24v COB rgb 840 strip controled by Raspberry Pi

 fskrzycki You have a big problem 24+ volts and logic do not never mix.  I have one better than my buddy shabaz has AND you only need 1 part to use an optio-isolator anything in the 4N44 etc (there are tons of them on eBay,) will work just look at the collector voltage and current. ~~ C. Harrison

Update 9/3 I Fibed!   It has 2 parts, see the comment below. 

I like opto-isolators too, there are some fancy ones that can directly drive MOSFETs. In the end I only sketched one approach with transistors since unfortunately I didn't have time to provide additional circuit options : (

The transistors are less exotic I guess!, but really all of these approaches are a bit excessive for the use-case, because the Neopixel-like strips would eliminate all of the circuit (well, almost all - might still need to adapt the 3.3V to 5V logic levels).

 shabaz  Well, there are more ways to skin this cat.  level shifters cost $ while, your approach used more than a few parts, The optocouplers on eBay are about a half a buck each, buy 10 and save. but you will have to use a resistor in between the coupler and your Arduino's output pin (5v) the iso normally turns on at 1.3 volts so a quick trip to Mr Ohms or E= 1*R  so  R =( 5 -1.2)/20ma = 180Ω. But most likely get away with 220Ω ~~Cris H 

BTW I use this approach in my sim where the lamps are 28 volts ouch!  REMEMBER DO NOT USE A COMMON GROUND OR POWER SOURCE, or you will lose your galvanic isolation - this could ruin your week!!

the logic mosfets i ordered are rated to be maxed out@30v so i was not to concerned. my Pi is controlling the DC24v COB strip with  a GPIO setup for each color. Im using ChangeDutyCycle to control the color. pretty cool. The color wheel runs from 0-255 and CDC controls the current cycle of the mosfets with ratings 0-100%. I can generate any color by dividing 255 (100% power) into any color number than multipling that result by the number again. Currently the program will run any 1 color super bright for me (why i went with 24v) in an endless loop using the "while true" statement. The LEDs turn off fine when i break the loop. the program ends without error. 

the rub is after ai end the program. if powere is left connected to the strip, The LEDs turn themselves on to a yellow color.  interesting. i think the Red and Green MOSFETs are leaking.

when the project is complete. the python script will loop endlessly from startup and only break when a button is pushed to execute an "else" statement. which will be a sound wav file and some cycle color changes.

should i be concerned abiut the yellow? 

I had a negative experience with PC817 opto-isolators and Raspberry Pi so I tend to shy away from using them.

The Pi 3.3 output didn't drive the opto-isolator into saturation. I spent considerable time trying to resolve the issue and in the end used level-shifters to obtain consistency.

I would be the first to say it could have been my lack of knowledge but I have a box of  modules that I hope to use someday:)

I can imagine incompatibility is a common problem with so many modules which are really sub-circuits,

The modules stores always feel like an attempt to turn electronics into a fast food place, where it doesn't matter what burger goes in which bun, it will all kind of work out. 

But electronics is not like that, since the devil is in the detail, there's no standard interface on many of these modules. I couldn't find a datasheet, but it appears from some online diagrams that there is an external LED in series with the optocoupler's internal LED, so the total forward voltage drop is quite possibly higher than 3.3V logic levels. A quick fix (would not need level shifters would be to short out the external LED (e.g. desolder it and place a blob of solder). Then, it would be compatible with the outputs from a Pi.

I like the idea. of using an optocoupler between the MOSFET and the 24v strip. I think it will provide electrical isolation and prevent signal leakage between the Raspberry Pi and the LEDs. This will be an effective way to protect my Raspi from potential voltage spikes... Loke, when I plug in the power cord for the Strip... lol. This should minimize the risk of signal leakage.   Thanks 

 colporteur did you look at the 817's datasheet??

Please look at Fig  14. Notice the resistor (R_D )before the input? The board look like there is none. So refer to my response above and use the formula but change the 5v to 3v and away you go.. 

you can take a look at this article here opens new tab.

The board has the resistor R_D, indicated with red arrow. However, there is an additional LED in series (yellow arrow).

The board manufacturer saved $0.0001 by eliminating an additional resistor this way (they could have had the additional LED in parallel, with it's own separate additional resistor.

Some stuff is great. AliExpress is excellent for some non-critical things that would not be affordable otherwise. For sure some stuff needs to be avoided. Similarly, I cannot imagine not using PCB manufacturers in China, for non-critical prototypes. I decided to try prototype 3D printing last week. I've not received the items yet (so it could be a big [but low-cost] mistake! : ) I hope not) but for now it saved me having to buy and learn how to use a 3D printer.

In this case, I too agree this board is not great. The lack of clearance is worrying too, some people may try to use it with mains.

I believe it would also be cheaper in this case, to buy just the optocoupler parts and self-solder them either onto perfboard or a custom PCB.