• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 16 subscribers
  • Views 629 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

CD4051B: CD4051 output has spur while the input is clean

CD4051B: Control signal falling edge spikes on output signal.

Kevin O'Neill
Prodigy 10 points
Part Number: CD4051B
Other Parts Discussed in Thread: MUX36S08

Hello, I am getting spikes on the output of a CD4051BM that appear to be inductive feedthrough from the control signals:

Top trace is the pin13 output, bottom trace is the A control signal at pin 11. Y-axis is 5V/div, I forgot to note what the X-axis is set at, but the frequency at the A input is around 1-2kHz, and the spikes are present at all speeds. Spikes in image are 2VDC, but spikes on other outputs vary up to 6VDC, depending on summing overlap of A, B, & C control signals. Control input signal is 0-7 binary, so frequency A = 2x frequency B = 2x frequency C.

All involved ICs are powered at  VDD: +12VDC, VEE & VSS: 0VDC.

Setup:

  1. Incrementing 1-8 decimal count input to IC5, a CD4532BM
  2. Encoded output goes to IC14, a CD4070BM (used as selectable inverter)
  3. IC14 output to a pin header to socket header on parallel PCB. 
  4. Socket header to control inputs of three different ICs:
    1. IC41, the relevant CD4051BM
    2. IC34, a second CD4051BM (outputs of which drive individual 2N3904 through 10kΩ resistors)
    3. IC40, a DG408 (used to route ±10VDC signals)
  5. IC41 has +12VDC connected to Common Out/In pin3, I/O 0-7 each go to individual pins on a pin header that is 0.15" away from the IC.
  6. Pin header has no further connection.

Attempts to remedy (none had any affect):

  • Remove output pin header (outputs now wholly unconnected save for  0.1"-0.2" traces
  • Remove IC34 & IC40
  • Reduce control signal voltage to 0-9VDC
  • replace control input traces with 28ga. wire (traces cut)

The only thing remaining that I can think of is the warning on the data sheet  of

"To avoid drawing VDD current when switch current flows into IN/OUT pin, the voltage drop across the bidirectional switch must not exceed 0.6V at TA≤25°C, or 0.4V at TA>25°C (calculated from RON values shown). No VDD current will flow through RL if the switch current flows into OUT/IN pin." 

Is this what we're seeing, even though there is no load? If so, what can be done in this situation? 

Thanks in advance!

  • 0
    TI__Intellectual 1166 points

    Hi Kevin,

    From what you described it could be inductive feedthrough but since you have no load and it seems like the inductance is pretty small. 

    Another possible explanation is digital  noise. If the transition on your A control signal edge is very fast you might see some crosstalk to the analog channel. 

    And another possible explanation could be charge injection. When analog switches transition the change in capacitance of the MOSFETs will result in a charge change on the input/output which will show as a voltage offset. 

    For all of these though my first recommendation would be to add some capacitance on the IN/OUT pin. 10-100nF should help suppress these spikes and if you are running at low speeds like this in the kHz, shouldn't effect your sampling.

    We also have a pin-to-pin replacement for the DG408DY with the MUX508ID as well! 

    Thanks,

    Stephen

  • 0 in reply to Stephen Scheramic
    Prodigy 10 points

    Thanks! I'll try that out and let you know the outcome.

    And thanks for mentioning the MUX508ID, I wasn't aware of that one. I'll check that out later on.

  • 0 in reply to Stephen Scheramic
    Prodigy 10 points

    This didn't help, unfortunately. 100nF reduced the spikes greatly, but not completely (~1VDC), and slowed the trailing edge too much. The outputs will be used to gate analog signals, so a crisp edge is required.

    I also tried adding capacitors at various locations - the Out/In and the control inputs - and actually had some improvement at the control inputs but, for each output that was corrected, another was made worse. The best outcome was 100nF at the A control input (pin 11), which removed the spike from outputs 0, 2, 4, and 6 (pins 13, 15, 1, 2), but doubled the spike levels of 1, 6, 5, and 7 (pins 14,12, 5, 4). Adding capacitors to the B and C inputs only compounded the issue.

    I checked the voltage at the Out/In, pin 3-R185-R186 junction and it showed voltage drop impulses corresponding to the output spikes. I removed the divider and shorted +12VDC to the Out/IN, but it only made everything +12VDC - both the output levels and the spikes.

  • 0 in reply to Kevin O'Neill
    TI__Intellectual 1166 points

    Hi Kevin,

    It sounds like these spikes are from charge injection if you are seeing an improvement from the capacitors. If you are trying to reduce these glitches completely you may need a more precision device like the MUX36S08. This device is part of our precision portfolio and has extremely low charge injection (<0.5pC) and low capacitance (10pF).

    Another idea is could you add a resistor to ground on pin 13? If this isn't charge injection but instead some crosstalk coupling, a pull down resistor should help keep the X0 pin stable. 

  • 0 in reply to Stephen Scheramic
    Prodigy 10 points

    Charge injection does look to be the culprit. I guess it's time to rethink the section instead of fighting it, thanks for your help!