4-20 mA to RPM calculation

vijay s · Joined: 20 Oct 2007 Posts: 17 Location: coimbatore,india

Hi Folks
I need a little clarity. I am using a 4-20 mA current loop transmitter for RPM measurement. The transmitter has a frequency range of 0.0005Hz to 50KHZ. I have a 250 ohm resistor connected to a current loop, which intend to produce voltage of 1-5v. I got confusion with calculating this to frequency.
Please I need help on this.
_________________
with regards
vijay s

draghi · Joined: 01 Apr 2010 Posts: 2

freq_in_Hz = 12499.999875 * U_in_volt - 12499.999375

Ttelmah · Joined: 11 Mar 2010 Posts: 19513

Some comments though:
First, the resolution is going to be very low 50000/819. over 61RPM.
Second, the accuracy will be poor, because of using 5v as the refernce. To get the 'best' accuracy, consider using a bandgap reference, and adding a little signal processing, to subtract the 4mA offset.
Then, to avoid float maths, and correct for some of the errors in the resistors etc., look at something more generic like:

count_from_adc = value read from ADC
ADC_4mA = ADC counts for 4mA in
ADC_20mA = ADC counts for 20mA

mr.njt · Joined: 30 Oct 2013 Posts: 4

Sorry but i could not understand the value "819" . how could you get that number? and what is that for ?

Ttelmah · Joined: 11 Mar 2010 Posts: 19513

The 4-20mA signal, without any scaling, would use 0.8 of the range of the ADC. 0.8*1024=819.2. So this is the range of ADC values being used for the measurement. So without external processing (as I said to subtract the 4mA offset), you get at best 819 useable counts out of the ADC.

Best Wishes

mr.njt · Joined: 30 Oct 2013 Posts: 4

Thank you so much.

asmboy · Joined: 20 Nov 2007 Posts: 2128 Location: albany ny

Using a PIC with a 12bit a/d would help you too, if your current to voltage translation is low noise.

This would give you 4 times better resolution.

Ttelmah · Joined: 11 Mar 2010 Posts: 19513

Yes.

However the inaccuracy of any resistor converting I->V limits this unless you spend a lot.

Personally, I have to ask 'why' this approach Far easier to use a really basic PIC being fed from a simple shaft encoder, and count the time between pulses. Use a crystal, and in a tiny fraction of a second, you have RPM to several orders of magnitude better accuracy than the 4-20 can manage. If the data need to be sent a long way, then use opto-couplers, or 4-20mA transceivers, and an accurate RPM reading can safely be sent far further than 4-20 can do...

Just seems an expensive/inaccurate way to read the RPM....

Best Wishes