Solved!

[001]

Photo 1 - two channels grounded but traces not parallel to grid (and not horisontal anyway)

Photo 2 - chanel 2 connected with 36" RG58 directly (without any dividers)  to build-in calibrator 2V 1kHz but fronts are ugly 

How  can I fix it?

THANX! FIXED!

[w2aew]

The tilt is often caused by an external magnetic field, sometimes even the earth's magnetic field.  Most scopes include a "trace rotation" adjustment to correct for this.  Usually, it's a screwdriver accessible trimmer accessible on the front panel.  For some older scopes, there were wingnuts on the back of the scope that would allow you to physically rotate the CRT to correct for the tilt.

The calibrator signal looks as expected.  Remember, the sole purpose of the calibrator signal (which is more accurately called the "probe compensation signal") is to adjust the low-mid frequency compensation adjustment on 10x probes.  It isn't really designed to drive coax directly.

[001]

Thanx a lot!

I trim my scope now and what I see:

photo 1 - calibrator and 1:10 probes
photo 2 - calibrator and 1:1 probes

Is it fine?

Signal front looks "reversed" isn`t it?

[w2aew]

Thanx a lot!

I trim my scope now and what I see:

photo 1 - calibrator and 1:10 probes
photo 2 - calibrator and 1:1 probes

Is it fine?

Signal front looks "reversed" isn`t it?

Your images look fine/normal.  The calibrator is *only* intended to be used on 10x probes so that you can adjust their compensation adjustment.  The rounded edges on the 1:1 probe (and your earlier direct coax image) is due to the higher capacitance that each of these connections have.

[Gyro]

@001:

https://www.eevblog.com/forum/repair/gw-instek-gos-653g-display-distortion/ 

[james_s]

Epic? This looks like a very minor issue that is part simple adjustment and part user error.

From the title I was expecting smoke damage and a crater in the PCB.