is a frequency divider HC4518 as test signal output. Note,

    recently instead of IC9 (I removed it) I use an SN74AHC1G04 drived from TA2 out of MSP430 not the ALCK out. IC8 is a dual OPA (RC4580) acts as an active attenuator+level shifter, IC6 is an analog 8-1 MUX (for attenuation setting).

    Regads,

    Joseph

  • Hi Joseph,

    When you took the two scope captures, where did you have the ground lead of the probe connected and how long is this ground wire on the probe?

    Next, in your first picture showing the noise on the square wave, if you re-program the MSP to stay in LPM3 (no CPU or clocks running) then probe this same point do you see noise and if so does touching the ground plane make a difference?

  • Hi Dennis,

    The graphs are not taken by an oscilloscope but waveform built based on read-out samples via USB (HID).

    However, I also checked ADC input by real scope and I saw similarly noisy input signal (scope reference gnd was the MSP ground plane).

    It is a good idea to enter into LPM3 mode! I can back with this experiment only on next week..

    Note, when I touch (by finger or a piece a wire) SPI lines or especially the SN6501 circuitry (Würth transformer primer) the noise highly increases. I would like to ask why R-C snubber not reuqired (not siggested in the datasheet) for '6501 in contrast traditional fly-back converters?

    It should also be noted, the analog input has high impedance (1MOhm): input signal is converted to current by 1MOhm resistor and it may be susceptible to pick up high frequency signals/harmonics despite

    the presense of solid ground plane...However, I exclude the USB D+/D- lines because when I touch them under comm the noise does not increase.

    Regards,

    Joseph