MSOs

[FrankT]

I'm a hobbiest, with limited budget.  I currently have a Dynon Instruments elab-080 USB scope.  It's an all in one device with 2 analog channels, 16 digital channels, AWG, 4 digital outputs, 2 output clocks and 2 output "power supplies" (total 60mA output).

It has served me well, but recently I've struggled with its limited capture memory, 32k, slow usb (version 1), and its been playing up a lot - I'm using the AWG, and I get some strange results.  The software is pretty dated, and doesn't have any protocol decoding.

So, I'm looking to replace it, either with a single device, or multiple, to provide 2 channel DSO, LA, AWG and digital outputs (and lots more sample memory).

I couldn't find much info on the forums about MSOs, except that some people that have MSOs say the wouldn't buy another.  The posters don't elaborate, so I'm wondering if they are just not useful, too expensive, or can you do mixed signal analysis with a DSO and logic analyser?  If so how do you combine the results? 

I like the USB scope, probably because I've never owned a standalone unit.  I rarely use it away from my desk, or if I do its easy enough to connect it to my lap top.  Panning and zooming on a 25" monitor is hard to beat.

And budget - the elab-080 was about $500, but I know I'm looking at over $1k. 

Any thoughts?

[Electro Fan]

Hi FrankT,

I don't have the answers to your questions, but I liked your post a lot.  I have many of the same questions.

I haven't read about anyone with a MSO saying they wouldn't buy another, but I'd be interested to hear more about such experiences and such a line of thinking.  Like you, I'm up for trying to understand if it is it because MSO's are somehow not useful (kind of hard/very hard to believe), or is it because they are too expensive (perhaps when compared to separate DSO and LA units)?  And similar to your question - how hard (and/or how useful) is it to combine the results if you use a DSO + a separate logic analyzer?

So far, it seems to me that a Rigol 2072 + an Intronix LogicPort (or some other similar DSO and LA) would be a very good and cost-effective combination, but I'd like to hear from people who have a $1500 to $3000 MSO that provides better utility to find out what that better utility consists of.

I'n my case I'm not as interested in an AWG (I'd like one, but except for the abitrary functionality I'm getting by with a BK4040A); and here again Rigol seems to make a pretty nice family of AWGs - including one for under $400 and about four others for under $1k.

I started a thread with some questions similar to yours that has now bumped down lower in this forum and there were some good replies and comments but we never really seemed to get to a definitive answer.  Maybe a more definitive way to ask the question is, "to beat a Rigol 2072 plus an Intronix LogicPort, how far up the food chain do you have to go on MSOs?"  And when you get there, what are the key advantages of the MSO?

Having tried to ask the questions(s), here are my guesses at the answers.  (I hope I don't get beat too bad if they are wrong - I am clearly stating they are just guesses.)

1. A high end MSO might provide more memory, but I can't quite see how it would provide more memory or storage than a PC attached to a LA; having said that, it looks like the LogicPort is limited to 2048 samples per each of 34 channels - so maybe for long signals that need examination or for signals with lots of changes that are difficult to compress 2048 samples could be a limitation?

2. A high end MSO might provide both timing analysis and state analysis, but for example so does the $389 LogicPort.

3. A high end MSO might run at a higher clock speed (the LogicPort is limted to 500MHz for internal clock with timing mode, and 200MHz external clock for state mode.)  Probably depends on what you are building, testing, debugging?

4. A high end MSO might support more protocols, but it appears that protocols often cost still more $ on MSOs; on the LogicPort about 10 or 12 popular protocols seem to be included; I think I saw another standalone LA that supported many more protocols for about $500.

Net, net:  the only reasons I can come up for not using a separate LA are 1) if you need to debug things that exceed the speed/sampling/memory limitations of a standalone LA, or 2) if there is a notable benefit from being able to see the analog wave forms and the digital signals at the same time on the same screen - I'm wondering in what types of applications and how often this 2nd situation turns out be valuable?

