My embedded hardware is S3C2440 Samsung ARM chip, and flash is 256M.

Before i using qt-4.6.3, i am always using qt-4.5, and without qt desktop. But
when i have used qt-4.5, qwt graphic has never display in my hardware, so i have
tried qt-4.6.3 and with qt4 desktop. If you do not change code 
qt-everywhere-opensource-src-4.6.3/src/gui/embedded/qscreenlinuxfb_qws.cpp as
what i said. Maybe generate segmentation fault error.

By the way, my old u-boot problem has not supported burned 256M, so need change
u-boot for supporting burned image of system greater than 64M

--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> Good to hear that you could fix the problem.
>
> With Maemo, MeeGo or Android such a problem dit not came up
> because there it was not necessary to recompile Qt by ourself.
>
> Is it really necessary to change the sourcecode in
> qt-everywhere-opensource-src-4.6.3/src/gui/embedded/qscreenlinuxfb_qws.cpp ?
> Isn't there a switch to "./configure" that will do the trick ?
> Or do you use embedded hardware on which Qt never has been run and tested
before ?
>
>
> --- In pvbrowser@yahoogroups.com, "oceancalmcool" <yizhang.018@> wrote:
> >
> > This problem has fixed
> > I have transfer QT4 desktop to ARM and rebuild qt4 source code
> > Need rebuild qt-everywhere-opensource-src-4.6.3 source code
> > Chang code
qt-everywhere-opensource-src-4.6.3/src/gui/embedded/qscreenlinuxfb_qws.cpp
> > if you do not change code, will cause segmentation fault error, this error
is about linuxfb.
> > 424 line:
> >
> > /*
> >     canaccel = useOffscreen();
> >     if(canaccel)
> >         setupOffScreen();
> > */
> >
> > 726 line:
> >
> > /*
> >     if (canaccel) {
> >         *entryp=0;
> >         *lowest = mapsize;
> >         insert_entry(*entryp, *lowest, *lowest);  // dummy entry to mark
start
> >     }
> > */
> >
> > ./configure -prefix /opt/qt-embedded -release -opensource -shared -fast
-no-largefile -qt-sql-sqlite -no-qt3support -no-xmlpatterns -no-mmx -no-3dnow
-no-sse -no-sse2 -no-svg -no-webkit -qt-zlib -qt-gif -qt-libtiff -qt-libpng
-qt-libjpeg -make libs -nomake examples -nomake docs -nomake demo -no-nis
-no-cups -no-iconv -no-dbus -no-openssl -xplatform qws/linux-arm-g++ -embedded
arm -little-endian -qt-freetype -depths 16,18 -qt-gfx-linuxfb
-qt-gfx-transformed -qt-gfx-multiscreen -no-gfx-vnc -no-gfx-qvfb
-qt-kbd-linuxinput -no-glib -qt-mouse-tslib -confirm-license
> > make && make install
> >
> > So after do that, you can copy libqwt library to /opt/qt-embedded/lib. This
is will work.
> >
> >
> > --- In pvbrowser@yahoogroups.com, "oceancalmcool" <yizhang.018@> wrote:
> > >
> > >
> > > I will try it immediately.
> > > Thank you for your support.
> > >
> > > --- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@> wrote:
> > > >
> > > >
> > > >
> > > > > > Do you have a pvbrowser client running on your Arm platform?
> > > > > > Does it display the non qwt widgets?
> > > > > > Try running the pvserver and the pvbrowser client
> > > > > > on different systems and connect them over the network.
> > > > > > (1) pvserver on Arm and pvbrowser on PC
> > > > > > (2) pvserver on PC and pvbrowser on Arm
> > > > > > What is the result?
> > > > >
> > > > > pvserver on Arm and pvbrowser on both ARM and PC
> > > > > 1) Without qwt widgets, ARM and PC pvbrowser can display widgets well
> > > > > 2) With qwt widgets, ARM pvbrowser can not display qwt widgets, but PC
pvbrowser
> > > > > can display it well, PC OS is ubuntu 10.4.
> > > > > I have successfully compile libqwt.a both in ARM and in PC.
> > > > > This is very strange.
> > > > > By the way, i have tried qwt code together with pvbrowser,
> > > > > and have download qwt
> > > > > 6.0 source code and using arm_linux_gcc 4.3.3 rebuild qwt 6.0 code. In
that
> > > > > source code directory, i have found some example, i have compiled, in
PC can run
> > > > > it such as dials program, but using arm_linux compilor to make dials,
there is
> > > > > no error happen, but download it to embeadded system, that is can not
work again
> > > > > except a cursor.
> > > >
> > > > 1) The problem must be within the client
> > > >    because you run the server on ARM in both cases.
> > > >
> > > > 2) The problem could be a different byte order on ARM and x86.
> > > >    But qwt already works on Maemo and Android where ARM is also used.
> > > >    Thus the byte order can't be a problem.
> > > >
> > > > 3) It could be a problem with a special GCC version but i do not think
so.
> > > >
> > > > On Maemo we use GCC on a Ubuntu VM which is cross compiling to ARM code
> > > > [sbox-FREMANTLE_ARMEL: ~/maemo-pvb] > g++ --version
> > > > sbox-arm-none-linux-gnueabi-g++ (GCC) 4.2.1
> > > > Copyright (C) 2007 Free Software Foundation, Inc.
> > > > This is free software; see the source for copying conditions.  There is
NO
> > > > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE.
> > > >
> > > > For Android we used
> > > > android-lighthouse-4.8.0 GCC for Android
> > > >
> > > > 4) You have tried qwt 6.0 and qwt 5.2 we currently use.
> > > >    The version should not be the problem
> > > >    Please use qwt 5.2 since we have already prooved it works on ARM
> > > >
> > > > -----------------------------------------------------------
> > > >
> > > > Thus i currently do not have a good idea
> > > >
> > > >
> > > > You said:
> > > > > there is no error happen, but download it to embeadded system, that is
can not work again
> > > > > except a cursor.
> > > > Did you use the static qwt library on the embedded system ?
> > > > If you used the dynamic library on the embedded system it might not find
the library ?
> > > > Obviously the Qt libraries are found on the embedded system because
otherwise pvbrowser would not work.
> > > >
> > > > I would continue as follows:
> > > > 1) Create a very small test application that only shows 1 simple qwt
widget.
> > > >    That is something like:
> > > > int main(int argc, char *argv[])
> > > > {
> > > >   QApplication app(argc, argv);
> > > >   QwtSlider test(0,1);
> > > >   test.show();
> > > >   return app.exec();
> > > > }
> > > >
> > > > 2) First test it on the PC
> > > >
> > > > 3) Test it on the embedded system using the static qwt library
> > > >
> > >
> >
>

