Posts by andyinfsnet


    Can anyone tell me about the available PWM output the following questions:

    -Frequency resolution(adjustable steps of frequency)
    -Frequency accuracy
    -adjustable Frequency range
    -adjustable duty cycly range(s.t. called load)

    thanks in advance


    Yes now it is working, thanks for helping. The trick was a bit that the update program itself did the boot delay setting.
    Problem is still exisiting that we cannot detect a rolling back. If you have an further idea for that, please tell me.
    For now this helps a lot.


    This is for any who needs it:


    I understand what you mean but I am not sure if this is true that it is only watching for a executable in sepcial place.
    I did following code and named it AutoStart.exe:

    I renamed your AutoStart.exe with MyAutoStart.exe.
    When I plug in the usb stick, the update process starts immediately, this is different behaviour to usual sequence where I am asked for reboot.
    Do you see what I do wrong?

    Thanks in advance


    I think we talk about not the same. With the USB update tool (user Application software update by USB stick) the end user needs to know when he should switch off/on the device and when to unplug the USB stick for final rebooting.
    So when he has no display, he cannot do this. But when I say to my customer(end user) that he shoud connect a PC with a terminal to the serial debug output of the netdcu, then he could see a textual message, like on a real display, if I would have one.
    Then he would know what to do.

    Is it a bit more clear? If not we can telephone.

    thanks for helping


    Ok, I think I understand. This would be the same like if I let the update tools name as it is and execute some own code(that writes on debug port) with the X feature of the update tool, correct?

    But how should I, means my program detect things like copy failures or a started rolling back action?
    Also this needs to be displayed somehow. Any idea or am I still wrong?



    Some device from us have no display(even it is a netdcu8). An Application Software update for customer/user without Display seems not really possible now.
    I have the USB update tool that works great, but needs display interaction, at least some information for plug/unplug the usb sticker.

    Would it be possible (new kernel...) that you display the strings that I see usually for for usb update at the display that you send it also to serial debug output? Would be very suficient for me.


    Hallo zusammen

    habe etwas mit Sprachumschaltung experimentiert. Funktioniert soweit gut.
    Nur ist es mühsam eine vom OS nicht unterstützte Sprache wie z.B. den locale "zh-CHS", also einfaches chinesisch per font nachzuladen und dann an alle controls in jeder Form die Texte aus der Sprach DLL zuzuweisen usw. Einfacher wäre ja wenn das schon im OS bzw Kernel mit drin wäre(falls das von der Grösse noch machbar ist).

    Ich bräuchte also einen Kernel für eine Netdcu8 (derzeit nutzen wir NK82_CF35_090423.bin) der die oder eine einfache chinesiche Sprache (simplified chinese)unterstützt.

    Wie groß würde der Kernel werden und wie kann ich dafür weitere Updates erhalten, wenn sich bei ihnen etwas ändern würde?
    Welche Schriftart wäre dann da mit drin(MS Song, You Yuan, simsun oder was auch immer)?
    Was kostet sowas?

    Gibt es evtl. auch die Möglichkeit das BSP für das board zu erhalten um dann über den Platformbuilder sich entprechend die Images zu generieren? Evtl. halt mal Japanisch oder sonst was far eastern mässiges mit dazunehmen.

    Gibt es vielleicht bessere Möglichkeiten? Wie machen das andere?

    freundliche Grüße


    Yes now it is clear. I got it working, but netdcu8 has range up 4095 not ox3ff!
    This is the neccesary additional code for c# if anybody needs it:



    here is a VS2008 CF35 sample how to adjust contrast under Windows CE, in C# written.
    The sample has a trackbar to change current contrast value, the button is used to store it permanent in registry. Registry value is loaded to the trackbar after init.

    I tested it with Netdcu8 and it works. Values are changing expected and are stored permanent in registry and are used after device power off.

    But for Picomod2 I got different behavoiur, sometimes the display goes black when trackbar is used and the reslution (should be 0..4095) is also not the same. Nearly only 3 values are to see: bright , a bit less bright and black display.
    When i look in the registry after a picocom2 trackbar change(whithout pressing the button), the value is not the value from the trackbar(0..4095) but something undefined. Lets´s say I change trackbar from 0 to 200, the value in the registry is may be 24 or whatever but not 200.

