You could use 20% of the unreacted litres as methanol but that will leave quite an excess in the glycerol and biodiesel.
I have successfully used 12% on each stage though.
On a 300 litre batch it generally goes like...
Glyc wash with 45 litres of glyc from last batch. Stop, drain glyc, add 30 litres of methanol + 6 litres of methylate, mix for 45mins, take sample, stop and wait for an hour.
Then I measure dropout of the nicely settled sample (normally is around the 2.0l mark) and drain the first stage glycerol.
Then multiply the dropout ratio (ie 20% normally) by the remaining batch size once first stage is drained.
It is important to do it that way I feel rather than just assume the batch size remains unchanged since, depending on moisture content etc, the batch size at the start of stage 2 could vary up or down.
Assuming it was 2.0ml dropout and remains at 300 litres batch size, the unconverted litres value is 60. Therefore I would mix up 4.0X60X5 = 1200ml of methylate with 12% of 60 = 7.2 litres of methanol and suck that in with the venturi.
I then start taking 10/90 samples from around 20mins mixing onwards and keep going until I get a perfect pass. If it's not perfect after an hour of second stage mixing I know I have messed up somewhere so just stop, drain second stage glyc and do a third stage. It is very rare that I have had to do this though (generally it has been down to me being in a hurry and having wet oil or not letting the glyc wash separate for long enough)
