I'm talking about the level shifter where you have a logic-level FET connected like so:
Source - Low voltage logic side
Drain - High voltage logic side
Gate - Low voltage supply
(where both high and low side are pulled up to their respective supply levels).
I was just thinking about this the other day, because I have a 3.3v board using the hardware SPI to control some APA102 LEDs, and I'd like to add another SPI device on the same SPI bus without screwing up the LEDs, which have no CS line.
If instead of tying the gate to the 3.3v supply, I could tie it to the CS line for the other device. Thus, when I assert the CS line, the gate is now at 0v, and the fets won't turn on, and I can use the other SPI device without messing up what's shown on the LEDs.
pwillard:
These are just signal lines... you can turn them on an off with an analog switch, Like CD4066
Of course I could - with a CD4066, I'd still need the level shifters, though. I could also use a level shifter IC with an enable pin. But I'd like to do it with fewer, cheaper parts if possible.
Also, many of those level shifting IC's aren't bidirectional, which is what's so cool about the fet level shifter - 3 parts (1 cheap fet, 2 resistors - probably 10-20 cents total), and you get a bidirectional levelshifter. Not that I need it in this case, though. (though, my proposed trick doesn't keep a low on the high side from pulling down the low voltage side )
123Splat:
or with a discrete FET, which, in essence, is what Azzy is proposing.
The point here, is that, unless I'm mistaken, I can use the same fets I'm using the shift the levels as an enable line.
don't the bi-di shifters hold high on both sides, until one side is pulled low by a signal source and then pulls the other side low? isn't that what you want for your level shift of CS ? Sounds workable to me...