Open up the meter and take a look inside or look at the youtube vids to see why it survived, and with
no fuse in 10 amp range. The meter measures via the voltage drop across a wire shunt that internally connects the 10 amp jack to -ve.
The data sheet's 21 amps is the useful number you want. It is virtual, determined graphically by extrapolation and referred up to 75 Celcius. By direct measurement you get a ball park figure. The complicating factors in the motor are magnetisation and copper temperature coefficient.
Another meter may be the easiest way for you. However you could:-
1) take a measurement at lower voltage and then refer it upwards (sensible and perfectly valid)
or
2) measure volt drop across a 0.01 Ω resistor, in a similar way to the internals of the meter connected in the 10 range.
Copper resistance/temperature coefficient is high. Multiplied by 100 C temperature rise shows resistance increases by the order of 40%. That is why:-
1) the meter shunt and resistor material is an
alloy.
2) as the motor winding heats up the current drops before your eyes.
3) a rough home made copper shunt is only within calibration in a particular current/ temperature point.