Understanding DAC Output: Can It Accidentally Spit Out Digital Values?

[lcr2139]

I had a question regarding DACs. Can they accidentally spit out digital values? The output I am looking at seems to be fluctuating by discrete values and I was wondering if it was related to the DAC.

 

[berkeman]

I had a question regarding DACs. Can they accidentally spit out digital values? The output I am looking at seems to be fluctuating by discrete values and I was wondering if it was related to the DAC.

That's what DACs do -- they output an analog voltage (or current) that is related to the digital input.

There are measures of the accuracy and precision of DACs. Are you familiar with how to measure these qualities? Which DAC are you using?

 

[lcr2139]

I am using the LTC2601. I would like to get rid of the noise that is coming out of it. How do I do this?

 

[berkeman]

I am using the LTC2601. I would like to get rid of the noise that is coming out of it. How do I do this?

That looks to be a pretty good quality DAC, and has a guaranteed monotonic output:

https://www.digchip.com/datasheets/parts/datasheet/262/LTC2601-pdf.php

What kind of noise are you seeing? I'm guessing it's coming from something else. How is the digital input to the DAC being generated? Can you post a schematic?

 

[lcr2139]

I can't post a picture, but basically the temperature is jumping around between around 0.2 degrees C, peak to peak. It seems to go straight from one value to the next, instead of gradually changing. That is why I think it is a problem with the DAC.

 

[berkeman]

I can't post a picture, but basically the temperature is jumping around between around 0.2 degrees C, peak to peak. It seems to go straight from one value to the next, instead of gradually changing. That is why I think it is a problem with the DAC.

It's a precision 16 bit DAC. The problem is elsewhere, most likely. Why can't you use the UPLOAD feature to post a picture or scan of the schematic?

What is the source of the temperature data? What SPI temperature sensor are you using? How many bits does that 0.2 degree jump represent for the DAC?

 

[donpacino]

I can't post a picture, but basically the temperature is jumping around between around 0.2 degrees C, peak to peak. It seems to go straight from one value to the next, instead of gradually changing. That is why I think it is a problem with the DAC.

if your resolution is 0.2 degrees, then that is what will happen. To find your resolution you'll need the circuit topology and know information about all the components of your system.

If 0.2 degrees is indeed your max resolution, then you will need to deal with the problem.

If your problem is simply the sharpness of the data, that is a fact of life when dealing with real components. If you don't like it for visual reasons you can pass allways the data through a low pass filter.

 

[jim hardy]

As everybody else has pointed out,
the DAC in all probability is faithfully reporting whatever number it is given by the computer.

Basic trouble shooting - encircle the trouble then tighten that circle using logical steps..
Use software to print the number that is being handed to the DAC. That way you have drawn a circle around the DAC and you know whether to look inside or outside that circle.
Or write a program that repeatedly hands DAC the same number .

With what are you reading the DAC output? Maybe it's fibbing.

Lastly, what is the nature of your Vref applied to pin 6 ?
This thing appears to multiply that voltage by whatever binary number you hand it
from the better datasheet at http://cds.linear.com/docs/en/datasheet/2601fb.pdf

so a sloppy Vref at pin 6 = a sloppy output. You didn't just tie it to Vcc did you ?

Do you see how much you've asked us to guess ?

 

[jim hardy]

As donpacino and berkeman both asked
0.2 degrees out of how many for full scale ?