At 9:46 You have to use 1023 not 1024. 1,1V on analog pin in, return value 1023, the highest number possible. Else u read (analog in =1,1V=1023) / 1024 * (Vref=1,1V) = 1,0989V. Ie (analog in =5V=1023) / 1024 * (Vref=5V) = 4,995V.
05:10 basically shows there is negligible advantage in using NRM in this application. There is virtually identical noise with and without it on. If you calibrated the reading in NRM mode and in normal mode, the results suggest that the difference would be negligible. It's very interesting to know about but I wonder when it's actually useful.
Hey just a suggestion, on your blog, if you added the youtube video link to the code it would be helpful, at least it would have been for me right now. I'm usually watching on my phone but coding on my laptop. I copied and pasted a few of your examples yesterday and saved them to a sketch. I should have copied the address too but didn't think about it. Over time, if I come back to this info through the saved sketch in the distant future it would be ideal to have the addresses embedded into the commented intro description. Also, if you are ever interested in furthering this content directly, this would make an ideal library example of how use the ADC in depth. I wish I could save these sketches in my library examples. I title them with "do not edit" in my sketch folder but I could always make a mistake. It probably doesn't matter for me as I am actively transitioning away from the Ardy ide either on platform io or eclipse already and will commit to one completely after my current project. Thanks for the super helpful series here. Especially the register level Rosetta Stone, that is extremely useful. -Jake
So, the moral of the story is that one can get about 3 significant digits of a good accuracy with just an Arduono. Would it be possible to push that higher, possibly to 4-5 very accurate digits, if one bought a quality ADC chip. Maybe with some temperature stabilization and EM isolation? It would be great to have a self made 6.5 digit voltmeter, for say $50-$100.
3 years late but IC's like the old Linear Tech LTC2412 are (24-bit ADC) are advertised on the datasheet applications for "6 digit DVMs." They also had similar advertised application specs for supporting ADC hardware like the driver/buffer/op amps and voltage references. I was actually shopping on AliEx with a seller that had reasonable prices on old linear tech stuff. It's iffy to shop AE for this stuff there but if the price is right I'm usually dumb enough to try it out. Personally, I picked up a 5.5 digit keithley 197 DMM from eBay for $100 shipped. If you get the original 197 they are about the cheapest high resolution meter available. People want at least the 197a because it has an iec cord connector for power (same as most computer power chords) instead of a permit cord. The 197a also came with a backlit display. The other big factor with this specific meter and price is that this is one of the last keithley models that came in the old turd brown color. It's completely hideous. So if you want a great high resolution meter it's the butt ugliest option around and usually the cheapest too :-)
I believe it is 10k but that is off top of my head. I am sure you can find it in datasheet of 328p. The Arduino analogRead function is not optimized for speed so you need to dig into the ADC register settings to speed things up
+David Reeves Correct the Due would be different. In this series I just focus on the Atmega 328p which is on the Uno and a couple other Arduino boards. I believe the Due has a 12 bit ADC and I am sure it has more functionality when it comes to the ADC. I should have made the Arduino subset that this covers more clear in the videos.
At 9:46 You have to use 1023 not 1024. 1,1V on analog pin in, return value 1023, the highest number possible. Else u read (analog in =1,1V=1023) / 1024 * (Vref=1,1V) = 1,0989V. Ie (analog in =5V=1023) / 1024 * (Vref=5V) = 4,995V.
05:10 basically shows there is negligible advantage in using NRM in this application. There is virtually identical noise with and without it on. If you calibrated the reading in NRM mode and in normal mode, the results suggest that the difference would be negligible. It's very interesting to know about but I wonder when it's actually useful.
Hey just a suggestion, on your blog, if you added the youtube video link to the code it would be helpful, at least it would have been for me right now. I'm usually watching on my phone but coding on my laptop. I copied and pasted a few of your examples yesterday and saved them to a sketch. I should have copied the address too but didn't think about it. Over time, if I come back to this info through the saved sketch in the distant future it would be ideal to have the addresses embedded into the commented intro description.
Also, if you are ever interested in furthering this content directly, this would make an ideal library example of how use the ADC in depth. I wish I could save these sketches in my library examples. I title them with "do not edit" in my sketch folder but I could always make a mistake. It probably doesn't matter for me as I am actively transitioning away from the Ardy ide either on platform io or eclipse already and will commit to one completely after my current project.
Thanks for the super helpful series here. Especially the register level Rosetta Stone, that is extremely useful.
-Jake
So, the moral of the story is that one can get about 3 significant digits of a good accuracy with just an Arduono.
Would it be possible to push that higher, possibly to 4-5 very accurate digits, if one bought a quality ADC chip. Maybe with some temperature stabilization and EM isolation?
It would be great to have a self made 6.5 digit voltmeter, for say $50-$100.
3 years late but IC's like the old Linear Tech LTC2412 are (24-bit ADC) are advertised on the datasheet applications for "6 digit DVMs." They also had similar advertised application specs for supporting ADC hardware like the driver/buffer/op amps and voltage references. I was actually shopping on AliEx with a seller that had reasonable prices on old linear tech stuff. It's iffy to shop AE for this stuff there but if the price is right I'm usually dumb enough to try it out.
Personally, I picked up a 5.5 digit keithley 197 DMM from eBay for $100 shipped. If you get the original 197 they are about the cheapest high resolution meter available. People want at least the 197a because it has an iec cord connector for power (same as most computer power chords) instead of a permit cord. The 197a also came with a backlit display. The other big factor with this specific meter and price is that this is one of the last keithley models that came in the old turd brown color. It's completely hideous. So if you want a great high resolution meter it's the butt ugliest option around and usually the cheapest too :-)
what is the sampling frequency of arduino uno adc?
i want measure maximum and minimum values of a sine wave which lies between 0-5v dc
I believe it is 10k but that is off top of my head. I am sure you can find it in datasheet of 328p. The Arduino analogRead function is not optimized for speed so you need to dig into the ADC register settings to speed things up
@@ForceTronics i want to read 100hz sine wave signal(max and min)
can you please help with this problem🙏🙏🙏
These are great :D
Have you tried this with the Arduino Due? ...wouldn't a Due require a different version?
+David Reeves Correct the Due would be different. In this series I just focus on the Atmega 328p which is on the Uno and a couple other Arduino boards. I believe the Due has a 12 bit ADC and I am sure it has more functionality when it comes to the ADC. I should have made the Arduino subset that this covers more clear in the videos.
Wow TOP
Make a robot