7e94ed4442b breaks flashing GT2560 boards

[diresi]

Hi,

I have a 3d printer running a GT2560 [1] board, which is essentially an ATmega2560 with a CH340C as USB-to-Serial converter.
I found that flashing works with avrdude 6.3 but not with 7.2.

It took me a while to trace that to DTR handling that was changed in #1309.
Flashing works again when reverting commit 7e94ed4 in wiring.c.

[1]
https://www.geeetech.com/wiki/images/d/d3/Hardware_GT2560_RevA%2B.pdf

[stefanrueger]

@diresi Yes, C9 is a bit high with 1 µF, so the 50 ms dis/charge time in the usleep() calls might be a tad bit low for this board? Could you play with both values in lines 175 and 183 to see if you can get the board to reset and synchronise?

[mcuee]

Another way is to change C9 to something like 100nF, and then see if avrdude 7.2 version works or not.

[diresi]

That was solid advice :)
It works with 80ms discharge time, I created a pull request for that w/o reading any contribution guidelines ... please feel free to reject.

I was playing with the idea of replacing C9 already (since I broke the original SMD one when attaching probes) but then again, maybe someone else is also using that board, right?

[stefanrueger]

OK, further tests have demonstrated that the board design and too high out-of-specs Vcc led to the board being put into HVPP mode. Tinkering with usleep() has ostensibly solved the symptoms but the usleep() times were not wrong per se.