    I use a Hitachi TX09 320x240 LCD.
    Can anybody see any errors here in coding or thinking?
    I need one and the same software to running on both devices netdc8 and picocom2.



    I use a NETDCU 8 with WCE5.0 and CF3.5 mostnew kernel and made a contrast adjustment trackbar in my C# application for a TX09 Display.
    The adjustment works fine while application is running.

    I use this following principal code when I press a button to save the contrast value changes to registry to save them after power off:

    What happens is that everything seems ok in the registry, means the key "contrastvalue" is saved correctly and it is also available after a power off when I read the number of the key(programatically or by reg editor, ok).
    But the value is not taken as contrast value, means the value before power off is not here.
    To make it more complex, every let´s say 10 times it is working fine and the stored value of the regitry key is used.

    Do you have ideas what I could do bette or check?


    Hello fsnet

    I have the problem that my Picocom2 is switching into bootloader mode by himself, LCD stays dark and on COM interface I see the following:

    The chapter with starting "32 MB Ram" then always is shows up again every let´s say 10 seconds.

    Then I have to insert MAC adresse and need to download the kernel again, and then it works. Unfortunately this happens 3-4 times a day, after switch off and switch on.
    (No, I didin´t load kernel into RAM)

    could you support me please?


    Hello coders

    This is a small basic example how to make a simple TCP asynchronous Communication in C# under VS2008. I use Windows CE 5.0 and CF3.5 here.

    The example let´s the client send a Telegramm and the server responds to that with a answer telegram including just a incremented counter value send back to the client. The Client was waiting to that and if received he starts next telegram and so on. So it is like a loop running forever.
    To get it running you need start SmartDeviceServConnTest.exe and then the SmartDeviceClientConnectionTest.exe and then input the Server IP and press Begin Button.
    The samples run on desktop and CE device(I have tested it with Netdcu8´s here). You see the number of send and received TCP Telegrams per second on both the Client and the server panel.
    The example does not include thread safety ot any other error handling just to let it be as simple as it can be.

    My question is to the people who have knowdledge and experience in socket communication like with this example. When I do this test between 2 Desktop PC´s I see about 2100 Telgrams per second. When I do the same between 2 NETDCU8´s I see only poor 100 Telegrams per second :(.
    Is this usual or did I make some basic error n my code? The CPU usage is never above 60% so there seems to be some reserve in the NETDCU regarding the CPU.

    I am about new to this issues so please help if you see any chance to improve performance becuase I wonder why there is factor 100 between desktop an NetDcu8 (about factor 10 I would have expected).
    But may be it is like it is.

    thanks a lot in advance


    The EQATEC Profiler is very good tool but has disadvantage that he calculates not the real times used by a thread(when interuppted by another thread) but the overall time inside a thread including the time wasted by interuption.
    So when you want make real measurement of your thread times there (at least what I know) is only the possibility to make some timing code inside your sources and stop out specific sequences that are in you interests.
    If you like, I did this once and could send you some example (this is done by p/invoke calls of some timers).


    As discussed with Mr Keller I tried now kernel NK82_CF2_080804.bin and it worked ok.
    So fs-net needs to create a new kernel for CF35 with these corrections.

    Bye the way, is it also possible to rising edge trigger interupt J5 pin1 (the predefined interupt pin)? There seems to be no registry setting for this pin like for the gen purpose IO pins.


    There is no difference between using the program first time and thereafter with out rebooting. I always get these tons of interupt. What I did now was to use an ohter NETDCU8 with I had. It is one with kernel 27.5.07 with CF2. With this kernel it works as it should. So this is what you meant when you say that you changed the interrupt behaviour with newer kernels.

    Here is the startup when it is not working:

    Here it works(it the older another board):

    Does it make sense that I call every time when I get an interupt the function InterruptDisable() or will it not help? Ok I will try it anyway. If it hepls I will post it here.
    But could you look further in the meantime please?

    Andreas Heß