decimation in fpga

Similar Questions

• FPGA Butterworth filter - why conversions to decimal fixed?

  I was looking through the screw FPGA Butterworth filter Express and stumbled on this code (after conversion the Express VI into a Subvi and then by opening the front panel):

  

  What is happening with the conversion to decimal fixed values that look like they must be integers (in the red boxes)? Why these conversions would be useful?

  Here is the micro-optimisations to help the synthesis tools to minimize the number of bits used in these code paths. The iteration Terminal is 32 bits wide and also involves a logic to do in the face of what is happening when the maximum value is reached. We have seen some benefits from the reduction of this path to the minimum width required both that which is written. I think it is probably still the case since the compiler usually does not know how many times a given line will run with a dynamic stop like this condition.

  The optimization of the meter width is similar, taking advantage of the fact that we know that we have limited storage space to use circular buffers power-of-2 size. Using the exact number of bits of the address allows us to implement a counter of reversal with no additional logic. otherwise, we would need to check the value, and add a mux (Select) to reset the counter when it reaches the maximum value.

  Synthesis tools usually do a good job of optimizing the unused bits, but here are two cases where we can help them by being more explicit about what we really need.

• Low pass FIR Filter on FPGA

  Hey guys,.

  I the DFD Toolbox and already built a few low pass FIR filter. I have a sampling rate of 50 kHz and I decimate it with a CIC filter with rate variable decimation. After the CIC filter, I need a high-grab for the low-pass filter FIR because I do want my DC signal. As I can change my rate of decimation, my FIR filter sampling rate changes also.

  The problem is that the coefficients of the filter cannot be changed during execution as the Butterworth IIR-filter function, including labview has implemented in the mathematics of the fpga function palette section.

  Are there examples how to build FIR filters where I can change the coefficients on the run?

  Think you can build a low pass FIR with the DFD toolkit and simply change the table with the coef. for a control? I could change them on the track...

  Greetz

  Slev1n

  Hi Slev1n,

  just found it in the community, maybe this might help you already:

  Polyphase Interpolation FIR Filter on FPGA with Diabaté and Coregen

  https://decibel.NI.com/content/docs/doc-16650

  Greetings

  Michael

• atan2 in FPGA of point sets

  I use the function atan2(x,y) implemented on a FPGA (OR-7851R) and have a simple question that frustrates me.

  See the attached vi, my two inputs (x, y) are fixed numbers in decimal, length of 16-bit words, 1 whole bit (this is to say, - 1.000000, 0.9999695 with 3.051758e - 5 precision).

  Since atan2 must be a value between - pi and pi and labview outings pi radians (and not simple radians), I would expect the output to be the same format as the input, a 16-bit number between - 1 and 0.9999695 (3.051758e - precision 5) with 1 single bit used to the whole.

  Instead, the VI requires to be a word of 16 bits, but with the whole part of 2 bits (that is to say, - 2.0000, 1.999939 with 6.1036e - 5 precision, see attached photo atan2_output_type.png). This earns me nothing and I have only half the accuracy I should have! I tried several entries and I confirmed it returns only numbers between-1 and 0.9999389 (6.1036e - 5 precision).

  I can't seem to be able to change it because all the options in the dialog box are grayed out (see attached photo atan2_dialogue.png), as he decided that's what he must return in light of my entries. Can someone explain why it does this? Is there something simple that I'm missing? I need that little extra to go on decimals, not the whole of my fixed point number side.

  Thank you

  Dan

  Hi Dan,.

  Normally, a little like that growth is due to a case of corner, I guess in this case it takes cover 1.0 pi radians as an output as possible. If you need more precise output, just to increase the word length; with a fixed whole word length, all the bits that you add will be fractions. If you don't want the whole extra bit, just throw it away using fixed point to set the length of whole word to 1 and the overflow mode to Wrap.

  Jim

• PCIe-7842R (series R FPGA) digital output does not work properly

  Greetings,

  I'm having some problem show TTL the correct voltage with my PCIe-7842R FPGA board.

  The block diagram of my code FPGA LV Moose appears in "analog - digital .png '. The idea was to convert an analog input (decimal value) to a binary code and 16-bit output by 16 DIO ports. I use the connection block SCB-68 has as the terminal and trendy on the FPGA 1 connector RDIO with SHC68-68-RDIO shielded cable.

