When I did my tests I learned that the brass needle will close off fuel level by its weight alone when the fuel level is at a point that the the tang on the float is in the "dead area" of the needle.
Now do your test with the float set as the Triumph Service Bulletin. You should make two observations: 1. The float will upset the needle, and thus allow fuel to flow, with very little drop in fuel level. This is because the float's tang sits in the needle groove at an angle. When the float is adjusted level the float has to drop about 1/4" for the tang to move the width of the "dead area." When it is adjusted by the bulletin the float only has to move a tad to upset the needle. 2. When the float is adjusted per the Bulletin when adjusted so it hits the float it almost immediately upsets the needle. When adjusted level the tickler has to move the float down at least 1/4" before it starts to move the float far enough to upset the needle.
While the concern has been about fuel level, and the effects of head pressure on the needle, there is more going on with the overall operation of the carburetor with the position of the float. The Bulletin levels seem to give more than ample fuel supply as I have run these float levels on a reasonably competitive 750 twin, and a fairly competitive 500 twin, running the high speed banks of Daytona, and never had a fuel starvation problems.
I also haven't experienced what ever symptom that has been seen by people who wish to run the float level with the top of the bowl. If we have to rely on an aluminum vs brass needle to make these carbs work we are all in trouble. What I have been saying over, and over, there is more to the positioning the float per the Triumph Service Bulletin than fuel level in the bowl. Doing it overcomes the problem of relying on fuel head pressure to upset the needle and get fuel flowing. If we had to rely on the float when it is set level, the fuel level would have to drop the float at least 1/4" before it would upset the needle. With is set down the .080" fuel level only has to drop .010-.020" to upset the needle and have fuel flow.What I wouldn't do
is to bend the tang of the Stay-up float if you wish to lower the float to levels shown in the bulletin. Doing this creates the same "dead area" situation you have when the float is level. If I was to change the float level I would still move the seat. Bending the tang also has the effect of limiting the height the float can lift the needle with the chances of limiting fuel flow.
Think about it...