I2C Testing (continued from Picoscope 6424E Road Test - I2C Serial Decoding )
Keen followers of my Picoscope 6424E Road test blogs will have noticed that I rather messed up on the I2C decoding by having the 6424E connected to my PC via a USB 2 only port.
Sorry about that !
But every cloud has a silver lining and the problems with the slow connection made me look a lot deeper into buffers and how they work than I might have done otherwise.
And I had meant to get round to looking at the impact of connecting via USB2 
The good conclusion from this is that most ordinary scope work is just fine with either connection.
First thing, lets round off the I2C testing:
Picoscope triggering once every 0.5s and storing to buffers using Rapid triggering mode.
The little multiple graphlet window in the Buffer Manager is activated by using the zoom buttons. With USB 3 it is acceptably fast even with 10Ms per buffer but its speed is affected by post processing operations like low pass filtering or serial decoding.
At 5Ms buffer size the scope could keep up in normal (not rapid) triggering mode even with low pass filtering and serial decode. The advantage of working like that is that the display is updated in real time.
The speed of serial decodes working on the entire set of buffers is much improved. These features are all affected by the USB link speed because almost all the processing is done by the PC, which saves the Picoscope itself from needing a Quad Core 4GHz processor with attendant hard drives, OS and power supplies.
So in conclusion, the serial decode offers some nice features, including the table view and the ability to work with multiple buffers and search across them for features within the data packets.
Buffers and Complex Triggering
I set up a 2 channel signal generator to generate 50ns pulses at 1us intervals (channel A) and 40ns pulses at 1ms intervals (channel B).
Then I used the Advanced Triggering menu (accessed via the pulse with an arrow key in the Trigger controls) to set up a trigger when both signals are high.
I'm storing 1MS in each of 1000 buffers, and using the Rapid triggering mode as well.
And to add a bit more excitement I enabled the Deep Measuring facility.
Complex triggering on faster signals
The Deep Measuring can trundle through your captured data and show the timings of pulse edges (there's a tool for setting it up.). The table of results can be dragged off the main window if you like. The data can be saved to a file. If you have 1000 lines in the table this makes sense - you can't eyeball 1000 lines of complex timing data in a useful way - so using MATLAB or the like on the exported data will be helpful.
You don't have to show the whole buffer set, you can just look at one buffer a a time, but I don't think you can set ranges.
It's a pretty neat capability and I don't see anything similar on other scopes.
I do have one gripe however, the Buffer Management window isn't that good when you have 1000 buffers - it can only show 24 at a time in the Graphlet array.
This is not a problem unique to the 6424E, my 1000 buffers of 1MS each total 1GS, and were set up and recorded with a few mouse clicks. But that is a lot of data, other scopes I know well can make pretty heavy weather of finding stuff in the 4 or 15 MS they can record.
None the less, once you've been spoiled by being able to record 1GS you then feel you ought to be able to navigate it easily. One possibility, offered by LeCroy, is to be able to search for complex patterns in the stored data - when you aren't working in real time multiple level triggers with maths operations are possible (like "after 76 rising edges on A arm C to trigger on pulse < 10ns wide") . Another one for the wish list.
I'm looking forward to trying to debug some disgusting FPGA problem with the Picoscope, I won't know for sure until I try, but I think it offers a better range of tools for that kind of work than anything else I have access to.
Pulses
I don't have the gear to do proper testing of bandwidth or rise time but I thought my ancient HP pulse generator (£75 at an auction) deserved an outing. It's from HP's slide switch era (which fortunately didn't last long) but it still works OK.
HP Pulse generator
I connected the pulse generator directly to the Picoscope and a Lecroy 610zi using a 50 ohm BNC lead.
Picoscope rise time 50 ohm direct connection
LeCroy 610zi rise time 50 ohm direct connection.
The result from the Picoscope looks pretty reasonable.
Picoscope rise time, supplied x10 probe using supplied springy BNC adapter
The result using the supplied scope probe is very good indeed.
This will probably be my last Picoscope 6424E blog - it's time to wrap it all up and do the formal Road Test report.
If there is anything you want me to look at then please let me know.
MK
