ok a poured exactly 1L water in a bottle and another 2L in another bottle and grabbed afew bags of 1kg flour n sugar in the kitchen.
placed the scales on a perfectly flat sheet of glass cos flatness affects their accuracy.
i found that one digital scale only increments at 0.2kg at a time so won't use that.
the other 200kg digital scale, which were used to corner balance the coilovers, increments in 0.1kg, has a lower reading limit of 2kg and after weighing various combo's of flour & bottles its seems quite accurate.
before adjusting the wrench allen screw i thought it was abit stiff so i thought heck, lets clean up the mechanism after several years. like any machinery, then cleaner it is the more efficient n accurate it'll operate.
so this is the insides of a sprung torque wrench
the roller pin end of the mechanism was full of old gunked caked brown grease, no wonder it causes the vague inconsistant results of the graph
mechanisms all shiny clean
regreased
reassembled.
i position the weights to where it'd produce 28Nm, set the wrench to 28Nm and adjust the allen screw till it just clicks at that distance.
after the first series of readings (purple) the cleaned mechanism calibrated to match at 28Nm produced a much steeper line
recalibrated to match at a higher 40Nm to see if that flattens the curve (yellow) but nope its just as steep after cleaning the mechanism
seems that the bad habit of storing the wrench forgetting to unload the spring has permenantly wrecked its springrate
woops
think its time to order that new digital wrench