So, built a 2nd 64 x 64 matrix using 16 x 16 panels.

fastled-support
gplus
(Mark Estes) #1

So, built a 2nd 64 x 64 matrix using 16 x 16 panels. I am finding this odd difference in color/ brightness which is unwanted.

These entire thing is powered by a pair of 5 V, 10 amp power supplied split right down the middle (left and right). I am running this on relatively low brightness so the voltage remains at or above 4.3 V even in the most power intensive patterns.

The power feed from each supply splits to power the upper and lower 4 panels on each side. So the power feed does not match the pattern.

The panels were from a stack that accumulated over time in the garage. At least 2 or 3 vendors on ebay or amazon so there could be differences in the chips/ board.

Questions:

Is the brightness/ color variation a vendor issue ?

Can anyone suggest a cure that does not involve me ordering new / sorting out the panels and rewiring.

(Marc Miller) #2

Wow that’s quite a large visual difference. I’m surprised it’s that varried.

(Mark Estes) #3

yep, that is why I am baffled. About to dig out the balance of my 16 x 16 panels and run a sort to see if I can find a difference.

(Ben Delarre) #4

Are you sure they don’t just have a different color order? Have you done a simple test with just pure rgb colors?

(JP Roy) #5

Did you actually check the 5V level on each of the 16 panels ?

If you are satisfied with the voltage difference between them being negligible and verified your sketch does not somehow send different brightness levels then I would assume you unfortunately got wildly different panels !

Now, apart from changing some 16X16 panels to match them all, the only suggestion would be to work out a separate global panel brightness level or maybe even a full color correction for each of your 16 panels to ‘tune’ your full display!!

(Mark Estes) #6

@Ben_Delarre i checked that, the color order is the same, when sending pure red, they are all red although with small differences in brightness. blue is all blue but the differences are more marked, same with green. suspect different batch of leds are involved. trying to change the color balance on select panels in code is my next step.

(Mark Estes) #7

@JP_Roy yep, checked the voltage. very close across every panel. This odd color/brightness mismatch even happens when I remove a bunch of panels from the load, what is left still has this pattern. despite higher voltages on the remaining. working on your last suggestion (adjust color balance by panel).

(Mark Estes) #8

made some progress with panel by panel color adjustment. seems like it will work with a bit more fiddling. thanks for the suggestions.

(chad steinglass) #9

@Mark_Estes sounds like you’re already making progress, but I’d say it’s almost definitely a vendor (or even batch within vendor) issue. I’ve found that when making big projects from strips, there are significant brightness differences occasionally between strips from different vendors, especially at low brightness. This difference tends to be less noticeable at higher brightness, but down at brightnesses in the 5-15 range, some strips are basically off while others are certainly noticeably on.

Only band aid I’ve found is exactly what you’re doing - once my project is done I run a couple test patterns and then write a correction function that brightens all of my dim pixels based on a XY map (they’re usually whole strips so can specify by X value) in order to make them more consistent.

I’ve noticed less of these issues with strips I’ve ordered in the last year, but in the last year I’ve also taken to ordering large batches of strips from AliExpress and marking all of the packs with an order date so I’m rarely mixing batches within a project, and have been more consistent in which vendors I order from, so that prob has a large impact.

(Marc MERLIN) #10

I think @JP_Roy may be right. This may just be a power problem. You need to re-inject power in each panel and even on 16x16 panels (or 8x32 which I use instead), I re-inject power in 3 places just to be safe.