hi *,
i was wondering how to best model and quantify the noise of the DAC124S085.
i guess it's somehow composed of noise of the internal resistor network plus the noise of the output amplifier/buffer but i couldn't find any hints/values which would lead me to an appropriate calculation.
any good ideas, links, appNotes to this problem?
thank you very much in advance!
kind regards,
fabian
Hi Fabian,
In the case of this family of DACs the main noise contributors are the resistor string, and the unity gain buffer driving the output. The story complicates as the effective impedance of the r-string seen by the buffer varies with code. The worst case in therms of noise is therefore at midscale where the effective impedance of the r-string is about 60k. R-string mainly contributes the thermal noise.
Unity gain buffer contributes the 1/f noise, and thermal noise.
I do not have test data for the output noise for the DAC, but I do have a simulation result which I have attached to this posting.
Vertical axis is spectral noise density in Vrms/rt(Hz).
Hope this helps.
Hi Tom
Thanks a lot for your answer!
If my rough calculation isn't too bad your plot tells me that the spectral noise power "integrated" over the whole frequency range finally leads to a Vrms amplitude smaller than 1uV (1E-6 Vrms). Is this correct?
Why is this not listed in the datasheet? Simply because the noise is so low and the error from the quantization is some magnitudes higher?
Fabian
My integration gives me about 81uVrms, integrated in the 1Hz to 100MHz band.
This is still pretty small compared to 1LSB (1LSB is about 660uV at VREF=2.7V), and maybe that is indeed the reason why this is not listed in the Data Sheet..... not to mention the fact that non-linearities are in the order of multiple LSBs.
Otherwise I don't know why this is not spec'ed.
Sincerely,
tom
Hi Tom,
Oops, my calculation was obviously a bit too coarse.
Thanks a lot for your help!