  The compiled code ok. But during the test, I noticed that some ports has no output TTL levels correctly. For example, for input 1000 decimal, I would expect binary code 0000001111101000. However, some ports (DIO #6, #7, #9, etc.), which are supposed to ~3.3V (1 digital) high TTL output, output actually 0.8V. I have attached the result measured in 'exit digital test.png '.

  To ensure that the question was not because of the code of the LV, I did some more tests on DIO #6 with a simple example (simple digital output.png). The output was ~ 1V this time at the digital 1.

  It's really confusing because of the digital Edition is supposed to be simple. I used the same FPGA card for controlling roller shutters with TTL signals before and it worked fine.

  Does anyone have similar problems? Any suggestions are greatly appreciated.

  iron_curtain wrote:

  DIOs are connected to a controller digital galvo Cambridge Tech. But I measured the voltage at the terminals of the connector block.

  If you unplug the controller galvo DIOs, do they look good (have the right voltages).  Do you know how many of these entries to the need for controller?  I think you hit the limit the total current available for EID within the Council.

• return only the decimal value

  What function can I use to return only the decimal value of a formula

  (97 * 2.2) / 14 value returns 15.24

  I only want to use le.24

  Salvation is

  Use the formula - INT (formula)

  In the special case above:

  =(97*2.2)/14 - INT ((97*2.2)/14)

  You see two places to the right of the decimal separator in reason to be rounded off to the value real, in the shape of the cell to only show as two decimal places, either by setting the width of the cell close enough to force the rounding of the displayed value.

  Note that if the result is rounded (using the ROUND function) the actual value of the cell will be the rounded value. For the other two cases, the value displayed indicates only two decimal places, but the real value in the cell (and the additional calculations used by referring to this cell) is 0.242857142857144, which could lead to surprises in the results of the calculations downstream.

  Kind regards

  Barry

• First HP: Hms to decimals then back, decimal to HMS

  Hi everyone, have the problem, when I use →HMS (53.1301), sometimes the answer is 53.0748.

  Sometimes it's 53 ° 07'48 ". 36, I don't know what I'm doing for it to change.both is correct"

  Also can not get HMS→ (53.0748) to work at all continues, 53.0748, RPN does not have it, so I guess only, does not work in the entry, manual RPN and algebraic → HMS works, so what's happening with HMS→. Have looked in the manual, can not find.

  Thanks for any help.

  surveying101bil

  Hi!, @surveying101bil:

  See, this image...

  

  See...

  Syntax: →HMS (value) converts a decimal to hexagesimal, format; in other words, in units divide, in groups of 60. This includes the degrees, minutes and seconds, but as hours, minutes, and seconds.

  Example: →HMS (54.8763) returns 54 ° 52′34. 68

  Syntax: HMS→ (value)

  Converts a value expressed in the format of hexagesimal, in decimal format.

  Example: HMS→ (54 ° 52′34. 68) returns 54.8763.

• Do not use the correct decimal value

  I am using a cell to calculate a simple equation, but the value is incorrect, because program seems to use the wrong decimal length to calculate the answer.

  A cell is formatted with 2 decimals with a value of 54.55%. (Formula for this cell is 90000/165000)

  B cell is formatted in a monetary value of $75,000 with 0 decimal.

  Cell C's function cell A = * B cell with currencies with 0 decimal formatting.

  Cell C should multiple 54.55% $75,000, resulting in $40 913 in cell C, but the result is always $40 909. When I manually type =. 5455 * 75000 in another cell, I will get the result of 40913. So I guess the program is multiplied by the actual value of the % and not the number rounded to 2 decimal places. However this change so the value will be correct?

  cklat wrote:

  So I guess the program is multiplied by the actual value of the % and not the number rounded to 2 decimal places. However this change so the value will be correct?

  Use the ROUND function:

  

  SG

• Satellite L505 - decimal button and '8' pad does not

  Hello

  My buttons decimal and 8 on the numeric keypad of my laptop does not work. It seems to work after a few minutes of bashing alt + ctrl + shift + fn + keyboard arrows keys, then as soon as I move away from my computer it stops working again. does anyone have the solution button to turn on? also, does anyone know how to disable this?

