Rigol DS1052D Logic Analyzer Question/Feature

[croyleje]

Hello Everyone,  I was hoping someone here might be able to answer a quick question about the Rigol MSO LA DS1052D.  I have been needing a logic analyzer lately and don't want to have to use my computer every time I need one.  My question is, does the DS1052D do decoding for SPI, I2C, CANbus ect. and if so does it only decode when you save the data and use the included software ie. defeating the purpose of getting a bench logic analyzer?  Also if anyone knows of an affordable bench logic analyzer?

Thank You,

Jason

[pascal_sweden]

The DS1074Z has 4 analog channels. Is it correct that you can use the analog channels to decode SPI and I2C?

What about the MSO1074Z? Is it possible to do the decoding on that scope either through the 4 analog channels, or any 4 of the 16 available digital channels?

I would like to have a better understanding on the decoding capabilities on the analog side (outside LA), and the decoding capabilities on the digital side (part of LA).
Which limitations/features are there on the analog side? Which limitations/features are there on the digital side?

If all the decoding can be done on the digital side, maybe then one would have enough with the MSO2072A? Maybe that is then in fact a better option, as it has better specs.
And maybe there is no need for these 4 channels, when the majority of your designs is around digital work.

[Ribster]

The DS1074Z has 4 analog channels. Is it correct that you can use the analog channels to decode SPI and I2C?

I can confirm that the DS1074Z has trigger and decode capabilities for SPI & I2C.
This is an option with a trial-time. But offcourse, this has workarounds as you will probably know.

[WesleyK]

The DS1074Z has 4 analog channels. Is it correct that you can use the analog channels to decode SPI and I2C?

What about the MSO1074Z? Is it possible to do the decoding on that scope either through the 4 analog channels, or any 4 of the 16 available digital channels?

I would like to have a better understanding on the decoding capabilities on the analog side (outside LA), and the decoding capabilities on the digital side (part of LA).
Which limitations/features are there on the analog side? Which limitations/features are there on the digital side?

If all the decoding can be done on the digital side, maybe then one would have enough with the MSO2072A? Maybe that is then in fact a better option, as it has better specs.
And maybe there is no need for these 4 channels, when the majority of your designs is around digital work.

As far as I know and as I can see the decoding functions are exactly the same for the analog channals as they are for the digital channels. You simply select CH1/CH2 or D1-D16 in the decoding menu.

[Electro Fan]

The DS1074Z has 4 analog channels. Is it correct that you can use the analog channels to decode SPI and I2C?

What about the MSO1074Z? Is it possible to do the decoding on that scope either through the 4 analog channels, or any 4 of the 16 available digital channels?

I would like to have a better understanding on the decoding capabilities on the analog side (outside LA), and the decoding capabilities on the digital side (part of LA).
Which limitations/features are there on the analog side? Which limitations/features are there on the digital side?

If all the decoding can be done on the digital side, maybe then one would have enough with the MSO2072A? Maybe that is then in fact a better option, as it has better specs.
And maybe there is no need for these 4 channels, when the majority of your designs is around digital work.

As far as I know and as I can see the decoding functions are exactly the same for the analog channals as they are for the digital channels. You simply select CH1/CH2 or D1-D16 in the decoding menu.

If it is correct that the MSO2072A can decode the popular/baseline protocols (such as I2C and SPI) equally well on CH1/CH2 or D1-D16 then indeed the MSO2072A would be a great choice for anyone primarily focused on digital signals.  However, I think the digital channels are limited to their digitally oriented voltage specifications (i.e., the digital channel voltage capacities are significantly exceeded by the analog channel voltage capacities) so while the digital channels could be very useful for digital signals that fit within those capacity thresholds, the digital channels won't help with other work that can only be accommodated with the higher voltage handling capacities of the analog channels.  There was another thread in the last few days in which Dave and others gave examples of using multiple probes (3, 4, and even more probes requiring more than 1 scope) to simultaneously monitor various relatively high voltage signals in order to isolate the cause of faults (for example on power rails, etc.).  No doubt, skilled circuit designers and troubleshooters can benefit from having visibility into multiple tests points at the same time - but only if the scope channels being used can handle the voltages being probed(i.e., if enough analog channels are available).