It appears that for relatively low end / low speed work (I2C on Arduinos, etc.) that a DSO like the 2072 and a LogicPort (or even a Saleae if you don't need more than basic triggering) that the only thing missing is the simultaneous view of the analog and digital signals?  Yes or No?  And if yes, how valuable or cool is it to have the simultaneous analog and digital view?

Thanks to any forum members who can shed light on this stuff.

[FrankT]

1. A high end MSO might provide more memory, but I can't quite see how it would provide more memory or storage than a PC attached to a LA; having said that, it looks like the LogicPort is limited to 2048 samples per each of 34 channels - so maybe for long signals that need examination or for signals with lots of changes that are difficult to compress 2048 samples could be a limitation?

There are a few logic analysers that have small amounts of memory and rely on a fast connection to the PC to provide storage.  The Saleae Logic analyser (http://www.saleae.com/logic) claims 24M samples per second on each channel, but that is dependant on how well your pc and usb bus is working.  The specs (http://www.saleae.com/logic/specs) say....

Note that achieving the highest sample rates requires low USB latency; this may not be achievable on all computers. Performance may improve with the removal of other USB devices, using a different USB host controller, or increasing the software's process priority.

They also claim 10B sample memory, but...

10B samples. Absolute max depends on data compressibility, available RAM and operating system. 10B samples assumes reasonably high compressibility.

I'm concerned the LogicPort has these issues.

Picotech make some nice usb scopes.  One has a USB 3 interface so is cable of very fast transfer back to the PC.

[Electro Fan]

Looks like the new buyers leading the new buyers here

but regarding this,

"The Saleae Logic analyser (http://www.saleae.com/logic) claims 24M samples per second on each channel, but that is dependant on how well your pc and usb bus is working.  The specs (http://www.saleae.com/logic/specs) say...."

If the USB connection is the only issue a user could have a dedicated "low latency" PC (and to your earlier point, for good panning and zooming it could have a big screen - much more so than what would come with any scope); it could simply have a dedicated USB port or at least a decent USB bus; or perhaps we will see a new generation of USB LAs with USB 3.0 if this is really a notable issue.

I just don't think the USB latency/reliability issue is the numero uno issue; I can believe that other factors related to the ability to support the required sample size could be an issue for users that require mondo sample sizes.  But if a USB LA was architected properly I think neither speed or sample size should be an issue.

Saleae gets lots of good reviews from users but their triggering options don't look quite as extensive as perhaps they could be.  Likewise the Intronix gets very good reviews but occassionally users indicate they might like more/better sampling.

Any way you cut it, I think the next generation of LAs (both USB based and MSO based) should be very good; and as each segment gets stronger that will create competition/stimulus for the other segment to keep pace.

We know hardware is going to get faster and less expensive, and it seems that we are starting to head toward some consenus on what ths software should do.  I think we are headed toward ever nicer abilities to surface stuff that was previously hard to see.

A little off topic, but here is an interesting post about an Arduino user chasing some issues with a LA (happens to be with a LogicPort):  http://arduinococo.blogspot.com/?m=1

[FrankT]

A little off topic, but here is an interesting post about an Arduino user chasing some issues with a LA (happens to be with a LogicPort):  http://arduinococo.blogspot.com/?m=1

That's freaky, because the project I'm working on at the moment is a device to read the tapes of a CoCo (TRS-80 color computer).  It was trying to play back some of the waveforms using the AWG and monitor the wave and use digital outputs for debugging.  The attached image shows the waveform, the zero crossing detection, and when I think I've located a bit.

[jpb]

You don't mention anything about budget, or bandwidth/frequency/sampling rate that you require (or want).

