I'm running some addressable LEDs with a FeatherMO LoRa. I have to level shift the 3.3V output form the feather to 5V for the LEDs with a TI SN74AHCT125.
The oscope shows about a half a volt jitter up and down of the ground and high traces, both on the output of the Feather and the shifted signal going to the LEDs. Is this a problem? How can I quiet this down? The LEDs seem to work fine, but I'd like my design to be right.
Thanks for any help.
Oscope Jitters with Feather MO LoRa
Moderators: adafruit_support_bill, adafruit
Please be positive and constructive with your questions and comments.
- ansenseale
- Posts: 6
- Joined: Wed Aug 22, 2018 2:38 pm
Oscope Jitters with Feather MO LoRa
- Attachments
-
- oscope.jpg (364.16 KiB) Viewed 130 times
- adafruit_support_mike
- Posts: 67446
- Joined: Thu Feb 11, 2010 2:51 pm
Re: Oscope Jitters with Feather MO LoRa
That looks like a phantom signal in the wire between the probe’s tip and the ground clip's connection to the circuit.
Noise like that usually comes from fluorescent lights, or increasingly from dimmable LED lights. The ground clip’s wire makes a loop whose inductance picks up EMF from nearby sources, and the high impedance of the probe makes even a small induced current visible.
To get rid of the noise, reduce the size of the loop from the probe tip to GND. That’s what the little spring-thing that comes with the probe is for: it creates a very small loop from the signal to GND. You can also use a pair of wires twisted together between the circuit and the probe:
The one above is made from a few inches of phone wire soldered to the ends of some 30-gauge wire, wrapped around the probe's tip and GND ring, then twisted together. Making it took maybe five minutes.
Solder the free ends of the wires to points that are close together in your circuit. That will keep the loop area small while giving you a lot of freedom to choose connection points.
Noise like that usually comes from fluorescent lights, or increasingly from dimmable LED lights. The ground clip’s wire makes a loop whose inductance picks up EMF from nearby sources, and the high impedance of the probe makes even a small induced current visible.
To get rid of the noise, reduce the size of the loop from the probe tip to GND. That’s what the little spring-thing that comes with the probe is for: it creates a very small loop from the signal to GND. You can also use a pair of wires twisted together between the circuit and the probe:
The one above is made from a few inches of phone wire soldered to the ends of some 30-gauge wire, wrapped around the probe's tip and GND ring, then twisted together. Making it took maybe five minutes.
Solder the free ends of the wires to points that are close together in your circuit. That will keep the loop area small while giving you a lot of freedom to choose connection points.
- ansenseale
- Posts: 6
- Joined: Wed Aug 22, 2018 2:38 pm
Re: Oscope Jitters with Feather MO LoRa
Mike,
That is a great theory, but the wire gizmo didn't help the situation.
I also noticed that the timing changes with each trace. That has to be coming out of the Feather, right? I wouldn't think that induction would change the timing of a signal, only the level. I tried a different Feather with the same result.
Maybe the DC-DC converter is funky. But the effect shows, even when I power the Feather with a LiPoly battery. So it's not AC ripple.
That is a great theory, but the wire gizmo didn't help the situation.
I also noticed that the timing changes with each trace. That has to be coming out of the Feather, right? I wouldn't think that induction would change the timing of a signal, only the level. I tried a different Feather with the same result.
Maybe the DC-DC converter is funky. But the effect shows, even when I power the Feather with a LiPoly battery. So it's not AC ripple.
- Attachments
-
- IMG_2461 copy.jpg (350.74 KiB) Viewed 110 times
- adafruit_support_mike
- Posts: 67446
- Joined: Thu Feb 11, 2010 2:51 pm
Re: Oscope Jitters with Feather MO LoRa
Yeah, that looks like about 20ns of jitter.
That's close to one clock cycle of the microcontroller (1/48MHz=20.8ns), so my first guess would be a hitch-step in the delay code. They average out the mismatch between the CPU clock and a true microcsecond clock.
It's well within the +/-150ns tolerance for NeoPixel signals.
That's close to one clock cycle of the microcontroller (1/48MHz=20.8ns), so my first guess would be a hitch-step in the delay code. They average out the mismatch between the CPU clock and a true microcsecond clock.
It's well within the +/-150ns tolerance for NeoPixel signals.
- ansenseale
- Posts: 6
- Joined: Wed Aug 22, 2018 2:38 pm
Re: Oscope Jitters with Feather MO LoRa
Thanks for your help, Mike!
Since it all works OK, I'm not going to worry too much about the level shifting.
Since it all works OK, I'm not going to worry too much about the level shifting.
Please be positive and constructive with your questions and comments.