Hello!
Is there any ways or sensors that can measure the voltage of nickel cadmium cell directly(wiithout losses)? I am suppose to measure the voltage and transmits the data wirelessly to my laptop using arduino UNO.
Thanks!
Technically you can't measure anything "without losses" as the very act of measuring requires you to steal something from the object being measured.
You can have very low losses, what can you afford?
Rob
He's right you know:-
It is one of the fundamental laws of Physics.
you can measure with extremely small losses using a voltage divider with very big resistances. but it's impossible to measure without.
Graynomad:
Technically you can't measure anything "without losses" as the very act of measuring requires you to steal something from the object being measured.You can have very low losses, what can you afford?
Rob
Thanks for the reply! I can afford up to an acurracy of at least 4 decimal places.
scorpio20:
you can measure with extremely small losses using a voltage divider with very big resistances. but it's impossible to measure without.
Thanks for the reply! I need to measure the voltage of 20 fixed 1.2V series cells, is voltage divider still possible?
Grumpy_Mike:
He's right you know:-
Observer effect (physics) - Wikipedia
It is one of the fundamental laws of Physics.
Thanks for the reply! Appreciated 
Thanks for the reply! I can afford up to an acurracy of at least 4 decimal places.
This is not a power loss.
You can't measure that accurately with the built in A/D converter in the arduino.
Having 20 cells in seriese makes things very difficult. There have been some long threads on this, search for them.
20 fixed 1.2V series cells
So you are measuring voltages ranging from 1.2 to 24, some cells (the ones > 5v) will need voltage dividers and some won't (the ones < 5v), maybe some should even be amplified. It is not a trivial problem.
4-digit accuracy is 10x what the Arduino is capable of so I think you need an external ADC in the order of 13-14 bits. But there has been many discussions about this subject, try searching the forum and get back to us if you can't find an answer.
Rob
Graynomad:
So you are measuring voltages ranging from 1.2 to 24, some cells (the ones > 5v) will need voltage dividers and some won't (the ones < 5v), maybe some should even be amplified. It is not a trivial problem.4-digit accuracy is 10x what the Arduino is capable of so I think you need an external ADC in the order of 13-14 bits. But there has been many discussions about this subject, try searching the forum and get back to us if you can't find an answer.
Rob
Okay. Is the external ADC some sort of IC chip like "TI ads1213"?
By the way, i'm suppose to measure individual Ni-Cad cells connected in series. If i were to use ADC, it is just the programming part i have to take care of in order to get the most accurate readings?
Is the external ADC some sort of IC chip like "TI ads1213"?
Yes, however getting 22 bit accuracy needs a good circuit and layout, you just can't throw a chip at it on bread board and hope it will work.
i'm suppose to measure individual Ni-Cad cells connected in series.
You will have to measure each node and subtract adjacent nodes to work out the voltage across each cell, you can not measure the voltage on each battery directly.
it is just the programming part i have to take care of in order to get the most accurate readings?
No as I said before good layout and good circuit design are also needed.
i'm suppose to measure
This sounds like an assignment.
Grumpy_Mike:
it is just the programming part i have to take care of in order to get the most accurate readings?
No as I said before good layout and good circuit design are also needed.
i'm suppose to measure
This sounds like an assignment.
It's a project So what i am suppose to do now is to implement the ADC(13-14bits) and interface it with arduino?
By the way, i'm suppose to measure individual Ni-Cad cells connected in series. If i were to use ADC, it is just the programming part i have to take care of in order to get the most accurate readings?
So you measure each 1,2 volt cell individual, so with 12 cells you need 12 ADC's on a Arduino!
Paco
I thought you posted a 22 bit A/D?
That one will handle 8 inputs. So you also need to expand that with a data selector like the 4051, then you need to calculate the resistors for the potential dividers you need for each cell so that you don't present more than 5V to tha A/D.
Then you need to write the software.
If you go down the 12Arduno path you need to optically isolate the serial output on each one to stop it all shorting out.
I hope I am getting good marks for this project.
Graynomad:
Technically you can't measure anything "without losses" as the very act of measuring requires you to steal something from the object being measured.
Uncertainty principle will stop you from getting a perfect solution, but way before that kicks in you would have to deal with self discharge rates and so on.
Assuming the goal is that the battery is not powering the sensing circuit, I suppose you could achieve that using something analogous to a wheatstone bridge. I mean, have the Arduino control a circuit that generates a voltage, and use some form of current sensing to determine the direction of flow between your output driver and the battery. When the flow is sufficiently close to zero, you know that the battery voltage is the same as your output voltage. Given that you [should] know the output voltage your driver is producing, you now know the battery voltage.
Or, taking a slightly more crude approach and assume that battery discharge is the thing you want to avoid, you could design your sensing circuit so that it supplies charge to the battery rather than taking charge from it. It'll affect the battery (slightly) but won't discharge it.
I think there are three ways to do this
- High-res ADC with various combinations of voltage divider.
- String of 12 intelligent monitors, each isolated from the cell it's monitoring.
- A MUX made from relays feeding an isolated ADC.
I prefer #2 but probably #1 is easiest.
All these have been discussed ad nauseam in the past.
Rob