One issue with low-end MSOs is lack of screen space - a dedicated logic analyser able to display on a laptop screen is perhaps more of a pain to  setup but is much clearer to view. The MSO options on scopes can be as expensive, or rather more expensive than a separate USB based logic analyser and are generally less powerful - you pay a premium for having everything neatly in one box. For example with an Agilent 2000X scope you can only have 8 digital lines, you need to pay for decoding separately and even with recent special offers the incremental cost is a lot more than say buying the Intronix separately:
http://www.pctestinstruments.com/
which costs $389 whilst the MSO upgrade on the Agilent together with two decoding options is about 4 times the cost and only provides 8 channels vs 32 and has far fewer options (no state decoding and much lower frequency). The Intronix can be ordered with trigger out so that analog waveforms can be viewed separately on a scope. (I should add I've neither an MSO nor a logic analyser at present - I've just looking to buy so have done a lot of research recently.)

The MSO approach is neater in being always there, and for an engineer working to a tight schedule and where time is money I can see that it makes sense to use whatever is most straight forward. But for a hobbyist I think it is a very expensive route to take.

There are quite a few discussion threads on this forum where the pros and cons are discussed if you do a bit of searching.

[Electro Fan]

Hi FrankT - that is a very interesting coincidence.  Nice image you posted with your project signals labeled.  It looks like maybe you pasted together two different screen shots (one DSO and one LA), or were they shown on one screen at the same time as you triggered and captured the samples?  This idea of analog and digital integration with simultaneous viewing has me fascinated.

Per a comment made by jpb, I'm wondering what the possiblities are of using outputs from both a DSO and a LA to simultaneously view (and manage?) triggered events on one (computer) screen?

Hi jpb,

It sounds like you are in the "separate LA's are not only less expensive but potentially better" camp.  I'm almost there but per my comment above I'm still trying to get my head around the virtues (or lack thereof) of simulataneously capturing and viewing analog and digital signals.  Having said that you ask very reasonable questions:  what is the budget, bandwidth, frequency, samplying rate required/wanted?

If I could answer those questions I would - but in my case I'm more on a journey than seeking a particular destination.  I have a moderately good understanding of IT but a lesser understanding of electrical engineering - so I'm coming at this not so much to build anything in particular but rather to learn more about the intersection between underlying analog discreet ciruitry, some IC-based digital circuity, plus firmware and software . 

My approach has been to dabble with some elementary projects (so far mostly with Arduino Uno breadboarding) and to slowly assemble some test equipment (Techtronix 2247A, Fluke 179, BK4040A, Korad 3005P) .  I'd really like a BK 8500 electronic load

I doubt that I'm going to do lots of high speed firmware development or debugging - but I enjoy the learning process and it revolves around basic electronics plus computing and networking hardware and software.  I find tools that help visualize cause and effect and that can confirm or dispprove a hypothesis are interesting and exciting. 

So, I'm not sure about what specs I am likely to need.  The Arduino Uno only operates at 16MHz but a Rasberry Pi operates at 700MHz (and for sure CPU and other devices are only going to get faster).  As for protocols, I'd like to examine I2C, SPI, maybe some basic RS232, maybe CAN - the problem is I don't know what I don't know and what I will want to know

As for the budget, I can get my head around a $400k DSO and if need be a $800-$1k DSO.  Likewise, I could potentially justify a $300-$500 LA.  If it turns out that a better solution is a $2500 or $3k MSO I could get my head around that too.  Some people buy vintage hifi gear, camera gear, or boats, or cars - when compared to these things a couple thousand dollars or so doesn't seem like alot, but I'd like to avoid investing in something I'm not likely to use - but I also don't want to dampen the learning process for lack of tools that can provide meaningful insight. 

Learning should be fun and sometimes you have to invest in tools

Long story short:  What are the chances that with an Intronix LogicPort and a Rigol 2072 a DUT could be simultaneously probed for analog and digital signals so that upon triggering of either the analog or the digital signals that both the analog and digital signals could be displayed (and ideally reviewed and analyzed interactively) in a synchronized manner on a nice landscape computer monitor? 

And if the DSO + LA (+ computer) isn't the right answer, what's a recommended entry point with a MSO that will get flexible single (MSO) screen viewing and interactive analysis?

Thanks for any questions or suggestions.



