Objective Development Forums

You can check for SOF frames which come in 1 ms intervals if the device is connected to a host. The easiest way to do this is to (1) wire INT0 to D- instead of D+, (2) update to the latest driver version and (3) define USB_COUNT_SOF in usbconfig.h. Then you can check whether the global variable usbS...

We have created a binary distribution of AVR GCC and associated tools for all Mac users. The package offers both, GCC version 3 and 4 because version 3 generates more efficient code. All software is compiled as universal binaries for Intel and PowerPC Macs.

See http://www.obdev.at/avrmacpack/ for more information.

Please verify that the definitions in usbconfig.h and the initialization in main() match your hardware. On the hardware side, you should probably consider adding series resistors and 3.6 V zener diodes to limit the bus voltage to 3.3 V. Not all hosts tolerate 5 V on D+ and D-.

Since you use the PLL for clock generation, the CPU actually runs on 16.5 MHZ and you should set F_CPU to 16500000. We currently don't have a good standard method to obtain OSCCAL. I would recommend that you find it by trial and error by measuring the actual clock frequency. Getting the HID descript...

This is different. If XP does not recognize the device, it's either a hardware problem (D+ D- swapped, missing zener diodes or wrong zener voltage, etc.) or a problem with the clock rate. I think, we should come up with an auto-calibration algorithm. It should be possible to take the host's USB cloc...

Es ist zumindest eine Version von avrdude im Umlauf, die mit AVR-Doper im HID mode nicht zusammen arbeitet. Ich glaube aber, dass er den AVR-Doper gleich nicht findet. Daran sollte es also nicht liegen. Probier' einmal mit Deinem hex-file, ob's am Programmer liegt. Unser hex-file wurde mit gcc 3.4.6...

That's interesting because we have only added new features between these releases, no bug fixes. Since the code structure has changed quite a bit, it's hard to find real differences. So I suggest that you stick with the new version and be happy that it works better

If you tell me which was your previous version and which is the new, I can tell you what was changed (or you can see the Changelog in the driver directory to find out). It's possible that a bug was fixed which affects HID.

The key point is that you publish the source code. You can make any profit you like with GNU software as long as the source code is available.

Getting a Joystick into 2 kB of flash can be a challenge. But if you still want to try it: The only thing that's different when you run from the internal oscillator is that you need to calibrate it to 8.25 MHz and set the fuses for internal RC oscillator and clock doubling.

Hmm....

Ich probiere jede neue release vorher auf genau diese Art aus, allerdings meistens nur am Mac (ausser bei groesseren Aenderungen). Bei mir hat's funktioniert.

Welche alte Version hast Du laufen?

Or look at the Automator project. It uses Windows HID functions to send and receive data and it contains an abstraction library based on Windows HID calls for Windows and libusb calls for Unix.

If you need to send data to the device, I would recommend that you also encapsulate this in reports sent to the device, so that you can stay fully compliant with HID. See the Automator project for an example. You don't need the interrupt-in endpoint. However, the HID specification demands that one e...

If you do the same thing in the main loop, usbPoll() is not called frequently enough. You CAN and SHOULD poll in the main loop, but you MUST NOT hang waiting for an event. If the UART is busy, just continue with the main loop and try again during the next cycle. Do this for every byte you send and f...

Just a quick note on report sizes: AVR-USB can handle longer reports than 8 bytes with a trick. If you want to send (say) 12 bytes, use usbSetInterrupt() for the first 8 bytes, then wait for usbInterruptIsReady() and then send the remaining 4 bytes with usbSetInterrupt(). This is not really intuitiv...