Yep 1Y is the output of the chip 1A is the input. we are connecting SPI clock of the Pi to LED Clock.
So whenever the pi pulls the CE pin low, the clock signal on 1A is forwarded to 1Y into the LED strip.
But I'm having issues too, that I can't get the NRF working. I think it's more of a software issue though. Haven't had time to tinker much lately though.
Looks good. I think the top is the side with the long strip printed on it. So yeah should be right.
Which library are you using for the NRF btw? I can't seem to get mine working, always crashes with TMRh20 RF24 library.
No resistors.
Yes:
GPIO10 MOSI will be connected to both WS2801 and NRF24
GPIO11 SCLK will be connected to both 74AHCT and NRF24
The raspberrypi will use the CE0 and CE1 pin to tell which one of the devices will listen.
I used the 74AHCT125D
It's SMD, my local shop only had those in SMD.
As penfold said though its important to use the T variant, so it will have TTO compatible threshold voltages.
The other conditions the chip needs to fulfill are 5V VCC so it can act as a level shifter, taking the 3.3V logic signals from the RaspberryPi and outputting 5V for the LED strip.
And an active low output enable pin because the Chip select pins are active low.
1 GND -> 25 (GND)
2 VCC -> 17 (3.3V)
3 CE -> 22 (GPIO 25)
4 CSN -> 24 (GPIO 8 SPI_CE0)
5 SCK -> 23 (GPIO 11)
6 MOSI -> 19 (GPIO 10)
7 MISO -> 21 (GPIO 9)
8 IRQ
Wiring for the WS2801:
1 GND -> GND
2 DI -> Raspberry PI GPIO 10 (SPI_MOSI)
3 CI -> 74AHCT 1Y (output1)
2 VCC -> 5V
Wiring for the 74AHCT:
1 1/OE (output enable not 1) -> GPIO7 SPI_CE1
2 1A (input 1) -> Raspberry PI GPIO 11 SPI_SCLK
3 1Y (output 1) -> WS2801 CI (Clock In)
7 GND -> GND
14 VCC -> 5V
In the Hyperion Config be sure to set spidev0.1 since the WS2801 will be connected to CE1.
Finally got around to it. And good news it seems to work. So it seems like the 3.3V were in fact the problem.
At least without the other SPI channel being used.
I'll report back once I get that NRF24 chip working for a final conclusive green light.
Thanks a lot for the suggestion with the 74HCT 
Thanks for the heads up. I'll go get one
So a 74AHC125 would do fine?
Yes I have.
Could it be the little extra voltage drop by R_DS(on)?
As the WS2801 is not really rated for operation at 3.3V. It worked without the Transistor, but that might have just been the little extra drop for it to not work anymore?
I've tried it and for now it doesn't work. The first LED ist erratically setting colors while the rest stays dark.
When I set it to clear in hypercon. It switches off the first led after a delay of several seconds. When I set a certain color it cycles through all sorts of colors. So it's obviously communicating. I don't have an oscilloscope to really look into the signal unfortunately.
I'm using a 1k pull down. Using the CE1 pin on the raspberry. The other spi device on CE0 is not wired up yet.
When setting spidev0.0 in the config, all LEDs stay dark. So that part seems to work.
I've set the baudrate to 122000. So a clock cycle should be around 8ms.
The datasheet of the transistor states a Turn-On Delay Time of 11ns and a 250ns Turn-Off Delay Time
Also a rise time of 32 ns and fall time of 210ns.
Any thoughts what might be the issue?
Isn't it more to dampen the ringing that might occur because of the resonator that's made up of the gate capacity and the wiring inductivity? Probably doesn't really matter in this case, you're right.
Thanks. I'll try this and post back with the results.
It's a WS2801.
After looking though the inventory of a local store I will probably use a IRLML6401 MOSFET
Would a gate resistor of 10Ohm be reasonable?
And you're right, it has to be a pull down resistor.
So all I need is a 10k pull up resistor to SCLK and a PNP transistor to disconnect the LED strip when CS* is pulled high like ?
Is it possible to add another SPI Device? I would like to add a NRF24 wireless transmitter which is a SPI device. The pi has two chip select pins which are not used for some reason.
