System Design with the TM4C123GH6PM 64-pin LQFP package (ADC)

May we note that this vendor supplies a great variety of highly accurate (external) ADCs - all of which (escape) the mixed signal limitations which plague (all vendor's) MCU-Based ADCs. 

Untold (still) is your (real) accuracy requirement.    And as poster Robert mentioned - (surprising as it may seem) - dependent upon your sensor type & class - high accuracy may prove of less value than you suspect.

Clear & detailed specifications are (usually) required to meet "high" (accuracy) demands...

External (i.e. proper) ADCs are unlikely to require similar limitations - nor "errata-driven" specification management...

Guidelines vendor's Amit mentioned - I believe - aim (only) at MCU's ADC - not at (unspecified) far higher performance, external ones...

Another key/critical "selling point" of "external ADCs" - most always it proves wise to, "Perform your conversion, "Near the source" not long inches (or even feet) away!   External ADCs may go - and fit - where an MCU (and its support) cannot.  

External ADCs enable you to determine (exactly) how much (real) accuracy your design requires.   In many cases - the exact footprint will accommodate ADCs of different conversion levels - enabling the optimal "ADC fit to your (unique) requirement."

Hi cb1_mobile,

I am looking at errata sheet Tiva C Series TM4C123x Microcontrollers Silicon Revisions 6 and 7 Errata (Rev. E) on page 63.

Thank you for your contributions, are very useful for me.

Regards,

Fernando.

Hi Fernando,

May I note that one hunts Tigers in India, Nile Crocks in Africa - and "accurate" ADCs anywhere but w/in "Mixed Signal" MCU.   (especially one plagued by errata...)

Hi Amit,

Amit Ashara said:

2. Using REF5033 to have enough guardband on voltage spikes and droop.

I find no reference to the REF5033, you can provide me a link to the datasheet of this component?

Thanks!

Regards,

Fernando.

Fernando Liozzi said:

I need accuracy because accuracy is limited to 12 bits.

Actually the precision is 12 bits.

The accuracy is limited by other factors to somewhat less.

I will note that getting 12 bits of accuracy is really difficult. It means everything in your system has to be better than 0.05% accurate. The total of the inaccuracies of the sensor, the conditioning circuit, the reference and the A/D must be < 0.05%. This is a non-trivial task. With some classes of sensors it's likely impossible.

Just on the signal conditioning consider the requirements on resistor accuracy that implies.

For it to be meaningful in most cases, it also means your noise has to be down at that level as well.

You haven't shared your application but my first step would be to question the need for this kind of accuracy. It's not often needed.

What is the sensor you are using?

Robert

Hi,
I am Sorry!, jaja, I wanted to write the sentence: "I need accuracy because the precision is limited to 12 bits." :P
I use sensors whose recommended power is 3V0, such as MP3H6115A (pressure sensor), MP3V5050. I am also using temperature sensors fed 3V0, such as, LMT87. For infotainment applications.

I just try to do my designs following some guidelines and suggestions, and also learning in the process. I think everything worthwhile to be done, has to be well done and consciousness. If I need or not so accurately, I think is not bad looking for her. In the process I learn and get advice from people who know more than me. As in the case of this forum, you help me to improve.

Regards,
Fernando.

Hello Amit,
The final application is a small system for variable sensing as tire pressure, temperature and humidity inside and outside the vehicle.
And I wanted to take advantage the internal ADC of microcontroller and with the addition of a referrer, have a better accuracy in measurements. I have the ads8326, the ads1213, the ads1018 and ads1204, but are intended to be tested in other applications.
I just want to have a little more accuracy with small aggregates. But without losing robustness in the design. The microcontroller can not be reset. If any ADC that can be useful for this application, I accept suggestions.

Regards,
Fernando.

Fernando Liozzi said:

If I need or not so accurately, I think is not bad looking for her. 

Accuracy costs.

Now, good design rules help. And the baseline accuracy of components is better than it used to be so the starting point is better than it used to be decades ago. Still getting even 10 bits takes hard work.

Now, to double check the sensors

Fernando Liozzi said:

I use sensors whose recommended power is 3V0, such as MP3H6115A (pressure sensor), MP3V5050. I am also using temperature sensors fed 3V0, such as, LMT87. For infotainment applications.

Infotainment? I doubt the accuracy requirements are high.  Cost is likely to be important.

So checking data sheets (I had  been looking at a couple of these a few months ago for a colleague as it chances)

MP3H6115A +/- 1.5% accuracy

MP3V5050 +/2.5% accuracy

LMT87 +/-2.7C or +/-1.4%

So the sensor accuracy for all of these is around 6 bits, although the MP3V5050 is getting closer to 5 bits. They do have more accurate sub-ranges but that only manages 7 bits for the temperature sensor.

This is why I questioned the sensor. Most sensors are less than 8 bits and as illustrated by these often considerably less. And note these are the top line datasheet numbers, detailed examination may find achieving this requires considerable effort although sometimes it's easy.

Using high accuracy A/Ds is a pointless expense and complication with these sensors.

If you actually need higher accuracy measurements of these values then first you need to invest in finding higher accuracy sensors and they are not going to be inexpensive if you need 12 bits of accuracy.

You could use a reference to calibrate your A/D periodically so you don't need to rely on a high accuracy power supply.

Robert

Hello Robert, hello cb1_mobile,

Thank you so much for the recommendations and observations. I will lend greater attention to these considerations and will have particular care in handling voltage in VDDA. I will follow your advice cb1_mobile =)

I really appreciate your advice, I always try to do things well, and often sin through ignorance.

Regars,

Fernando.

Robert gives unusually good advice: "Attend the nearest Analog Seminar" when provided by "Analog Leaders.""  

You can reduce (but not eliminate) such, "sins of ignorance" by a serious, regular read/review of the key/critical data available from such "leaders:"

• this vendor
• Analog Devices
• Linear Technology

(to name a few - and your attendance at one of their seminars will enable you to meet (others) - some "much like you" - and others who "dwell" w/in hallowed analog halls...)