Strong internal pulses showing up at the Keithley digital voltmeter input

[bb1]

Unlike analog voltmeters (multimeters), digital ones have internal pulses.
It is natural to expect that to some degree these pulses will leak to the input of a digital voltmeter.

I connected input of Keithley 199 multimeter/scanner by 50 ohm terminated BNC cable to
the Channel 2 input of a Keysight scope. The Keysight channel 1 input was connected to the Keithley "Meter complete" output.
The Keithley was set to measure DC volts at the 30 volt range.
I set the Keithley measurement interval to 115 mS to avoid confusion of seeing dummy Keithley integration signals.
Otherwise the Keithley had all the default settings. The Keithley LEDs showed 00.0000 volts.

Attached Keysight screenshot KeithleyInput_50ohm_135mV.png demonstrates that the
"Meter complete" signal indeed had period of about 115 mS.
The screenshot shows four regions of pulses, each region of 16.67 mS duration.
Obviously, these regions are related to the Keithley four integrations, each of 60 Hz period duration.
According to the Keithley manual, they are "SIG ZERO", "CAL ZERO", "CAL" and "SIG" integrations.

What is remarkable here is the amplitude of the signals. It is 135 mVpp on 50 ohm load.
So the signals are not exactly feeble "ghost" ones.

As is shown on the KeithleyInput_50ohm_20.8uS.png the pulses during the integration regions mostly come with period of 20.8 uS (48 kHz).
During each 20.8 uS time the integrator output goes up and then down.
The pulses appear at the moments of change from up to down and from down to up.
The period is determined by the U17A and U18A on the Analog board, which divide
3.84 MHz crystal Clock signal by 80.
The 20 mVpp background seen on the screenshot is this 3.84 MHz signal,
which is gated on the Analog board, and then counted on the Digital board.

It is not clear how much these short duration pulses of noticeable amplitude leaking from the Analog Board to the Keithley input change the measurement results.
They may affect the circuits inside the multimeter.
They also may affect devices being tested.

[bb1]

Screenshots