Precision problem on GC-FID

[GalantVR]

The signal of injections could not be reproduced.

GC-FID i am using: Agilent 7890A
autosampler G4513A

Injection mode : split 150:1

I have changed the syringe (gas tight),
septa,
split liner,
tried split ratio 50:1,
tried slow injection,
tried cut column for improperly cut column,
cleaned jet with acetone.

gold-plated seal has used for 1 yr.

Does precision have any relation to the column?

Thanks in advance for reading this.

[Don_Hilton]

Some compunds are easier than others. Can you tell us what the compund is?
How much of a problem is the problem? Are we looking at an RSD of 5%, 50%...

How is the peak shape, and what does the integration look like?

Is this a standard mixture or a sample with matrix present?

Is the irreproducability a trend across several runs or is it random?

Are we talking about area, height, retention time, peak width... ?

[Hornet]

It could be either a leak, active site in the column (not in the liner since you changed it) or maybe injector temperature is too low and doesn't vaporize all the sample.
Which compunds are you injecting? Is the column old/heavely used? When a column is at the end of it's life active sites (polymide surface) start to appear and they will adsorb the sample, this happens especially if you have a thin film.