 

[jpb]

Long story short:  What are the chances that with an Intronix LogicPort and a Rigol 2072 a DUT could be simultaneously probed for analog and digital signals so that upon triggering of either the analog or the digital signals that both the analog and digital signals could be displayed (and ideally reviewed and analyzed interactively) in a synchronized manner on a nice landscape computer monitor? 

And if the DSO + LA isn't the right answer, what's a recommended entry point with a MSO that will get flexible single screen viewing and interactive analysis?

Thanks for any questions or suggestions.

I have a LeCroy WaveJet oscilloscope so was looking at the LeCroy LogicStudio and this does precisely what you're asking about - i.e. the LA triggers the oscilloscope and the software runs on the pc and controls both the LA and the oscilloscope and shows both analogue and digital waveforms on the screen. Unfortunately from my point of view it requires a WaveJet 300A series and mine is an original 300 series. In my interchanges with LeCroy trying to see if I could get the system to work they told me that there was a 100nsec trigger delay so presumably the software compensates for this.

The Intronix Logicport can be ordered with a trigger out - they don't advertise the fact but I e-mailed them and was informed that they do it and it costs $15 extra. I would guess that there would be a trigger delay and their software as it stands won't transfer the analogue waveforms from the oscilloscope and compensate for the delay.

So to see all the traces on the pc scheme you'd need to write your own software to read the data from the oscilloscope and from the LA (which I think produces a csv file) and then plot both together shifting the oscilloscope trace to allow for the extra trigger delay.

Personally I wouldn't bother. Provided the scope had captured the right bit of waveform (having been triggered from the LA) you could inspect the pulse shape on the scope screen and read off the digital information from the pc screen.

To answer your question more specifically, you would need to trigger from the LA which would then trigger the oscilloscope - this makes sense as the LA can trigger off complex data patterns. The oscilloscope would be set to trigger off the external trigger channel attached to the LA. You could trigger it separately but then you wouldn't be able to trigger the LA.

On the MSO front, Agilent were selling bargain versions of their 2000X series on their e-bay store recently (there was a thread on these forums) but when I looked recently they seem to have finished but it is worth looking out, e.g.
http://www.ebay.co.uk/itm/Agilent-MSO6012A-Mixed-Signal-Oscilloscope-8ML-/190806934357?pt=BI_Oscilloscopes&hash=item2c6cfac355
but note that you may have to pay quite a bit for decoding (though again search the forums as on this) and also if the item is not "Premium" used it won't have much of a warranty.


As a caveat - I'm no expert, I've not yet bought myself a LA but I've followed various threads on this forum and have been investigating for sometime. There are quite a few others with more practical experience.

[jmole]

FYI, the serial decode on the Agilent MSOX2000 series only works on the analog channels. Just picked one up, and I've been using the 30-day trial of all features, and was very underwhelmed when I discovered that you can't use the digital channels for protocol decode 

[nctnico]

IMHO the Intronix logicport is not a good choice because of the severely limited amount of memory. I have used the logicport at my former employer but for real problems I brought my own logic analyser along. With 32Mpts of memory taking one measurement is often sufficient. The rest comes down to analysing the data.

I don't know if I would buy a logic analyser nowadays. The gap between a 'real logic analyser' and an MSO is closing fast so I would probably buy an MSO. Preferably one which allows to trigger when a condition (combination of input levels) is present or absent for a certain period of time. Decoding SPI and I2C is a big plus.

[jpb]

FYI, the serial decode on the Agilent MSOX2000 series only works on the analog channels. Just picked one up, and I've been using the 30-day trial of all features, and was very underwhelmed when I discovered that you can't use the digital channels for protocol decode 

That is a bit of an epic fail!

Perhaps the hardware was never designed for it - they originally didn't have decoding at all on the 2000 series and have only added it to make it a little more competitive relative to Rigol and Hameg scopes I guess.

