weird thing,please help

[creageng]

I inject my new column (luna 5u C18(2)) with the standard solution (Uracil,acetophenone,banzene,toluene,anphthalene), the running conditions as column tested in the comany:ACN 65% isostatic.254nm,agilent 1100.
Here is my tested result:
Toluene Rt=4.5 peak height 0.6mA
Naphthalene R=7.0 peak height 0.4mA

I can not obviously see other three peaks because baseline 0.1mA fluctutian overlapped them.

In the Certificate provide by the company all the five peaks are discrete and the height is 70~490mV,baseline flat and at least below 10mV.

My problem is that my signal is very weak. And I tried several ways by replacing the column with just guard column or et.al ,injecting the same amount of sample.here are the results:
1)guard column+analytic column: two peaks as I mentioned above ,0.6mA
2)analytic column:two peaks are clear (Toluene and Naphthalene),0.6mA
3)guard column: one spike peaks,<1mA
4)connection tube: one spike peaks 1000mA
all the tested baseline are almost the same.

AND injecting the BLANK (5% ACN) did not have a peak >1mA bypassing the colomn with connection tube in 4)
so that is the sample peak (1000mA), andS/N is very good,
BUT THE PROBLEM is that S/N is very poor if connecting the column/or only guard column.
what is most likely the culprit,
I am appreciated your help.
creageng

[tom jupille]

If your system has a so-called "needle in loop" autosampler, I would look for a leaking seal where the needle docks into the injection valve. When you bypass the column, there is much less back-pressure and the injection works OK. When the column is connected, the higher back-pressure means that most of your sample leaks out.