TM4C123 AIN0 problem

Richard Bland

Expert 2050 points

Other Parts Discussed in Thread: TM4C123BH6PGE, TM4C123GH6PGE

Hi,

I am using 22 analog inputs on a TM4C123BH6PGE. My prototype board works perfectly. My production board, running the same code, has a problem.

The prototype board has chip marked....

980 YF TM4C123B H6PGEI 35AVELW G4

The production board has chip marked....

980 YF TM4C123B H6PGEI 35AVENW G4

I can't find out how to read the silicon revision on the chip.

I am using PE3 for AIN0, this is driven directly by an opamp which is buffering a pot. It appears that the voltage presented to this pin affects all the other adc channels as if the input multiplexer is not working properly. This happens even if I do not sample this channel. Cutting the track to this pin stops the problem happening, so it seems it is not a board error.

I can't find a reference to this in any of the errata, but may be I missed something?

Can anybody help please?

Thanks, Richard

I put these two lines at the top of each file. Is this right?...

#define PART_TM4C123GH6PGE
#define TARGET_IS_BLIZZARD_RA1

My ADC initialisation code is....

ROM_SysCtlPeripheralReset(SYSCTL_PERIPH_GPIOE);
ROM_SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOE);

        ROM_SysCtlPeripheralEnable(SYSCTL_PERIPH_ADC0);
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_7);       //ch20
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_6);       //ch21
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_5);       //ch8
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_4);       //ch9
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_3);       //ch0
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_2);       //ch1
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_1);       //ch2
        ROM_GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_0);       //ch3

        ROM_ADCSequenceDisable(ADC0_BASE, 1);                                                   //disable sequence before we change it
        ROM_ADCSequenceConfigure(ADC0_BASE, 3, ADC_TRIGGER_PROCESSOR, 0);                       //select processor (software) trigger
        ROM_ADCSequenceStepConfigure(ADC0_BASE, 3, 0, 0 | ADC_CTL_IE | ADC_CTL_END);
        ROM_ADCIntClear(ADC0_BASE, 3);                                                           //clear the interrupt status flag
        ROM_ADCSequenceEnable(ADC0_BASE, 3);                                                   //enable sequence

My ADC sampling function is....

unsigned long getADC0(unsigned long channel) //channel is bit mask eg ADC_CTL_CH0
{
uint32_t ADC0_value[1];

    //ROM_ADCSequenceDisable(ADC0_BASE, 3);                 //disable sequence 3 before we change it; doesn't seem to be necessary
   ROM_ADCSequenceStepConfigure(ADC0_BASE, 3, 0, channel | ADC_CTL_IE | ADC_CTL_END);
    //ROM_ADCSequenceEnable(ADC0_BASE, 3);                 //enable

   ROM_ADCProcessorTrigger(ADC0_BASE, 3);               //trigger
    while(!ROM_ADCIntStatus(ADC0_BASE, 3, false))       //wait complete
    {
    }

   ROM_ADCIntClear(ADC0_BASE, 3);                       //clear the ADC interrupt flag
    ROM_ADCSequenceDataGet(ADC0_BASE, 3, ADC0_value);   //read ADC Value

    return ADC0_value[0];
}

over 10 years ago

0 cb1 over 10 years ago

Guru 47900 points

Is not your part a, TARGET_IS_BLIZZARD_RB1?

That said - somewhere I recall others running afoul when using PE3 as ADC... (forum search may illuminate...)

0 Richard Bland over 10 years ago in reply to cb1

Expert 2050 points

Thanks cb1-
I just changed all to TARGET_IS_BLIZZARD_RB1, made no difference.
Where do I find out about the TARGET_IS_????? I don't understand what it is for or why some examples have it etc

0 Amit Ashara over 10 years ago

TI__Guru**** 244440 points

Hello Richard,

The code seems to be sequencer 3 alone and not with a mix of channels.

Richard Bland said:
Cutting the track to this pin stops the problem happening, so it seems it is not a board error.

Where exactly is the cut made?

