Precision of formula node?

I need to calculate a simple formulae.
But different results between two program.
Can you check which one is correct using other program?
or tell me the principle of calculation with formula node?
Enclosed is the example.
LabVIEW: 8.0614613965
Excel: 8.0614616898
-labmaster
Attachments:
excelexam.xls ‏584 KB
LabVIEW exam.vi ‏7 KB

If you do all the calculations in regular LV functions rather than in the formula node, you can use Extended Precision data types. That changes the 17th digit right of the decimal point.
When I look at the data at intermediate steps (the individual terms summed to get e_ITS90) I see that the spreadsheet matches to 7 to 13 significnat digits, depending on the exponent. Note that for the third term (with exponent = 0), the spreadsheet calculation does not return the g2 coefficient: g2 = 19.542636 and g2*x^0 = 19.54263612. I set up a column in the spread sheet with the values equal to the g[i] coefficients but displayed with as many digits as the spreadsheet allows. g1, g2, and g3 differ from what is entered in column C in the first several digits to the right of the last digit entered. When I do the same thing in LV, the differences tend to be in the 15th or 16th digit.
All of these things are the result of the way numeric data is represented in the computer, as Jeff pointed out earlier. It appears that you can get somewhat more precison in LV than in the spreadsheet. I did not do any accuracy checking.
Lynn
Attachments:
LabVIEW exam.2.vi ‏11 KB

Similar Messages

Extended precision variable declaration in formula node?

I'm looking for a way to define my variables inside my formula node as extended precision but everytime I attempt this, I get an error. To my understanding, the declaration should be "long double" but it does not seem to work. Anyone have any ideas?
Greycat

Thanks Dennis ... I guess sometimes it seems quicker to ask instead of looking... but then I found what you were referencing very quickly. Just being lazy I suppose!!! Anyway, thanks for the help.
Greycat

Functions in Formula Node

Hello:
I have written a simple formula node in the VI in annex. It uses two functions that should be called for R, G and B. Because I do not know how to implement functions (the way I implement C functions) I had to repeat 3 times the same code (for R, G and B) which makes the code not so readable and maybe I loose on performances? Could you correct the formula node by implementing functions, please? I was not able to find this tip in the various documentation that I have nor on the net. Thank you very much in advance,
Chris
Attachments:
myRGBtoLAB.vi ‏32 KB

Wow Chris!
That is one impressive formula node! In terms of computational efficiency, I wouldn't expect there to be a lot of difference between the formula node and separate VIs. Stll, I have attached two sub VIs that I think implement the two repeated areas of the formula node code.
Something to do though is change the type of all the outputs to doubles. The calculation is being done in double precision anyway, so you might as well preserve it. Also, the single frame sequence around the whole thing isn't performing any function so it can be deleted.
As far as readability goes, remember that readability (like beauty) is in the eye of the beholder. If you can read the code and know intuitively what it does, I wouldn't worry too much about it...
Mike...
Certified Professional Instructor
Certified LabVIEW Architect
LabVIEW Champion
"... after all, He's not a tame lion..."
Be thinking ahead and mark your dance card for NI Week 2015 now: TS 6139 - Object Oriented First Steps
Attachments:
n_to_n1.vi ‏22 KB
the_other_calculation.vi ‏25 KB

Formula Node Floor(x) Produces Different Result

Hi, A search didn't find anything about the Floor(x) function, so... I'm using LabVIEW 6.0.2, and the Floor(x)function in a Formula Node seems to be producing different results depending on previous calculations. I've got the following 2 lines in a Formula Node:
MSS = Ref / RefDiv;
MDN = floor(RF / MSS);
Ref is always 20.0M, RefDiv always 500.0, and for this calcualtion RF is always 1539.4, all numbers Double with 6 digits of precision. I generate an array of frequencies given a start, step, and frequency count. These frequencies then go to a subVI with a Formula Node that calculates the byte values to send to a couple PLLs.
If Start = 70.1, Step = .025, and Count = 20, at frequency 70.2 the Floor function gives 38.485.
If Start = 70.0, Step = .025, and Count = 20, at frequency 70.2 the Floor function gives 38.484.
I've omitted some calc steps here, but I've verified the starting values in the subVI are the same in both cases. Why the result changes I'm hoping someone can tell me...
Thanks,
Steve

