Reading multi-channels with ADC10

Thomas Kirchner said:

unsigned long temp_V = ((ADC10MEM/ 1024)*2.5);    //Ref 2.5 V Spannungsmessung
    unsigned long temp_C=(((2.5*((float)ADC10MEM + Toffset))/1023)-0.986)/0.00355;    //Temperatur berechnen

ADC can't do two parallel measurements at the same time. In IRQ routine you shall check which exactly channel is measured at the moment - voltage or temperature, then use ADC10MEM value accordingly. One more tip:  unsigned long voltage_mV = ((long)ADC10MEM * 2500) / 1024; Temperature conversion also can be done using fixed point math similar way.

From p557 of MSP430 User Guide slau144i:

"22.2.6.4 Repeat-Sequence-of-Channels Mode
A sequence of channels is sampled and converted repeatedly. The sequence begins with the channel
selected by INCHx and decrements to channel A0. Each ADC result is written to ADC10MEM. The
sequence ends after conversion of channel A0, and the next trigger signal re-starts the sequence.
Figure 22-8 shows the repeat-sequence-of-channels mode."

It doesn't look like you can just do Ch11-Ch10 because the sequence will continue to Ch0.

What I do know, but my  problem is  that, in my mesurments when it begins by channel 11 and its mesured for example 0x2EA and goes then down to channel A0 it should start when again with the  mesurment and if it didn't change it should again have the value 0x2EA but that isn't the case. Or if I do it with DTC and put the mesurments of all channels in an arry the values of temperature and voltage supply should be on the two first memory locations of the arry.

Or is my understanding of the DTC wrong?

If I'm correct while the ADC mesures the values will be first put in the ADC10MEM when transfered in the memory location I defined befor. And triggers an Interrupt when he has finished the mesurment of channel A0. And when starts anew.

" ADC10DTC1 = 0x0B;"

You're transferring 11 conversion results, but the A/D is reading 24. (Edit - Should have said 12) If I'm understanding the state diagram in Fig 22-10 (p560), the last balloon shows that when your count of 11 goes to 0, the ADC10IFG is set to 1, and also the DTC  transitions to the balloon where x=n (i.e. the count is again set to 11, but also the start address is set to the initial value causing the A/D which is still on its way to channel 0 to overwrite data)

why does the ADC read 24 conversion results? There are only 11 channel. Could you explain that to me.

My mistake. I should have said 12 channels (Ch 11 to Ch 0).

OK, but thats unfortunately not the mistake thats causes my problem.

Is my approach generally right?

I think if it runs correct the mesurrment of channel 11 must be in the arry memory location [0], that from channel 10 in [1] and so forth. And by the next mesurments it will place the values of the channel mesurments in the same array memory locations.

Thomas Kirchner said:

"I think if it runs correct the mesurrment of channel 11 must be in the arry memory location [0], that from channel 10 in [1] and so forth. And by the next mesurments it will place the values of the channel mesurments in the same array memory locations."

Edit: Removed my response from the quote;

The way I see it is that on the first go-around, the 12th A/D measurement (Ch 0) will overwrite the starting (first) ram location after the A/D has completed its first cycle count, then start with CH 11, CH 10, CH 9 ... Ch 2 going in the second, third, fourth, ... eleventh ram positions. Then on the third A/D cycle CH 1, CH0 will overwrite the first, second ram locations, because the remaining memory was filled when the A/D converted Ch 2.  And it will just ripple up with each cycle.

I set ADC10DTC1 = 12 in your program and I get 722 in res{0] and 439 in res[1] using MSP430g2553 and halting the debugger. Starting the debugger and halting again, I get res[0] = 722, res[1] = 441 which tells me that I'm overwriting the same positions with the same channels.

Joseph Raslavsky said:

"I think if it runs correct the mesurrment of channel 11 must be in the arry memory location [0], that from channel 10 in [1] and so forth. And by the next mesurments it will place the values of the channel mesurments in the same array memory locations."