So, to take this discussion down one particular path - which is where I think this thread was headed, I believe the digital channels on the Rigol MSO series are going to be mostly useful for digital logic.  I'm not sure but I think the digital channels can do CMOS, TTL, and maybe some other lower level logic - maybe around 1.2 volts.  So, for example, if it is correct that the triggering and decoding work the same for CH1/CH2 and D1-16 I would still wonder if the digital channels can - for example - decode RS232?  I think RS232 coming out of a PC port is on the order of 13 or 14 volts peak to peak which I believe exceeds the D1-D15 specs.

My guess is that for users who want to do work primarily with IC2 and SPI but who have been concerned about the DS2000's limit of 2 analog channels the MSO2000 now alleviates such concern.  For users who need more than 2 channels to go above the voltage ceilings of the digital channels on the MSO the 4 channel MSO1000 is going to be more attractive.  (I don't have the MSO manual available to me at the moment but I think the MSO2000 digital channels are limited to +/- 5 volts, or maybe less - if someone has the spec, please let us know.)

Maybe Mark_O can shed some insight on this, Thanks, EF

PS, found this:  http://www.tequipment.net/assets/1/26/MSO2000A_DS2000A_data_sheet-EN.pdf
- see the Specification section "Vertical (Digital Channel)"; looks like maybe the max voltage is on the order of 5 volts?

[Mark_O]

I have been needing a logic analyzer lately and don't want to have to use my computer every time I need one.  Also if anyone knows of an affordable bench logic analyzer?

Jason, do you know what you want to use it for?  You started asking about a logic analyzer, and those have a different focus than a protocol analyzer ("decoding for SPI, I2C, CANbus ect.").  While there is some overlap, recommendations would depend on which you really need for your applications.  Also, do you need to see correlated analog signals, along with the LA or PA signals?  And if so, how many?

Lastly, what would you consider to be "affordable"?

[pascal_sweden]

I am not an analog design freak anyhow. Power analysis, etc. is really not my thing.

From that perspective I think I have finally made up my mind. It is simply going to be a MSO2072A

Is it possible to upgrade this baby to 300MHz, ALL options enables, WITHOUT physically opening up the unit. That is by simply entering the S/N in some online tool to get the option code, and not needing to send memory dump captured through the JTAG interface internally?

Hopefully this is supported, or will be supported in the Riglol tool soon! MSO users count on Riglol experience/dedication to make it happen!

Still it would be good to get feedback on the above question regarding accepted voltage levels on the digital channels!
I would like to get that answer in place, before I can make up my mind 100%

[David Hess]

Still it would be good to get feedback on the above question regarding accepted voltage levels on the digital channels!
I would like to get that answer in place, before I can make up my mind 100%

The specifications in the manual say that the threshold for the digital channel has an adjustment range of +/- 20 volts with a dynamic range of +/- 10 volts (and greater transient overload capability) and a minimum voltage swing of 0.5 volts which I assume is do to built in hysteresis.  That would exclude the lower voltage variations of LVDS but 1.8 volt CMOS would still work and most implementations of RS-232 as well.

[Electro Fan]

Still it would be good to get feedback on the above question regarding accepted voltage levels on the digital channels!
I would like to get that answer in place, before I can make up my mind 100%

The specifications in the manual say that the threshold for the digital channel has an adjustment range of +/- 20 volts with a dynamic range of +/- 10 volts (and greater transient overload capability) and a minimum voltage swing of 0.5 volts which I assume is do to built in hysteresis.  That would exclude the lower voltage variations of LVDS but 1.8 volt CMOS would still work and most implementations of RS-232 as well.

David, where in the manual (what page) did you find the voltage specs?  Thanks, EF
- David, per your reply below - yes, those specs are in the manual, indeed; after I looked closer at the spec sheet they are there too.
     - The MSO2000 should be an excellent scope.

[David Hess]

David, where in the manual (what page) did you find the voltage specs?  Thanks, EF

I am looking at the MSO2000A/DS2000A user guide available here:

http://beyondmeasure.rigoltech.com/acton/attachment/1579/f-0509/1/-/-/-/-/MSO2000A%26DS2000A_UserGuide.pdf

Page 18-2 shows the overload limits for the digital inputs and page 18-4 shows the operating parameters.

[Electro Fan]

Ok.... so since the digital channels have lots of voltage headroom, the key thing they are missing would be the A to D converters that are included on the analog channels (hence the digital channels are only useful for inherently/original digital signals)?

And in turn the analog channels can do everything the digital channels can do (and more)?

David, Mark_O, anyone want to give us a tutorial?  Thx

[WesleyK]

Yep pretty much.
Digital channels won't tell you the signal integrity, so your missing that part. But other than that, I'm pretty confident with my choice for a MSO2000 over the DS1000Z as I'm doing mostly digital work and sometimes some basic analog stuff.