I want to thank those involved again for their help. With ideas and hints from others I found a solution without scaling.
In recap, what had bothered me was it *appeared* like the same subVI was giving correct results one time and incorrect results only randomly. While I understand binary fractional imprecision, I wasn't doing any looped calculations 100+ times or anything.
I did some more checking though. The problem was indeed introduced by cumulative fractional addition. In this case 10 additions were enought to cause the error. Apparently, floor(71.199_94) produces 72.0. However, using a shift register and constant fraction to add an offset to produce an array introduces enough error in under 10 iterations to be a problem. By the time the loop got to what was supposed to be 72.0, it was actually 71.199_84 or something, enough to throw the floor function. Now I understand why the error occurred, and why it wasn't a problem before.
I fixed this problem by changing the real frequency number to a I32 before introduction to the formula node. This corrected the error introduced by the fractional addition by forcing 71.199_84 to 72, instead of letting it propagate through the rest of the calculations. And it was a whole lot easier than changing all the VIs to allow scaling! Also, I prefer to know where and why the problem happened, instead of just scaling all my calcualtions. Maybe I can recoginse potential problems in the future.
My boss wants to go back and look at his program to see if HPVee somehow bypassed the problem or if he did the calculations differently.
Thanks again for the insight and help,
Steve

IF statement in Formula Node

I have the following statement in a Formula Node in LabVEW 7.1:
if (x=1) y = 3; else y = 1;
When running the VI, whatever the value of X is, I obtain Y = 3 !!!!
Why ???
Tom

Try x == 1 (instead of x = 1).
Using LV8.0
Don't be afraid to rate a good answer...

If-then statement in formula node

**Note - The attached file shows the current block diagram.
I have a formula node inside a for loop. Inside the formula node I am trying to program the if-then statement:
If T1<2.0E-5 OR T1>5.5E-4 THEN GOTO 3
U7=U7+V
I have 4 items I need help with:
(1) Is 'II' the correct operator for 'OR'?
(2) Is 'goto' an acceptable operator within LabVIEW?
(3) 'U7' is building a 1D array. How do I need to define it? Is it simply 'int U7'? Also, I will add an output value named 'U7' on the formula node wall with an indicator wired to it from outside the for loop.
(4) Is my use of the equal sign appropriate or do I need to use '=='?
Thanks for the help.
Philip
Attachments:
if-then_in_FN.jpg ‏18 KB

Thanks for the help. I went ahead and rewrote the program so I could omit the 'goto's. Another problem that I am having is that I want 2 statements for 1 assignment:
if (T1>=.002004 && T1<=.002584)
B1=U
A1=V;
I want to assign both B1 and A1 values when the above condition is true. I tried a comma between the U and A1 but it doesn't work as well. Should I define A1 and B1 separately? Or would this create problems?
if (T1>=.002004 && T1<=.002584)
B1=U;
if (T1>=.002004 && T1<=.002584)
A1=V;
Thanks,
Philip

Can I use a formula node to start the execution of a case structure?

Hi,
I am working on editing a VI to make it much easier to understand (for colleagues and non LV users). For one reason or another, the VI's creators did not fully make use of the power of case structures and other structure types. In the application we need it for, the instrument cycles and repeats measurements on several objects. We are interested in measuring the motor currents for several motors in the system. As it stands now, for each sample object's run, there is a separate sub-vi diagram which displays its waveform trace and several indicators such as max current, time at max, etc. That code was contained within a T/F case structure and copied and pasted 20 plus times.... Obviously the vi became extrememly cluttered and needed a huge screen to see. The trigger for the current case structures is an EQUALS comparison between one input (which is the sample object counter; and this part I will likely leave unchanged since it makes a lot of sense already) and a constant which was defined 1 through 20+ for case.
I have since made one case structure and 20+ instances of that case and have labeled them (at the top selection box near the detent arrows) "1", "2", 3, etc. I am wondering if I can use a formula node to act as the "trigger" which sets each case structure running? I believe that a simple IF statement should work. Please see a snippet of my attempt at making this in C below.
For the node, I defined X as the input and Y as the output. The input to the node, X, is connected to the sample object counter. The output, Y, is connected to the case selector of the case structure. My attempt at the code is below:
int32 y
For(x == 1)
y = "1";
For(x == 2)
y = "2";
etc, etc.
Is that above code snippet correct? Do i need something like "ENDIF" or "end if" at the end? Does "y" have to be defined as "int32" or can it be something else?
Thanks for the help!

Is your formula node doing anything else besides what is shown?
Why don't you just wire the value that is going into it at X directly into the selector of the case structure?
Attachments:
Example_VI_BD.png ‏2 KB

How do I get an array output on a Formula Node?

My problem is simply that I cannot figure out how to get an output on a Formula Node to be an array. Documentation states that "you must declare local arrays and output arrays in the Formula Node" but doesn't say anything more than than. Attempts to put something like "float32 out[100];" for an array output called "out" fail.
If anybody knows how you can do this or even knows if it is possible I would appriciate your help.
Thanks,
Naveen

I found a typo in the formula node I was doing that was making the output not be an array. I don't think I was getting the error before but I was going to put together an example VI to attach here and I found it. So, thanks for your help even tho it was a stupid little mistake on my part.

