if you connect a signal greater than approx 7V DC or ACrms to that channel, you will burn that resistor and thus damage the oscilloscope, with an expensive repair as a result. I have seen quite some broken scopes for this very reason.
And since this is the "Bugs / Missing Features / Feature Requests thread, I'd call again for Siglent to 'fix' this by monitoring and automatically disconnecting the 50R input when the signal level gets too high.
Define 'too high'. You can't just add a comparator or so that looks at the voltage level. It would need a circuit that monitors power dissipation over time instead of just a voltage level. At some point I've fed pulses into a DSO for which I needed the 20V/div range in 50 OHm mode in order for the signal to fit on the screen vertically.
This is one important point.
Now in some Siglent models what have internal 50ohm load resistor before normal Hi-Z amplifier/attenuator,
display vertical sensitivity is limited to max 1V/div. It do not protect anything but limits usability strongly. If it "protect" something, it only happen in human (user) mind but nothing inside oscilloscope - this "protection" effect I do not believe so much. After 50ohm parallel load resistor there is nothing to protect but just exactly normal oscilloscope 1Mohm /xx pF input.
So this 50ohm load resistor(s) is component what is first in circuit and important to protect (of course also other things if want dive more deep to these things)
SDS models what have internal 50 ohm and what can also set for DC50 ohm or AC50 ohm (not all models where is 50ohm input available) need note that DC connected
50ohm load is before DC block. DC block (AC coupling) is there in 1Mohm input circuit after 50 ohm load. So 50ohm input is DC coupled to GND, independent of AC or DC selection, if AC selection is available.
If think this 50 ohm load and signal what it can handle it also good to look some common basics
https://www.vishay.com/docs/48516/_ms9702509-2003-vishaychecklistpulseload.pdfIt also depends highly about what resistors there are used and also what kind of drift we can accept in long time use or even due to single overload peak (Also resistors manufacturing quality and also front end PCB thermal design are important factors).
But it is extremely clear it do not fail if I have
just one simple example:24V positive pulses. Pulse width is 50ns and whole cycle is 5us (200kHz 50ns pulses from 0 to 24V)
Cycle power 5000ns, and and long time average is
0.115WPeak power(50ns) is 11.5W
Peak V is 24V
Peak I is 480mA
And now I think what ever resistor there is used... it do not break with this.
I believe it last well and without even long time meaningful drift or damage risk in practical user cases.
But user can not look this simple pulse with SDS oscilloscope using internal 50ohm... displayed full vertical is 8V with 1V/div.
This limit do not protect any single thing but rejects usability. It only torture users. Also can calculate with more extreme cases...
But also then need be careful, user can not know this 50ohm load resistor type, thermal dynamics and failure/permanet drift characteristics.
Naturally one good solution is external direct feed thru or attenuating terminator what can handle more power and if fails no need repair scope and with it, full V/div range can use.