About configuration of qt source code, i think this is only need my hardware
requirement, maybe other hardware need change of this configuration.


--- In pvbrowser@yahoogroups.com, "oceancalmcool" <yizhang.018@...> wrote:
>
>
> My embedded hardware is S3C2440 Samsung ARM chip, and flash is 256M.
>
> Before i using qt-4.6.3, i am always using qt-4.5, and without qt desktop. But
when i have used qt-4.5, qwt graphic has never display in my hardware, so i have
tried qt-4.6.3 and with qt4 desktop. If you do not change code 
qt-everywhere-opensource-src-4.6.3/src/gui/embedded/qscreenlinuxfb_qws.cpp as
what i said. Maybe generate segmentation fault error.
>
> By the way, my old u-boot problem has not supported burned 256M, so need
change u-boot for supporting burned image of system greater than 64M
>
> --- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@> wrote:
> >
> >
> >
> > Good to hear that you could fix the problem.
> >
> > With Maemo, MeeGo or Android such a problem dit not came up
> > because there it was not necessary to recompile Qt by ourself.
> >
> > Is it really necessary to change the sourcecode in
> > qt-everywhere-opensource-src-4.6.3/src/gui/embedded/qscreenlinuxfb_qws.cpp ?
> > Isn't there a switch to "./configure" that will do the trick ?
> > Or do you use embedded hardware on which Qt never has been run and tested
before ?
> >
> >
> > --- In pvbrowser@yahoogroups.com, "oceancalmcool" <yizhang.018@> wrote:
> > >
> > > This problem has fixed
> > > I have transfer QT4 desktop to ARM and rebuild qt4 source code
> > > Need rebuild qt-everywhere-opensource-src-4.6.3 source code
> > > Chang code
qt-everywhere-opensource-src-4.6.3/src/gui/embedded/qscreenlinuxfb_qws.cpp
> > > if you do not change code, will cause segmentation fault error, this error
is about linuxfb.
> > > 424 line:
> > >
> > > /*
> > >     canaccel = useOffscreen();
> > >     if(canaccel)
> > >         setupOffScreen();
> > > */
> > >
> > > 726 line:
> > >
> > > /*
> > >     if (canaccel) {
> > >         *entryp=0;
> > >         *lowest = mapsize;
> > >         insert_entry(*entryp, *lowest, *lowest);  // dummy entry to mark
start
> > >     }
> > > */
> > >
> > > ./configure -prefix /opt/qt-embedded -release -opensource -shared -fast
-no-largefile -qt-sql-sqlite -no-qt3support -no-xmlpatterns -no-mmx -no-3dnow
-no-sse -no-sse2 -no-svg -no-webkit -qt-zlib -qt-gif -qt-libtiff -qt-libpng
-qt-libjpeg -make libs -nomake examples -nomake docs -nomake demo -no-nis
-no-cups -no-iconv -no-dbus -no-openssl -xplatform qws/linux-arm-g++ -embedded
arm -little-endian -qt-freetype -depths 16,18 -qt-gfx-linuxfb
-qt-gfx-transformed -qt-gfx-multiscreen -no-gfx-vnc -no-gfx-qvfb
-qt-kbd-linuxinput -no-glib -qt-mouse-tslib -confirm-license
> > > make && make install
> > >
> > > So after do that, you can copy libqwt library to /opt/qt-embedded/lib.
This is will work.
> > >
> > >
> > > --- In pvbrowser@yahoogroups.com, "oceancalmcool" <yizhang.018@> wrote:
> > > >
> > > >
> > > > I will try it immediately.
> > > > Thank you for your support.
> > > >
> > > > --- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@> wrote:
> > > > >
> > > > >
> > > > >
> > > > > > > Do you have a pvbrowser client running on your Arm platform?
> > > > > > > Does it display the non qwt widgets?
> > > > > > > Try running the pvserver and the pvbrowser client
