3.7V Lipo booster noise

[wolf12]

I'm using the following tps61200 module. It makes lot of noise in ADC reads of the arduino. How to cancel the noise ? I'm working with a pulse sensor. I need to make it small. What are my other options?

https://www.sparkfun.com/products/11231?

Thank you.

 

[tgootee]

Make sure that you have proper decoupling capacitors everywhere in your circuit, i.e. a small capacitor connected very close to each power pin and then to ground, and also an electrolytic (>= 10 uF) from every power pin to ground. If any power pins try to pull transient currents through conductors, you will get spikes, just from that.

Also be sure that grounds that need to be quiet do not share any length of ground-return conductor with fast-changing transient or large currents.

If the tps supply already has spikes, maybe you could add a CLC pi filter.

 

[wolf12]

Make sure that you have proper decoupling capacitors everywhere in your circuit, i.e. a small capacitor connected very close to each power pin and then to ground, and also an electrolytic (>= 10 uF) from every power pin to ground. If any power pins try to pull transient currents through conductors, you will get spikes, just from that.

Also be sure that grounds that need to be quiet do not share any length of ground-return conductor with fast-changing transient or large currents.

If the tps supply already has spikes, maybe you could add a CLC pi filter.

When I'm using the PC USB supply or if I connect DC power pack to arduino, there is no noise. I think the noise is from tps module. I don't know about CLC or PI filters. Can you explain a bit?

Thank you.

 

[tgootee]

A CLC Pi filter is a Capacitor from power supply rail to ground, then an inductor in series with the power supply rail, then a capacitor from the power supply rail to ground. It makes a low-pass filter on the power supply rail, for current going in either direction. You can also just use an LC filter, with a series L and then a C to ground (i.e. a shunt C) on the downstream side of the L.

In practice, there will be an LC resonance, so a little resistance might be needed, either in series with the inductor or one or both of the capacitors.

If your maximum current is not too high, you will have a larger choice of inductors. Mouser.com has plenty to choose from, as do others.

You could probably try a 10 uH to 100 uH inductor, with something on the order of 1000 uF to ground after the inductor. But you will probably have to experiment.

If you know how an RC low-pass filter works, an LC is similar. They are both frequency-dependent voltage dividers.

If you don't work with filter equations every day, it's probably easiest to use a simulator like LT-Spice (free from linear.com), and run an AC Analysis.

 

[wolf12]

A CLC Pi filter is a Capacitor from power supply rail to ground, then an inductor in series with the power supply rail, then a capacitor from the power supply rail to ground. It makes a low-pass filter on the power supply rail, for current going in either direction. You can also just use an LC filter, with a series L and then a C to ground (i.e. a shunt C) on the downstream side of the L.

In practice, there will be an LC resonance, so a little resistance might be needed, either in series with the inductor or one or both of the capacitors.

If your maximum current is not too high, you will have a larger choice of inductors. Mouser.com has plenty to choose from, as do others.

You could probably try a 10 uH to 100 uH inductor, with something on the order of 1000 uF to ground after the inductor. But you will probably have to experiment.

If you know how an RC low-pass filter works, an LC is similar. They are both frequency-dependent voltage dividers.

If you don't work with filter equations every day, it's probably easiest to use a simulator like LT-Spice (free from linear.com), and run an AC Analysis.

For the CLC filter, what capacitor range of value can be added?

I'm making a single layer PCB for this. Can a ground plane help ?

Thank you.

 

[dick_freebird]

Some questions to chase down:

conducted or radiated coupling to the "victim"?

baseband or harmonic / ringing interference?

The character / source of the noise leads you to the
solution.

 

[tgootee]

Good point and questions.

Wolf12, A photo of your setup could be very helpful.

Can you describe the physical connections you have, for power and ground, in each of the three alternative cases you mentioned? i.e. USB, Arduino, and tps61200. For example, how long are they, are the power and ground conductors close together, are they in a single cable, is there more than one of either of them (power and ground), etc.