Oscilloscope noise. mV without any input signal ?!

[Hatimoooo]

Hello everyone!

I'm having a trouble with a Tektronix TDS 3045B oscilloscope.
Here is a screenshot of a signal where I connected the probe pin with its ground.

I can't understand where did that voltage come from. I need this measure to be very precise because I want to detect 4.88 mV signals
Should I use some filtering? If so, let me know some of your tips.
Thanks a lot
PS: The probe I'm using is a Tek P6139A

 

[barry]

I would try looking at the input with NO probe and see if you've got the same issue.

 

[Hatimoooo]

Hello again.
When I remove it, it gets a bit better 1.10mV instead of 11.3 mV.
The probe I'm using is a 10X one, that means that the oscilloscope can go to 10mV/div. But without it, it can go to 1mV/div, and this makes the noise even lower (something like 680µV).