Regards

Amit

0 Roberto Romano over 10 years ago

Guru 27285 points

Hi Richard, when this happen processor is running fine or overheat too? Can you measure current when issue is in place?

The issue was CB1 talking about can be this, different chip version I am not able to reproduce but still monitoring for, try check for reference:

http://e2e.ti.com/support/microcontrollers/tiva_arm/f/908/t/361224#pi239031349=2

0 Richard Bland over 10 years ago in reply to Amit Ashara

Expert 2050 points

Hi Amit,

Thanks for looking.

It is sequencer 3 alone, but notice sampling function takes channel number as parameter.

The cut is made right close to pin12 to isolate any board fault.

Best regards

Richard

0 Richard Bland over 10 years ago in reply to Roberto Romano

Expert 2050 points

Hi Roberto,
I have it on the bench PSU and the current is monitored all the time. I can confirm no extra current draw is observed at any time at any setting of pots. The temperature is monitored continuously too - nothing getting hot.
I was hoping BLIZZARD_RB1 would sort it, but I don't properly understand this and how it works with the settings in CCS.
Thanks for the pointer though - could have been it!
Best regards
Richard

0 Amit Ashara over 10 years ago in reply to Richard Bland

TI__Guru**** 244440 points

Hello Richard,

Are the other channels connected to a similar OPAMP buffer as well? I think there is the ADC#13 side effect that you may be seeing.

Regards
Amit

0 Richard Bland over 10 years ago in reply to Amit Ashara

Expert 2050 points

Hi Amit,
Yes all channels are buffered by an identical opamp. What is this ADC#13 side effect?
Richard

0 Amit Ashara over 10 years ago in reply to Richard Bland

TI__Guru**** 244440 points

Hello Richard,

The ADC channel 0 is the idle state and causes the S/H capacitor to be charged. When you switch the channel the next channel has 4 clocks for sampling and may not discharge/charge the cap causing an erroneous reading. An experiment to confirm the same would be keep the AIN0 at 3.3V and take a reading of the other channels held at 0V to see the deviation and vice-versa.

Regards
Amit

0 Roberto Romano over 10 years ago

Guru 27285 points

Richard Bland said:

I am using 22 analog inputs on a TM4C123BH6PGE. My prototype board works perfectly. My production board, running the same code, has a problem.

Hi Richard, here is the errata document where you can find silicon errata: http://www.ti.com/lit/pdf/spmz849

I also found an incongruence, is this a mistype?

Richard Bland said:
#define PART_TM4C123GH6PGE

this part is tqf64 not tqfp144.

0 Richard Bland over 10 years ago in reply to Roberto Romano

Expert 2050 points

Hi Roberto, you're right the "G" was a mistype. I changed it to "B" in my project but it made no difference.

I'm not clear on the significance of the PART_ and the TARGET_IS_ statements. In CCS General project properties I have the Tiva part correctly selected and "Manage the project's target-configuration automatically" ticked, so I suspect this overrides the statements, but I'm not sure.

I have studied the silicon errata document you pointed me to (thank you), but as far as I can see, my particular problem is not listed.

Can anyone show me how to read the silicon revision (ie read register DID0), there doesn't appear to be a library function for this.

Thanks, Richard

0 Richard Bland over 10 years ago in reply to Richard Bland

Expert 2050 points

Hi All,
It turns out the board manufacturer put the wrong component on the board where the opamp buffers were supposed to be. Unfortunately the marking on the chip was very nearly the same as for the correct component. And surprisingly the system very nearly worked.
Hey ho - that's life. Thanks for all your help and suggestions though. This forum is a fantastic resource.
Best regards
Richard

0 Amit Ashara over 10 years ago in reply to Richard Bland

TI__Guru**** 244440 points

Hello Richard,

And very few get the information to the forum. Appreciate the feedback.

Regards
Amit

0 cb1 over 10 years ago in reply to Richard Bland

Guru 47900 points

One hopes that you used "quads" - even w/those you'd require SIX quad op-amps. (in support of > 20 ADC channels)

Arm-based microcontrollers

Arm-based microcontrollers forum

TM4C123 AIN0 problem