My IF2OF2 analog input isnt doing full scale....?

[frenchy]

Hello,
I am using an IF2OF2 analog IO module in conjunction with a Micrologix 1100. I have both analog inputs set up for voltage mode with raw resolution. I thought I should expect 0v = 0counts and 10.5v = 32760 counts.... but, when I use a power supply to feed a test voltage into the analog inputs, I am seeing a maximum of 32672 counts on both inputs with an input voltage of approx 10.57 volts. The counts never rise above 32672 even when my input stimulus is up to 10.8v. Why are my modules "capping out" early?

So, when I set up my SCP functions in RSLogixMicro, should I use the "truth" or the expected?
Truth = 0-10.5v = 0-32672counts, so 0-10v should be 0-31116counts.
Expected = 0-10.5v = 0-32760counts, so 0-10v should be 0-31200counts.
My gut tells me to use the truth, but someday the IF2OF2 might wake up and do its job properly!

Thanks for any input!
Respectfully,
Steve French
814.584.1220.office
814.730.0003.cell

President, Volt Vision
www.voltvision.com

 

[Mickey]

Are you reading this voltage with a precision volt meter across the analog cards input terminals or using the power supplies output meter? Always go with the truth.

 

[Gil47]

You asked Why are my modules "capping out" early?

The module you have is a 0 to 10 Volt module
They scale them out to about 10.5 Volts and cap the value to stop it exceeding 32767 which would return you a negative value if it took one more value count, this would read inside the PLC as a step from full scale 10.5022 volts to Nothing 0 Volts when you actually have 10.5025 Volts on the terminal, because the nearest value to your negative value will be 0, or if it only reconciled the 15 bits that does come out as 0 .
Thats 0.3 Milli volts increase

 

[frenchy]

Mickey,
Thanks for the reply. I was reading my voltages with 4 different meters, the power supply meter, an old abused fluke, a yellow radioshack and one I just got from Omega as a gift. The power supply one didnt have enough resolution, so I ignored it. The omega and fluke were always within 20-30mv and the radioshack was usually 40-60mv low. Since the range is what is bothering me, lets shift the dicussion to this.... "Why did my counts stop @ 32672 and never go above that despite the input voltage"? This question takes all 4 of the meters and their accuracies out of the equation. Have you ever seen an IF2OF2 not be able to go full range? Do you automatically use the spec'd 32760 range or do you always "see what it really does"?

thx,
frenchy

 

[frenchy]

Gil47,
Thanks for your reply from many miles away! I follow what you are saying, but mine is never going above 32672 counts. It is never making it to the theoretical 32767 maximum that you discuss. Is it common to "never get there"? This means my SCP (scaling) needs extra attention because the theoretical max is not reality. Am I missing something? I did what Mickey suggested and I used the "truth" not the theoretical max.

thx,
frenchy

 

[Gil47]

Your scaling I believe should be 0 Volts = 0
10 Volts = 31200
The idea of the capping is to protect the Electronics of the card should you for example put 24 Volts on the terminals.
It also has the advantage of stopping your PLC value going Negative in value

 

[Mickey]

If this is a new card then the accuracy should be better then that. If you have another one then replace it and compare. If not contact your local AB rep and see if you can exchange it.

 

[frenchy]

Thanks again Gil47 for your response, although I am sorry that I cannot agree with you. If my unit was able to actually reach 32760counts @ 10.5v stimulus, then yes I would agree with you that 10v=31200counts.

But since my unit is only reaching a max of 32672counts @ >10.5v stimulus, then I would hypothesize that 10v should be 31116counts.

I agree that analog input clamping is useful to protect the electronics from overvoltage or negative voltage, but regardless of that, 10.5v should be 32760, not 32672. I am being shortchanged in resolution by 88counts. I know how to deal with it, but it surprised me and makes me think that the 32760 is merely a "guideline" rather than an engineering spec. I hope I am missing something.

respectfully,
frenchy

 

[Mickey]

Double check the dip switches. Don't just look at them, move them. It can't hurt.

 

[Gil47]

What value do you get when you put exactly 10 Volts on the terminal, I ask this because the capping is probably controlled by a Zenner Diode or something similar with a ? % of tolerance, 1 % to 5 %


EDIT
I see 88 counts as being +- 0.27 % tolerance on the Zenner

 

[frenchy]

Mickey,
Yes I do have two of the IF2OF2 modules sitting right next to each other...one for two temperature inputs and one for two motor drive feedback inputs. I think tomorrow I will try to hook my power supply test stimulus up to all 4 at the same time and see what they all do @ 10.5v. Up until now I had blindly assumed that 10.5v = 32672 counts. If one of them is 88counts low I wont take it back, I will just account for it........unless it changes over time........hmmmm....

thx,
frenchy

 

[rdrast]

88 counts out of 32672 ("Spec Full Scale") is only 0.27% in error at full scale. How much accuracy do you really need?

Aside from which, unless a device is Standards Traceable, you should always calibrate it if the application requires it.