bootloading the VNC1L with the USB Master Kit

[electronicks]

Hi all,
I'm trying to upload a new firmware into the VNC1L (USB Host controller) that's on my USB Master Dev Kit. Yet, even with the functions of CCS, the firmware is corrupted. By "making a printf" of the address variable (the one that tells the VNC1L's bootloader where to put the data in its flash), I saw that not all bytes of the file got into the flash. I've tried sending the *.rom file with siow, hyperterminal and Advanced Serial Terminal: all said they transfered the file completely but the VNC1L didn't got it all.
A test program I wrote proved that, when transmitting a file on the serial port, by the time my PIC sends a Xoff, there has been an overrun and data is lost. I've tried reducing the baudrate: it's better but I need perfect to make it work.
By the way, I know I could have changed the firmware by putting the file on a flash disk but I need to do it with the UART. Also the firmware is corrupted now...
I wandered if anyone else got this software handshake issue, and mostly if there's a way to fix it without having to write a C program to do it. Hardware handshake is not an option because of the kit's interface.

I need help... I'm out of debugging ideas...

thanks

[dbotkin]

Good luck.

I tried for quite a while to get this to work and wasn't able to. Handshaking on the VNC1L seems to be broken. In fact, a number of things on the VNC1L seem to be broken. Good luck getting any help from FTDI; they don't seem interested in fixing it.

In case you're wondering, the VNC2 hasn't been much better. They say you can develop your own code with their toolchain, but it seems you have to depend on their pre-compiled driver binaries, which are about as buggy as a roach motel. I have a VNC2 eval board and debug interface I'll sell you at a discount.

[electronicks]

Sad news! I was afraid of something like that.

Hyperterminal let me space each byte sent with 1 ms which corrected the software flow control issue (and extended the download time by 50!). Still doesn't work...
Yet, I hope you're wrong, because that would totally mess up my project. Maybe that's why FTDI sells modules like the VDRIVE, because their chip is s# %.

thanks for the reply BTW.

[dbotkin]

On the bright side, uploading VNC1L firmware from a USB drive *does* work. Most of the time. Of course when it fails, like if the users pulls the drive or cycles power or whatever, then you're screwed. Back to the hardware programmer. I mean, seriously... the bootloader overwrites itself?? Even I am smart enough to not make that mistake.

For my next version of my product I'm looking seriously at 2 PICs. A PIC32 for the host interface, and something else for the slave interface (I run VDPS firmware in the VNC1L). At least I'd be able to fix problems myself rather than just living with them. And it will probably still be cheaper than a VNC1L.

If only the VNC2 would work as advertised... sigh. But it doesn't.

[zorrolabinette]

For all of you, and mostly for Electronicks, that are trying to program a VNC1L chip:

*Start by downloading the software from ftdi web site;
*get all your material plugged and ready (better with a hub between your PC and your target);
*Use the Install program that came on the CD with the USBMASTER kit, this will create a folder with all you need, like Ex_VNC1L_flasher.c;
*Open that same program and comment out this line «dataptr=&datastream[0]» , because you will get a compile error from that leftover...;
*Compile and download into the PIC, then run! (the program I mean)

I have done this a numerous number of time and it works great each time.
Happy PICking!