[jpb]

IMHO the Intronix logicport is not a good choice because of the severely limited amount of memory. I have used the logicport at my former employer but for real problems I brought my own logic analyser along. With 32Mpts of memory taking one measurement is often sufficient. The rest comes down to analysing the data.

I don't know if I would buy a logic analyser nowadays. The gap between a 'real logic analyser' and an MSO is closing fast so I would probably buy an MSO. Preferably one which allows to trigger when a condition (combination of input levels) is present or absent for a certain period of time. Decoding SPI and I2C is a big plus.

The logicport's memory is very small at 2k but it does do real-time compression which must help a bit. It does seem to have a lot of happy users. There are other USB LAs with much more memory though they don't seem to have the same features nor as well designed as the logicport from what people say.

I don't think that low cost MSOs are closing fast. They cost much more, are lower frequency, don't do state analysis, have limited number of digital channels (generally only 8 with 16 on the more expensive scopes) and limited screen space for display. Perhaps Rigol will change things, but looking at current costs adding MSO options plus channel decodes can double the cost of the scope. I don't have practical experience with either but if you have a limited budget you don't seem to get much for your money when going the MSO route. I guess it depends on what you wish to do. If it is just decoding serial buses then perhaps just hardware decode and trigger such as Hameg do are quite sufficient and much more convenient than a separate device.

Perhaps I'm slightly biased as I've already bought a 4 channel scope without decode options (a WaveJet) so my only route is a separate LA, but I did consider MSOs when I was picking my scope.

[nctnico]

Nowadays logic analysers have very little use. All the high speeds stuff is serial at hundreds of MHz to several GHz. Finding problems there has more to do with signal integrity (impedance matching) than problems in logic. Microcontrollers have their flash and SRAM inside and no external busses. What is left are SPI, I2C and UART interfaces. If you are into FPGA design you might want to hook up a logic analyser to a couple of debug pins to get to some internal signals. Over a decade ago I worked on a PCI design (the parallel bus version). I think that is about the last time I needed more than 16 pins connected to a circuit for testing. I did lots of FPGA designs after that but those where all debugged using 8 pins fed by an internal 'debug signal' multiplexer and an Agilent MSO.

[FrankT]

Nice image you posted with your project signals labeled.  It looks like maybe you pasted together two different screen shots (one DSO and one LA), or were they shown on one screen at the same time as you triggered and captured the samples? 

That's the elab-080 software - the DSO and LA outputs are shown in separate windows, but they can snap together and all activity is synchronised between the two - display cursor, timing cursor, time scales, etc.

[jpb]

Nowadays logic analysers have very little use. All the high speeds stuff is serial at hundreds of MHz to several GHz. Finding problems there has more to do with signal integrity (impedance matching) than problems in logic. Microcontrollers have their flash and SRAM inside and no external busses. What is left are SPI, I2C and UART interfaces. If you are into FPGA design you might want to hook up a logic analyser to a couple of debug pins to get to some internal signals. Over a decade ago I worked on a PCI design (the parallel bus version). I think that is about the last time I needed more than 16 pins connected to a circuit for testing. I did lots of FPGA designs after that but those where all debugged using 8 pins fed by an internal 'debug signal' multiplexer and an Agilent MSO.

You have much more experience than me in this area. But I can see some circumstances when it might be nice to have quite a few digital inputs. For example my scope has connecters for external LAN or GPIB cards which have 64 connectors (presumably a lot of them grounds). I am curious to look at the bus. OK this is an old design but as a hobbyist rather than cutting edge engineer this sort of thing I might want to do from time to time. Also what if you're working with multi-bit D2As or DACs looking at complete values would be good. The software that comes with the Intronix Logicport will provide an analogue plot from a digital signal from multiple lines (essentially act as a DAC) - you couldn't do this easily on most MSOs.

