1

Do I need a separate capacitor bypass (network) for 74xx14 Schmitt Trigger sharing power with Op Amp?

Power supply is 500mA but entire circuit is using around 70-80mA. And sub-circuit with OP AMP + Schmitt Trigger is between 1-3mA max. Random interval pulses going every few seconds from Op Amp (1kHz~65kHz and 0-3.3Vpp) to Schmitt trigger and out to MCU input capture pin (10kOhm).

Also, when I turn on the power, during the first 3-5 seconds the Op Amp's output spikes HIGH before settling down to a baseline of near 0V. Any way to prevent that? The sensor that connects to the Op Amp draws 40mA, but the signal of the sensor is below 1mA. The sensor does not have any bypass capacitor (network).

Thank you.

enter image description here

EDIT: The answer I would accept is in CL's comment/link to TI's specs that were not listed in the datasheet I looked: "The power supply can be any voltage between the minimum and maximum supply voltage rating located in the table. Each VCC pin should have a good bypass capacitor to prevent power disturbance. For devices with a single supply, TI recommends a 0.1-μF capacitor. If there are multiple VCC pins, then TI recommends a 0.01-μF or 0.022-μF capacitor for each power pin. It is ok to parallel multiple bypass capacitors to reject different frequencies of noise. 0.1-μF and 1-μF capacitors are commonly used in parallel. The bypass capacitor should be installed as close to the power pin as possible for best results."

TommyS
  • 135
  • 10
  • Why not try it and see? Most folk would use one irrespective of whether there was a perceivable functional or performance benefit. If you are not fitting one because of cost then you might be on a slippery slope. – Andy aka May 30 '21 at 16:21
  • It depends from a lot of factors. For example, what is the distance to the nearest capacitor? – Justme May 30 '21 at 16:33
  • 2
    Possible duplicate of [Does the output bypass capacitor of an LM7805 double as a decoupling capacitor?](https://electronics.stackexchange.com/questions/319947/does-the-output-bypass-capacitor-of-an-lm7805-double-as-a-decoupling-capacitor) – CL. May 30 '21 at 16:37
  • @Andy aka It works, but would prefer to see if there is anything wrong fundamentally with the design instead of just testing it a few dozens times manually for the anticipated millions of cycles – TommyS May 30 '21 at 16:54
  • @Justme the distance is yet to be determined because the Schmitt trigger is an addition to the existing board and real estate is scarce. But if I add it without a bypass capacitor network, it should be within 10-15 mm from the capacitor network of the Op Amp that it would share. – TommyS May 30 '21 at 16:57
  • @CL thank you for the link. It answers two issues: short distance and duplicate capacitor value can be shared. In my case the 3 capacitors for the Op Amp cover pretty much most of the spectrum I believe. What confused me is that the 74xx14 datasheet does not mention any bypass capacitors. However, they mention not leaving floating the unused pins and rather bypass LOW or HIGH. Which may explain why I may get that few spikes at start up. – TommyS May 30 '21 at 17:07
  • A similar Schmitt trigger, the CD40106 mentions: "The power supply can be any voltage between the minimum and maximum supply voltage rating located in the Recommended Operating Conditions. The VCC terminal must have a good bypass capacitor to prevent power disturbance. A 0.1-μF capacitor is recommended to be used on the VCC terminal, and it must be placed as close as possible to the pin for best results" – TommyS May 30 '21 at 17:17
  • @TommyS Depends on which datasheet for which manufacturer and which logic family you read. For example, TI datasheet for SN74HC14 does include about a page for power supply and PCB layout suggestions. Which family type will you be using, HC, LS, etc? – Justme May 30 '21 at 18:08
  • @Justme The actual part is the SN74LVC2G14IDCKR https://www.ti.com/lit/ds/symlink/sn74lvc2g14-q1.pdf?HQS=dis-dk-null-digikeymode-dsf-pf-null-wwe&ts=1622333864701 – TommyS May 30 '21 at 19:04
  • 1
    That datasheet is not for that part number. The SN74LVC2G14 datasheet [mentions capacitors](https://www.ti.com/document-viewer/SN74LVC2G14/datasheet/power_supply_recommendations#SCES3514291). (TI has begun adding this to recent datasheet revisions; the SN74LVC2G14-Q1 datasheet is older.) – CL. May 30 '21 at 19:27
  • @CL Thank you, that gives the answer that I was looking for. The link you sent also addresses the multiple VCC pins, as I had the Schmitt trigger originally on a separate VCC pin before sharing it with the op amp's. – TommyS May 30 '21 at 19:41
  • You’ve asked two questions- the answer to the first depends on the amount of noise you can tolerate on VCC, which you haven’t specified. The answer to the second part may be down to C8, but without knowing what J11 is connected to it’s difficult to say for certain. – Frog May 30 '21 at 20:25
  • @Frog Thank you for the inputs. J11 is connected to a doppler sensor that is drawing 40mA from the 3.3V/500mA power. The schematics shows only the doppler's signal line going to the op amp's C8. That signal level is around 60mV-250uVrms A/C. Noise level 20uVrms at 100-1kHz. I measured 430uA. – TommyS May 30 '21 at 20:42
  • @Frog The noise I currently have upon start up of the MCU which has an LDO that provides the power for the op amp and the schmitt trigger and sensor is rail to rail spikes within the first 3-5 seconds. After it settles, then its minimal at about 10-20uV. Part of random spikes after the warm-up/settling time of that 3-5 seconds may be attributable to lack of high-pass (and possibly low-pass) filtering. The doppler may picking up noise from the environment. I guess I need to fix that by looking at the op amp first. – TommyS May 30 '21 at 20:55
  • 1
    @TommyS it struck me that C8 and (effectively) R1 have a time constant of 30ms, so if J11 has a significant DC offset then it could take many time constants before the op amp stops driving hard against one rail or the other. However, since R4 is only 10R there isn’t much scope to reduce C8 without limiting the low-frequency response, which I imagine would be unacceptable. If you can do anything to keep J11 as close to ground as possible then the startup will be quicker. – Frog May 31 '21 at 07:06
  • @Frog I am on wobbly grounds here with my knowledge, but I think I understand that I should bring the DC offset from the sensor as close to zero as possible (i.e. J11). R4 was selected previously where I needed little resistance to current from an inductive sensor (eddy-currents) before I adapted it to the current doppler sensor. The frequency envelope I want to detect for the entire circuit is between around 1kHz and 200kHz. More realistically, I am happy with between 1 Khz and 60 kHz. I only tested it so far up to 26kHz. – TommyS May 31 '21 at 17:21
  • 1
    As it stands the RC has a time constant of 30us so you’ll see significant attenuation at 1kHz. On the one hand you might want to make the time constant longer to improve the low frequency response but on the other hand that would make the startup time even longer. Minimising the offers between J11 and J12 will help the startup. – Frog May 31 '21 at 19:43
  • @Frog Thank you for the suggestions. I will adjust and measure the results to tune it further. – TommyS May 31 '21 at 20:55

0 Answers0