Dotstar power/control circuit

I have a Dotstar strip one it's way to me, and I'm just trying to tie a circuit together so I can control it. The power regulator in the diagram is actually an Adafruit 5v 4a switching supply (also on it's way!), and the RGB LED is the dotstar strip.

Also the switch at the top of the diagram (between +5 and the live rail) is actually a board jumper, so I can disconnect the controller from the positive rail to make sure I get no voltage leak into my computer while I am writing/testing.

Does the controller need a ceramic capacitor between +5 and G? If so, which side of the jumper should it go? Any other suggestions or advice? I'm trying to get it all ready in my head for when the strip arrives!

Many thanks in advance!

That circuit is bad. When the jumper is removed the Arduino is unpowered but you are applying a voltage to the unpowered Arduino through the pots. This is a bad thing and could cause latchup.
Move the power to the pots to the other side of the jumper.

Wokstation:
..., so I can disconnect the controller from the positive rail to make sure I get no voltage leak into my computer while I am writing/testing.

Not needed. The Nano has a backflow protection diode between the 5volt pin and the USB supply.
D1 on the schematic diagram.
Leo..

Grumpy_Mike:
That circuit is bad. When the jumper is removed the Arduino is unpowered but you are applying a voltage to the unpowered Arduino through the pots. This is a bad thing and could cause latchup.
Move the power to the pots to the other side of the jumper.

Awesome, that's exactly the kind of thing dozy old me misses. I had kept in my head I needed to do that, but managed to miss it.

Another reason for the jumper is to prevent the controller from trying to power a 30-led Dotstar strip from the USB!

Wawa:
Not needed. The Nano has a backflow protection diode between the 5volt pin and the USB supply.
D1 on the schematic diagram.
Leo..

Very good to know, thanks! I may include it anyway for the "stop it trying to power the strip" reason above. I think 1.8a peak may just exceed the power regulator capacity, and certainly that of my USB ports!

Some Karma all round, methinks! Thank you both!

I'm guessing the ceramic capacitor isn't really needed...? (I suppose I can always add it later if it proves neccesary)

I'm guessing the ceramic capacitor isn't really needed.

If that regulator is in fact a module then it should be on the board already. However, if you are going to supply a few amps of LED current the I would suggest a large capacitor, something like a 1000uF across the 5V supplying the LED strip.

Grumpy_Mike:
If that regulator is in fact a module then it should be on the board already. However, if you are going to supply a few amps of LED current the I would suggest a large capacitor, something like a 1000uF across the 5V supplying the LED strip.

The regulator in the circuit diagram is a 5v 4a swiching power supply, mains in one end, 2.1mm jack the other (couldn't find a suitable object in Fritzing, probably overlooked one).

I take it you refer to a capacitor like the one mentioned in the Neopixels uberguide? The Dotstar guide said they didn't think it was needed, but could be included anyway just in case... but they don't mention the orientation of the capacitor (and the capacitor in the images is unclear which way round it is).

Would it be anode to + and cathode to ground? (I only have electrolitic ones that big, all my ceramic ones are tiny) I know, these are really basic questions, but I'm learning as I go!

Your circuit also burns out the red LED immediately on switch on. You don't mean that!

MarkT:
Your circuit also burns out the red LED immediately on switch on. You don't mean that!

I'm confused. The RGB LED is just there to represent the four ports of the dotstar strip, it's not actually the setup. +5 will go into vin, g to ground, the pins going to data and data clock. Sorry if the diagram confused you! (Unless I missed something and pixel one of the strip will be blown?)

Ok I've redone the diagram (found a jumper, it's called a header pin in Fritzing), and moved the pots power to the same side of the jumper as the controller, and I've also added a capacitor (diagram says the wrong valie though).

I've also edited the RGB LED object to make it clearer what the connections are.

Hopefully this circuit is better!

Hopefully this circuit is better!

Sorry no.

Fritzing is a very bad choice to draw your schematics, especially for a beginner. You can't find the right parts and you just end up substituting some other parts and the result is a mess that no one can understand.

Imagine trying to write a story about cars, but every time you wanted to put the word car the word banana was substituted. Then "road" got changed to "sewing machine". The results would be rather incomprehensible, rather like your so called circuit diagrams.

Please simply draw them on a piece of paper and photograph the results, then post it.

Grumpy_Mike:
Please simply draw them on a piece of paper and photograph the results, then post it.

Third try lucky :wink:
Paint is my pen & paper... new diagram attached.

And thanks for your patience!

:slight_smile:
Much better, now it can be read.
Yes that capacitor looks fine and is the right way round.

