T4D #83 - Agilent DSO 2000 memory upgrade - DSOX2MEMUP

[MJLorton]

Hi Folks,

I hope to be actively online soon and getting back to communications. The  new lab is almost ready...it's been crazy few weeks...more of that in a T4D soon.

Here is an email thread related to this video which might be of interest:

-----Original Message-----
From: Martin Lorton  Martin@MJLorton.com
Sent: Saturday, November 16, 2013 4:11 PM
To: ASAY,BRIGHAM (A-ColSprings,ex1)
Subject: Lifting the lid on oscilloscope triggering

Hello Brig,

My name is Martin and I recently read your article "Lifting the lid
on oscilloscope triggering" (thanks, great read) while doing some
research to understand a triggering "issue" I have on my DSO-X2002A.

I run a Youtube channel on "electronic measurement and solar power"
(www.youtube.com/mjlorton). I've just upgraded the memory of my
DSO-X2002A (DSOX2MEMUP) to improve the sample rate and I hoped the
triggering on infrequent pulses or bursts.

I have a far cheaper ISO-TECH (re-branded GW Instek) IDS 6072A-U
which is able to trigger on such pulses.

I have a video demonstrating this (http://youtu.be/Dts-9nRH87s) and
I hope you can either point out what I might be doing wrong...or
explain technically why the Agilent is only able to trigger once the
frequency of pulses / bursts is increased?

The video noted above is current set to unlisted so only you can
view it with the link.

I look forward to any input you may be able to provide.

Cheers,
Martin.

On 11/18/2013 3:11 PM, Richard wrote:
> Hi Martin,
>
> My name is Richard Markley - I'm the product planner for Agilent's
> InfiniiVision line of scopes.  My colleague, Brig Asay, forwarded me
> your note.
>
> Thanks for writing to us about your question (and thanks for using
> an Agilent scope!).  I believe I know why you are seeing what you
> are seeing, but I'm clarifying it with our design team just to be
> certain.  In the meantime, I wanted to let you know we had received
> your message and will reply shortly.
>
> By the way - I watched several of your video blogs this morning.
> Very nice.
>
> Take care,
> Rich

Hi Martin,

I just got out of a meeting with the R&D team.  A couple comments:

1.  We believe we know what is causing the inability to trigger in your video.  It isn't related to memory depth, but is likely more related to how quickly we rearm the scope in "auto" mode.  While we don't have an IsoTech scope to compare, here is what we do when working in our autotrigger mode:

1.   An acquisition completes.
2.   The pretrigger delay needs to fill. If an valid trigger event happens here, then we'll miss the qualified trigger (since we're not ready yet).  All scopes work like this, regardless of memory depth.
3.   Now we start looking for a trigger event, and start Autotrigger timer
4.   If a trigger event occurs before autotrigger time times out, then it is a regular triggered acquisition , go plot, and go back to number 1
5.   If a trigger event does not happen before autotrigger timer times out, then to an autotrigger acq.
   a.   This means, stop the acquisition, and plot what is in the acquisition memory
   b.   Go to step 1 again.

2.  You are probably thinking to yourself "That's great.  Why does the IsoTech trigger more consistently?"   Like I said, we don't have one, but we suspect it is the autotrigger timeout window is much larger than Agilent.  It is likely purposely made much larger since the plotting of acquisition data is much more resource intensive operation in the IsoTech than the 2000x (because of our MegaZoom ASIC) and would likely cause more dead time if they were attempting to plot more frequently (which would mean they would miss many events during that dead time).  Because we have such a fast update rate and can plot data so quickly, we purposefully had the autotrigger timeout window very short.

Having said all the above, the team is actively looking at ways to enhance our autotrigger timer.  I'd like to give them some time to work on it and test different ideas.  Would an update on our latest thoughts in the middle of next week be reasonable to you?

Thanks,
Rich

[Cdngreybeard]

The Fluke reading of 8-9 waveforms/second (wfms) is typical for a DSO with a 10 ms timebase setting. Once a DSO triggers on a waveform, it cannot trigger again until the display update completes.  This is a part of what makes up a DSO's 'dead time'.  While in the 'dead time' zone, the DSO is blind to all other input signals.

The theoretical maximum wfms update rate may easily be calculated using the time base setting x the number of screen divisions.  For example, 10 ms x 10 divs = 100 ms. 1/100 ms = 10 wfms.   The actual wfms may be less than this maximum because of various scope settings and the scope's processing time.  One parameter that can have significant impact on wfms performance is memory.  As more memory depth (tracking points per sample) is selected, fewer wfms will result.

This is a nutshell explanation of the DSO's wfms rating, but it should clear up some of the confusion around why this numer is often so much lower than the published rating.

Other types of scopes, such as the DPO, get around this limitation by implementing parallel processing techniques, but they can be significantly more expensive than DSO scopes.

[JuanPC]

Well, decided to do some test with my GDS-3352 & ProTools.
and record some videos...

I think "I know what the problem is..."
in my GDS-3352 FW.v1.21
the Trigger Menu has several options that behave like Agilent DSOX Trigger when activated... for example:
Trigger (Roll or Normal) Mode, Coupling (AC,DC) Reject ----> (Off, HF, LF*),
Noise Reject (On, Off), Type (Edge, Delay*, Pulse Width, Video, Other), Source (CH1, CH2).
Etc...

The Trigger in the GDS-3352 can detect very slow & v.fast pulses, but... nothing is perfect.
misses some pulses between >100bpm <400bpm at <10ms/div. in Auto-Mode (Untriggered Roll).
At >=20ms/div. Works OK.
In Normal Mode also works OK, but Waveform display does not auto-fade like Auto Mode, until a different waveform is detected.
Like for example the Click track in ProTools or any other Audio software, the 1st click is different than the other ¾ clicks of a 4/4 measure.
Any Pulse faster than 400bpm or slower than 100bpm is detected OK in Auto Mode @ <10ms/div.

20ms VPO Auto vs. Off
http://www.youtube.com/watch?v=GiodjwUPxMM

20ms vs 10ms.
http://www.youtube.com/watch?v=4GSWboL9g7s

Missing pulses
http://www.youtube.com/watch?v=ktB6uWxMFSA

Scopes not perfect, specially under $1million dollar.
So, all comes down to need, what i need “the most!.”

In my case I wanted 5GSa/s RT to measure jitter in 12/24/48Mhz clock signals from soundcards.
Also very accurate (“low ppm”) hardware frequency counter. *Must have feature.
High BW as I could, for calibrating old FM radios & looking SQR waveforms, measuring Rise Times.
and as low $$$ as possible.
The Answer was Instek.
Did I mention Bigger Screen? 800x600

Any other option was too much $$.

It has a Small memory, 25K points, I wish it had 50K, so it would not slow down at 50ms/div.
and Pause & Zoom was longer, But... Logic Analyzers are much better for that.
does not have the 10Mhz Atomic Clock input like others in the 5GSa/s range, with similar specs, twice the price$$, & some with smaller screen.

Instek GDS-3000 is not an All.In.One, Do it All scope, but what it does, does very well.
¿The new GDS-2KA?, has a bigger memory, a Logic Analyzer, but... not 5GSa/s, nor 350Mhz & higher ppm.
Very powerful 500MHz 32ch Logic Analyzers can be purchased later at $500usd.
That was not a worry, but higher ppm, not 5GSa/s, not 350MHz was a deal breaker.

[w2aew]

For what it's worth, I ran this same test on the Tek MDO3000, and it was able to trigger reliably on the infrequent pulse, even in Auto trigger mode.