I'm curious to know exactly how you derived these measurements. Specifically, did you destroy a new 80ml P800 cartridge by prying it apart and emptying the contents of the bladder into a measuring cup and found exactly how many ml poured out? Or did you derive this in some less direct manner?
for this specific case - the measurements were somehow mix of direct and indirect ways:
1) the full, new cartridge weights 143 g (average of 11 different cartridges +-2 g, weighted using two different devices, however not laboratory ones)
2) completely empty, with all ink manually removed from inside weights 63 g --> so, there was 80 g of ink inside (average of 11 measurements)
3) empty in Epson understanding, when needs to be removed weights 75 g (again, average)
4) today I have specifically opened "empty" VM cartidge to check, and there were 15 ml of ink left, that weighted 17 g what gives 1.13 g/cm3 of density (the weighting device resolution is 1 g so it could be also less or more; capacity measured with medical syringe, opened with blocked ending.
assuming even 1.1 g/cm3 (closer to widely used value of 1.08 g/cm3) those 80 g transfers to 73 ml of the TOTAL ink capacity including unused ink and 62 ml of usable capacity (even worse than I have written in previous post).
now, I wait for an occasion to buy cheaper, outdated cartridge for 3880 and open it unused and measure capacity in ml
I am ex-metrology guy, so have some understanding of potential sources of errors and measurement protocols
all the values for points 1)-4) are the same for 3880 and P800 cartiges including P800 starting ones (in that case, with except of 1))