Objective Development Forums

It supports HVSP and ISP only, not HVPP. There are not enough pins on the ATMega8.

AVRMiniProg has a variation which can do HVPP, see http://www.simonqian.com/en/AVRminiProg/

That's 1k words which is 2k bytes, enough for the boot loader. Please use the lock bits to protect the boot loader from overwriting itself. There's no additional check in the boot loader.

The hex file may be 8k in size because it starts at 0 and goes up to the boot loader. The actual code is only 2k.

The Netburner SBL2e looks fine.

The question how many keyboards you can emulate means how many USB connectors you can serve with this solution, right? I'm afraid I have to disappoint you: One AVR can only handle one single USB device in firmware. For multiple USB connections you need multiple AVRs.

See http://www.usb.org/developers/hidpage/

If you want to send arbitrary control messages, you need libusb. Otherwise you can only send those controls defined in the USB HID paper.

Since the PC recognizes the device by name, USB communication is working as such. There must be an issue with I/O ports for key inputs.

If you have a real HID device, you don't need libusb. See the hid-data example. The only limit here is that you have fixed size descriptors for your data.

If you want more fine grained control (like in e.g. hid-custom-rq or custom-class), you need libusb.

You need some kind of Ethernet to serial converter. Search the net, there are some available. Then it should be relatively straight forward.

It should be sufficient to create the tmp directory on the same partition where sensord is started. Yes, windows uses backslashes instead of slashes, but fopen() should convert it accordingly.

See http://metalab.at/wiki/Metaboard . Not exactly boot-loader compatible, but it can be integrated with the arduino IDE.

What's the advantage of using a resonator? You'd need 12.8 or 16.5 MHz due to the tolerance. If the point is price, use the internal RC oscillator at these rates. That's even cheaper. Otherwise use a cheap crystal at 12 or 16 MHz.

Your problem depends heavily on timing. Since USB 2.0 hubs implement low speed communication on their own, you are effectively dealing with a different USB host implementation when you insert a hub. This different implementation may use different timeouts, retry intervals etc.

If you are unsure, remove the zeners and lower the supply with two series diodes. If that works, look for other zener diodes.

12.8 MHz and 16.5 MHz are a matter of constant definition (F_CPU) plus additional calibration code from the libs-device directory.

8 MHz is not possible, in my opinion. However, if somebody is able to implement an 8 MHz module with PLL, this would be a complete rewrite of the assembler part.

You probably have included assembler parts multiple times or multiple assembler parts. Please see the Makefile provided with our reference projects to see which object files must be linked.