Alrighty guys, been getting it hard from the fuel starvation problem. Mine kicked in at 55mph........ (There was other weirdness too.)

Yeah the pain.

Preliminary tests have shown my fixes to work. (It's kinda snowing here on and off, so I rode as far as I could stand.)

First thing I did was drain my tank and rinse. Twice. I never took the "All the crud settles to the bottom of your tank" stuff seriously before this. I do now. Clean it, it's a 30min to 1hour procedure (even for those of us with extreme wrenching problems) The ammount of crud in there was beyond amazing. (Make sure to pull tank valve and all that good stuff.)

Mke sure your tank valve is properly aligned. It must be dead on, a few degrees cuts fuel flow by 70% or so.

Clean petcock.

Redo all fuel lines.

Disassemble and clean carbs, do not lose part number 26 http://houseofmotorcycles.bikebandit.com/partsbandit/assets/schematics/Suzuki/SU0083_013.gif.

Reconnect all things, use prime, remember to shut off fuel when you turn off the bike.

(Oh yeah, before this te prime trick did not work on my bike.)

Wash, rinse, repeat for each GS you own. LOL.

So maybe the problem isn't the diaphragm after all...it's just the tank valve that need to be absolutely verticle to provide the adequate flow needed.  Thanks!  Good work!

Circles Center, your case may very well have been a clogged or partially blocked tank valve, but I don't think that is the case for most people.  It is certainly good to  know that the tank valve being misaligned causes a substantial reduction in flow.  However, if it the tank valve or tank blockage was the problem for everyone, then putting the petcock in PRIME wouldn't help.  In my case, it solves the problem completely to put the petcock on PRIME.

I almost always have the problem when I am trying to go 75+ mph in high wind conditions or up a long hill.  My suspicion is that there is a lack of engine vacuum when the engine is under extreme load.  Since the fuel petcock is vacuum driven, it does not pass sufficient fuel when there is very low vacuum for an extended period.  Putting it in prime solves the problem in that case.

Somewhere I have a vacuum gauge that would (if I could find a place to mount it and splice it in the vacuum line to the petcock) tell me if my theory holds.  Observing the engine vacuum with that gauge AND testing the petcock with a Miti-Vac (which I do not have) would test my hypothesis.

I can think of a fairly simple way to test the "tank-petcock-alignment-off-by-a-few-degrees-drastically-reduces-the-flow" theory.

1) Make sure there is plenty of fuel in the tank.

2) Turn the tank-mounted petcock OFF (slot is horizontal).

3) Detach one of the hoses from the tank petcock.
    (Shouldn't matter which one, but the front REServe hose is
     easier to get to and is the one used by the PRI position.)

4) Attach a length of hose or tubing that you can direct into a large
    see-through (glass?) container.  Clamp the hose in hemostat
    fashion so that no fuel can flow through it.

5) Mark a "target line" near the top of the container. 

6) -- TEST 1--


* Turn the petcock to the perfectly-vertical ON position.

* Unclamp the hose, start your stopwatch, and let fuel drain
   from the tank into the container.

* When the fuel level in the container reaches the "target line":

# Stop the stopwatch
# Reclamp the hose
# Pour the fuel back into the fuel tank

* Record the elapsed time and reset the stopwatch.

7) -- TEST 2 --


Repeat step 6, but turn the petcock to a not-so-vertical
position.

8 ) -- TEST 3 --


One more time, with the petcock turned a few

more degrees
toward the horizontal.[/list]

9) If any of the times were longer than a minute, convert them
    to seconds.
    (eg. from 2 min, 27 secs to 147 secs)[/list]


At this point you can say "The flow rate for TEST X was

([/color]TEST 1 time * 100) / TEST X time[/list]

percent of the flow rate for TEST 1"


PS - It would be great to have pictures of the 3 petcock positions to go along with the numbers. 

PPS - If you don't have extra tubing to use in steps 3 and 4, leave the hose attached at the fuel tank end and detach it from the frame-mounted petcock end.  Something like in this photo, but with only one hose detached and no couplers or "extender hoses".

Uggghh. This reminds me of related rates problems in Calculus and Diff EQ.

The petcock outlet is a circle -- actually two circles side by side. Anything less than dead on gets no circle and starts making a elipsoid iris. Call it a "cat's eye" for lack of a better name.

The two arcs of the cat's eye do not change, but the distance between the former diameter now becomes two changing values -- a short distance between the arcs and a longer distance at the arc intersections.

Here's a break away shot of the  petcock --


I took this shot as an analysis of why my petcock was leaking around the valve "handle". The o-ring that seals the handle was cracked. Also there was decrepitude around the ball of the valve allowing petrol to seep past the interface.

BTW Kerry - it is great to see you here again! 

Has anyone ever tried to adjust or disassemble the vacuum part of the petcock? 

If my theory is correct (that the problem is caused by insufficient engine vacuum when the engine is under extreme load), then a modification to let the petcock flow with less vacuum would fix it.  There's got to be a spring in there that regulates the vacuum necessary to let fuel flow.  However, I think I remember a post where someone said it couldn't be taken apart....if true, that's a bummer.

The automatic petcock is a great thing, but a PITA on the highway when it doesn't work right.  I grew up riding in the old days when EVERY petcock was manual.  You had to switch it ON before you rode off or the carbs would run dry.  It's a shame to have to use our modern petcocks in the same way.