[rose9090]

My HPLC is 15000 theor plates but the baseline is really noisy. I was thinking of either increasing the response time or decreasing the attentuation. im not sure which one would help more

[Ligte]

Let assume you are using RP or NP with UV detection.

A noisy background can be a result of:
1. Mobile phase contamination which in turn will contaminate the detector cell from working sloppy or with low quality materials
2. Leak in system
3. Your column could be leaking packing material or the connection made by the peak tubing to the column is bad.

[voidSetup]

If you're doing normal phase with a buffer(s), you might want to filter the buffer first before adjusting the pH.  However, sometimes this does not always help and can make it worse depending on what your buffer is.

[Arkcon]

rose9090:, you really haven't given us enough usefully information, and everyone has started flailing randomly.  To summarize what two people have already said, and to try and direct the conversation usefully ...

Make sure your system is set up properly.  That it doesn't have a slow leak, the pump isn't failing intermittently, your mobile phase is fresh, your column is clean, etc.  Having adequate plate counts doesn't mean all of that is OK.  Also, check your detector:  make sure your lamp isn't bad, there isn't air in the flow cell, the flow cell is clean, etc.  Make sure your chosen wavelength is appropriate for your mobile phase (this is important information you left out.) 

Generally, we work with these sort of things, before we adjust detector parameters.  If you're sure they're OK, you should analyze a known standard, at a dilute level, trying to mimic your lowest analyte sample.  Under those conditions, you can see if you can cut attenuation, and still get useful data.  Likewise, if your peaks are very small and narrow, increasing responce time can adversely affect the detection of peaks.

[marquis]

Can you check the age of your lamps in the UV/VIS detector?

As the lamps age, the noise goes up, sometimes dramatically.