Not very many people know what those two 510R resistors are doing. It is a bit advanced and I would recommend you not worrying about it. But just to say it has nothing to do with limiting the current to the LEDs.

I think a story about bananas driving down the old sewing machine is rather more
interesting than something about cars, but it won't get you your driving licence that's
for sure!

Grumpy_Mike:
:slight_smile:
Much better, now it can be read.
Yes that capacitor looks fine and is the right way round.

Cheers - it's a curious omission from the "uberguide"... or I just wasn't looking closely enough!

Not very many people know what those two 510R resistors are doing. It is a bit advanced and I would recommend you not worrying about it. But just to say it has nothing to do with limiting the current to the LEDs.

I did a lot of reading up before making the choice between neopixels and dotstars, and whether those two resistors were even needed... then I found this which seems to suggest it's about smoothing off the waveform (or something like that).

But as the uberguide said, if unsure, add a resistor anyway - so I did :slight_smile:

Unfortunately I don't know what connectors the strip will come with (Adafruit say it varies, I can't tell yet because the postman has it), but I've found my local chandlery stocks wire rated for 8a at 24v, it's just a bit... thick. Might have some issues soldering it on... so to my point;

How thick a wire would you suggest for a 1.8a peak current bridging just a few inches (just a little longer than a perfboard, in fact)?

And again, thanks for helping!

@MarkT Sounds like a cross between Noel Fielding and Cbeebies. Terrifying!

then I found this which seems to suggest it's about smoothing off the waveform (or something like that).

Interesting, they were attacking the symptoms of the problem without knowing the actual cause. What causes this ringing is the reflections at the impedance discontinuities on the signal. There is something called "time domain reflectometry" which uses impedance discontinuities to detect things like faults on a line, telling engineers where to dig the road up. What those resistors are doing is to absorb reflections and give a better impedance match on the line.

it's a curious omission from the "uberguide"

Well not really these things are not written by engineers and it is rather dependent on exactly what power supply you have, but the extra cap is certainly a good idea.

my local chandlery stocks wire rated for 8a at 24v,

The idea of wire only being rated for 24V is quite comical. The thought that you can get insulation breakdown at 30V is absurd. As a rough guide see this:-

Grumpy_Mike:
Interesting, they were attacking the symptoms of the problem without knowing the actual cause. What causes this ringing is the reflections at the impedance discontinuities on the signal. There is something called "time domain reflectometry" which uses impedance discontinuities to detect things like faults on a line, telling engineers where to dig the road up. What those resistors are doing is to absorb reflections and give a better impedance match on the line.

Though I don't quite understand the principles there, the explaination makes sense. Cheers :slight_smile:

The idea of wire only being rated for 24V is quite comical. The thought that you can get insulation breakdown at 30V is absurd. As a rough guide see this:-
American Wire Gauge Chart and AWG Electrical Current Load Limits table with ampacities, wire sizes, skin depth frequencies and wire breaking strength

I'd misheard the fellow, or got it conflicted with "gauge 22" (can't even remember why that was mentioned, tbh), either are possible - I have some cognitive issues which means it can take a few tries to understand something. That's probably what's happened there. Might also be why I'm having trouble making heads or tails from that table you've linked to... I'll keep trying. Went back to the chandlers today, grabbed some of the cable - it just said 8a, no voltage. I also hit the DIY shop for some 2a 2-core. Just got to wait and see what connections I'll be dealing with now.

And again, thanks!

The Dotstar strip arrived yesterday, and the connections were much smaller than I anticipated... beyond my soldering capacity, so I took them to a local PC shop where they soldered a molex connection onto it for me with AWG 18 wire.

Today I built & tested that circuit, minus the potentiometers and button, to run strandTest on. Plugged in the power pack, flicked the switch...

...held my breath...

...And it worked! Wooo! I've been stressing over that since the weekend, heh! Only issue is, instead of showing red first, it showed green first, so I need to tweak the order of the colour declaration for the strip.

Thanks again for your help!

Yes strips go with green first.

..And it worked! Wooo!

Like this:-

Grumpy_Mike:
Yes strips go with green first.

Apparently not always - the strandTest sketch definitely starts it's run with red, but the strip started with green.

Another production batch variation is the order in which color bytes are sent. Your own code can always use red, green, blue order when specifying colors, but the library needs to know the LED color sequence so it can adapt…you can see this as an optional last parameter to the constructor (change 'BRG' to 'GRB' or whatever data order your LEDs require):

So it's just a difference in the order the chips take it in, easy fix :slight_smile:

Like this:-

http://www.youtube.com/watch?v=9saYHg_JmWE

Almost! Though my first (finished) project was this: https://youtu.be/j4PkkIVC9Lg (For halloween!)