> > > > > > > on different systems and connect them over the network.
> > > > > > > (1) pvserver on Arm and pvbrowser on PC
> > > > > > > (2) pvserver on PC and pvbrowser on Arm
> > > > > > > What is the result?
> > > > > >
> > > > > > pvserver on Arm and pvbrowser on both ARM and PC
> > > > > > 1) Without qwt widgets, ARM and PC pvbrowser can display widgets
well
> > > > > > 2) With qwt widgets, ARM pvbrowser can not display qwt widgets, but
PC pvbrowser
> > > > > > can display it well, PC OS is ubuntu 10.4.
> > > > > > I have successfully compile libqwt.a both in ARM and in PC.
> > > > > > This is very strange.
> > > > > > By the way, i have tried qwt code together with pvbrowser,
> > > > > > and have download qwt
> > > > > > 6.0 source code and using arm_linux_gcc 4.3.3 rebuild qwt 6.0 code.
In that
> > > > > > source code directory, i have found some example, i have compiled,
in PC can run
> > > > > > it such as dials program, but using arm_linux compilor to make
dials, there is
> > > > > > no error happen, but download it to embeadded system, that is can
not work again
> > > > > > except a cursor.
> > > > >
> > > > > 1) The problem must be within the client
> > > > >    because you run the server on ARM in both cases.
> > > > >
> > > > > 2) The problem could be a different byte order on ARM and x86.
> > > > >    But qwt already works on Maemo and Android where ARM is also used.
> > > > >    Thus the byte order can't be a problem.
> > > > >
> > > > > 3) It could be a problem with a special GCC version but i do not think
so.
> > > > >
> > > > > On Maemo we use GCC on a Ubuntu VM which is cross compiling to ARM
code
> > > > > [sbox-FREMANTLE_ARMEL: ~/maemo-pvb] > g++ --version
> > > > > sbox-arm-none-linux-gnueabi-g++ (GCC) 4.2.1
> > > > > Copyright (C) 2007 Free Software Foundation, Inc.
> > > > > This is free software; see the source for copying conditions.  There
is NO
> > > > > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE.
> > > > >
> > > > > For Android we used
> > > > > android-lighthouse-4.8.0 GCC for Android
> > > > >
> > > > > 4) You have tried qwt 6.0 and qwt 5.2 we currently use.
> > > > >    The version should not be the problem
> > > > >    Please use qwt 5.2 since we have already prooved it works on ARM
> > > > >
> > > > > -----------------------------------------------------------
> > > > >
> > > > > Thus i currently do not have a good idea
> > > > >
> > > > >
> > > > > You said:
> > > > > > there is no error happen, but download it to embeadded system, that
is can not work again
> > > > > > except a cursor.
> > > > > Did you use the static qwt library on the embedded system ?
> > > > > If you used the dynamic library on the embedded system it might not
find the library ?
> > > > > Obviously the Qt libraries are found on the embedded system because
otherwise pvbrowser would not work.
> > > > >
> > > > > I would continue as follows:
> > > > > 1) Create a very small test application that only shows 1 simple qwt
widget.
> > > > >    That is something like:
> > > > > int main(int argc, char *argv[])
> > > > > {
> > > > >   QApplication app(argc, argv);
> > > > >   QwtSlider test(0,1);
> > > > >   test.show();
> > > > >   return app.exec();
> > > > > }
> > > > >
> > > > > 2) First test it on the PC
> > > > >
> > > > > 3) Test it on the embedded system using the static qwt library
> > > > >
> > > >
> > >
> >
>

