YouTube movie about LA performance in Rigol MSO1074Z and MSO2072A

[pascal_sweden]

Would be cool if there is any user out there with a Rigol MSO1074Z or Rigol MSO2072A, who can make a YouTube movie where the LA features are demonstrated in any of those models.

Ideally a user who has both models, so that he can even compare the relative performance of the LA between the Rigol MSO1074Z and Rigol MSO2072A.

Thanks for making a YouTube movie!

[pascal_sweden]

Based on previous postings, there are many people out here with an MSO1074Z and an MSO2072A.
Go make that YouTube movie on the LA features

[AntiCat]

I can't offer a YouTube video. However I've prepared a few pictures to illustrate the MSO2000 series.
The MSO is delivered with two 10x 350MHz Probes and a 16Channel LA "Probe" set:


Below you see a screen grab with 3 of the 16 channels switched on:

• As you can see on the bottom of the screen, statistics are available per LA line.
• You can configure any logical voltage level you ever need. I tested 1.8V, 3.3V and 5V.
• Triggering options are similar to the future set of the analog channels.
• Every digital line can be switched on and off on its own. With all 16 channels switched on the screen space is very tight. Especially if you also use Channel 1 and 2.
  I wrote a script that uploads the data using the VISA interface. I prefere to analyze the data in PulseView (http://sigrok.org/)

The second screen grab shows D3 and Channel 1 connected to the same signal line:


If you have any specific questions feel free to ask.

[Gandalf_Sr]

Not sure I could come up with a good video but there are a bunch of videos out there.  This is one that shows how to use I2C decode on a RIGOL scope (the menus are all the same as my MSO2072A) and there are many more.  If you have a logic analyzer (LA) too, you can use any LA channel in place of an analog channel for decoding purposes - e.g. where the user selects SCL to be Ch1, you could use Ch1 on the LA.  The LA channel is not going to give you the same detail so you should use an analog channel to look at the suspect lines.

For SPI, you will typically need 3 or 4 channels, as data is always sent simultaneously between master and slave using 2 lines, MISO and MOSI (Master In Slave Out and Master Out Slave In) plus SCLK and some form of chip select usually marked SS.  On a DS2XX2A scope you only have 2 channels but the MSO2XX2A has 8 or 16 extra LA channels that can be used.

The main reason I wanted an MSO2072A was to add these channels for serial protocol decoding purposes.  I found it pretty intuitive but it's a bit clunky that you have to set the trigger channels up for I2C and set the trigger levels - then do the same for the decode.  I'd also comment that the MSO2072A is a bit slow in presenting the decode data when you crank up memory depth and sample rate; you have to hit [Run/Stop] and then trigger a sample set of data, then you'll see the scope chug through that data set decoding the data.  However, this is not a problem as that's what you really need to do when trying to figure out what's going on on an I2C or SPI bus; grab a big chunk of data, and look at it in slow time.

At work, I have a Tek MSO3024 with I2C/SPI decode and it is a bit faster but that's all; the list price for that scope is $4250.

[pascal_sweden]

There is a video now finally on YouTube. Interface looks quite neat, but seems there are some bugs.

[Mark_O]

There is a video now finally on YouTube. Interface looks quite neat, but seems there are some bugs.

I let the video run while I was doing something else (I didn't have 20 minutes to watch it in detail), so I didn't try to follow his description of the false triggering on logic states too closely.  However, his comment at the end about driving high-current LED loads directly from the outputs of his FPGA would certainly be of some concern.

But he also neg'ed the Rigol at the end on the digital and analog channels not being precisely in sync.  Which is true, but not unexpected, and not uncorrectable (as it was on older Rigol MSO's in the past).

If you look at the lengths of the leads on the A and D probes, I think you'll understand why that is the case (differential propagation delay).  Rigol built in an adjustment setting for trimming the skew between an analog channel, and the trigger point (Channel Delay Calibration).  Similarly for the digital channels.  I believe this can be used to time-align the A & D channels quite precisely (+/-20 ns ps), by sync'ing both to the scope's internal trigger reference point.

