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.

TLV3702 comparator glitches with slow signals

Other Parts Discussed in Thread: TLV3702

Hi all,

I have an interesting problem with a fairly simple comparator circuit that converts a low-level sine wave to 3.3V logic levels to measure frequency. Part of the system is shown below, and was reverse engineered from an obsolete design.

The input signal is a sine wave in the range 200mVpp to several volts p-p, ground referenced. The hysteresis is set at 70mV to cater for the low amplitude signal. The output feeds a microcontroller timer input to measure frequency/timing.

The input RC filter rolls off around 15Hz because the signals of interest are only a few Hz or tens of Hz and amplitude increases with frequency. The sensor is effectively a small AC generator.

Our problem is that this circuit is in close proximity to a GPRS modem and the power supply is shared, although with separate regulation and ground planes on a 4-layer SMD PCB. We've tried to reduce grounding issues during layout, but still suffer some false comparator transitions when the GPRS modem is operating.

Removing the input filter capacitor (C20 on attached schematic) usually clears up the glitches, but we're not sure why. The capacitor ground is direct to the comparator ground. Is it possible there's some sort of input instability with a capacitor like this, and relatively high source impedance?

Does anyone have ideas on what to look at in terms of decreasing noise sensitivity? We suspect that we're getting some RF breakthrough/demod too, and figure it's likely to be a combination of things...

 

Any thoughts welcome!

Regards,

Matthew

  • Hello Matthew,

    This sort of problem is tough problem to track down. The TLV3702, being a nanopower comparator, is a relatively slow switching device and most likely the false triggering is the result of the GPRS RF being rectified and moving a threshold point on one of the inputs. If rectification is at the root of the problem it is most likely taking place in the differential input stage, the input offset protection circuit, or the input ESD protection cell. Another possibility for rectification would be the external components, D12 and D13.

    One would think that C20, the 100nF, capacitor would be effective in shorting out any RF appearing at the inverting input; but likely, its ESR and ESL are high at the multi-Gigahertz GPRS frequencies. If that is the case the capacitor would appear as a non-zero impedance at such frequencies and an RF voltage still could be developed at the input. C19, the power supply bypass capacitor, would exhibit the same characteristics. Since it shares the same power supply line as the GPRS circuit RF might enter the TLV3702 through the supply pin. The offset protection circuits have normally-off ,diodes junctions in the path from the power supply pin back to an input.

    If the rectification scenario is correct, keeping the GPRS RF from entering the TLV3702 would be key. I understand why you wouldn't want to remove the input filter and power supply pin capacitors; they are essential for their intended functions. Possibly another capacitor, effective at the GPRS frequency, could be added from the TLV3702 inputs to ground and another at the power supply pin. Also, if the RF is being conductor on the power supply line then an RF choke, or power supply EMI filter - effective at the GRPS frequency - should be considered.

    Making changes to a PC board is a pain but often required to resolve unexpected problems of this nature.

    Regards, Thomas

    PA - Linear Applications Engineering

     

     

  • Hi Thomas,

    Thanks very much for the ideas and advice. It appears that we did have some RF breakthrough and/or demod happening. We've added some low-value RF decoupling caps close to the comparator and modified a couple of PCB traces, and that seems to be enough to reduce the interference to acceptable levels even with a poor GPRS antenna position (highest transmit power).

    We know what to look for next time!

    Regards and thanks,

    Matthew.

  • Matthew,

    Thank you very much for the feedback. We are working hard here in Precision Analog to be as knowledgable about EMI/RFI issues as is possible. Our customers are encountering EMI/RFI problems more often because of the close proximity of analog integrated circuits and powerful RF sources in a common circuit space.

    If you would like some additional information what we are doing about this matter we have a newly published Applications Report, sboa128. Just search on "sboa128" from the TI website to obtain it. I think the report will prove insightful.

    Regards, Thomas

    PA - Linear Applications Engineering