AW4 control again...

so having read about the manual shift aw4, i decided i had to
build one. i reviewed several designs and started on my own.
unfortunately i couldn't find any serious info on brettM's device,
as it sounded good. mine initially featured a pic 18f458, and a
pair of switches, one for up/down and one for TC lock. didn't want
to get too ambitious too quickly, but the 458 would leave me a lot
of room to expand, and features all the hardware needed to process
all of the same signals as the factory TCU and then some. as i was
plotting and planning i figured i'd do well to test this out
before i implemented a full on replacement of the TCU. i decided
to use a rotary switch and diodes. picked up the switch(275-1386)
at shack for $3. it's current rating was low, but it was rated for
120VAC, so i figured it might be ok. added four diodes from my
spare parts (1N5406 rated 3A at 1KV, but anything better than
maybe 1.5A and 15-20V should do) some spare wire, an hour to
build, an hour to install, and i had a functional shifter. with
all new parts, this could probably be done for about $15-20 if you
purchased connectors and wire. when i was done, i had a 6 position
switch with half rotation that gave me a pattern of 1, 2, 3, 3
locked, 4, 4 locked. i pigtailed quick disconnects into my TCU
harness after unplugging it, plugged in my switch and took it for
a test drive.

i should mention something here. i keep reading that the trans
locks in first if the TCU fails. this was definately not the case
for me. mine locked in fourth just as the progression of the
transmission solenoids said it should.

my first attempt failed. it would seem that the 93 TCU harness
diagram i have doesn't quite match my 95 (should have checked that
more carefully). after working out how to get power to my switch
having run out of tap splices, success. having not seen anyone do
one quite like this, and given how simple it was, i figured you
all might be interested.

some notes:

on construction: this switch was 2P6T, but since i didn't need 12
positions, i bridged the two halves(poles and throws wired to
corresponding pins) to increase it's current capacity. i'm not
sure how long it'll last before the current kills the contacts or
wipers.

there is no TC unlock with this switch. if you try to brake, you
will probably run into problems. switch into unlocked first.
i built this only to test out my trans. your milage may vary.

ok... i'll post pics when i find somewhere to host them unless someone can tell me how to do it through here.

pics - imageshack.us, it'll give you code to paste (you want the forum thumbnail one.)

I still owe you a bunch of links and threads, haven't had time to find em yet sorry

no problem. i've still got a bunch of work to do. it's been way too long since i coded asm. it's probably gonna take me a few days just to get my controllers up and working properly again, then i'll have to start coding all the stuff i've already learned about them. i've probably got a week or so at least of work.

anyway, here are the pics and schematic. sorry about the poor graphics.

the schematic:


the whole thing:


close up of switch:

ok, so after some reading and planning, i've got the first part of the design done. i've decided to use two seperate units to simplify construction, and in the jeep spirit, harden the design to allow for failover. currently the first part i've got going is the source selection and solenoid driver board. this one will allow just about any electronic circuit to drive the solenoids, and switch from stock TCU to/from custom/manual control. the one i'm currently building also has a set of lines to inform the custom controller what the stock unit thinks it should be doing, which should allow for a seamless switch from stock to custom under any running conditions (there may be 1/10th or so second of 4th unlocked during switch, but i don't think the trans will work as fast as the relays will engage). the failover part comes through the source selection relays. if the controller fails, the relays drop to their default position, and connect the trans to the stock TCU. since the stock unit retains all it's default function, it shouldn't be possible to take the trans out of comission with any sort of controller failure while using this design. total count of 8 major parts. if anyone's interested, i'll post schematics and pics.