Ran across an individual selling LED panels at the Denver Maker Faire. When I pointed out the WS2803 drivers he was using are slow compared to the LPD8806, he started shaking his head, saying “not the 2803, they are the fastest” … I’m sorry, but I have a strict policy against people who are close minded: I walk away. Was not worth my time and energy to talk to. If I did, I would’ve had some choice phrases for his 'I know best." attitude …
derpity derp derp
It looks like he’s right. The WS2803 has a max clock frequency of 25Mhz - http://www.jarzebski.pl/datasheets/WS2803.pdf - and the LPD8806 has a max clock frequency of 20Mhz - https://solarbotics.com/download.php?file=1889 - (and I need to go update the library chipset defs 
For that matter, so does the WS2801, 25MHz input clock. Their refresh rate still sucks balls though.
Unfortunately - not all WS2801’s are 25Mhz data rate. There’s a bunch of WS2801’s out there that are a 1Mhz data rate 
It looks like world-semi upped the WS2801 data rate quietly at some point. So - the question is - do I bump the WS2801 to 25Mhz in the library - and tell people who are using the WS2801 and seeing data problems to bump back down to 1Mhz?
Though, the datasheets also say:
The transmit the data over long distance by cascading, the WS2801 integrates push-pull output stage with strong driving capability which enables the data and clock can be transmitted up to 6 meters at 2MHz clock frequency.
So, I wonder if this means that cable length affects the max datarate of the chip?
(It’s been a couple years since I looked at the WS2801 chipset datasheets - Also, I hate that these specs can fucking change - annoyingly, neither the WS2801/WS2803 datahseets document refresh rate)
Ah - there’s a 1.2Mhz oscillator on the chip, the question is does that mean a ~4khz refresh rate, or is it 4khz/num output channels?
Cable length does very much affect the datarate, unfortunately. I would not be surprised if that refresh rate is not per channel …