Multiple SPI in "parallel" ?

[yosh1426459906 0]

Hello everybody,

 

I'm totally new to microcontroller programming, so please excuse my question ;-) I am planning to build up a thermocouple reader for 4 thermoelements.

The stellaris datasheet states, that there are four SPIs available. I would like to use these, to connect four MAX31855 thermocouple ICs on seperate "lines", so that the four MAX31855 do not have to share a single channel. According to some reading in the net, a single temperature reading needs about 100ms of time, so I could make 10 readings a second. Is it possible to make 10 readings a second on all four MAX in parallel, so that I get 4 x 10 readings a second? I would like to use Energia for programming, but I'm not sure how to achive the "parallism" - my guess is, that if one reading needs 100ms, I could only get 2,5 reading per second per MAX31855 if I use four of them, since the temperature reading functions have to be called from inside the loop()-Function and the code examples I saw used delay(100)-calls!?

 

Kind regards,

yosh

[bluehash 1,581]

You could read them one after another on the same SPI port, but different select lines. They would be off by a few ms this way.

[yosh1426459906 0]

Hi bluehash,

 

thanks for your reply. The problem is not, that the readings of the four thermocouples could "be off by a few ms". The problem is, that I want 10 readings/sec on each thermocouple, so 40 readings total per second. So from my understanding this could only be done, if I use four seperate SPI ports - or am I wrong? The 100ms delay in the code examples confuses me ;-) Or do you think it would be sufficent to make the 4 "readTemperature"-calls and then wait the suggested 100ms?

 

Kind regards,

yosh

[bluehash 1,581]

I would do it this way:

- Setup all 4 sensors on one SPI port, with differetn chip select lines.

- Setup a timer to interrupt every 100ms.

- Once the timer interrupts, poll sensor 1, then 2, then 3 and then 4.

- done. Repeat when timer interrupts.

[hvontres 22]

Also, one other thing to keep in mind is that one of the configurations for the SPI pins confilicts with the pins used for the on-board RGB LED. So in order to use all four SPI ports, you would need to remove some 0-ohm resistors.

[yosh1426459906 0]

Thanks for the replies,

 

@bluehash: ok, I'll try it this way - I think I was on the false track regarding the 100ms. Your suggestion makes sense, since the 100ms are only needed for the MAX to "settle down", but a single reading of course will be shorter than this time. So after four polls all sensors have the mentioned 100ms until next reading.

 

@hvontres: Thanks for the hint, I will keep this in mind, e.g. when wiring a SPI-LCD.

 

Kind regards

yosh