 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Math Forum / Math Software / MATLAB / July 2008RTW doubt n simulation time problem
Hi, I have made a simple DC motor model in simulink with current and velocity loop. The position output is converted into 17 bits and is sent serially to one pin of parallel port. Now I want to make this simulation in real time so I started reading documentation of RTW. After reading first few pages I get a picture that it requires a host PC and a target PC. The host PC will have Matlab which will generate a code for target PC. Then the target PC will run it in real time. Am I right? But I just want the simulation to run in realtime on my current PC having MATLAB. I don't have any target PC. Can I do this using RTW? Also the other problem I am facing is that the simulation is very slow. The DC motor model and loops are fast but converting position output to 17 bits and send them serially to parallel port seems to take a lot of time. This ia how I do it: 1) send the postion value to sample and hold block. 2) use a quantizer block and send the output to integer to binary conversion block. 3) The 17 bits are send to embbeded matlab function block. here I have written a long switch function which sends out one bit at each clock pulse. After all 17 bits have been sent a pulse is sent to sample and hold block to take the next sample. The quantizer, integer to binary converter and embebbed matlab function are in a triggered subsystem. While he sample and hold block is outside the triggered subsystem so I have to use data read write and store blocks to send pulse to sample and hold (Direct feedback to sample and hold from triggered subsystem gives an error) My question is: Is there a faster and better way to do this? Thanks Trupti Do you want your DC motor model to run in real time or the whole simulation (DC motor + controller) ? If you are building a so-called H(ardware)-I(n)-(the)-L(oop) simulator out of the motor, you might want to try xPC or real-time Windows option. The xPC requires a Target pc as you mentioned; the real-time windows uses your current pc as "target". hth GZ > Hi, > [quoted text clipped - 37 lines] > Thanks > Trupti Thanks.. I want to use the whole simulation in runtime (DC motor+ controller + encoder which has turned out to be a big block) Trupti "Guoliang Zhang" <zhangl@mathworks.com> wrote in message <g62lfu$4cs$1@fred.mathworks.com>... > Do you want your DC motor model to run in real time or the whole simulation > (DC motor + controller) ? If you are building a so-called [quoted text clipped - 47 lines] > > Thanks > > Trupti Sorry for mispelling: I want to run the whole simulation in realtime and not runtime. Trupti "Trupti Ranka" <trupti_ranka@yahoo.com> wrote in message <g64245$knj$1@fred.mathworks.com>... > Sorry for mispelling: I want to run the whole simulation in > realtime and not runtime. > Trupti Hi Trupti, Sorry if I am confusing things further, but if you want to communicate with an instrument/device via a serial port, TCPIP or UDP you can try using Simulink blocks that come along with Instrument Control Toolbox. However this won't work if your goal is to just simulate and not communicate with an external device. -Ankit Then i would suggest trying out the real-time windows target (a toolbox) for your described problem. but depending on how your model is constructed (complexity), the simulation may or may not be able to run in real-time. hth GZ > Thanks.. > I want to use the whole simulation in runtime (DC motor+ [quoted text clipped - 58 lines] >> > Thanks >> > Trupti I checked the RT block by Leonardo Daga. Using this you can run a fast simulation in real time but not a slow simulation. Is this true even with realtime workshop toolbox of Matlab? i.e. a faster simulation can be slowed to realtime but a slower simulation cannot be sped up to realtime. If this is the case then I will have to work on making my simulation faster. Okay I will give more detials of my work. We are working on servo modeling of an antenna. The present servo computer to control the antenna is very outdated so we want to upgrade it to PC104. Instead of testing PC104 directly on antenna we are thinking of testing it first on antenna simulation on MATLAB. The current and velocity loops for antenna are there in hardware (electronic cards) but the position loop is there in software (PC104). So my task is to simulate the anntena, current and velocity loops in matlab. The postion output will be given to PC104 which will complete the position loop. As we are not working on actual anntena we don't have encoder so I will have to simulate that also in Matlab. At present instead of whole anntena we are only considering only the motor with the loops and encoder simulation. This itself is becoming very slow for required clock frequency of encoder (about 15KHz). With the whole anntena model the things become infinitely slow!! Also we are not using TCP/UDP or any such communication protocol "Guoliang Zhang" <zhangl@mathworks.com> wrote in message <g651j7$afr$1@fred.mathworks.com>... > Then i would suggest trying out the real-time windows target (a toolbox) for > your described problem. [quoted text clipped - 4 lines] > hth > GZ The type of simulation you've described is called Hardware-in-the-Loop simulation: running a model that describes the physical system in realtime and test out the controller hardware. what RTW is capable of doing is to generate C code that describes your system to be modeled. Other add-ons from Mathworks (including xPC, real-time windows) or third-party vendor tools (OPalRT, dspace) provide solutions (hardware+software) to compile, download and run the RTW generated code onto certain real time target (PC, special hw, os etc.) and provides digital/Analog IO to interact with the controller circuit. These are different solutions that are priced differently with different capabilities and performance. If you decide to stick with RTW and matlab, you can start with xPC, which might be the best cost/performance for a starter project like what you are doing. just my 2 cents GZ >I checked the RT block by Leonardo Daga. Using this you can > run a fast simulation in real time but not a slow [quoted text clipped - 39 lines] >> hth >> GZ Thanks "Guoliang Zhang" <zhangl@mathworks.com> wrote in message <g681jj$gj0$1@fred.mathworks.com>... > The type of simulation you've described is called Hardware-in-the-Loop > simulation: running a model that describes the physical system in realtime [quoted text clipped - 14 lines] > > GZ |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2010 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.