[David Hess]

I am just following the discussion out of curiosity and have not done anything more than a quick evaluation of any Rigol oscilloscopes so someone else will have to go into detail.  I would recommend reading the user manual though to see if it covers what you want to know.

[pascal_sweden]

The digital inputs can even support LVDS:

Vertical (Digital Channel)

TTL (700 mV)
5.0 V CMOS (+2.5 V)
3.3 V CMOS (+1.65 V)
2.5 V CMOS (+1.25 V)
1.8 V CMOS (+900mV)
ECL (-650 mV)
PECL (+1.85V)
LVDS (+600mV)
0 V

[Mark_O]

The digital inputs can even support LVDS:

Vertical (Digital Channel)

TTL (700 mV)
5.0 V CMOS (+2.5 V)
3.3 V CMOS (+1.65 V)
2.5 V CMOS (+1.25 V)
1.8 V CMOS (+900mV)
ECL (-650 mV)
PECL (+1.85V)
LVDS (+600mV)
0 V

Good work on checking the manual or data sheet.

The digital inputs can even support LVDS... LVDS (+600mV)

Yes, that's correct.  That's why David wrote "That would exclude the lower voltage variations of LVDS".  I.e., there are some versions, such as SLVS-400, which the Rigol would be unable to handle.  LVDS with a Vcmo of 1.2V (for example, commonly found on systems with a Vcc of 2.5-3.3V) should work OK.  In other words, the vast majority of stuff you're likely to run into.

The +/-20V threshold range on the MSO2000 (a bit smaller +/-15V on the MSO1000Z) is really very good.  The compact USB "logic analyzers" typically don't handle anything over +5V (or anything negative), which means direct monitoring of RS-232, which can have bipolar excursions of +/- 3 to 15V would fry them.  Therefore you have to somehow tap in on the TTL side, after the level-shifter.  Which isn't always convenient. 

Even the +/-6V of RS-422 is too much for them to directly handle.  Meaning a secondary interface is required to use them (which can frequently be found fairly inexpensively on the bay, and elsewhere).

The Rigol MSO's are really quite flexible in their digital input capabilities.  Thresholds can be adjusted in 10mV increments (relative), with absolute accuracy spec'ed at +/-100mV.  This is important, because you want to be looking at the same threshold levels as the circuit you're testing.  Otherwise you won't be "seeing" the same transition points, and you could arrive at incorrect conclusions as a result.

[pascal_sweden]

If I would go for an MSO1074Z after all, instead of an MSO2072A, what would be the most annoying limitation or most missed feature?

Why would I regret that I didn't go for the MSO2072A after all, if I would not need the BW and the big knob?
Can some one tell me this in a very good high-level summary?

[David Hess]

Ok.... so since the digital channels have lots of voltage headroom, the key thing they are missing would be the A to D converters that are included on the analog channels (hence the digital channels are only useful for inherently/original digital signals)?

I do no not remember the make and model offhand but at least one series of older mixed signal oscilloscopes made by one of the big names used like 4-bit ADCs for the digital channels so it was possible to see signal details.  I guess it was too gimmicky as I do not see anything like that available now.

[Mark_O]

If I would go for an MSO1074Z after all, instead of an MSO2072A, what would be the most annoying limitation or most missed feature?

Why would I regret that I didn't go for the MSO2072A after all, if I would not need the BW and the big knob?
Can some one tell me this in a very good high-level summary?

I suspect you can get along with the bandwidth, and the "Big Knob" certainly makes Navigation simpler.  But if you haven't gotten used to it, you probably won't miss it that much.  You certainly can get around without it.

I already shared with you the 2 biggest annoyances/limitations for me of the 1000Z, which I actually like a lot.  No CAN, which may not be of any interest to you.  But I do a lot of embedded comms work, using all the protocols Rigol supports (and a handful of others), and CAN is about 30% of that work.  For me to buy another scope, with protocol capabilities, and NOT get CAN would seem a bit foolish.  That may not apply to others though.

The other is the lack of time-stamps in Segmented acquisition mode.  Many times you just don't need them.  I.e., it doesn't matter if 2 consecutive occurrences happened 1 msec apart, or an hour apart.  But sometimes it does.    The 2000 has that, and the 1000Z doesn't.  That doesn't mean though that if a situation does arise that you DO need them, there are no options to work around it.  I.e., you could use the trigger-out the Rigol generates on each occurrence, and send it to something else to time-stamp it.  Another scope, a PC, a plotter, etc.  But that's extra work, and often a hassle to correlate afterwards.

