Bytes missing when decoding UART with Siglent and Rigol, but Magnova is ok

[H.O]

OK, that's a lot of bytes...can you try doing a single shot capture at a slower timebase to see if there are delays between each 'row' of text? I'm starting to think that you're experiencing the 're-triggering' issue I described after all.

[Sorama]

OK, that's a lot of bytes...can you try doing a single shot capture at a slower timebase to see if there are delays between each 'row' of text? I'm starting to think that you're experiencing the 're-triggering' issue I described after all.

here you go:
a screenshot of a single trigger where you can see the the end of line 1 and the start of line 2.
Have a look at line 38 in decode list which is a space, followed by " (line 39, a LF?).

Line 40 is the very first character of the next line (ASCII J).

Is that what you wanted to see?

Settings of scope Acquire:
Interpolation: Sinc
Acq Mode:Slow
Acquisition: Normal
Max Mem depth:20k

Tnx!

[Sorama]

The same exercise, but with a slower timebase and mem depth set to 2M.

Now the crossover between end of first line and start of second line is at line 5, which is either a comma (,) or double quote (") as can be seen in the decoder list at line 5.

The character J is the first letter of the second line (Junior).

Edit: not sure what that first character is , a squared o at the very beginning of the waveform?

[Sorama]

The same exercise using a Rigol scope (MHO98).
That reveals the CR and LF in the decoding list.

(Why doesn't the Siglent do that?)

Sam.

[H.O]

OK, I think what's happening is that the scope triggers on the startbit of the first byte, fills the memory, re-arms the trigger and triggers again on the startbit of a byte somewhere in the middle of that long string.

The message is about 800 bytes long. At 115200 that's roughly 70ms. If you run the scope at a much faster timebase (as you showed in your first post) it will fill the buffer WAY before the message ends and it WILL retrigger in the middle of it.

Does that make sense?

(Yeah, the decoder could use some work, for sure)