WARNING: All protocols shoul dbe testing on stock board again.
Fixed a number of problems relating to the (dis)enabling of hardware
switched PPM_OUT for the G9XV4 board. I also broke a lot of stuff along
the way over several previous commits, as I slowly learned how all the
code went together and the 'whys' for some stuff. My bad.
In other words, this commit finally has DSM2=PPM mode working on the
G9XV4 board without breaking a bunch of other stuff -- hopefully.
!! STOCK BOARD TESTING NEEDED !!
I am pretty sure I have not broken anything for the stock board code --
BUT it MUST be tested before the next release, because I have made quite
a few changes in and around that code.
added to DSM2=SERIAL code.
Hopefully also fixed DSMX mode bind omission from previous commit.
IMPORTANT: I do not have a DSMX module for testing -- only DSM2. Can someone
with DSM2=PPM mode and DSMX please test and confirm that both bind and range
check mode are working. Thanks.
I have modified the DSM2=PPM code to implelent bind and range check the
same way as for DSM2=SERIAL.
NOTE: I have checked for clean compile but do not have hardware to test
that the protocol is working correctly. PLEASE TEST. Thanks.
Tidied up Timer1 ISR code -- numerous little issues resolved, comments corrected.
Obviously, anything to do with this ISR requires thorough physical testing.
I have tested DSM2_SERIAL and the PCBV4 PPM_out is working properly. Someone
should double-check PPM/16/sim and I guess DSM2=PPM, on the stock board.
Thanks.
However, I don't like that the only place I could find to turn off range check mode when the SETUP menus is exited, was in popMenu ...
void popMenu()
{
if (g_menuStackPtr>0) {
#ifdef DSM2
s_rangecheck_mode = 0;
#endif
...
}
It was the only way I could find to trap the event of [EXIT long] being pressed
while in the RANGE field and s_editMode==true. But this solution gets executed
for all menu exits -- not just menuProcModels. It works, but there must be a
more eloquent method?
(It only took me about an hour and a half! :-/)
Also increased max filename length handled by "Restore Model" to 16 chars
(only when long filename support is configured in FatFs/ffconf.h)
Bertrand: I had to comment out a call to sdPollmS() as it doesn't seem to exist.
Hopefully the system freeze when I try to select "Restore Model" is
just a symptom of this same issue ... though I fear not.
Half -- because I don't yet know how to make warning sounds in this
new code. HELP! :P
... otherwise it's working. I have confirmed that
the range check bit appears on the scope, while watching the DSM2
data stream.
Some DSM2 modules take longer than others to power up. Thus, sending a
single BIND packet is not sufficient in some cases. I have implemented
here an approimately 4 second period (after power-up) during which
time the trainer swtich can activate BIND mode.
This is in fact not strictly needed, since the DSM module will by
itself reject bind packets after a short time. (I have confirmed this
again with my DSM2 module today.) However, we like to be sure, so the
limited time window for BIND activation remains.
The word "BIND" also now appears as confirmation in upper-right of
the main view, whenever bind packets are being sent to the DSM module.
(I did not consider it important to add this "BIND" word to the
language files. Please feel free to do so though, if warranted.)
Gruvin.
bsongis