We just set up an HP 1090 HPLC. We injected a standard (biphenyl) and noticed the peak was broad and it tailed. We transferred the column and tubing to our HP 1100 and injected the same sample (all other conditions remaining the same). The peak width and symmetry on the 1100 was fine. Here are some numbers:
HP 1090 - Width:0.147 Symmetry:0.59
HP 1100 - Width:0.094 Symmetry:0.95
We just replaced the rotor seal but it didn't make a difference. Even tried injecting less and it still looked awful. The flow cells are also the same.
Any ideas?

Without knowing the exact method this is just a guess but have you checked that the pumps are performing correctly? i.e. that the amount going from pump to column is correct? If one of the pumps is not pumping the right amount of eluent through this may cause your peak to tail and broaden.

Thanks for replying. I guess I should have included more details. The MP is 95:5 (v/v) ACN-Water, flow 1.0 mL/min. The retention times are exactly the same. I measured the flow of each pump (separate and together) and they are delivering the appropriate amounts.

Assuming this is a well-maintained system, it sounds like time to troubleshoot.

Your description sounds like a dead volume problem. Dead volume can exist anywhere in the system.

You've already checked the column, so that's not it.

Now put some capillary tubing (to provide backpressure for the pump) between the injector and the detector to confirm that an injection without a column also gives a tailing peak. If it does not, then the problem is in some connection that you've already removed. A poorly installed ferrule, leaving a dead volume in a fitting, can cause this.

If you do find peak broadening without a column, then you need to start looking for the offending component. This can be subtle. I have found, for example, that the very convenient cartridge guard columns sold by P-E/Brownlee will cause poor peak shape if the seals get loose. (Change the seals. If tight, no problem.)

Poorly installed fittings will broaden peaks. A sample loop is a particular offender because it is tempting to install it, tighten the fittings and go with it. You'll never know that way that you've got dead volume. You have to install one end and swage the ferrule on. Then remove the loop and install the other end and swage the second ferrule on. Remove it again and reinstall both ends. Also, don't reuse a sample loop because it's too easy to confuse loops from different valves, and one valve may require a different extension of tubing beyond the ferrule than another valve of the same brand and model. (By the same argument, change all tubing to a valve when you change the valve.)

One far-fetched notion: I vaguely recall that the HP1090's we had (gone now, and I never used them much) had two alternative column connection lines. One was smaller ID than the other. I believe the provision was meant as a convenience to the user, but it was a bit confusing. You might want to check that you connected up to the proper line.