[jmole]

Nowadays logic analysers have very little use. All the high speeds stuff is serial at hundreds of MHz to several GHz. Finding problems there has more to do with signal integrity (impedance matching) than problems in logic. Microcontrollers have their flash and SRAM inside and no external busses. What is left are SPI, I2C and UART interfaces. If you are into FPGA design you might want to hook up a logic analyser to a couple of debug pins to get to some internal signals. Over a decade ago I worked on a PCI design (the parallel bus version). I think that is about the last time I needed more than 16 pins connected to a circuit for testing. I did lots of FPGA designs after that but those where all debugged using 8 pins fed by an internal 'debug signal' multiplexer and an Agilent MSO.

Agreed. Standalone logic analyzers are mostly useless. MSOs on the other hand are incredibly useful when debugging mixed signal systems when you need to look at 2-4 analog channels and trigger on a particular (set of) digital signal(s).

[Electro Fan]

"To answer your question more specifically, you would need to trigger from the LA which would then trigger the oscilloscope - this makes sense as the LA can trigger off complex data patterns. The oscilloscope would be set to trigger off the external trigger channel attached to the LA. You could trigger it separately but then you wouldn't be able to trigger the LA.

On the MSO front, Agilent were selling bargain versions of their 2000X series on their e-bay store recently (there was a thread on these forums) but when I looked recently they seem to have finished but it is worth looking out, e.g.
http://www.ebay.co.uk/itm/Agilent-MSO6012A-Mixed-Signal-Oscilloscope-8ML-/190806934357?pt=BI_Oscilloscopes&hash=item2c6cfac355
but note that you may have to pay quite a bit for decoding (though again search the forums as on this) and also if the item is not "Premium" used it won't have much of a warranty."

- This is what I'm trying to figure out - how to smoothly synchronize the analog and digital signals and be able to inspect/analyze/manage them from one interface

[Electro Fan]

Nowadays logic analysers have very little use. All the high speeds stuff is serial at hundreds of MHz to several GHz. Finding problems there has more to do with signal integrity (impedance matching) than problems in logic. Microcontrollers have their flash and SRAM inside and no external busses. What is left are SPI, I2C and UART interfaces. If you are into FPGA design you might want to hook up a logic analyser to a couple of debug pins to get to some internal signals. Over a decade ago I worked on a PCI design (the parallel bus version). I think that is about the last time I needed more than 16 pins connected to a circuit for testing. I did lots of FPGA designs after that but those where all debugged using 8 pins fed by an internal 'debug signal' multiplexer and an Agilent MSO.

Agreed. Standalone logic analyzers are mostly useless. MSOs on the other hand are incredibly useful when debugging mixed signal systems when you need to look at 2-4 analog channels and trigger on a particular (set of) digital signal(s).

- what would be really cool would be if someone could draw a "use case diagram" to show where the MSO shines and where the DSO + LA shines

[nctnico]

"To answer your question more specifically, you would need to trigger from the LA which would then trigger the oscilloscope - this makes sense as the LA can trigger off complex data patterns. The oscilloscope would be set to trigger off the external trigger channel attached to the LA. You could trigger it separately but then you wouldn't be able to trigger the LA.

- This is what I'm trying to figure out - how to smoothly synchronize the analog and digital signals and be able to inspect/analyze/manage them from one interface

I doubt this is very useful. There is at least some delay from de D/A or A/D converters and filters not to mention buffering. Like said before you can trigger the scope from the logic analyser and vice versa but that won't show the problem area immediately.

[AndyC_772]

FYI, the serial decode on the Agilent MSOX2000 series only works on the analog channels. Just picked one up, and I've been using the 30-day trial of all features, and was very underwhelmed when I discovered that you can't use the digital channels for protocol decode 

You're kidding, right? That's just bizarre, I wonder if it's a hardware bug in the 2000X that means it can't be done.

You can certainly use the digital channels for serial decoding on the 3000X.

[jpb]

"To answer your question more specifically, you would need to trigger from the LA which would then trigger the oscilloscope - this makes sense as the LA can trigger off complex data patterns. The oscilloscope would be set to trigger off the external trigger channel attached to the LA. You could trigger it separately but then you wouldn't be able to trigger the LA.

