I am looking for some literature/tech support that investigates the variations of human and instrument mobile phase mixing as well as variation of solvent mixing between instruments

In short can a a single chemist produce more reproducible data mixing his/her own mobile phase or can an instrument(s) do a better job

instrument do not drink coffee...

A modern HPLC will be more reproducible day to day, week to week than a single chemist.

Reproducibility? The instrument is much better. Accuracy? Most instruments have no clue what a mixture of 49/51 is. Sometimes it is 45/55, on another instrument it may be 52/48.

The issue is more complicated than any of the commenters have discussed. If the method is sensitive to the amount of organic then it is important to know what the true ratio of mobile phase components actually is. This is more critical for a method that would have to work in many different laboratories. It appears that there are zero methods in the usp that advocate instrumental mixing of mobile phase. In addition it would seem that someone should have conducted a study on this but I have not been able to find one.

Which method is not sensitive to the amount of organic?

We test as a routine in our validation protocol the differences between mixing automatic and prepared by an analist.

Only in one method we saw significant differences, and then we include in the S.O.P.

"It appears that there are zero methods in the usp that advocate instrumental mixing of mobile phase."

There are so many poor methods in the USP it makes it hard to to consider them an advocate for anything.

The USP is the official legally enforceable method for products labelled ,USP. I have seen a lot of chemists claim that the USP doesn't work. Usually the issue relates to how the individual tries to run the USP method or the companies that submitted the method to USP neglected to include critical information such as the need for instrumental mixing. Therefore, can anyone make a scientific comment to how well the methods tend to behave on instruments by different manufacturers? Anon 3:24 how well do the methods behave(numerically) and under what validation section(robustness vs.intermediate precision) do you include this in your submissions?

This whole discussion is really quite amusing. We depend on our HPLC hardware to deliver accurate and precise flow and mobile phase composition when we use gradient methods. However, as soon as the method becomes isocratic, many users suddenly think the LC's can no longer deliver an accurate, precise, reproducible solvent composition. I understand the concerns when moving methods from instrument to instrument, but remember that the same model from the same vendor are vitually identical as they leave the factory. This means that untill those HPLC's are modified or change by users, you will get the same result LC to LC.

The same precise results, reproducibly? Yes! But accurate? Where do you think the changes between vendors are coming from?

I will agree, some vendors do a better job at accuracy than others. Most often, however, changes in RT between 1 vendor and another are due to differences in system volume and column heater control.

System volume is a factor in gradients, not in isocratic separations. Differences in isocratic separations between instruments and between the instrument and manual mixing are due to inaccuracies in mixing. I don't think that the inaccuracies between instrument and man are due to the imprecision of man.

Post injector volumes vary from vendor to vendor. Post column volumes differ. Detector cell designs and volumes differ. And as was mentioned before, column heater control. All of these factor contribute to vendor to vendor RT differences in isocratic chromatography. How inaccurate do you (anon 04/14/7:59) belive instruments from the major vendors are? 1%, 5%, 10% I am truly interested in your thoughts on this matter.

We are not talking about microliters in connection tubing. Here is what I suggest that the doubters do. Take three HPLCs in a single room. Your new one, one from another manufacturer, and one of these old clunkers from the QC department. Put them into the same room to eliminate the question of temperature. Take a reasonably new column. Make big buckets of mobile phase A and mobile phase B. Then do the same separation on all three instruments on the same column, letting the machines do the mixing. You will easily get 10 to 20% difference in the retention times for an assay with 30 to 70% organic. Then run a second round to see if the column has changed. The retention times in each instrument will be within 1%. Now take your buckets and mix them by hand, and do the same experiment. Now the retention times will be within maybe 3 to 5% from instrument to instrument, due to slight differences in the flow rates.
After you have done this exercise with the high organic content, do the same thing with an assay that requires 5 to 10% organic. Now your differences in retention times between the different instruments may easily be a factor of 2.
Of course, you might say that you do not recommend to have the machines mix only 5% organic. Well, wasn't this the point?

Another factor you may wish to consider (for isocratic elution) is that, if you require maximum sensitivity, your baseline noise will typically be lower using only one mobile phase component (i.e., human mixed).

These appear to be the first examples that demonstrate the superiority of humans over machines. The machines beat us in chess, but apparently not yet in chromatography.