Hi! I'm failing to get an ATMEGA32U4-AU be detected by my OS. My windows doesn't detect anything at all, it's just silent. Same on Macintosh. Here is the circuit:
Voltage is fine, as far as I've checked several VCC pads in the circuit. And on D+ I've measured 1.37v, and on D- 0v. I guess this is fine too.
Can You give me a hint how to proceed with troubleshooting?
No pullup resistor on reset?
Hi Kartman. Yes, I have no pull-up on RST. I think a floating RST pin shouldn't prohibit windows to detect a new USB device, or am I wrong?
Take the Pepsi challenge - add that resistor and try. I'd suggest that if the device is in reset, then no USB action will happen.
Kartman: Ah, I see .. I can hardly change the circuit, since it is a PCB :-( I measure 0.08v between GND and RST, how to interpret this?
Duh! What do you think? How long does it take to add a 10K resistor compared to writing a post about doing it?
Hello Valusoft! OK, I will build a wire bridge from RST to VCC with a 1k resistor inbetween. But first I would like to know how to interpret those 0.08v? Is this a float value?
Sounds like a float value. Put a pullup resistor in and measure again.
Thanks, Kartman. Now I'm going to solder ...
Alright, I pulled RST with 1k up. RST is on 5.15v now, as my multimeter say. Still Windows doesn't detect any device. Moreoever, I built a wire bridge from HWB to GND (no resistor). Also I temporarily shorted RST with tweezers. Still, nothing happens.
Does your chip have a bootloader loaded?
Is the crystal oscillating?
If you don't know my whole story, keep your mouth shut.
If you know my whole story, you're an accomplice. Keep your mouth shut. 
Hi Torby! A preinstalled bootloader comes with every ATMega32U4. As for the crystal, I can't say if the crystal is working, is there a chance to check it without an oscilloscope?
Get an Arduino Leonardo (the Arduino built around the 32U4) then you have a working 32U4 to compare against. You can compare the behaviour of the signals and see what is different between your design and the one that works.
Yes, a Leonardo as reference wouldn't hurt.
Hi! Krümelmonster
I am lucky than you in this case.
Do you have ISP programmer? (avrisp mkii)
It can access your chip no matter having bootloader or not.
If success, your chip/ power supply & reset is function.
Then try download bootloader and test again.
If not success, your chip may preset with external clock.
It is very easy to wrong in fuse bit programming.(I have experience in atmega32 in before).
In this case necessary to make chip return to atmel default.(HVSP)
Hope it can help you.
Hi lc2047. I've ordered the ISP programmer, and it should arrive about next week. Maybe some fuse bits are wrongly configured, but I wonder why other report me their ATMega32u4 was detected on the spot by Windows, and they use a pretty similar circuit, just as me.
I have another question: I measure 2.33v at my pulled up RST pin, what could be the reason for this? 2,33v? Is this correct?
It was +5 before and now its 2.33?
Loose / bad wiring?
Poor solder connection?
Power supply faulty?
Multimeter out of calibration?
Bad multimeter leads?
Battery going dead in meter or project?
I pulled RST with 1k up
I think 10K would be a better choice.
Some programmers might struggle with a 1 K load.
JC
You cannot simply connect the Mega32u4 to a USB port of your PC and recognize it. If you read the datasheet you have several ways to get the bootloader to activate. Both in your firmware which would have to be loaded via ISP, or activate the bootloader using the /HWB pin which also needs to have it's corresponding fuse programmed.
Again, you cannot just connect the USB port of the AVR to a PC and see it in Device manager.
How do I know this? I tried it myself 
In the datasheet the Bootloader section begins on page 314, and the actual implementation explanations begin on page 318.
Here is a link to an Appnote on using the bootloader on AVR's with built in USB interfaces:
http://www.atmel.com/Images/doc7...
Also. LISTEN to what everyone is saying and solder the 10k resistor from reset to Vcc or the AVR will never operate reliably....EVER!
Jim
