Understanding 4-20mA inputs and raw values

[robclay]

I am not just a bit rusty, but majorly rusty...

I am using an AB 1756-L55 ControlLogix 5555 Controller.
It has a remote 1794AENT/A Adapter.
I have an Analog Input Module - 1794/IE8
I have an Analog Output Module - 1794/OE4

My Analog Input Module is configured for (0-10V) / (0-20mA)
My Analog Output Module is configured for 4-20mA

For Inputs: I have been told that a 4mA signal will produce a value of 6242 (I figure it is exactly 6241.6). And a 20mA signal will produce a value of 31208.

For Outputs: 0mA = 0 value, and 20mA - 30840.

1. Can someone confirm this?
2. Where did they find this? In the manual for the analog cards, the PLC?
3. I forced an output to 15420 and noticed that the output was not 12mA. Should it have been?
4. For a PID Loop, in RSLogix v11.15, I am setting up a PID loop and am looking at the scaling tab. I have a Pressure Transmitter that is ranged from 0-300psi. I have a Control Valve that opens from 0 to 100%.
Should my Values be:
Process Variable:
Unscaled Max = 31208
Unscaled Min = 6241.6
Engineering Unit Max = 300
Engineering Unit Min = 0
Control Variable:
Max at 100% = 30840
Min at 0% = 0

I am sure this is such basic stuff, I am just trying to gain a better understandning.

 

[jonesy29847]

First of all the AB analog modules have the capibility of using engineering units when reading or writing.

The caveat is that you must configure all channels, a small price to pay for the conveinence, I would highly recommend it. Now your program will just use Engineering units and you will find it much easier to understand.

One count will be equal to Engineering Units/Resolution. For example lets assume we have a 0 to 10V signal, which represents 0 to 1000 psi, and assuming a 12 bit resolution 1 count = 1000/4095 or .244psi.

4-20MA signals are a little different since the signal never goes to zero, and I have not done this in 20+ years. But maybe this will help.

 

[NWBeaver]

Robclay, those 1794 modules use a 12 bit data format, except that the data is shifted to the left...

There is an analog data format table in the appendix of the manual which should help explain things.

http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1794-um002_-en-p.pdf

 

[avguy]

You config you AI as 0-20mA,
"My Analog Input Module is configured for (0-10V) / (0-20mA)"

while you were using 4mA as min value of the range,
"For Inputs: I have been told that a 4mA signal will produce a value of 6242 (I figure it is exactly 6241.6). And a 20mA signal will produce a value of 31208."

Similarly, in your AO you config 4-20, but you used 0-20mA
"My Analog Output Module is configured for 4-20mA
-For Outputs: 0mA = 0 value, and 20mA - 30840."

 

[jonesy29847]

I guess I just don't understand where the 30840 comes from. If 20 mA is full scale it should be 4095 assuming 12 bit, which is 4095 or 32767.

NWBeaver states that the values are 12-bit and shifted left, I would assume 4-bits in witch case 40950 would be the maximum count.

4mA is 25% of full scale for a count of 1024, 10240 shifted to the left. I can't get on Rockwells website right now, so I am making assumptions.

The information is in the manual for the particular module.

As far as Engineering units I would use 3000 or even 30000 for better resolution, unscaled max would be 4095 and 0 for min assuming the value will alwasy be positive.

The analog output scaling would be the same.

 

[jonesy29847]

Delete 32767.

 

[NWBeaver]

jonesy,
The value 30840 comes from the Data Table in the AB manual that I posted a link to above.

AB used a range of 4 to 21ma for the 4 to 20 mode, and 0 to 21ma for the 0 to 20ma mode...

So, for the 4 to 20ma mode, 20ma turns out to be about 30840.

The math goes something like this...
With 12 bits, 4095 represents full scale.
The value is shifted left 3 bits (multiplied by 8)
So 4095 becomes 32760

For a span of 0 to 21 ma, 32760/21 = 1560
20ma is then 1560 x 20 = 31200 or 79E0h

For a span of 4 to 21 ma, 32760/17 = 1927
20ma is then 1927 x 16 = 30832 or 7870h

 

[jonesy29847]

Thanks, I was assuming a 4-bit shift to keep things on even nibble boundries. I tend to think that way.

I tried to get to your link, but was unable to.