If you want to look at 720P/1080i/1080P HD video signals, the 1000Z will not allow that.  It maxes out at 480P/576P.  That may be something you never need.  But it's unfortunate that component video, which has 3-channels (Y,U,V) would be a good match for the 4-channel 1000Z, but the scan-rate is too high.  While the 2000 handles the extra lines, but comes up 1-channel short.

The biggest protocol limitation of the DS2000 is it doesn't really do SPI, but that problem is completely eliminated with the MSO2000.  It can actually do two full 4-signal SPI busses (as can the MSO1000Z).

There were at least 2 people here, and ElectroFan was one of them, that started out with a DS1000Z, then switched to the DS2000.  They may be able to give you their reasons for doing so, which should also apply to the MSO versions.  Or you could just do a search, because they already explained their reasons in detail (quite a while back).  I don't have time to try and dig it up again.

The extended LA capabilities of the 2 MSO units don't differ much at all (very, very close), though the RPL-2316 probes with the 2000 are a bit nicer than the RPL-1116 probes with the 1000Z.  However, functionally they're both very good, and a large step up from the LA section of the previous generation of DS1xxxD models (or the DS1102CD that I have).  The digital-section differences are very minor (+/-15V threshold setting on the 1000Z, vs. +/-20V on the 2000 is the major one), compared to the analog sections.

[Electro Fan]

I think Mark_O summed it up well.  If you have a 1000 and haven't experienced the additional features and control that comes with the big Nav knob you might not miss it but once you have the knob you will like it.  The time stamps are also pretty valuable.  The key capability I was looking for at the outset was the ability to work with both analog and digital signals and see them time correlated (and be able to easily navigate among them) - unfortunately the MSO versions weren't available when I bought my scope.  While it would of course be nice to have a 2000-class scope with 4 analog channels in this price range the ability to do 16 digital channels plus 2 analog is probably sufficient for many users.  And as Mark pointed out the ability to have the relatively high peak to peak voltage capacity (+/-20 volts?) on the digital channels is well beyond what is supported by typical PC Logic Analyzers that go to about 5 volts.  In my case I'm more interested in being able to learn/do SPI and I2C than I am probing more than 2 analog test points on a circuit at one time.  It just depends on what you are trying to learn and do.  Some people are going to spend more time in the analog domain and the 4 channels are more important than the Nav knob and time stamps.  Additionally there are some other differences between the 1000 and 2000 you can probably find in the threads here; for example, the size of the displays and how some of the menus works.  I like the way you can move the menus off the displays on the 2000 (can't remember for sure but I think this was different on the 1000).  Either way you can't go wrong and all in all Rigol has delivered great value and a nice choice for users.  The MSO1000 offers a ton of value.  For people working primarily in the digital domain or a mixed domain the MSO2000 looks like an excellent scope for very little more cost than the MSO1000.  At the end of the day in this general price range I think all you have to ask is "do I need 4 analog channels?"  If so, buy a DSO or MSO1000.  If you don't need 4 channels and want an even nicer scope (imo), buy a DSO or MSO2000. 

[Mark_O]

Thanks, EF.

The other big difference I left out of my summary above is that the 2000-series units provide a powerful Analysis function that the 1000Z units lack, along with the Record/Play segmented captures (which it has).  Being able to auto-scan through thousands of segments, and quickly step through those that deviate by N% from a template waveform can be very useful, and save a huge amount of time.

[PedroDaGr8]

Don't forget the sample rate on the 1000Z series is 1GSa/s total. So two channels drops it to 500MSa/s per channel and 3+ drops it to 250MSa/s  per channel.

[Mark_O]

Don't forget the sample rate on the 1000Z series is 1GSa/s total. So two channels drops it to 500MSa/s per channel

Thanks, Pedro.  I don't think anyone has forgotten that.  The specs on the 2000-series are clearly higher.  This was focusing on a feature/capability comparison.

and 3+ drops it to 250MSa/s  per channel.

Right.  And 3+ drops the rate on the 2000 even farther, to 0 Sa/s.     250 MSa/s is infinitely better than 0.

As a side note, since the emphasis here is on the LA capabilities, both the 1000Z and 2000 share the same 1 GSa/s rate with 1 Group of 8 digital channels enabled, and 500 MSa/s with both groups on (all 16-channels).  So unless there is a way to split the rates, when sampling in Mixed mode (analog and digital simultaneously), the 1000Z is just as fast as the 2000.