  Thank you

  > can someone give me a link for the drivers for the keyboard? satellite l505d? I can't locate them.

  Keyboard drivers are part of the Windows operating system. You can not download this driver separately
  Usually, if remove you the keyboard from the Device Manager, then reboot again, the Windows operating system should recognize and automatically install the keyboard driver.

  > the key 8 has worked now... I was just overwhelming the shiftaltctrlarrows buttons.
  But it seems that the buttons work now

• First HP - bug of Fraction / decimal?

  I bought my first HP Calculator about a month ago and its pretty awesome summer for most. I noticed a problem last week trying to get a decimal of a fraction like 1/101, where the answer it gives me is 9.90099009901E - 3.  On a standard calculator UNITS, I would get the correct decimal: 0.00990099009901, that tells me that the comma is disabled by three digits. I tried both CASES/nonCAS environments, and its giving the same answer. Is there something I can do on my first to correct this problem, or a workaround?

  Thank you

  Joshua

  Hello

  The comma is not off here at all. Part E-3 at the end means that it is your result with a 10 x 12-digit ^ power-3.

  1E3 is the same as 1000. 1E-3 signifie.001. Not how in positive E3, he moved the value 10 s 3 places. The E - 3 he moves down to 10 s 3 places.

  In "standard mode", rather than affichage.00990099009901, it removes the first few zeros and adds the exponential part. Why we do this? Well, imagine if you have the number 6.022E23 or 6.022 x 10 ^ 23 (it is a very famous number used in chemistry in almost every calculation!). It would be really tedious and likely to cause errors, who write in the form 602200000000000000000000 - wouldn't it?

  It is very common to write numbers in this way it makes it actually easier to compare, to write and to understand. What do I mean? Try this - which is the smallest of these numbers: 0.0000009347374, 0.00000543422 or 0.000003542? Have you had to count the zeros?

  What now? 9.347374E - 7, 5.43422E - 6 or 3.542E - 6

  Well Yes, the first is the smallest! In short, you can see who is the smaller number. (see https://www.mathsisfun.com/numbers/scientific-notation.html if you want to know a little more)

  However, if you really don't like it, you can set your screen to be the ' round '. Go to your home settings ([SHIFT] [HOME]) and change the 'Standard' to 'round. " This will adjust your something numbers comme.000012345678 displayed comme.00001235 instead if you prefer this type of display.

  You still want to learn what E < number > consisting of a medium however, number because even in rounded form he will need to show results in scientific notation when the number becomes sufficiently large or small that she does more!

• RM-YD022 send any decimal point. What is the solution?

  RM-YD022 send any decimal point. What is the solution?

  Hello
  Welcome to the community of Sony.

  The button of the decimal point can be used with the number buttons 0-9 and press ENT to select digital channels. For example, to enter in 5.1, press [5] [.] [1] and ENT.
  The television receives digital channels?

• Decimal years sum in figures

  Hello.

  I have two questions that work together:

  1. How can I make figures show '1.6 years"years, months and days? As '1y 7 m 3d. "

  2. in addition, how I sum "1.6" years until 2016, for example?

  Thank you very much

  Luiz

  Regarding (1)

  Integrated numbers duration format is not the same thing as you want, so you'll have to create your own format or use a formula.

  to create a format with a formula, you can use the following text:

  

  Enter the decimal number of years in cell A2

  B2 = A2 × 365

  It's shorthand dethrone select cell B2 and type (or copy and paste it here) the formula:

  = A2 × 365

  C2 = INT (B2÷365)

  D2 = INT ((B2−C2×365) ÷7)

  E2 =(B2−C2×365−D2×7)

  recommend that you try the YYYY MM DD format because the number of months of change and the formulas will be more complicated.

  WRT (2)

  I do not understand the question

  If you mean how you add up the values in A W J format (that I presented above), the answer is that you want to convert A M J entry each day, and then add the days then convert back

  

  in order to add decimal years entered in the column, you can add the decimal years, then convert or add days, then convert them

  line 7 is a footer line

  B7 = SUM (B)

  Fill the formula C2, D2 and E2 down by selecting the cells C2 to E2, place the cursor on the lower edge of the selection, and then drag the yellow control downwards as needed

• Possible to use the main clock FPGA as a basic Source of time for the NI 9234?

