Objective Development Forums

Sorry, I overlooked that Windows recognizes the device. If you have configured libusb-win32 to attach to the HID device, then bootloadHID will probably not be able to connect to it. Maybe the result of a previous experiment? The first thing bootloadHID does is to enumerate the devices, ask them for ...

It's hard to analyze the Windows problem with only this message. Is the device recognized as HID at all? You should see some kind of bubble info when you connect it.

Sorry, I don't know the project so I can't comment on the function of the "test supply voltage" LED. However, if Windows does not bring up ANY alert, there is no 1.5k pull-up between the supply and D-, or the AVR outputs low level on D+ and D-.

The new version should be online already...

If Windows does not even display the "device not recognized" message, you have a hardware problem. A simple pull-up of 1.5 k from Vcc to D- is sufficient to get this message. I don't understand your other question, though. If you connect your device to USB, the power LED does not light. Bu...

bootloadHID was compiled with gcc 3.3 and a more recent avr libc than you used.

I see that the most recent version is 2007-06-25. We have done some code optimizations since that version so that it fits into 2k even when compiled with gcc 4. I'll make a release soon...

This look like a general problem with the MinGW installation. The make utility complains that it can't find the compiler (as far as I understand the error message). Either the make utility does not match the rest of the installation (I remember there were two "make" packages for MinGW, one...

Have you ever been able to compile anything with this MinGW installation? This looks like a broken make utility or a missing compiler.

Have you modified the Makefiles? This is an error message created by the make utility itself. It looks as if it tried to find a program named "gcc C:\MinGW\include -O -Wall -c sensord.c" instead of running "gcc" and passing the rest as parameter. This can happen if you have quote...

CDC mode does not work reliably on all operating system / hardware combinations. Osamu Tamura maintains a list of compatible and incompatible systems in his AVR-CDC project.

The error messages says that the macros USB_INTR_CFG and USB_INTR_ENABLE expand to invalid left-hand values for an assignment. These macros are defined in usbdrv.h to MCUCR and GIMSK (or similar registers). You may override these values in usbconfig.h, though. Please run your code through the prepro...

The clock rate must be correct within a couple of 100 ppm (parts per million). 14.75 MHz is definitely out of this range. Please note that asynchronous protocols usually tolerate clock rate deviations of more than 1%. You can usually trim the baud rate precisely enough from any high frequency clock,...

Have you tried this on another operating system, e.g. Linux or Mac OS X? The most likely cause for your problem is that Windows has stalled the bulk endpoint after an error. The error could be that an USB packet was lost due to high CPU load (and long interrupt disable time) on the AVR. Other operat...

You should not need to modify usbdrv.c. The macros you mentioned should expand to no-ops unless you have configured a HID device. When you use the Mega88, please make sure that your fuse settings are correct. You must ensure that the processor runs on external crystal clock. And you may have to chan...

If you stick with the words of the data sheet, yes, then this is overclocking. There are AVRs with a more precise specification of supply voltage vs. allowed clock (in a diagram). Since the rating is consistent with the values given in other AVR data sheets, we assume that the other AVRs have a simi...