I think Rob was asking for even more details:
- The analytes are proteins, right?
- Column dimensions?
- Peak widths?
- Temperature?
To add to Rob's comments:
- a 1 minute separation between peaks should give you lots of separation (like, Rs > 5) --
if you have enough efficiency (plates). This is where the modeling software comes in handy, because you can play "what if" games.
If this were my problem, the first thing I would do is to run an isocratic standard through the system and measure the plate count. A somewhat optimistic rule of thumb for expected plate count is N ≈ 3000*L/dp, where L is the column length in cm and dp is the packing particle diameter in microns (yes, I know the units don't cancel; this is only a rule of thumb

). If you're getting anything above half that value, your system is OK. If you're getting much lower efficiency, you have to address the reasons.
If you're using a narrow-bore column (which I suspect from the flow rate), then take a hard look at the extra-column volume in your system.
If the extra-column volume is reasonable, look at your peak shape. Do your peaks tail or look "funny". Proteins can be particularly problematic because you can have multiple conformers of the same protein, each retained a bit differently which merge to give you a single broad distribution. Lack of sufficient buffering can also contribute to peak shape problems.