"Eval Parsed Formula Node VI" does not return outputs in predefined order

I make a data analysis program, where the data consists of some million events and each event has e.g. 4 channels and 1-5 hits on each channel.
I would like the user to select different expressions of these channels to give coordinates to plot in a 2D histogram (increment a bin in Intensity Graph), e.g. for some experiment you want to show x=ch1-ch2; y=ch1+ch2+ch3+ch4; while in another experiment you want x=ch1-123; y=123-ch2;
There are other VIs that use static LabView-code for the normal things, but now after a few years of adding to this program I find that it would be quite good with a general plotter and let the user specify simple expressions like this. Also with the "normal" static plots, there is a need to filter out bad data and then it would be much simpler both to program and use if you could write text expressions with boolean logic to combine multiple filters. Making a LabView code and GUI that allows AND/OR/parenthesis combinations of expressions will be quite an effort and not very reusable.
So, with the above motivation, I hope you see that it would make sense to have a useable string formula evaluator in LabView. I find some info about MathScript's user-defineable functions, but haven't managed to get MathScript working in LV2010 yet (maybe some licensing or installation issues, I think I had it in 8.6). But I think it would be possible to do most of what I want for the display-part (not the filtering) with the simpler Eval/Parse Formula Node VIs and suitable use of the limited variable name space. So I started testing, and found a quite annoying issue with how the evaulator works.
To the parser, you are expected to send an array of output variable names. But then it ignores this array, and returns one output per assignment/semicolon in the formula, in the order of the formula text. Since the static parts of my program need to know what the output values mean (which of them is x and which is y), I would have to rely on the user not using any intermediate variable and defining x before y. The attached screenshot demonstrates the problem, and also that it has been solved by NI statff in the "Eval Formula Node VI" which does the appropriate array-searching to extract only the pre-defined outputs, in their expected order. But using that VI is about 100 times as slow, I need to pre-compile the formula and then only use the evaulator in the loop that runs over a million events.
I don't know if I'll take the time to make my own tweks to the parsing stage (e.g. preparation of array-mapping to not have to repeat the search or maybe hacking the output list generated by the parser) or if I'll have to make it in a static Formula Node in the block-diagram (which supports more functions), so that the user has to run with development environment and stop the program to change the plotting function. But I wanted to share this trouble with you, in hope of improvments in future LabView versions or ideas from other people on how I could accomplish my aim. I have MATLAB-formula node possibility too, but is far as I have seen the only place the user could update the formula would then be in a separate .m file, which is re-read only when typing "clear functions" in the Matlab console window. (Having this window is also an annoyance, and perhaps the performance of calling Matlab in every iteration is not great.)
Besides this issue, it also seems very strange there is virtually no support for conditional expressions or operators in Formula Node evaulated formulas (called Mathematics VIs in the documentation). Maybe using (1+sign(a-b))/2 you can build something that is 0 when a<b and 1 when a>b, but it is not user friendly! Would it really be diffcult to add a function like iif(condition, return_value_if_true, return_value_if_false) to replace the unsupported "condition ? if_true : if_false" operator? Would it really be difficult to add support for the < <= >= > == || && operators? Although compiled to an assemply language, this can't exactly be difficult for a CPU.
Attachments:
LV script test.png ‏62 KB
LV script test.vi ‏18 KB

(1) You can put any kind of code inside an event structure with the limitation that it should be able to complete quickly and without user interaction. For example a while loop for a newton-raphson method is fine, but an interactive while loop where the stop condition depends on user iteraction is not recommended. By default, event structures lock the front panel until the event completes, so you would not even be able to stop it.
(2) Yes, you can do all that. LabVIEW has no limitation as a programming language. Use shift registers to hold data and state information, the manipulate the contents as needed.
(3) I would recommend to use plain LabVIEW primitives instead of formula nodes. Show us your code so we can better see what it's all about. Obviously you have a mismatch betweeen scalars and arrays. It is impossible from the given information where the problem is.
(4) Yes, look inside the nonlinear curve fit VI (you can open it an inspect the code!). One of the subVIs does exactly that. Just supply your model VI.
LabVIEW Champion . Do more with less code and in less time .

