Post Go back to editing

LTC6229 Output Signal Spikes

Category: Hardware
Product Number: LTC6229

I am currently working on a SiPM detector design. The photocurrent is converted into a voltage by a two stage transimpedance amplifier which is based on a single LTC6229 OPV (Schematic shows LTC6228, since the 6229 is not available in LTspice):

After successfull simulation, I have designed a PCB. The oszilloscope shows small unexpected periodic peaks on the output signal of the detector:

My first guess was that those peaks were caused by distortions on the power supply line from e.g. a switching regulator. But both, positive and negative 5 V supply lines do not show any distortions. The peak-peak voltage of both rails is approximately 10 mVpp without any significant spikes. So the problem has to be somewhere else. I recognized that the frequency of the distortions depends on the supply voltage of the photodetector. So I thought that the spikes are actual "current spikes" of the SiPM detector. To double check this, I disconnected the supply voltage of the SiPM, and soldered a 2k4 resistor between the inverting input of the first stage OPV and ground. So that the first stage is an inverting amplifier, instead of a transimpedance amplifier. However, the oszilloscope still shows some siginficant peaks on the output signal (Frequency and amplitude of the peaks have changed):

Do you have any idea, where those spikes come from? What can I try to prevent this behaviour?

Thank you in advance.

Best regards.

  • Hi  

    Have you implemented using bypass capacitors on the power supply pins? You can check the page 20 of the datasheet regarding this matter.

    Hope this helps.

    Regards,

    JE

  • Hi jeobico,

    thank you for your response. The datasheet does recommend a 100 nF ceramic capacitor in parallel with a 1 nF ceramic capacitor for each supply rail. I decoupled each supply rail with 10 uF of size 1210, and two 100 nF of size 0603. Since I do not want to use smaller capacitor sizes for technological reasons, a 1 nF capacitor should not bring any advantage, since the parasitic impedance does mainly depend on the capacitor size. Here are the theoretical impedance and ESR values for the 100 nF (blue) and 1 nF (green) capacitor:

    However, I changed the closest capacitor to 1 nF. Which had no effect on the output signal of the amplifier.

    Meanwhile, I recognized that those spikes might come from electromagnetic distortions of a small power module on the PCB that generates the supply voltage for the SiPM. One edge of the SoC does contain an unshielded inductor. The distance between the inductor and the amplifier is less than 1 cm. If I place a piece of metal between the module and the amplifier, I can reduce the amplitude of the spikes. So I made a new PCB with the following changes:

    • Move the supply SoC as far away from the amplifier as possible
    • Rotate the supply SoC so that the inductor is located on the other side of the amplifier
    • Add a metal case around the supply SoC for shielding
    • Shorter signal traces within the amplifier cicruit

    The PCB is in production now. I hope that those changes might solve the problem.

    Best regards.

  • Hi  

    Have you tried to use the evaluation board of this part? The eval board can be the benchmark on your design. For reference, see the link below.

    EVAL-LTC6227MS8E/EVAL-LTC6229MS8E (Rev. A) (analog.com)

    Hope everything will work out. If you have any questions or concerns, kindly contact us again.

    Regards,

    JE

  • Hi,

    I believe that the questions have already been answered so I am closing this thread. Thank you!

    Regards,

    Mae