Received my 1st test batch of the new SK9822 Pixel. It is as previously stated a direct drop in replacement for the APA102 pixel, with some minor differences. I can confirm its maximum stable data rate is 30Mhz. I ran 31Mhz and it worked fine, but 32Mhz make it go completely crazy (The APA102 could not go above 20Mhz). In standby mode the pixels also consume WAAAY more power than the APA102. I had a test strip of 72 pixels on 1 uninterpreted strip. Both ends had the same 5V 40A power supply attached. When no LEDs were ON, the strip drew 0.3A. However running a test animation drew the same amount as the APA102 would at its identical intensity levels. I have an English datasheet here http://www.elec-tron.org/wp-content/uploads/2016/04/SK9822-LED.pdf with thanks to @Jurgen_Skrotzky The driving system is exactly the same as APA102! (Tests were done using an RPI and using my own SPI driver )
@01001101_01010111 At the moment is pretty much imposible to source APA102 If you can get them, do it. If not, I think these will be the market winner. I got these from http://www.szledcolor.com/ get in touch with Alan -> admin@szlcoe.com and find out
@01001101_01010111 hard to say - someone did some digging a while back and determined that the apa102 had a 20khz PWM rate - the question is, do these? Or is it more like the 400hz that we’re used to elsewhere (which sucks for POV)
Any idea why there is an APA102 shortage?
Because everyone is buying them.
Ah! That explains the price increase I observed. Thanks!
@Daniel_Garcia How would one test this?
@Leon_Yuhanov
Hi Leon, I found what you say interesting, however I am a bit of a laymen. Thus, what is the interest point between 20mhz for the APA102 and 30mhz for the SK9822? I am just keen to know more to see if I should maybe invest in the SK9822…
thank you
@Tejkaran_Samra The Serial data transmition rate is important when you have a huge number of LEDs connected. For example if you have 2000+ LEDs and you want to render very smooth animations, a fast data rate allows for this. If you have 2000 LEDs each time you render a frame you transfer 2000*4+8=8008bytes per frame. The faster you dump the frame the smoother you can animate
@Leon_Yuhanov
This is very interesting! I am working my way to work with around 10000 LED’s on the APA102 strip, for which accurate timing is critical. Therefore, going by what you say, the SK9822 would be a better investment?
Are there any other differences between the two?
I’m going to google search this, but would be nice to know your thoughts too.
4mA is a ridiculous amount of quiescent current. Hopefully APA can ramp up their production, because for a wearable, 4mA/px is ridiculous.
EDIT: I noticed the datasheet lists 1mA/px “static power consumption” – what was your measurement precision for the “0.3A” figure, @Leon_Yuhanov ?
@Luminous_Elements I have a feeling my measuring device is very wrong. its one of those cheapo USB through things
@Tejkaran_Samra from my tests, there is 0 difference in driving the SK9822. Exactly the same
@Leon_Yuhanov thank you very much for sharing your tests with us. This really helps me. I’ll order some LEDs next week for my tests. But I’m really shocked about the power hungry side effect…
@Leon_Yuhanov , you mentioned that the 4mA draw was measured by simply hooking up the meter without sending any data to the pixels. Does that measurement change if you send “black” to the strip with FastLED.showColor(CRGB::Black); ?
Also, I’d like to have a poke at these myself; where did you source your SK9822 strip?
@Luminous_Elements When the strip i tested was powered ON with no data sent, the meter i used(one of those USB pass through things) displayed around 0.3A. When i sent black, there was no change. However when i sent data the power meter showed the same usage as the same data sent to the same length APA102 Strip.I think the cheapo usb power meter i have may be wrong
@Luminous_Elements i got them from http://www.szledcolor.com
Also - the chips are not exactly compatible with the APA102 - they follow slightly different rules for latching the data (e.g. when the led takes the data that has been written to it and starts showing it). I’ve just checked in a fix for this on master@HEAD, adding an SK9822 led type that will do the right thing for end frames.
@Leon_Yuhanov Ah, right. I had a similar experience using a $10 DMM to measure the draw from a 4-pixel strip. If you have a bench supply, use its current reading. If you don’t have a bench supply, get a bench supply.
Basically, get a bench supply. You’ll wonder how you ever got by without one.
@Luminous_Elements Would you like to gift me a bench supply 
I Currently use a cheap Chinese 40A brick that buzzes when its drawing too much current, That is how I know to turn it off 