About MODBUS data aquisition.

I think the reason is our linux kernel has not supported RS485 driver, because i
design modbus protocal over RS485. So i need change linux kernel code first to
support RS485.

By the way, our linux kernel only support RS232.

Is it right ?

--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> <a_la_wachinwei@> wrote already good suggestions
>
> Here is another suggestion from me:
>
> Use the command "stty" to show and set the tty parameters.
> For example show the parameters with:
> stty -F /dev/ttyXYZ
>
> Use a null modem cable to connect 2 PC, or your embedded device with the PC.
> On one side use a terminal emulator like
> http://en.wikipedia.org/wiki/Minicom
> and on the other side type
> ls > /dev/ttyXYZ
>

There is no such thing like a "RS485" or "RS232" driver.
You only have a driver for a serial tty device.
Example: /dev/ttyUSB0 is the tty device i get when i connect a USB / RS232
adapter.

The difference of RS485 / RS232 is a different physical layer according to the
ISO OSI model.
RS232 is for point to point communication with +-12V signal voltage.
RS485 is for a multi drop bus communication with twisted pair cables and a
differential signal voltage.

Normally you have a RS232 line on your computer.
Now you can use a RS232 / RS485 adapter.
http://www.rs232-converters.com/rs232_to_rs485_converters.htm

The data link layer according to ISO OSI model is the same in both cases. It is
defined by baudrate/bits per char/start bists/stop bits/parity.

The network layer is empty because there is no routing at all.

We can say that we have a transport layer according to ISO OSI.

The difference between is
with RS232 you can access the wire without a collision.
with RS485 we use a maste/slave protocol so that no collisions on the wire will
occur.

Now Modbus can be thought as an application protocol (ISO OSI layer 7) that uses
the transport (ISO OSI layer 4) for communication.
Our rlModbus class would be the API for the application protocol.

If you use a Null Modem Cable
http://en.wikipedia.org/wiki/Null_modem
you can connect a PC to your box.
Then you must verify that you can exchange data over that tty device.
This could be done with http://en.wikipedia.org/wiki/Minicom under Linux or
Hyperteminal under Windows.
It should also be possible to simply copy text to/from the tty device as long as
baudrate etc. is the same on both sides.

1 computer:
cat send.txt > /dev/ttyXY

2 computer:
cat /dev/ttyYZ > receive.txt

> About MODBUS data aquisition.
>
> I think the reason is our linux kernel has not supported RS485
> driver, because i design modbus protocal over RS485.
> So i need change linux kernel code first to support RS485.
>
> By the way, our linux kernel only support RS232.
>
> Is it right ?

Hi Master,

I think i do not describe my testing environment clearly.

I have one temperature sensor which use RS485 connect to my embedded box and
take one serial port. And my box has two serial port, another serial port
connect to my PC using RS232, PC control embedded box by this serial port.

When system startup,  in /dev directory, will generate two virtual device such
as tq_serial0 and tq_serial1, but in linux kernel this two device both only
suppor RS232, so when modbus deamon program starting, it will open one serial
device, anyway, neither these two serial device can not support RS485.

Up till now i do not want to use RS232/RS485 adapter. Because most physical
device of industry use RS485.

I only want to modify linux kernel code of serial driver to support RS485

I think it is feasible.

