Indirect Square root? Help?

[waterboy 7]

Can someone walk me through the Indirect Square root formula as it pertains to Differential Pressure transmitters? I know this isn't strictly PLC related but this is a pretty smart group surely you have all dealt with this before? This is for diff pressure transmitter calibration and I would to prepare a spreadsheet with multiple input points of pressure in whatever value (in. H20 or PSI ) and the expected mA output from the device.

[panic mode 246]

i don't understand the question. it seams that you want to know how to linearize analog input. if that is the case, try collecting few points (each point will have both sensor output and correct value). use something like CurveExpert 1.3 (http://curveexpert.webhop.net/) to find suitable fit (square root, polinomial or whatever).

[Mickey 71]

Like panic I'm not sure i understand your question. Most transmitters have build-in the abilitiy to scale the output ( 4-20ma) to repersent H2O or to extract the square root. Check your transmitter configuration setup.

[waterboy 7]

I'm sure I'm mangling the question too. I would like to measure the output of the device to see if fits what the mathmatical calculation comes up with. Say I have an pressure input of 50% of the scale of the transmitter . Because the output in mA isn't linear,I won't get 12mA output, I have to use the Indirect Square root formula to calculate the expected mA output from the device. If I understood the formula I could build the spreadsheet easily enough but the terms aren't yet familiar to me. The bottom of this shows the formula: It's going to be elementary once I get familar with what values to put into the terms but I can't get that part. i.e. Out_Scale.EU0 is what value? same for Field_Val and Out_Scale.

[panic mode 246]

well this should be obvious, let me see if I can explain it. your help file screen shot seam to focus on 0-100% range (ti is everywhere in your example). EU is Engineering Units (V, A, gallons/min, psi, miles, whatever) EU0 and EU100 are EI (engineering units) for 0 and 100% So if your pressure is 30-75psi, then 30-75psi is 0-100% or 0%=30psi and 75psi is 100% so those values in this case are just the range limits (30 and 75). Thise values are fixed (they normally don't change). An output example they have shown is for temperature in degC, range -100...400degC, Input range is +/-50mV (EU_0=-50, EU_100=+50). Note that all of this is LINEAR. As an example there are three possible output types and only the last formula (Indirect Square Root) is not linear. Is this what you really have (L_TYPE= Indirect Square Root)?

[newpageboba 7]

It is very basic, but remember the units. use 0-100% and remember 100%=1 therefore 50% =.5 So, if your pressure is 0% of input range, your square rooted output will be 0% of output range (4ma) if your pressure is 25% of input range, your square rooted output will be 50% of output range (12ma) if your pressure is 50% of input range, your square rooted output will be 70.7% of output range (15.3ma) if your pressure is 75% of input range, your square rooted output will be 86.6% of output range (17.8ma) if your pressure is 100% of input range, your square rooted output will be 100% of output range (20ma)

[waterboy 7]

That is helpful. Can you show me the math you used. (and what is "indirect" about it?)

[newpageboba 7]

for 25% i took the square root of .25 which is .5 or 50% for 50% - the square root of .5 is .707 or 70.7% hope this helps. Is this for a flow measurement? If so, you would need to know what 100% Diff Press represents in flow. It looks like indirect means not direct with square root added???? Edited 29 Nov 2006 by newpageboba

[waterboy 7]

yes, a vortex flowmeter for plant influent

[newpageboba 7]

Using Excel, put in the values from 0.00(0%) to 1.00(100%) using the increments you want. I used .05 (5%) Use this formula - sqrt. FlowvsPress.bmp FlowvsPress.xls

[newpageboba 7]

Using your indirect squareroot formula, XD_SCALE.EU0 = 4ma, XD_SCALE.EU100 = 20ma (or use what you scaled your analog input) this make FIELD_VAL = 0 - 1 (or 0-100%) OUT_SCALE.EU0 = 0 & OUT_SCALE.EU100 = 100 (or use the actual flow range) the divisor of 100 in the square root does not make sense.(I think it is actually 100%, which is 1) The FIELD_VAL formula gives a value of 0 to 1 (not 0 to 100). If the divisor is actually 100, then the XD_SCALE.EU0 & EU100 values should be 100 times greater.

[waterboy 7]

Well now..... That certainly wasn't what I was building it up to be! Thank you for that. What exactly makes it "Indirect". Square root is simple enough but throw the word "Indirect" at it and I just crawled into a corner and grabbed my teddy.

[newpageboba 7]

I think what they meant by indirect is 'not direct'

[JRoss 84]

FIELD_VAL is a number from 0-100 (in percent, of course). The divisor of 100 simply converts it from percent to a numeric ratio.

[waterboy 7]

Very funny . I see in some more reading that direct and indirect refer to the signal path and not to a particular math function. Would you agree?

[newpageboba 7]

I am a stickler for engineering units in math formulas. I see many people get burnt in their calculation by not using engineering units. 0%-100% equals 0-1 0-100 does not equal 0-1 it equals 0-100 Edited 29 Nov 2006 by newpageboba

[waterboy 7]

That makes sense, It's starting to clear up now. Thank you very much for your help, all of you.

[JRoss 84]

Good point! I remember the days. Physics, Statics, Dynamics, etc., etc., ad nauseum. So that portion of the formula should read: FIELD_VAL 100%

[newpageboba 7]

Yes, then the formulas make sense.

[waterboy 7]

Just when I get the math, now the application gets me....How do you use this non-linear value in a PLC5 and convert it accurately? The existing value that is received from this transmitter is currently treated in the PLC like it was a linear value. i.e. 12mA input to the module is treated as 50% of the scale of the instrument. If I understand right, the real value for 12mA input to an analog module should be about 25% of the full scale of the instrument. If this is right, how do I handle this non-linear value from the instrument?

[panic mode 246]

read first reply

[waterboy 7]

But there is no "curve fitting" in a PLC's language.

[rpraveenkum 5]

using the Function Generator (FGEN) inst in RSL5k you do linearisation The FGEN instruction converts an input based on a piece-wise linear function.

[panic mode 246]

:sigh: TGIF one more time - read first reply... the idea is to use your own math skills or help of a tool (excel, curvefit, mathlab or whatever) to find the formula which is close fit or good approximation and works for the application. not all solutions are accurate enough and then, (once you know what the formula and parameters/coefficients are), you can use math instruction to implement it in language your controller supports (controller happens to be a PLC in this case). this may require series of rungs and ADD, MUL, DIV, CPT or whatever blocks your controller supports and furmula needs.

[waterboy 7]

Ok I see what you mean now.