Their only difference is the max digital Depth on the MSO1000Z is 24M, vs. 28M on the 2000.  Hardly worth mentioning.

[Mark_O]

And as Mark pointed out the ability to have the relatively high peak to peak voltage capacity (+/-20 volts?) on the digital channels is well beyond what is supported by typical PC Logic Analyzers that go to about 5 volts. 

Just to clarify, the +/-20V is not the peak voltage capacity.  Those are the maximum selectable logic threshold levels (where a 0 becomes a 1).  Peak levels (dynamic range) are +/-10V more on both, so the 1000Z handles a signal range of +/-25V, and the 2000 is +/-30V (in both cases, as long as the Vrms doesn't exceed 40V).

[Mark_O]

I like the way you can move the menus off the displays on the 2000 (can't remember for sure but I think this was different on the 1000).

The menus are locked down on the 1000Z, and as a result there are 12 horizontal divisions, vs. 14 on the 2000.

Either way you can't go wrong and all in all Rigol has delivered great value and a nice choice for users.  The MSO1000 offers a ton of value.  For people working primarily in the digital domain or a mixed domain the MSO2000 looks like an excellent scope for very little more cost than the MSO1000.  At the end of the day in this general price range I think all you have to ask is "do I need 4 analog channels?"  If so, buy a DSO or MSO1000.  If you don't need 4 channels and want an even nicer scope (imo), buy a DSO or MSO2000.

It can be a tough call for many, trying to decide between the two.  And stepping up to the MSO4000 to get both has a huge cost premium. 

But another option if you want both, is the Siglent SDS2000.  It provides the high perf of the MSO2000-series, combined with the 4-channels of the MSO1000Z, along with LA capability, and protocol decoding on all channels as well.  For less than the cost of an MSO2000 + a DS1000Z, you can get both in one unit... albeit with only 8 digital channels. 

And if you value the time-stamps on segments with the 2000, you'll really like the scrollable list of date/time-stamps on the SDS2000, with direct-jump capability.  Which even the MSO2000/MSO4000 don't have.

[Mark_O]

Getting back to Jason's original question about the DS1052D and its LA capabilities...

While that unit was very good for its time, even though it had no protocol decode (or trigger) capabilities, the digital channels did suffer one limitation.  I.e., there was effectively no way to time-correlate the analog and digital sides.  While both signals could be on-screen at the same time (a true MSO, vs. the so-called MSOs from Owon that can only do one or the other, but not both), you couldn't sync them properly.

This was because the 1000D family used an active logic-head (the LH1116, same as the 1000CD before it).  In many ways an active pod can be superior (although more expensive), but it introduces a serious problem.  You run your digital lines out on flying leads to micrograbbers, and the threshold levels set by the differential comparators in the pod latch and sample the signal right in the pod.  Those signals are then sent down the LVDS ribbon cable to a high-density SCSI connector on the scope.  So the distance from board signal to sampling is less than 1-foot.  However, on the analog side you have a 3-ft or 5-ft probe cable (depends on what you're using), and sampling is done at the other end, inside the scope.  Due to propagation delays of the signals (limited to 0.6-0.7 the speed of light), this introduces a timing difference of several nanoseconds between the analog and digital signals. 

Most folks never understood this, which led to YouTube videos demonstrating how the Rigol was "broken", because they attached both digital and analog probes to the same point on a circuit, and while the "real" analog signal was correct, the broken Rigol displayed the digital signal magically appearing 3-4 ns "too soon".  Since that was "impossible", there must be some bug in the Rigol display logic.  Of course, the Rigol was displaying exactly what it's sensors were reporting to it.  The cables and acquisition points were different.

And to compound the problem, there was no way to "dial out" that time difference or skew, so you were stuck.  However, the LA sections of the new Rigol MSOs have addressed both issues.  First, they're now using passive probes on the digital side, which don't get sampled until the signal arrives inside the unit... just like the analog signals.  And secondly, they provide a handy Delay offset, which can be used to time-align the analog and digital sides exactly... even when the cables lengths vary (which they usually will).  Thus the new scopes can claim to be "time-correlated", and they do, on their spec sheets.

This is a another big win over the older 1000D series.


But Jason still hasn't responded yet on what he considers "affordable", so it's hard to say if any of this discussion is helping him or not.  Even the best MSO doesn't do you any good, if it's out of your price range.