Formula node

Hi everyone.
I have a problem with the formula node.
I want to calculate the following function:
Ux=xdot+Omegat*(y-y0)+thetax
where xdot, Omegat, y, thetax are all 1D arrays ( honestly these are waveforms but I convert these in arrays) while y0 is a constant. Also the output Ux has to be a 1D array.
When I implement the formula in the appropriate box I receive a message about
"array indexing indices".
For more information I link my .VI.
Some idea how to solve this task?
Second question:
Using the formula node is faster than realize all these operations in a
different way?
I ask it because I will need to include this .VI into 5
different "for loops"?
Thanks
matomato
Attachments:
formula node.vi ‏87 KB

Hi matomato,
fastest is usually to use LabView primitives! See attachment...
Edited: made another small change to your vi: changed y0 to DBL to avoid unneccessary datatype conversions!
Message Edited by GerdW on 02-27-2008 05:11 PM
Best regards,
GerdW
CLAD, using 2009SP1 + LV2011SP1 + LV2014SP1 on WinXP+Win7+cRIO
Kudos are welcome
Attachments:
formula node1.vi ‏87 KB

Formula node uInt32 overflow

I'm having a weird behaviour in formula node. The overflow of unsigned integer operations simply doesn't match the results of standard LabVIEW blocks as shown in the attached picture.
How can I deal with this problem? I need the behaviour of the block diagram. I could not do a right logical shift in unsigned integers in block diagram, so I'm using formula node (I could not figure out how to do it because I can't negate a unsigned value).
I'm using LV2012SP1. Were there any changes in this behaviour in newer versions?
Thank you.

I found a problem in hexadecimal and octal numeric representations of the formula node. The picture illustrates the issue. Three variables are assigned with the same value but outputs different results.
The hex and octal representations only work for half range of the 32 bits unsigned integer. When the value assigned gets beyond this limit the variable receives zero. This doesn't happen with a decimal value neither with 8 bits variables. Seems the problem doesn't only happen with overflow operations.

File Handling in formula node

Hello All,
I am trying to open a file and write some data to a file within the formula node. A declaration like
file *in;
produces this error:
Formula Node:undefined variable.
Error on line 1 is marked by a '#' character: "file# *out;"
How does one do file I/O operations within the formula node?
Thanks,
KB

Hello KB,
it seems you are trying to write a LabView program by
inserting C-code into a formula node... That's not the way it should be.
As far as I know there is no way to open a file inside the formula node. (It's a "formula" node!) So you have to open the file with LabView for reading and/or writing. Use your prefered subVi from the file palette. Perhaps you have to convert the data to a format suitable for your formula node (like string to U8-array).
In my experience it's better to use Labview-code instead of the formula node. In my programs a formula node was always slower than Labview code. So you should think about converting the C-style code to Labview.
Best regards,
GerdW
Best regards,
GerdW
CLAD, using 2009SP1 + LV2011SP1 + LV2014SP1 on WinXP+Win7+cRIO
Kudos are welcome

Parenthesis in formula node

I'm using the formula node and it appears that parenthesis don't work. Is this correct? What is the order of operations?
The Formula Node Syntax page in the help is extremely brief and sparse, bordering on useless for anyone that doesn't already now what is going on. Is there a more thorough description of the Syntax anywhere?

Thanks Dennis,
The error is just an incorrect calculation.
I put together a simple node and parenthesis did work. So something else must be amiss. I stripped down my VI and equation and made two formula nodes implementing the same equation with and without parenthesis. One gives an exponential decreasing curve, the other a constant array. Where is the error?
Attachments:
parenth_ex.vi ‏49 KB

How can I increment the index of an array in a formula node?

Hi,
I need some help with the formula node. I need an array, that increment every iteration of the code in the formula node. I need it to save amplitudes.
I thought i could do this with i++; i is the variable for my index.
Martin
Solved!
Go to Solution.

Hello,
does it sound well to you ?
@+
Tinnitus
ps: sorry posts mixed
Message Edité par tinnitus le 12-15-2008 03:05 AM
CLAD / Labview 2011, Win Xp
Mission d'une semaine- à plusieurs mois laissez moi un MP...
RP et Midi-pyrénées .Km+++ si possibilité de télétravail
Kudos always accepted / Les petits clicks jaunes sont toujours appréciés
Don't forget to valid a good answer / pensez à valider une réponse correcte
Attachments:
inc_node.vi ‏6 KB

For loop in formula node

Hi,
I wanna do some signal processing after I got the magnitude and phase infos from FFT. But the errors say that there is undefined variables. I am just wondering that here absP and angleP are both vectors. Can I express them like what I did in the attached vi? 3x.
Huihui
Attachments:
Untitled 12.vi ‏9 KB

Hi huihui,
do you just want to know the real and imaginary values of your signal?
Why use formula node at all?
Best regards,
GerdW
CLAD, using 2009SP1 + LV2011SP1 + LV2014SP1 on WinXP+Win7+cRIO
Kudos are welcome
Attachments:
Untitled 12.vi ‏13 KB

Precision of formula node?

Similar Messages

Maybe you are looking for