How to capture serial data with a Rigol Scope

[maketronix]

Has anybody taken the time to write up instructions / do a video on how to capture a serial bitstream (I'm thinking I2C as I write this) with a Rigol scope, so you can visualize the bits being passed / debug the protocol?  Just asking, because I tried and failed to successfully do this last week, and ended up buying a logic analyzer.    It seems like a lot of us have this scope, it would be nice to have such a training aid

[marmad]

You might get more responses if you mention what model/series of Rigol scopes you're asking about. I'm assuming the E series, but I don't know for sure (although obviously you don't mean the new Rigols, since all of the UltraVision scopes come with I2C triggering and protocol decoders built-in by default).

[maketronix]

Well personally I'm using an e series (DS1052e) but I'd certainly be interested in other models if 'similar techniques apply'.

[Harvs]

I'd say the reason behind the question about the E vs D is because the D has the logic level channels.

Obviously the DS1000 series doesn't have serial decode built in - which is a shame.  However, most of the time all I need to look at is whether the SPI clock polarity, _enable is timed correctly etc for the specific device I'm trying to talk to.  To do this, just a decent capture of the signal is more than enough.

So there's two parts to that, one is triggering on the right bit of the signal you're after, and the other is about capturing enough data.  Dave has explained DSO triggering in a number of videos and really I don't have anything to add.  Obviously a signal shot capture is what you're generally looking for to manually analyse the signal.

As to capturing the data, first thing is to turn on long memory depth (if needed).  Then bring up the horizontal menu and note it has the current sample rate. Assuming you know what you're bit rate is, then you can change the time base until you get the sample rate to a level that will capture the number of bits you're after (remember the 1MEG sample depth, which will be split between two channels.)  I keep to about 100 samples per bit as a minimum, and that should give you a 5000 bit capture at rates that can use the full memory depth. Obviously far more than you're going to be scrolling through on a 5" display.

Then once you've completed the capture, wind in the time base to zoom in and use the horizontal offset to scroll through the waveform.