How did you measure the current? if you have all four wired up already you need to measure the current at the tailcap.. if you attempt to measure current as if you'd measure voltage.. what you are actually doing is shorting out the emitter through your ammeter and just pulling max current from the battery.
In-light current measurement:
remove tailcap
turn on light
set meter to DC A.. (10A).. make sure leads are in the correct jacks
touch Neg lead to bat - inside light
touch Pos lead to bottom edge of battery tube.
(press quick and hard.. if it's high-resistance it'll generate a lot of heat).
The batteries are all in series.. whatever current you pull from one it's from all of them... if.. you are measuring 2.3A from the battery.. you are only getting 2.3/4 = 575mA to each emitter.
looking at the C-size
datashee from duracell.. we can find out a few things..
1) 150mohm resistance.. means that for every amp you lose 0.150V.. if you pulled 8A from a C size battery you'd get (1.5 - 8 x 0.150) = 0.3V remaining.. so you didn't pull that one off.
2) The only rate up to 2A rate.. but at 2A.. they show 0.3hours to 1.0V and 0.5hrs to 0.9V/cell.
So.. let's stay within the rated values as an example.. if you tune the thing 'til you get say 2.0A from the batteries and 500mA to each emitter.. the best you can expect is 30 minutes of sharply declining output.
Looking at the D size.. you'll get over 1 hour to 1.0V and more like 1.7hr to 0.9V/cell.. but that means.. with 3D.. the voltage will drop to 2.7V in about 90 minutes.. and the voltage will drop continuously from the start.
I built a light similar to this concept with 3 luxe's and tried to run it with 3D and they didn't have the voltage to get 700mA to each lux.. I re-worked it to use 4xC in a 3D host.. and now it will regulate (with fatman) 700mA with fresh batteries.. but the math works like this:
at 700mA it takes about 2.5W per emitter.. and that means 8.26W from the batteries.. at 1.5x3=4.5V that means 1.83A to pull that off.. bump that up to 4 emitters.. now you need 11W from the batteries and at 4.5V that means 2.44A from the batteries.. the problem is.. that's fresh batteries no load.. more realistic.. you'l get 1.25Vx3 = 3.75V fresh and 3.0V toward the end.. and that means starting current of 2.93 and ending current of 3.67 for 4 emitters.. and 2.2 to 2.75A with 3 emitters.. so.. with fresh 3D you can JUST regulate 700mA.
now.. using 4xC instead:
with 3 emitters.. the input current will range from 1.65 to 2.29A but with 4 emitters the input current will range from 2.24 (upper limit) to 3.11 (not doable).
Those are all values figured with fatman driver..
running unregulated.. you will not be getting nearly the output you deserve... i most certainly would not try to run more than 3 emitters.. the more current you try to pull from the bats.. the faster the votlage will drop and output will suffer a LOT.. You will likely get more light from 3D x 2 emitters than 3D x 4.
Get some accurate tailcap current numbers and wait for the current to 'settle in'.. the voltage on the bats will drop dramatically when you first turn it on.
Also.. measure the voltage at the emitters (all will be the same if parallel so just measure one)..
What is the Vf of the emitters again?
From my experience, the happiest solution you'll find will either be a 3 emitter fatman solution.. or a 3D x 2 emitter direct-drive solution. It does depend on the vF of the emitter.. but with 2 emitters.. DD from 3D.. vBat will range from about 3.75 down to 3.0.. and with Jbin emitter that would be a range of like 1A / emitter down to 122mA/emitter (with dead bats).
With 2 emitters.. you can use the PQS and have the alignment be symmetrical.. also.. you will push the emitters at 1A and get on the order of up to 90L/emitter.. if you TRY to run 4 emitters DD from 3D.. you can expect to see the voltage start closer to 1.1V/cell average and get more like 45L/emitter = exactly the same output.. but for much less time.
you can also still use PQS not with the alignment marks and put three bare emitters on 120° apart...
It boils down to this.. alkaline do not have the POWER capability to run this much load.. NiMH not have a problem.. the trade-off is to use fewer emitters and maybe some resistance to keep the initial current below about 1.2A/emitter.
-awr