Thank you for your information

--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> There is no such thing like a "RS485" or "RS232" driver.
> You only have a driver for a serial tty device.
> Example: /dev/ttyUSB0 is the tty device i get when i connect a USB / RS232
adapter.
>
> The difference of RS485 / RS232 is a different physical layer according to the
ISO OSI model.
> RS232 is for point to point communication with +-12V signal voltage.
> RS485 is for a multi drop bus communication with twisted pair cables and a
differential signal voltage.
>
> Normally you have a RS232 line on your computer.
> Now you can use a RS232 / RS485 adapter.
> http://www.rs232-converters.com/rs232_to_rs485_converters.htm
>
> The data link layer according to ISO OSI model is the same in both cases. It
is defined by baudrate/bits per char/start bists/stop bits/parity.
>
> The network layer is empty because there is no routing at all.
>
> We can say that we have a transport layer according to ISO OSI.
>
> The difference between is
> with RS232 you can access the wire without a collision.
> with RS485 we use a maste/slave protocol so that no collisions on the wire
will occur.
>
> Now Modbus can be thought as an application protocol (ISO OSI layer 7) that
uses the transport (ISO OSI layer 4) for communication.
> Our rlModbus class would be the API for the application protocol.
>
> If you use a Null Modem Cable
> http://en.wikipedia.org/wiki/Null_modem
> you can connect a PC to your box.
> Then you must verify that you can exchange data over that tty device.
> This could be done with http://en.wikipedia.org/wiki/Minicom under Linux or
Hyperteminal under Windows.
> It should also be possible to simply copy text to/from the tty device as long
as baudrate etc. is the same on both sides.
>
> 1 computer:
> cat send.txt > /dev/ttyXY
>
> 2 computer:
> cat /dev/ttyYZ > receive.txt
>
> > About MODBUS data aquisition.
> >
> > I think the reason is our linux kernel has not supported RS485
> > driver, because i design modbus protocal over RS485.
> > So i need change linux kernel code first to support RS485.
> >
> > By the way, our linux kernel only support RS232.
> >
> > Is it right ?
>

> I only want to modify linux kernel code of serial driver to
> support RS485
>
> I think it is feasible.

If your system software (linux driver) supports RS232 it will support RS485 also
(I am sure it does). For the driver both are the same.

The difference is you need different hardware on your port.

Either a port with RS232 like MAX232
http://www.maxim-ic.com/datasheet/index.mvp/id/1798
see the "diagram" there.

Or a port with RS485 like ADM485
http://www.analog.com/en/interface/rs-485/adm1485/products/product.html
also see "Functional Block Diagram for ADM1485" there

If your port has only a RS232 driver you can convert the electrical signals from
RS232 level to RS485 level by using an external adapter like
http://www.rs232-converters.com/rs232_to_rs485_converters.htm

hello all,
i need someone/s to assist with the development of a mining application. to
monitor processing plant with 8 different conveyor belt outputs and 5 heavy duty
vehicles ie. bulldozers, frontend loaders and dumptrucks.
if anyone has experience in this field plse let me know with basic detail of the
development that was done. also, i will appreciate having a developer from south
africa (cape town) on the team. i hav a mis background and not development
experience.
thankyou kindly
mario

> i need someone/s to assist with the development of a mining
> application. to monitor processing plant with 8 different conveyor
> belt outputs and 5 heavy duty vehicles ie. bulldozers,
> frontend loaders and dumptrucks.
The software doesn't care about the area in which you want to deploy your
solution in.

You will have to select the hardware that is suited for mining.

For example the
http://en.wikipedia.org/wiki/IP_Code
that your hardware must be approved to.
Think about wiring in copper or fiber optics.

Hello:
We were doing some tests with a Modicon PLC-Telemecanique, we read WORD type
variables to position 1024 (% MW1024 in the PLC).
The question is how we read the positions are on these,%MW1025 to %MW2999 ?.
Also could not access the variables of type DWORD, REAL AND EBOOL.
We use rllib configuring the daemon with MODBUS over TCP.
If anyone has had any experience with this type of configuration I appreciate
any suggestions.
What is striking is that the remote management is working with P-CIM and read
these variables.

Thanks Carlos




--- In pvbrowser@yahoogroups.com, "kurt_braun" <kurt_braun@...> wrote:
>
> Hello,
> I'm not able to read data from my Modicon M340 PLC via the modbus_client
daemon (TCP Socket).
>
> I'm just trying read one boolean register for starters ->  modbus addr: 100102
(%M102).
> Is this correct addressing for the ini file?
>
> [CYCLES]
> NUM_CYCLES=1
> CYCLE1=2,inputStatus(1,100102)
>
>
> Thanks
> Kurt
>

In Modbus there is sometimes confusion about
register number versus addresses

And it is not clear how to communicate values other that
1 bit and 16 bit.
Only coils and registers are within the original modbus specification.

Here is a good documentation about this:
http://www.csimn.com/CSI_pages/Modbus101.html

Read:
"Modbus: When 40001 really means 1, or 0 really means 1"
"40001: Modicon convention notation for Modbus registers"
"If registers are 16-bits, how do I read Floating Point or 32-bit data?"

