R70/R71 on debug v3 board ?
Werner Almesberger
werner at openmoko.org
Fri Feb 15 03:10:28 CET 2008
Harald Welte wrote:
> It was, your schematics just aren't the latest version (see private
> mail). ADBUS4 is connected to nNOR_WP.
Thanks, that explained it :-)
I've written a little utility to set that bit.
trunk/src/host/norwp.c
Testing this on a board I've reworked with 5% resistors on R70 and R74,
and using 82R for R74 instead of 80R, I get 0.025V on nNOR_WP when set
low, and 1.27V when set high.
Unfortunately, the FT2232 resets ADBUS4 to high :-( OpenOCD sets it low.
The M58WR016QT data sheet says that we need at least 1.3V for a "high"
on VPP. So those 1.27V are slightly too low. They're close enough to
make this work on my board, though.
VPP can be fed with up to 3.3V (M58WR\ detail\ spec.pdf, table 20, Vpp1)
and is actually 14V-tolerant, so there's no need for the voltage divider.
So if we'll make a new batch of debug boards for GTA02 anyway, we could
just omit R70, and be well within the specified range.
Unfortunately, we shouldn't make R2505 bigger. It's already nominally a
bit too large at 100k, with Ipp2(max) = 5uA and Vpplk(max) = 0.4V.
I don't think R2505 = 100k will be a problem, but I'm less confident
about the ~1.27V (which could be as low as 1.237V, if we get resistors
just from the wrong edge of the tolerance range).
If we'll ship a yet to be made batch of debug v3 boards for GTA02 anyway
(which I think may be the case), just omitting R70 would solve the issue.
People who already have a v3 board will probably not experience problems.
And in case they do, R70 is pretty easy to unsolder.
- Werner
More information about the openmoko-kernel
mailing list