SPI/MMC signal via <6.5 Ohm path

[permal]

Hi,

For various reason I need to connect my MCU and SD Card via both SPI and the MMC interface, though only one at a time. To achieve this I've opted for a MX4948 to switch between the two, but I'm concerned that the extra resistance of <6.5 Ohms of the MX4948 will interfere with the signals.

Am I right to be concerned, or can this work without issues?

[rstofer]

When signals leave FPGA boards, it is fairly common to have 330 Ohm series resistors to protect the device.  This sometimes shows up in uC boards as well.

So, I wouldn't spend the least bit of time worrying about a 6.5 Ohm resistance.  If you want to do the math, find out what the input current is on the SPI/MMC devices and compute the voltage drop across the resistance.  Then compare the result with the switching threshold voltage.  You can do all this and still come up with the fact that 6.5 Ohms is insignificant.

What I don't understand is how the turn-on and turn-off times can be so slow (on the order of 800 ns each) and the device still has a -3dB frequency of 300 MHz.  I'd need to think about that.  Even so, it might not matter.

[permal]

Thanks!

What I don't understand is how the turn-on and turn-off times can be so slow (on the order of 800 ns each) and the device still has a -3dB frequency of 300 MHz.  I'd need to think about that.  Even so, it might not matter.

Lets see if I undertand that. You're saying that the 300MHz is unexpectedly high for a -3db signal reduction? (I need to watch Dave's video about signals again)

[rstofer]

I think what I am saying is that the total of turn-on plus turn-off is about 1.6 us.  Given 0 ns actual on-time or off-time, the maximum frequency would be 625 kHz.  That was my original thought as I just glanced at the datasheet.

The thing is, I also think I am wrong!  I suspect the turn-on/turn-off times apply to switching sources and this makes sense.  It generally wouldn't matter much how long it took to select devices.  Then the 300 MHz frequency also makes sense because, once a source is selected, the actual signal can be quite fast.

I like the second explanation.  That and I should really read the datasheet in detail to understand what they're getting at.  I think you should disregard my original comments re: speed and move on.  It's going to work fine!