Our Modbus functions use addresses not register numbers.

Error opening file for wrinting

C:\Program Files\pvb\LICENCE.GPL

Click Abort to stop installation,
Retry to try again, or
Ignore to skip this file.

Is there a problem with your rights in this folder?

Try installing to somewhere else.
c:\test

May be you need to become admin first.

The installation procedure will do the following:
- copy all files to the destination
- create shortcuts for pvbrowser and pvdevelop on your desktop
- define the environment variable %PVBDIR% to point to the installation
directory of pvbrowser
- add %PVBDIR%\win-ming\bin to your %PATH% environment variable

Yes, it seems that we need admin right to do something in "C:\Program Files".
This is new in windows 7.

Installing to an other folder like "test" work well.

More than this, with right click and then "execute as admin" it also work well
in "C:\Program Files".

Thank you for you help.

Regards

CRL


--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> Is there a problem with your rights in this folder?
>
> Try installing to somewhere else.
> c:\test
>
> May be you need to become admin first.
>
> The installation procedure will do the following:
> - copy all files to the destination
> - create shortcuts for pvbrowser and pvdevelop on your desktop
> - define the environment variable %PVBDIR% to point to the installation
directory of pvbrowser
> - add %PVBDIR%\win-ming\bin to your %PATH% environment variable
>

Anyone using a driver to communicate to a omron PLC via serial or ethernet
(FINS)??

> Anyone using a driver to communicate to a omron PLC via serial
> or ethernet (FINS)??
No,
but on the following page you can find a library that supports FINS on Linux and
Windows.
http://www.dastec.com/

Hi

I am having trouble with the rlController class. I wanted to setup a controller
but I am not sure  how to link the measurement and the output variables to the
rlController object. The following code-example for explanation:

#include "rlcontroller.h"
rlController              *regler                   = NULL;
double                     messwert                 = 55;
double                     ergebnis                 = NULL;

int main(void)
{
   regler = new rlController(messwert,ergebnis);
   return 0;
}

Sorry for my bad understating of c++. I hope you can help me.

Takiyuki

rlController uses callback functions for reading the measurement and writing the
output.

When you call
   regler.start();
a new thread will be started that runs the controller algorithmn.
Within that loop the callback functions will be called.
Somebody who wants to use the rlController class must supply these 2 callback
functions.

Example:


#include "rlcontroller.h"

double getMeasurement() // callback funktion
{
   double wert = 1.0; // TODO: wert von einer schnittstelle lesen
   return wert;
}

void writeOutput(double output) // callback funktion
{
   // write to the output port
}

rlController regler(getMeasurement, writeOutput);

int startController()
{
   regler.setReference(1.0);   // sollwert setzen
   regler.setPI(0.001,4,0.1);  // PI regler einstellen
   regler.start();             // thread für regler starten
   return 0;
}

PS: Please have a look on the "control" function here:
http://pvbrowser.org/pvbrowser/sf/manual/rllib/html/rlcontroller_8cpp_source.htm\
l

Thank you! My controller Program works!

I am not sure but shouldn't the controller class have an limitation?  If now for
some reason the process can not reach the reference the output will grow
indefinitely. To help me I added the following code in rlcontroller.h/cpp:

rlcontroller.cpp:

rlController::rlController(double (*_getMeasurement)() ,void
(_writeOutput)(double output))
              :rlThread()
{
   .......
   yk_max = 0.0f;
   yk_min = 0.0f;
   limited = 0;
   .......
}

switch(c->type)
{
   .......
}
if (c->limited)
{
   if (c->yk > c->yk_max) c->yk = c->yk_max;
   if (c->yk < c->yk_min) c->yk = c->yk_min;
   if (c->y1k > c->yk_max) c->y1k = c->yk_max;
   if (c->y1k < c->yk_min) c->y1k = c->yk_min;
   if (c->ydk > c->yk_max) c->ydk = c->yk_max;
   if (c->ydk < c->yk_min) c->ydk = c->yk_min;
}


void rlController::setLimits(double _yk_min, double _yk_max)
{
   rlThread::lock();
   yk_min = _yk_min;
   yk_max = _yk_max;
   limited = 1;
   rlThread::unlock();
}


rlcontroller.h:

/*! <pre>
Set the limit values for the controller
</pre> */
void setLimits  (double _yk_min, double _yk_max);

And the Variables:
int limited;
double yk_max,yk_min;

With this changes I can give my controller output directly on the hardware.

