spi problem breadboarding two msp430g2553.

Gautam Bhat said:

The clock line wire is slightly long. Is it catching interference?

It could be. Just a guess: have you shorted grounds of both boards?

Regards,

Peppe

I am programming the bread boarded msp with a launchpad. I was using a breadboard power supply but now I am using the power from the launchpad which I am programming it. I have not shorted the grounds of both the launchpads as power is the same for both the launchpads right? i.e. from the USB of my computer so common ground?

Right... but if the point of common ground is too far (e.g. you have two long cables from PC to launchpads) you could still have issues. I'd suggest a try with a short wire from launchpad ground to breadboard ground.

I have a pretty short wire from the launchpad pin to the breadboard ground now. Same problem.

On second thought, have you already configured the master SPI as such before connection? If not, both SCLK ends are inputs without pull-up/down, so it could be an issue.

Gautam Bhat said:

I have connected 2 seven segment displays and it shows characters coming on the lines and it goes wild. The clock line wire is slightly long. Is it catching interference?

On third thought this could not happen if master is holding CS# high.

Regards,

Peppe

Both are msp's. Why do I need a chip select?

Gautam Bhat said:

Both are msp's. Why do I need a chip select?

SPI is just serial bitstream transfer interface, SPI data or clock does not embed any information where word (byte?) starts and where it ends. Imagine what happens when at the very beginning (at power-up?) slave is one clock pulse off? How would slave get in sync with master?

You mean to say if I miss sensing the clock edge in the slave? Makes sense now. So to fix this I will setup gpio pins which when asserted will remove the SPI controller out of reset in the slave.

Even if I don't connect anything to the pins a receive interrupt gets triggered. Not sure if this is normal. Also tested in loopback (UCLISTEN) mode and I get the interrupts fine.

Gautam Bhat said:

Even if I don't connect anything to the pins

Define 'connect anything'. To do so you should hold CS# to Vdd on the slave.

Regards,

Peppe

I had left it floating. Also I have setup a GPIO to be a CS. Now if I don't start master SPI msp430 I get a lot of interrupts on the GPIO pin. Only after I program does the pin become stable. I enabled an internal pull up on the input line and now its stable. I have setup and LED to detect any incoming interrupts. If the pin is left floating the LED blinks rapidly.

*Update*: Got it working now partially! Learnt how bad floating pins can get. Somehow there is still a unwanted interrupt happening when powered up.