TL;DR: Try a high quality cable with ferrite bead(s)
Hello,
as I had some problems with programming my Mega clone with CH340 chip, I googled for the solution and found that others have similar problems too. So this is the description and solution to my problem:
The error message people are having after Arduino IDE tries to upload the sketch but failed (after quite long waiting):
avrdude: stk500v2_ReceiveMessage(): timeout
avrdude: stk500v2_ReceiveMessage(): timeout
avrdude: stk500v2_ReceiveMessage(): timeout
avrdude: stk500v2_ReceiveMessage(): timeout
avrdude: stk500v2_ReceiveMessage(): timeout
avrdude: stk500v2_ReceiveMessage(): timeout
avrdude: stk500v2_getsync(): timeout communicating with programmer
I have a genuine Uno R3, a Mega2560 clone with 16u2, a Uno clone with a CH340T and the newest one and problemcausing: Mega2560 clone with CH340G. As with all new boards I tried to programm it with the AnalogReadSerial sketch to test the programmability and the serial connection. It failed with the above message on my windows 10 laptop with Arduino IDE 1.6.5. I tried other drivers and many things in all possible combination with the following parts:
Laptops:
- Windows 8.1 64 bit
- Windows 10 64 bit
USB A to B Cables:
- 28/28 AWG (gauge) without ferrite bead
- 28/24 AWG with ferrite beads
Boards:
- Uno clone with CH340T
- Mega2560 clone with CH340G
CH340 Driver:
- version 3.4.2014.8
On both laptops, the CH340 serial interfaces are recognised by Windows and is shown as working in the device manager windows. The COM ports were selectable in the Arduino IDE aswell.
Conclusion:
It was the USB cable IN COMBINATION with my laptop that runs Windows 10 (note: I don't know if it's Windows 10 related, maybe just the USB 3.0 chip in that laptop)! The 28/28 cable could be used to programm all other boards except the Mega clone with the CH340G chip on the Windows 10 laptop, on the Windows 8.1 laptop the 28/28 AWG cable works.
With the 28/24 AWG with ferrite beads: works on both laptops with both boards/CH340 variants.
So it seems like some problems that people have programming clones with CH340 chip COULD be solved with a higher quality USB cable. At least this is a solution that I haven't seen mention anywhere. I would never guessed that the USB cable was the culpit as it worked fine with all boards (and other usb devices) before.
I hope this will help reducing trouble shooting time and frustration 