Takiyuki

--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> rlController uses callback functions for reading the measurement and writing
the output.
>
> When you call
>   regler.start();
> a new thread will be started that runs the controller algorithmn.
> Within that loop the callback functions will be called.
> Somebody who wants to use the rlController class must supply these 2 callback
functions.
>
> Example:
>
>
> #include "rlcontroller.h"
>
> double getMeasurement() // callback funktion
> {
>   double wert = 1.0; // TODO: wert von einer schnittstelle lesen
>   return wert;
> }
>
> void writeOutput(double output) // callback funktion
> {
>   // write to the output port
> }
>
> rlController regler(getMeasurement, writeOutput);
>
> int startController()
> {
>   regler.setReference(1.0);   // sollwert setzen
>   regler.setPI(0.001,4,0.1);  // PI regler einstellen
>   regler.start();             // thread für regler starten
>   return 0;
> }
>
> PS: Please have a look on the "control" function here:
>
http://pvbrowser.org/pvbrowser/sf/manual/rllib/html/rlcontroller_8cpp_source.htm\
l
>

> Thank you! My controller Program works!
>
> I am not sure but shouldn't the controller class have
> an limitation?  If now for some reason the process can not reach
> the reference the output will grow indefinitely. To help me I added
> the following code in rlcontroller.h/cpp:

You can add the limitation in your
writeOutput()
callback function.

Do you see any advantage to limit the output directly within controller ?
The behaviour will probably be different when your system includes a delay.

For the moment I use the limitation in controller directly. But I am not sure
this is right way to go. Right now it works for me. I bought the book you
mentioned in the documentation. I will study a little and come back to you when
I am sure.

--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> > Thank you! My controller Program works!
> >
> > I am not sure but shouldn't the controller class have
> > an limitation?  If now for some reason the process can not reach
> > the reference the output will grow indefinitely. To help me I added
> > the following code in rlcontroller.h/cpp:
>
> You can add the limitation in your
> writeOutput()
> callback function.
>
> Do you see any advantage to limit the output directly within controller ?
> The behaviour will probably be different when your system includes a delay.
>

Hi everyone,

I've just started using your suite, which I consider really helpful and well
made.

I'm working on a Modbus-RTU project in which I need to export the content of the
shared memory file to a web server.

I have divided the problem into two steps:
1. CSV export;
2. FTP upload.

Does anyone know how to implement the two steps above, especially the first one?
At what point of the project should I put the appropriate code (Main? Masks? ...
or Slots?)?

Thanks in advance.

See you.

> I'm working on a Modbus-RTU project in which I need to export
> the content of the shared memory file to a web server.
>
> I have divided the problem into two steps:
> 1. CSV export;
> 2. FTP upload.
>
> Does anyone know how to implement the two steps above,
> especially the first one? At what point of the project should
> I put the appropriate code (Main? Masks? ... or Slots?)?

- You can execute external programs / shell scripts (or batch files in case of
Windows) from your pvserver with the normal system() function.
#include <stdlib.h>
int system(const char *command);
There you could run the above steps.

- You can write CSV with a normal fprintf()

- You can save tables in CSV format with the write(filename) method
http://pvbrowser.org/pvbrowser/sf/manual/rllib/html/classrlSpreadsheetTable.html

These are only examples.

Instead of ftp i would suggest using sftp.

Tip: When you show data within a table in pvbrowser you can click the table with
the right mouse button and download the content as CSV file.

Your will write your own code within the slotFunctions.
There you respond to the events.
Within the other places you will not insert code very often.

Hi all,

I'm new to this group, and I know that pvbroser can run now on Android systems.
But can the pvserver run on Android? The idea is to have all in the same device,
like a Tablet.

Thank you very much

> But can the pvserver run on Android?
It has not been tested up to now.
At least the pvserver compile on Android.
Thus pvserver should run on Android.

There seems to be a terminal emulator for Android which you might use for
testing the pvserver.
http://www.appbrain.com/app/android-terminal-emulator/jackpal.androidterm

A better solution might be a ssh server for your Android device over which you
connect from a PC (SSHDroid).
With the ssh server you should be able to interactively login to the Android
command line from the PC and to copy (scp) files between PC and Android device.

After this is working you need a cross development environment for Android on
your PC. Here i use a Ubuntu virtual machine with:
http://developer.android.com/sdk/ndk/index.html
and
http://sourceforge.net/p/necessitas/home/necessitas/

Please build a "Hello World" application and run it on Android using the above.
Once you have done this please come here again.

