Capacitor education question...

[Hippogriff]

I have purchased some Neopixel Sticks (https://www.adafruit.com/products/1426) which I've got to work fine, no different than the WS2812B strings or individual LEDs... I was looking over the instructions for using these things at https://learn.adafruit.com/adafruit-neopixel-uberguide/power and I came across the following:

"Before connecting a NeoPixel strip to ANY source of power, we very strongly recommend adding a large capacitor (1000 µF, 6.3V or higher) across the + and – terminals. This prevents the initial onrush of current from damaging the pixels."

Now, I've never done this before... I now power my circuits via USB and, I dunno, I just never saw the need and I've never had any problems - yet. I note it's only a recommendation, even if it's a strong one, but I do have some capacitors laying around so I grabbed a 100μF 16V one and plonked it in there.

My question is - the website says 1000μF 6.3V or higher... if I'm using a 100μF 16V what are the practicalities of that?

Will it help not one jot as it's not big enough?
Will it help a little bit but it will not be optimum?
Could it even be a negative addition to the circuit?

I'm suspecting that it'll help a little bit... but I have ordered some 1000μF 6.3V ones from eBay anyway.

My question is really about the choosing of an appropriate capacitor, I guess, and whether any capacitor (in this scenario) is better than no capacitor.

 

[MrAl]

Hi,

It will help a little but not as much as a 1000uf cap, and a 1000uf cap will not help as much as a 2000uf cap, etc. My guess is that they are banking on the power supply to have at least some impedance so that the capacitor loads it down temporarily so the device does not see the full voltage at first and this means less surge to the connected device.

One way to be sure is to use a current probe and oscilloscope, or a series resistor and scope to measure the current surge. There is the possibility that the power supply can not deliver much of a surge anyway so a smaller cap would work ok too, but a scope would verify this or tell you that you need an even bigger cap.

I wonder if they also make a 'panel' with a square matrix grid of LEDs too, with each LED addressable. It would take several of the regular strips to make up a square grid of RGB LEDs.

 

[cowboybob]

With any LED circuit, it is always wise to include a current limiting resistor, as the site you reference: (https://learn.adafruit.com/adafruit-neopixel-uberguide/best-practices) suggests. Its value depends entirely the max allowable current limit of the LED(s) and on your brightness desires.

The limiting resistor(s) pretty much assure (assuming a regulated voltage source, such as a USB) appropriate current levels. You can, of course, overload the max allowable current load on the supply, something you should avoid. The NeoPixel Stick series clearly notes the current levels needed per pixel. Other LEDs have datasheet limits stated.

A 1000uF cap across the power input to your LED string will, at power up, actually increase the current draw on the PS (by as much a 50%), thereby slightly reducing the initial voltage level seen by the LEDs. So, in that sense, it does have a beneficial effect. Downside is that if you're pulsing the LED(s), timing might prevent the LED(s) from firing.

 

[Hippogriff]

The resistor is for the data line, as opposed to the power. I find that interesting. I've never used one of those either. These addressable LEDs seem to be quite happy without a capacitor (across power) or resistor (on data)... and I've played with a lot of them... strip, circle, single and, now, stick.

 

[KeepItSimpleStupid]

USB should have inrush limits.

 

[Tony Stewart]

This recommendation only applies to poor quality unregulated power sources that have much higher voltage with no load but have large capacitance.

Disregard if this does not apply to you.

 

[Hippogriff]

Well, I'm using 240 volts AC into this... https://www.tlc-direct.co.uk/Products/AA3346U.html ...and then the USB socket from there.

 

[cowboybob]

Should be OK.

Noticed this admonition:

• Avoid connecting NeoPixels to a live circuit. If you simply must, always connect ground first, then +5V, then data. Disconnect in the reverse order.
• If powering the pixels with a separate supply, apply power to the pixels before applying power to the microcontroller.

 

[Tony Stewart]

"Before connecting a NeoPixel strip to ANY source of power, we very strongly recommend adding a large capacitor (1000 µF, 6.3V or higher) across the + and – terminals. This prevents the initial onrush of current from damaging the pixels."

A CV USB charger will have a big Cap at source.
This RGB LED has a programmable CC source.
Unless it has poor power on Reset interactions with an unstable transient, I don't see any problem with yours.