It's described on Page 2-8 of the MSO2000 manual I have, for the analog channels, and 7-8 for the digital channels. 

ADDED:  I don't have an MSO2000 to test this on.  And of course, you'll never notice this if your timebase is in the usec range.  He started off in ms/div, then zoomed to ns-timebases, when commenting on the lack of sync, near the end.  Also, the point where the triggers occur on a state-transition will be dependent on how he set his threshold levels on the logic channels.

His criticism that 'It takes a very long time for the analog channels to settle', overlooks the rise/fall times of the Cyclone outputs, when driving the loads he has there.  Lots of things will look "slow", when you're viewing at 1 ns/div.

[pascal_sweden]

The interface looks very neat. In fact Rigol has good taste for making the UI look nice.
I would even say that it looks nicer than Agilent and Tektronix, but then again I have not seen all models out there =)

[Ivan7enych]

There is a video now finally on YouTube. Interface looks quite neat, but seems there are some bugs.

Concerning the false triggering issue, in that video LA works in slow capture rate, but LA trigger can work internaly at full speed (1Gs\s) and can capture very tyni transient true state event (when some signals go from 0 to1, some go from 1 to 0, some pins make it a bit faster, some other a bit slower). In trigger menu there should be some setting, how long the logic state should be to raise trigger event.



I've played a bit with decoding a simple 4 bits LCD protocol, here are some pictures

D1 is sync (from 1 to 0),
D0 - defines commands (=0) and text (=1),
D2-D5 - 4 bits of data
decoder is set to show 4bits of data as a digit on every falling edge of the sync.

In the picture controller is sending strings " -T0- ", " -T1- ", " -T2- ", " -T3- " to different places of the screen.
First byte (a pair of sync pulses) defines a place on the screen,
Next set of bytes is a string in hex format, for example string " -T0- " = (20 2D 54 30 2D 20)

[David Hess]

There is a video now finally on YouTube. Interface looks quite neat, but seems there are some bugs.

Concerning the false triggering issue, in that video LA works in slow capture rate, but LA trigger can work internaly at full speed (1Gs\s) and can capture very tyni transient true state event (when some signals go from 0 to1, some go from 1 to 0, some pins make it a bit faster, some other a bit slower). In trigger menu there should be some setting, how long the logic state should be to raise trigger event.

I am not expert on the Rigol MSO oscilloscopes but there apparently is such a setting.  Page 7-7 of the user manual describes "jitter reject" which is suppose to prevent false triggering on asynchronous inputs.  Indiastarter who made the video was apparently unaware of this or confused it with holdoff.

The following page, 7-8, describes how to compensate for analog to digital probe delay.

However, his comment at the end about driving high-current LED loads directly from the outputs of his FPGA would certainly be of some concern.

While I consider it poor form to drive LEDs directly from highly integrated logic like an FPGA, it looks like his drive current is somewhere in the 2 to 4 milliamp range and this by itself would not explain the slow transition times shown.  It is more likely that he has unterminated long transmission lines or the oscilloscope probe ground leads are too long.

[Ivan7enych]

I am not expert on the Rigol MSO oscilloscopes but there apparently is such a setting.  Page 7-7 of the user manual describes "jitter reject" which is suppose to prevent false triggering on asynchronous inputs.  Indiastarter who made the video was apparently unaware of this or confused it with holdoff.

Manual tells, that "jitter reject" works on analog channels only.

There is a "Duration" trigger mode, where you can set pattern and a minimum time duration to ignore spikes (or you can set maximum time to trigger on spikes only).

The only drawback of this trigger mode is, it triggers on the end of the event, not on the begining. It looks like oscilloscope measures time from start to the end, and if measured time meets criteria, it raises trigger here.

On first screenshot oscilloscope triggers on small spike, because minimum time is too small (2ns).
On second screenshot the minimum time is set to ignore small spikes, and oscilloscope triggers on required place, but it triggers on the end of the event, not on beginning.

