what factors consider as data loss

[shaik.khajababa.00@gmail.]

hey guys iam currently working on the dac7578 to read the digital data from accelerometer and convert it into the analog to read in DAS, for proof of concept iam sending a sine wave through my microcontroller and converting using this dac and the out put is feed back to the dso , but iam confused that what factors would effect the data loss in the real accel. and this too

 

[IRstuff]

I'm thoroughly confused as to why you would convert already digital data into analog, only to convert it back to digital. Unless you have tons of bandwidth to throw away, the D/A to A/D process will result in a ton of spurious noise without a ton of low-pass filtering.

What sort of data loss are you referring to? If there is loss in the original digital data, the D/A - A/D processing will also result in erroneous data being captured.

 

[shaik.khajababa.00@gmail.]

Hey, thanks for the input! You're right that D→A→D is wasteful in most signal processing chains—but in my case, I'm not doing that for processing. I'm using the DAC to emulate analog output from a digital accelerometer, mainly for proof-of-concept and to interface with a DAQ system or DSO that only accepts analog input.


The idea is:

• Generate a known waveform (e.g., sine wave) using the microcontroller.
• Send it through the DAC (DAC7578) to verify the analog output path and observe signal integrity.
• Later, replace the test wave with real accelerometer data, to simulate an analog sensor signal for external analog-only systems.

I understand the importance of low-pass filtering post-DAC to suppress stair-step and quantization noise—I'll be adding an RC filter accordingly.

 

[Eufalconimorph]

Each component can have data loss. I'd guess that the analog signal integrity will be the hardest to deal with, though a lot will depend on how long the analog cable is & how much environmental EM noise there is. Signal bandwidth might also be a concern, depending on how fast the edges of the accelerometer signal can be. The book "Electromagnetic Compatibility Engineering" by Henry W. Ott is the classic text's updated version, and well worth reading.

There's not enough info provided to give a much better answer. Forum posts are short, and you've essentially asked "what are all the sources of error in two entire fields of engineering: electromagnetic compatibility engineering & signal processing". If we had more info (datasheets, schematics, proposed layouts, link budget, etc) it'd be possible to give more detailed answers. Though that becomes work, so don't expect someone to run a field solver for you!

 

[IRstuff]

to emulate analog output from a digital accelerometer, mainly for proof-of-concept and to interface with a DAQ system or DSO that only accepts analog input.

OK, if you're doing this project just for giggles and experience; you can easily buy an accelerometer data logger for very little outlay, like barely 3x the cost of the DAC you're specifying. Moreover, you run the risk of chasing rabbits unrelated to basic problems, because you might have bugs in your hardware emulation.

 

[shaik.khajababa.00@gmail.]

Each component can have data loss. I'd guess that the analog signal integrity will be the hardest to deal with, though a lot will depend on how long the analog cable is & how much environmental EM noise there is. Signal bandwidth might also be a concern, depending on how fast the edges of the accelerometer signal can be. The book "Electromagnetic Compatibility Engineering" by Henry W. Ott is the classic text's updated version, and well worth reading.

There's not enough info provided to give a much better answer. Forum posts are short, and you've essentially asked "what are all the sources of error in two entire fields of engineering: electromagnetic compatibility engineering & signal processing". If we had more info (datasheets, schematics, proposed layouts, link budget, etc) it'd be possible to give more detailed answers. Though that becomes work, so don't expect someone to run a field solver for you!

I’ll grab Ott’s book (thanks for that!) — but if you’ve got practical heuristics or even "rules of thumb" for analog front-end shielding/layout/filtering that have worked for you in noisy environments, I’d love to hear them, what iam finding difficult is that when iam generating a 500hz sine wave using esp32 c3 and convert into analog via DAC7578 and connect it to the my oscilloscope, what iam seeing is 40hz singal instaed of 500hz, i dont know what is goind on here

• DAC: DAC7578
• Output is a smooth sine wave, but much slower than expected
• Sampling rate set to 10 kHz in code
• Using a sine table of 25 samples per cycle
• MCU is an ESP32 running at normal speed
  
  P.S : needed some guidance, not gonna make others to do my work : }

 

[shaik.khajababa.00@gmail.]

OK, if you're doing this project just for giggles and experience; you can easily buy an accelerometer data logger for very little outlay, like barely 3x the cost of the DAC you're specifying. Moreover, you run the risk of chasing rabbits unrelated to basic problems, because you might have bugs in your hardware emulation.

yeah i will look into that , but what iam finding difficult is that when iam generating a 500hz sine wave using esp32 c3 and convert into analog via DAC7578 and connect it to the my oscilloscope, what iam seeing is 40hz singal instaed of 500hz, i dont know what is goind on here

 

[IRstuff]

connect it to the my oscilloscope, what iam seeing is 40hz singal instaed of 500hz,

This is neither data nor communication loss; it's more likely an error in your code or misunderstanding of processor specifications.

 

[Eufalconimorph]

That's not data loss, that's almost certainly a code bug. First guess would be you messed up a timer configuration, e.g. got the wrong prescaler setting. You might want to use a logic analyzer on the I2C bus, and verify that you're sending samples quickly enough. I'd bet you aren't.

I'll also note that pre-made 500kHz DDS signal generators are <$30 on Amazon, cheaper elsewhere. Fine if all you want to simulate are sine/triangle/square/sawtooth waves. I don't expect you need an arbitrary waveform generator to measure the signal integrity of your setup.