OK, thank you very much for your guide. I will try what you are talking about
and let you know.

Also, is Necessitas always needed? I mean, if you use Lua for the server, is it
still needed? I would need also Modus TCP communication, so I suppose I will
need it anyway, but I ask that for curiosity.


--- In pvbrowser@yahoogroups.com, "pvbrowser" <pvbrowser@...> wrote:
>
>
>
> > But can the pvserver run on Android?
> It has not been tested up to now.
> At least the pvserver compile on Android.
> Thus pvserver should run on Android.
>
> There seems to be a terminal emulator for Android which you might use for
testing the pvserver.
> http://www.appbrain.com/app/android-terminal-emulator/jackpal.androidterm
>
> A better solution might be a ssh server for your Android device over which you
connect from a PC (SSHDroid).
> With the ssh server you should be able to interactively login to the Android
command line from the PC and to copy (scp) files between PC and Android device.
>
> After this is working you need a cross development environment for Android on
your PC. Here i use a Ubuntu virtual machine with:
> http://developer.android.com/sdk/ndk/index.html
> and
> http://sourceforge.net/p/necessitas/home/necessitas/
>
> Please build a "Hello World" application and run it on Android using the
above. Once you have done this please come here again.
>

> Also, is Necessitas always needed?
> I mean, if you use Lua for the server, is it still needed?
> I would need also Modus TCP communication,
> so I suppose I will need it anyway, but I ask that for curiosity.

In principle you only need the gcc compiler from the Android NDK toolchain (arm
cross compiler).
But it is much more convenient to have Necessitas also because then you can use
the Qt Creator IDE. Also qmake is included in Necessitas. We use qmake to
generate our Makefiles. Without Necessitas you would have to rewrite the
Makefiles.

Once you have compiled everything you will be able to use Lua without the need
for a gcc compiler.

You will need shell access to your Android device.
I would suggest to start a sshd on the Android device and ssh to it from your
PC.

In the first step compile the famous "hello world" program using Android NDK on
your PC.
Then scp the hello_world executable to the Android device and execute it from
your ssh window.
I have read that there is only a simple shell on the Anroid device but that bash
is alreday there and can be started. So i would use bash within the shh window.

If you managed to run hello_world it will be straight forward to compile our
serverlibrary and rllib and finally pvslua.

The normal Android apps work within the dalvik virtual machine.
You want to run native code.

Hallo pvbrowser,

ich möchte mit dem pvbrowser auf die COM-Schnittstelle zugreifen und bekomme es
nicht hin.
Normal muss ich anstelle von /dev/ttyXXXXXX beim Device doch nur COM1 angeben.
Ich bekomme aber auf einem Terminal keine Anzeigen und ebenfalls keine
Fehlermeldungen.
Vielleicht liegt ja auch ein grundsätzliches Problem bei mir vor, wo kann ich
eine Step by Step Anleitung evtl finden?
Mein Ziel ist es das Modbus-Modul auf mein CAN-BUS Modul umzuschreiben.

Ich habe schon das CControlSDK erfolgreich unter Linux verbunden, aber
auch hier bekomme ich keine Verbindung mit Windows hin.
Evtl. habe ich etwas übersehen. Ich dachte man braucht nur das Device ändern und
dann klappt es auch mit dem Zugriff unter Windows.

Ich hoffe auf einen kleinen Anschub ;-) Danke

Hallo free2msch,

es sind mir etwas wenig Information, um richtig antworten zu können.
Im Prinzip ist es richtig, dass man "/dev/ttyXXXXXX" nur durch "COMx" ersetzen
muss.

(1) Wenn es um die direkte Benutzung der rlSerial Klasse geht, bei openDevice().
http://pvbrowser.org/pvbrowser/sf/manual/rllib/html/classrlSerial.html

(2) Wenn es um die Verwendung des Modbus Daemons geht, der die Daten in binärer
Form im Shared Memory ablegt, würde ich gerne "modbusdaemon.cpp" und
"modbusdaemon.mkmodbus" sehen.

(3) Wenn es um die Verwendung des Modbus Daemons geht, der die Daten als lesbare
ASCII Zeichen im Shared Memory ablegt, würde ich gerne die INI Datei sehen.

Dokumentation zu Modbus ist hier
http://pvbrowser.de/pvbrowser/index.php?lang=en&menu=4&left=9
und in
http://pvbrowser.org/pvbrowser/doc/pvb.de.pdf
zu finden.