[David Hess]

I am not expert on the Rigol MSO oscilloscopes but there apparently is such a setting.  Page 7-7 of the user manual describes "jitter reject" which is suppose to prevent false triggering on asynchronous inputs.  Indiastarter who made the video was apparently unaware of this or confused it with holdoff.

Manual tells, that "jitter reject" works on analog channels only.

There is a "Duration" trigger mode, where you can set pattern and a minimum time duration to ignore spikes (or you can set maximum time to trigger on spikes only).

...

The MSO2000A manual that I am reading does not say that.  The page discussing Jitter Detect and Jitter Time is in the section about the logic analyser and digital channels:

Press Jitter Reject to turn on the jitter rejection function. Jitter refers to the
short-term deviation of the signal relative to its ideal time position at certain
time point, as shown as T1 and T2 in Figure 7-1. When no reference clock is
selected for the bus, the jump status of each channel would cause the variation
of the bus data and undesired data would be generated due to jitter when the
bus data varies. When the jitter rejection is turned on, the bus does not display
the variation of the bus data caused by certain jitter time and still holds the
effective data.

The MSO1000Z calls the same thing Noise Rejection on page 8-4.

[Ivan7enych]

I am not expert on the Rigol MSO oscilloscopes but there apparently is such a setting.  Page 7-7 of the user manual describes "jitter reject" which is suppose to prevent false triggering on asynchronous inputs.  Indiastarter who made the video was apparently unaware of this or confused it with holdoff.

Manual tells, that "jitter reject" works on analog channels only.

There is a "Duration" trigger mode, where you can set pattern and a minimum time duration to ignore spikes (or you can set maximum time to trigger on spikes only).

...

The MSO2000A manual that I am reading does not say that.  The page discussing Jitter Detect and Jitter Time is in the section about the logic analyser and digital channels:

I was a bit wrong, "Jitter Reject" option has only Bus decoder, I've made 2 pictures with different Jitter Reject time.

But the Pattern trigger has no such option, only Noise Reject mode, which is applicable to analog channels only. I called it wrong as "jitter reject".

Note: Trigger coupling and noise rejection are grayed out and disabled automatically when the source is set to any channel of D0-D15

[pascal_sweden]

Is there any functional difference (apart from physical difference) between the LA connector on the MSO1074Z and the MSO2072A?

The LA connector on the MSO1074Z is very wide, and only 1 row.
The LA connector on the MSO2072A is shorter, and seems like 2 rows.

Why does Rigol make the difference?
Although a bigger connector on the MSO1074Z, the actual cable seems more neat, as it is two cables next to each other.

The cable for the MSO2072A seems folded double. Not so nice solution I believe.

Any experiences/thoughts/preferences about the LA connectors and the respective cable connections to the PODs?

[Mark_O]

The LA connector on the MSO1074Z is very wide, and only 1 row.
The LA connector on the MSO2072A is shorter, and seems like 2 rows.

Why does Rigol make the difference?

Mostly to break the 16-channels into 2-groups, with a separate board-connector for each.  This allows you to plug each independently into separate headers on a board.  Or you can attach an adapter to either, then connect flying leads to that.  The LA on the 1000z can't do that.  It certainly costs more to do it better on the MSO2000, which partially explains the $400 vs. $250 additional cost for the LA.

Although a bigger connector on the MSO1074Z, the actual cable seems more neat, as it is two cables next to each other.

If you want neat, then get the 1000z.  It's also cheaper.  The 2 separate cable assemblies on the MSO2000 are more flexible though.

The cable for the MSO2072A seems folded double. Not so nice solution I believe.

Not quite sure what you mean, but maybe it's just the picture you were looking at.  The connector at the DSO is more compact, and has two independent ribbons coming out, sandwiched together.  But there is no 'folding'.

Any experiences/thoughts/preferences about the LA connectors and the respective cable connections to the PODs?

Well, I already commented on these issues before, when someone (you perhaps) asked about differences between the two units.  You might try looking that up.

