These are chat archives for Makuna/NeoPixelBus

5th
Nov 2017
IVData
@IVData
Nov 05 2017 02:01
Hi again. So with the Uno (5v gpio) and the delay I get the same problem =\
Michael Miller
@Makuna
Nov 05 2017 04:18
Just for clarity, you are sure they are rgbw pixels?
IVData
@IVData
Nov 05 2017 04:18
yes i am
Michael Miller
@Makuna
Nov 05 2017 05:13
The strange thing here then is, my code for Uno is Adafruits code. And you stated Adafruits works?
p.s. Adafruits esp8266 is my old bitbang code (we share).
IVData
@IVData
Nov 05 2017 05:13
Yep that's right. I can check again
IVData
@IVData
Nov 05 2017 05:22
fuck, nope. the adafruit library has the same issue. I think I tested some different colours with it
Some colors work better than others
Michael Miller
@Makuna
Nov 05 2017 05:22
What is your power supply rated at?
How many pixels again?
IVData
@IVData
Nov 05 2017 05:23
300 pixels, 5v 10A
Michael Miller
@Makuna
Nov 05 2017 05:23
try something, change sketch to 32 pixels, and see if that helps
(ps. did I calculate this right, 300 pixel could draw 24 amps if all RGbW are set to 255
IVData
@IVData
Nov 05 2017 05:27
0.o really?
Michael Miller
@Makuna
Nov 05 2017 05:27
RGB at full white is 60ma, rgbw at full white is 80ma. times 300.
this is the reason to try 32 pixels. If it works, try 64, 128, etc.
10A should be able to run 125, maybe a few more
IVData
@IVData
Nov 05 2017 05:32
Damn yeah you're right. Although the white LED is off for all of mine atm
32 looks the same, because the gradient is very gradual. I'll try 100
Michael Miller
@Makuna
Nov 05 2017 05:34
its not unusual for people to run just rgb part or just the w part. I was even thinking of having smart RgbwColor that would just set the W channel if R=G=B; but there was never must demand.
You might be able to get upto 200 pixels before your power supply becomes an issue.
If you only are using RGB
IVData
@IVData
Nov 05 2017 05:36
Na, 128 looks exactly the same as 300
Michael Miller
@Makuna
Nov 05 2017 05:36
but 32 was ok?
IVData
@IVData
Nov 05 2017 05:36
The difference in color over 32 is basically not seeable
Michael Miller
@Makuna
Nov 05 2017 05:36
65, 96?
IVData
@IVData
Nov 05 2017 05:37
ill try.
The thing is, I had the same issue with colours that were almost entirely one-channel
but not as obvious
Michael Miller
@Makuna
Nov 05 2017 05:40
Something you might try, inject the power in the middle of the strip. If the strips power wires (ribbon on the flexible strip) is not thick enough, then it can't source the amps across that distance; by putting the power in the middle, it now runs along less pixels (middle of 128 would be the first thing to try if you could).
IVData
@IVData
Nov 05 2017 05:40
This is a waterproof strip so that's hard =\
and it's in my garden..
Michael Miller
@Makuna
Nov 05 2017 05:40
hehe, is it raining also ;-)
IVData
@IVData
Nov 05 2017 05:41
not riiight no
w
Michael Miller
@Makuna
Nov 05 2017 05:43
sometimes long installations of pixels, they inject power along the strip from different power supplies rather than one large power supply.
This test is all about trying to figure out why even 128 doesn't work (which should if your power supply is truly rated for 10A)
IVData
@IVData
Nov 05 2017 05:43
I turned the red channel off, still using 128 leds, and it looks like it's working now
Which is weird, because the issue was ONLY the red channel was working at the end of the strip
Michael Miller
@Makuna
Nov 05 2017 05:43
Oh, one other thing, did you read the Adafruit connection notes, about putting a big capacitor at the power supply?
IVData
@IVData
Nov 05 2017 05:44
I saw that, but i suspect my supply has a large cap at its output
also a resistor on the arduino output? I haven't tried that
Michael Miller
@Makuna
Nov 05 2017 05:45
the resistor isn't as required anymore (earlier chips were more sensitive than the newer generations). But you could try it.
IVData
@IVData
Nov 05 2017 05:47
big cap it is
Michael Miller
@Makuna
Nov 05 2017 05:49
what gauge wire are you using between your power supply and the strip? And how long is it?
IVData
@IVData
Nov 05 2017 05:49
its about 2-3m long and... no idea what guage =3
22awg
Michael Miller
@Makuna
Nov 05 2017 05:50
is the wire thick like an AC power cord, or lamp cord?
that's too thin
that might explain the issue with 128 pixels.
IVData
@IVData
Nov 05 2017 05:52
Hmm. Alright shit. Ill bring the strip inside and test with no distance to the PSU
BTW the psu voltage is 5.03v so i dont think it's that that's struggling
Michael Miller
@Makuna
Nov 05 2017 05:53
it really shouldn't struggle at only 128 pixels being lit up.
You should be using at least 16g wire, I really would recommend 14g wire (look for low voltage landscaping lighting wire).
Lowes/Homedepot carry it.
let me know what happens next ;-)
IVData
@IVData
Nov 05 2017 05:56
We dont have those stores here =) but I have some silicone RC wire that might do
Michael Miller
@Makuna
Nov 05 2017 05:57
yeah, that stuff works really well (but expensive).
Where is "here" btw?
IVData
@IVData
Nov 05 2017 06:02
New Zealand
Ok, I removed /most/ of the wire length, and its better. So i think that's the issue. Still 0.5v drop between the psu and the strip though
Michael Miller
@Makuna
Nov 05 2017 06:05
Most non-switching power supplies will have some drop when under load (they often read half a volt above spec if not more). If its a switching power supply and you are seeing drop; then you might suspect the power supply.
p.s. I like using old PC power supplies, switching, high amp, and normally they would have gotten thrown away.
IVData
@IVData
Nov 05 2017 06:06
Na it's .5v over the small amount of 22awg i have still in the circuit.
I've got 16AWG speaker wire, and otherwise I dont have anything long enough
So ill have to buy some
IVData
@IVData
Nov 05 2017 07:15
Thanks for your help @Makuna