Only getting negative values in ADS1231

Hi Wesley,

If you're getting 7FFFFFh with your AAA battery positive polarity, and 800000h with a negative polarity it would seem that you're capable of getting positive and negative voltage swings on your data converter at home, the trouble is it doesn't show much in the way of resolution. I think the nominal voltage of a AAA battery is around 1.5 V, assuming VREFN is GND with your AA supplying VREFP should give your a reference around 3.0 V so you're railing the converter pretty hard since full scale for this converter is 0.5Vref/128.

If your VREFP connection was disconnected during testing it's hard to say what you could expect to see. You would have had a free floating reference voltage for your ADC which is just going to produce unpredictable, unreliable results. If you couldn't tell any difference in your data when your reference was connected or wasn't connected, that would seem to indicate you have a very noisy system. Is this on a PCB? Milled Copper Board? What are you using do supply your reference voltage at the lab?

Thanks Kevin for the reply.

The ADC is placed on a PCB bought from Radioshack and I used a DC power supply outputting 3.3V(verified with a DMM) to provide my reference voltage at the labs. I use the 3.3V to power my AVDD and DVDD as well at the labs. I was planning on going to the labs today and feeding a square wave into my ADC to check the values I get. Is there anything I can do to limit the noise?

Can you post a schematic with notes of what inputs you're applying and from what source at the lab? That will help understand precisely what you've got built up now. When you say PCB from Radioshack do you mean a breadboard? Perfboard? Or one of those general purpose SOIC boards?

Wesley,

In addition to what we've already offered you, a coworker and I noticed some other observations that you may find useful. From looking at the code snippet you posted, you're giving the device 32 SPI clock cycles. When you issue a fourth cycle of 8 clocks to toggle DRDY, the data you collect seems to get appended to your "result data" numbers in the attachment you posted (noise.png). What you've called "packet 3 data" should not be included in your resulting data.

Additionally, I'm not exactly sure what will happen if you continue issuing that many extra clock cycles, maybe Bob has some more experience with this device in particular. Once you've issued the 25th clock cycle, you still issue the device another 7 clocks afterwards. This is mildly concerning because your 25th SPI clock should move DRDY/DOUT high, so the data you store in "packet 3 data" should read FF each time. The data in "packet 3 data" reads 255 most times, but I see a few 127, 63, 31, etc. that have me a little worried. Maybe go back and check your SPI polarity/phase or issue only one more clock cycle through a GPIO or something. Again, hopefully Bob can give us some insight on what happens if you issue more clocks, it could be a moot point.

Wesley,

Just to fill in the blanks for Kevin and whoever else might read this post, due to the construction of the MSP430 peripheral, a fourth series of clocks are needed to force the DOUT/DRDY pin high when polling the DRDY/DOUT pin for new data.  The ADS1231 has no problem receiving the extra clock cycles as long as the clocks appear before the next conversion result is posted to the output register.

As Kevin mentioned, the fourth series of SCLKs will post invalid data to the RX buffer of the MSP430, and that data should be ignored.

Best regards,

Bob B

Thanks Kevin and Bob,

For Kevin's post: I actually don't do anything with data3. I read it but I don't do anything else with it. Here is my code that packs the data into their packets.

    firstBuffer = *(adc + 0);
    secondBuffer = *(adc + 1);
    thirdBuffer = *(adc + 2);    
    sign = (firstBuffer&0x80)?0xff:0;         //sign extension

Where packet[0] is sign and packet [3] is thirdBuffer. Bob actually helped me write this part of the code...so I think it should be fine. If you were still wondering about my PCB, here is the PCB used: http://www.radioshack.com/product/index.jsp?productId=2103799. I don't have a sch on me unfortunately.

For Bob's post. I thought about that. But I'm not sure because then when I put a AAA battery across the AIN's shouldn't it read zero? Since I'm putting about 1.5 V across AIN? It steadily reads the maximum positive value for the ADC so 7FFFFFh.  Anyway, I'm going to head to the lab later and try what you said.

Thanks for the advice!

Wesley,

With the battery "across the AINs" you're still not in the common-mode range for the device. AIN Negative has GND or a floating voltage from the battery and AIN Positive has 1.5 Volts from the battery. For the common mode range you'd need to apply a 1.5 V offset on your negative terminal as well to read "in theory" 0 from the converter, the battery is probably not outputting exactly 1.5 V. Right now your common mode voltage is floating where-ever the negative terminal of the AAA is.

Hi,

Regarding the code for sign-extend  cptr[3]=(cptr[2]&0x80)?0xff:0x00;  I have some doubts.

It fullfills cptr[3] with 1s if bit 8 of cptr[2] is 1 (negative value). Is this correct?  Shouldn't we insert a 1 only at bit 31?

Best regards,

Theo

Hi and sorry for my silly question.

Thats the way 2s complement work.

Best regards,

Theo