  Hello

  I am running a NI 9234, and NI 9222 module on a cRIO-9114 chassis with a real-time 9022 controller, data acquisition on the 8 channels on the full sample rate, using the FPGA.

  My problem is the synchronization of the data, where the NI9234 is usually a late acquisition, compared to the NI9222. I tried manually delay the NI9222 to take the first samples, but so far without success.

  To solve this problem and also to ensure a rate of data more comparable on both cards, I was wondering is it possible to create a Source of time Base master for the NI9234 WITH THE FPGA (opposite to import from another NI9234 or similar module)? The NI9222 I already provide with sampling of the impulses from the FPGA (as expected with this module).

  Another issue would be the characteristics of the exact chronology of the two modules, including in terms of synchronization between the beginning of the acquisition and the first samples of data. I couldn't find any information about it...

  Thank you in advance,

  O Hoppe

  Hello!

  I've never heard of a possibility to use a time base FPGA with a module of Delta-Sigma.

  But I think that the question should be what has not worked to delay the signal and how you did it. Have you read this article?

  How can I compensate for delays of different group with the C Series Modules in LabVIEW FPGA?
  http://digital.NI.com/public.nsf/allkb/74EB238E1BCADD528625735300681A7D

  Can you give a more detailed description of your trial to delay acquisitions? Can you provide a code example?

  Best regards

  Christoph

• LabVIEW FPGA SPI accident - SPI OR IP address for example - R series OR

  Hello

  I am trying to run the series R - sample project NI SPI FPGA Simulation.lvproj that comes with the SPI IP OR on a real FlexRIO FPGA SMU-7976R target with an attached digital adaptation NI 6581 B Module. The example is for a PCIe-7841R but I wore during my target FPGA, follow these steps and made additional changes to try to make it work with my set-up. I learned that FlexRIO FAMs CLIPs do not work with nodes in office had so I know I can't simulate the project originally planned so I will try to use FPGA to e/s node host side (open FPGA vi reference) to implement the actual hardware.

  My first question concerns my configuration of the adapter module e/s and selection. I added the IO Module (NI 6581 B: NI 6581 B channel) for my project and selected the channels as shown in the table below. I have a real physical hardware connection as described below using two NI SHC68-C68-D4 cables and a break-out Board.

  

  I changed the names as well:

  

  I selected these DIO channels because I wanted the DDCA connector to be the master and the DDCB connector to be the slave. In addition, in this CLIP every eight channels of i/o has a write enable signal. I have not used the Port configuration because I needed 4 available DIO channels and I saw DIO0-3. Is my logic of selection of channel vs correct Port here?

  Following the same strategy that examples FlexRIO/NI6581B, I changed the FPGA.vi to include initialization outside of timed loops:

  

  And in the Host.vi I have a node reference FPGA and wired loops of feedback accordingly:

  

  When I compile the FPGA and try to run Host.vi LabVIEW inevitably crashes with Crash Reporter below and must restart:

  

  Does anyone know what I'm doing wrong here? My guess is that it has something to do with the CLIP/IOModule. Any help is appreciated.

  Thank you

  Scott

  Hi Scott,.

  I suspect that the problem might be related to the fact that your node open FPGA VI reference is in the while loop and trying to open a new FPGA reference at each iteration.
  If you move outside of the while loop, it does not solve the crash?

  On channel vs port question, your logic seems reasonable to me, but I recommend to try it since this feature could depend on device.

• SMU 7857R FPGA and DAQ task

  I'm looking at the SMU 7857R.  I would use the 8 analog inputs and the FPGA to do some annalysisy and DIO based on the best data.  I was wondering if it is possible to also define a DAQ task so that all analog data will be made available to the host.  Using a data acquisition task would be easier for me to have to write code fpga for use a DMA fifo to send back data.

  Hi bcat!

  A data acquisition task is only for devices Data Acquisition product supported by driver DAQmx as X-, M-, S-, maps E-Series. The boards of the R series are only supported by the driver OR RIO.
  If you don't need pre-treatment on the FPGA so that you can distribute simple data through DMA on your host. If you are looking for an API for the comparable FPGA with the DAQmx API then please check the NI CompactRIO waveform reference library. You can also use the library for R Series boards.