- This is what I'm trying to figure out - how to smoothly synchronize the analog and digital signals and be able to inspect/analyze/manage them from one interface

I doubt this is very useful. There is at least some delay from de D/A or A/D converters and filters not to mention buffering. Like said before you can trigger the scope from the logic analyser and vice versa but that won't show the problem area immediately.

According to LeCroy, in their LogicStudio the delay is 100 nsecs. I would have thought it was similar for other LAs with trigger out. So the scope would need to have 100 nsecs of pre-trigger storage. If the delay was to be used then, on my scope at least, you're limited to 20 nsecs a division or slower (the delay is up to half a screen or 5 divisions). You can, of course, then zoom in to see detail.
Once set up, I'd have thought that this would work ok - not as good as an MSO but ok. Assuming that you'd want a clock tick to be one division before zooming then you'd be ok up to around 50MHz clocks.
The LogicPort may produce a much faster trigger out than the LeCroy LogicStudio - it goes to 500MHz to the LeCroy's 100MHz, in its specs it states the trigger sequencer goes to 250MHz.

Certainly, 200 nsecs is short enough for the scope to operate at 2GS/s without filling the memory so the information would all be there, just a pain to get at it perhaps and for fast clocks you wouldn't be able to see much until you zoomed.

[alm]

A delayed timebase/sweep (if available) should allow you to compensate for any trigger delay without wasting memory on pre-trigger information.

[nctnico]

"To answer your question more specifically, you would need to trigger from the LA which would then trigger the oscilloscope - this makes sense as the LA can trigger off complex data patterns. The oscilloscope would be set to trigger off the external trigger channel attached to the LA. You could trigger it separately but then you wouldn't be able to trigger the LA.

- This is what I'm trying to figure out - how to smoothly synchronize the analog and digital signals and be able to inspect/analyze/manage them from one interface

I doubt this is very useful. There is at least some delay from de D/A or A/D converters and filters not to mention buffering. Like said before you can trigger the scope from the logic analyser and vice versa but that won't show the problem area immediately.

According to LeCroy, in their LogicStudio the delay is 100 nsecs. I would have thought it was similar for other LAs with trigger out. So the scope would need to have 100 nsecs of pre-trigger storage. If the delay was to be used

I wasn't too clear  Say you have an FPGA with an ADC. If you measure the analog input and the data inside the FPGA there will be a significant delay between the analog input and the data inside the FPGA.

[500in1]

FYI, the serial decode on the Agilent MSOX2000 series only works on the analog channels. Just picked one up, and I've been using the 30-day trial of all features, and was very underwhelmed when I discovered that you can't use the digital channels for protocol decode 

You're kidding, right? That's just bizarre, I wonder if it's a hardware bug in the 2000X that means it can't be done.

You can certainly use the digital channels for serial decoding on the 3000X.

Looks like it won't. Here's a quote from the user manual for the 2000X series. (page 105)

"Digital channels and serial decode cannot be on at the same time. The
[Serial] key takes precedence over the [Digital] key. Serial triggers can be
used when digital channels are on."

[grego]

FYI, the serial decode on the Agilent MSOX2000 series only works on the analog channels. Just picked one up, and I've been using the 30-day trial of all features, and was very underwhelmed when I discovered that you can't use the digital channels for protocol decode 

You're kidding, right? That's just bizarre, I wonder if it's a hardware bug in the 2000X that means it can't be done.

You can certainly use the digital channels for serial decoding on the 3000X.

Looks like it won't. Here's a quote from the user manual for the 2000X series. (page 105)

"Digital channels and serial decode cannot be on at the same time. The
[Serial] key takes precedence over the [Digital] key. Serial triggers can be
used when digital channels are on."

Woah, for real?  That's crazy - what's the point of having all the digital channels if you can't decode off it (assuming you buy the option).  Yes, I know - triggering, etc.  But if I've got the digital channels I'd want to be able to decode off them.