[Ivan7enych]

Is there any functional difference (apart from physical difference) between the LA connector on the MSO1074Z and the MSO2072A?

The LA connector on the MSO1074Z is very wide, and only 1 row.
The LA connector on the MSO2072A is shorter, and seems like 2 rows.

Why does Rigol make the difference?

Ask the Rigol. I don't think anybody here have both MSO scopes, have disassembled both...

The cable for the MSO2072A seems folded double. Not so nice solution I believe.

No folding, there are 2 cables. Each consists of 8 coaxical cables in a row to remove any cross-interference. On the end of the cable there is a circuit to to make input capacity 8pF.

I've measured input resistance and capacity with and without this detachable part,

without detachable part it is 10kOm, 80pF
with detachable part it is 100kOm, 8-9pF

There surely is a 1:10 divider, just like in many 1:10 oscilloscope probes.

[pascal_sweden]

Can not find the section here on the forum where they compare LA performance between MSO1074Z and MSO2072A.

Anyhow, use the opportunity to check for the LA performance in MSO4014 as well

LA performance, sample rate, memory depth
MSO1074Z
MSO2072A
MSO4014

My understanding is that use of LA on MSO1074Z sacrifies 2 analog channels.
On MSO4014, there is no sacrifice there at all?

[Mark_O]

Can not find the section here on the forum where they compare LA performance between MSO1074Z and MSO2072A.

Perhaps you should look a little harder.

Anyhow, use the opportunity to check for the LA performance in MSO4014 as well

LA performance, sample rate, memory depth
MSO1074Z
MSO2072A
MSO4014

I'm unsure what you mean by this.  It sounds like you're asking somebody else to look it up in the datasheets and post a table here, so you don't have to?  Those datasheets are readily available.  And I doubt anyone but you is trying to decide between all 3.

My understanding is that use of LA on MSO1074Z sacrifies 2 analog channels.
On MSO4014, there is no sacrifice there at all?

Correct.  Well, mostly.  The MSO1074z sacrifices only 1 analog channel, per group of 8 digital channels.  So there is the in-between 8D + 3A mode.

[pascal_sweden]

I looked at the user guide from the MSO4000 series, from the Batronix website.
But there is no info there on the sample rate for the LA part. Really.

[Mark_O]

I looked at the user guide from the MSO4000 series, from the Batronix website.
But there is no info there on the sample rate for the LA part. Really.

Fair enough.  I usually pop over to Tequipment's web site.  They have the info there organized very well.  But I can also dig it up on Rigol's own site too (though they usually make me sign in before giving me any access, which is a PITA).

http://www.tequipment.net/RigolMSO4014.html -- look in the Specifications tab.  Right near the top it says:

Max. Sample rate    Analog Channel: Max. 4 GSa/s single channel, 2GSa/s dual channel
                                Digital Channel: Max. 1 GSa/s per channel
Max. Memory Depth    Analog Channel: Std. up to 140 Mpts single channel,70 Mpts dual channel
                                        Digital Channel: Std. up to 28 Mpts per channel
Waveform Capture rate    Up to 110,000 wfms/s(Digital channel Closed); 85,000 wfms/s(Digital channel Opened)

Real Time waveform Record,
Replay and Analysis function    Analog channel:Up to 200,000 frames(Std.)
                                                Digital channel:Up to 64,000 frames(Std.)

I think that's most of what you were looking for.

[pascal_sweden]

There is a video now finally on YouTube. Interface looks quite neat, but seems there are some bugs.

It would be nice if we can get the author of the YouTube movie on this forum.
Is he already a member on the EEV Blog?

[Mark_O]

There is a video now finally on YouTube. Interface looks quite neat, but seems there are some bugs.

It would be nice if we can get the author of the YouTube movie on this forum.
Is he already a member on the EEV Blog?

Why in the world would you ask that question here?  Why not just ask the guy directly at his YouTube stream?

But keep up your senseless postings.  They're good reminders of threads I need to turn Notifications off on.