- Ask a related questionWhat is a related question?A related question is a question created from another question. When the related question is created, it will be automatically linked to the original question.
Hi Team,
We are using Performance Demonstration Kit for the ADS130E08. After logging the data there are two files generated - Device_0_Codes and Device_0_Volts.
To calculate ENOB out thought is as follows,
1) Open Device_0_Codes file and convert this HEX data into decimal format.
2) Take log of this so that it will give ENOB.
Please confirm above our understanding OR suggest what is the use of Device_0_Codes file. I am attaching both files in the post.5706.Device_0_Codes.xlsDevice_0_Volts.xls
Regards,
Yogesh
Yogesh,
ENOB is a measurement of the noise in the measurement as a comparison agains the full scale range of the device. You can see the equation for ENOB in equation 1 the datasheet:
For your measurements, you have put in an input signal that is some sort of sinusoid. There are some errors in the datafile for the raw data, but for the voltage data, you get something that looks like this after you compile it into excel:
If you try to use that data, then you have an input signal that you're using to try to calculate the noise.
Normally, when you want to look at ENOB, you only look at the noise. In that case, you would take the ADC inputs and short them to a mid-supply voltage and you're looking at the noise in the measurement (which would be near 0V after you remove the offset). After you collect the noise data, then you would calculate the RMS value and plug it into the first equation.
Joseph Wu
Yogesh,
As I mentioned in the previous post, this is a measure of the noise in comparison to the full scale range of the input. Because of this, you don't want to measure the ENOB with a signal (unless you can completely remove the signal from the measurement.
When I opened your file of data, I made the conversion to codes and got this:
I'm not sure where the square wave comes from, but you won't be able to extract the noise from it, unless you choose a stable input. You might be able to do this if you only look at points that aren't changing (with a square wave, you may be able to do this but in this case, I'm not sure if your signal has it's own noise contribution.
Here, I took and ADS131E08EVM and made some of my own measurements with the inputs shorted, using 1kSPS as the data rate. I compiled this into an excel file here:
Device_1000_Codes.xlsx
Taking the data, I calculate the standard deviation of noise, which becomes the rms noise value. Then I calculate the ENOB from the equation shown in the datasheet:
This gives me the result for the 8 channels:
and I can compare this with the noise from the datasheet in table 1.
In the end, the values are not too far apart. I'm using different supplies, but I'm using the 2.4V internal reference.
Regardless, this was the process I used to calculate the ENOB. I would say there may be different ways to calculate ENOB based on the rms of the noise. In many cases, calculations will use the peak-to-peak noise and in others some will use a different scaling of the rms noise.
Joseph Wu
Yogesh,
I did something similar. Here's my board:
For this EVM, I used bipolar supplies (±2.5V). I tied all the inputs together and then I tied the inputs to the ground. This sets the inputs for all channels to mid-supply. After that, I just collected the data for the measurement.
Joseph Wu
Yogesh,
TP12 is connected to the analog ground. From the schematic, you can see the connection:
With the supplies at ±2.5V, the analog ground would be at mid supply.
Joseph Wu
