DSO Reliability

[tautech]

While the feature set of the latest scopes look inviting, you should have a good think of just what you may need.
IMO 4 channels is a winner.
Considerable added versatility with 4 channels, particulary differential measurements.
You will find more use for this than many of the other features.

[J-D-H]

While the feature set of the latest scopes look inviting, you should have a good think of just what you may need.
IMO 4 channels is a winner.
Considerable added versatility with 4 channels, particulary differential measurements.
You will find more use for this than many of the other features.

My differential measurements have usually involved AC line troubleshooting.  Could be wrong, but I don't recall having a need to scope two AC line sources at the same time.  However, who knows, maybe you are correct and maybe this would be handy in future.

[nctnico]

Well 2 channels differential and one for current adds up to 3 channels.

[J-D-H]

FWIW, You can hack the DS1054Z to 100MHz, and enable all of the software options (hack how-to).

Regarding the DS1102E, there's no contest between it and the 1000Z series. Larger screen, persistence display (1102 doesn't have this at all), more memory, and others. And don't discount the extra channels (very useful to have). 

Oh hacking....  Certainly it's an attractive thought (free lunch, etc.), but some here have advised against attempting it.  I read the URL you gave and others as well, and the hacks seem simple enough.  Yet some say this can fail in a way that completely disables the scopes (so-called bricking), so I would not want to try it unless I knew for sure that the hacking is ~always~ reversible.  If I can be always get the scope back to stock, great, however I believe the term "bricking" implies otherwise.  Is there any way to back up the OEM firmware in these scopes in a reliable way?

[J-D-H]

Well 2 channels differential and one for current adds up to 3 channels.

Oh sure, there are numerous scenarios which would make a 4 channel scope useful .  It's just that this feature isn't free.

[David Hess]

I cannot say that I have ever absolutely needed to make more than one differential measurement simultaneously although a couple of my oscilloscopes can be configured to do it.

What I have done though is one differential measurement combined with one or two single ended measurements.  Even more often the differential signal was too noisy for a clean trigger so another channel or external trigger input was needed for a trigger signal.  A 2 channel oscilloscope can handle that though assuming it has an external trigger input which I think they all do.  4 channel oscilloscopes often lack separate external trigger inputs.

[nctnico]

Well 2 channels differential and one for current adds up to 3 channels.

Oh sure, there are numerous scenarios which would make a 4 channel scope useful .  It's just that this feature isn't free.

OK 4 channels are not free but very well worth the extra money  O0 I wouldn't want to go back to a 2 channel oscilloscope.

[andrija]

I rarely need 2 channels, let alone 4, for analog use. It all depends on what do you need and how do you use it. As a hobbyist I never needed 4 channels but I'm sure a professional would. So it's not a cut and dry issue.

Speaking of reliability, I don't feel my Rigol 2032A is all that reliable. It locks up regularly. The extra features over my other scope, Tek 2335, are however well worth it to keep it around. But for quick checking of waveforms, the cleanliness and speed of use of my old analog scope are unmatched by Rigol. Just having to move waveform up and down on Rigol compared to nice and smooth operation in analog... brrr. But Rigol easily measures the whole thing for you, and is much more precise as well.

[TunerSandwich]

4 channels is a must, for measuring even the simplest SMPS.  Assuming you are doing the "poor mans differential pair".  To determine efficiency it's helpful to measure i+o of a circuit in test.  the Rigol Z series is a no brainer IMO.  4 chan, usable screen, decent front end.....also even if you don't do the pseudo-differential thing, it's nice to have one pair on i/o and the other for logic/controller.  Also a lot of serial signals are going to need a minimum of 2 chan for clk and packets (granted ext trigger can be used for clk)

[TunerSandwich]

I rarely need 2 channels, let alone 4, for analog use. It all depends on what do you need and how do you use it. As a hobbyist I never needed 4 channels but I'm sure a professional would. So it's not a cut and dry issue.

Speaking of reliability, I don't feel my Rigol 2032A is all that reliable. It locks up regularly. The extra features over my other scope, Tek 2335, are however well worth it to keep it around. But for quick checking of waveforms, the cleanliness and speed of use of my old analog scope are unmatched by Rigol. Just having to move waveform up and down on Rigol compared to nice and smooth operation in analog... brrr. But Rigol easily measures the whole thing for you, and is much more precise as well.

My 2101A was also locking up.....firmware fixed the issue.  So did alterations in mem depth....not exactly ideal, but hardly "unreliable".  I don't have any fear that the hardware itself is going to fail anytime soon.  If we look at software bugs, nearly EVERY DSO has them, regardless of price....

[andrija]

4 channels is a must, for measuring even the simplest SMPS.  Assuming you are doing the "poor mans differential pair".  To determine efficiency it's helpful to measure i+o of a circuit in test.  the Rigol Z series is a no brainer IMO.  4 chan, usable screen, decent front end.....also even if you don't do the pseudo-differential thing, it's nice to have one pair on i/o and the other for logic/controller.  Also a lot of serial signals are going to need a minimum of 2 chan for clk and packets (granted ext trigger can be used for clk)

If I wanted to do SPMS I would buy a differential probe or two (which would be mandatory for mains devices anyway, don't they produce 600-800V, well beyond oscilloscope input rating?) rather than get a 4 channel scope but fair enough your arguments are certainly valid . As for reliability of Rigol, I am on latest firmware and just yesterday the scope kept locking up (as in auto stopping updating) at least 20 times while looking at function gen output and once it did a hard lockup requiring a full reboot. That's not what I call reliable but I agree that it's not that big of a deal if I have to press run button to get the thing to start again (as long as it doesn't do a hard lock up). I could also be using it wrong, I am certainly not an expert in using DSO's, this is my first one. I did have some p-p measurements turned on, as well as frequency and the network was plugged in too, plus it was "unlocked" but it looks like there are still serious bugs in there.

[TunerSandwich]

4 channels is a must, for measuring even the simplest SMPS.  Assuming you are doing the "poor mans differential pair".  To determine efficiency it's helpful to measure i+o of a circuit in test.  the Rigol Z series is a no brainer IMO.  4 chan, usable screen, decent front end.....also even if you don't do the pseudo-differential thing, it's nice to have one pair on i/o and the other for logic/controller.  Also a lot of serial signals are going to need a minimum of 2 chan for clk and packets (granted ext trigger can be used for clk)

If I wanted to do SPMS I would buy a differential probe or two (which would be mandatory for mains devices anyway, don't they produce 600-800V, well beyond oscilloscope input rating?) rather than get a 4 channel scope but fair enough your arguments are certainly valid . As for reliability of Rigol, I am on latest firmware and just yesterday the scope kept locking up (as in auto stopping updating) at least 20 times while looking at function gen output and once it did a hard lockup requiring a full reboot. That's not what I call reliable but I agree that it's not that big of a deal if I have to press run button to get the thing to start again (as long as it doesn't do a hard lock up). I could also be using it wrong, I am certainly not an expert in using DSO's, this is my first one. I did have some p-p measurements turned on, as well as frequency and the network was plugged in too, plus it was "unlocked" but it looks like there are still serious bugs in there.

did the entire UI lock up, or it just stopped triggering?  My guess is this is a trigger issue, and not a real lock-up.  My issue was a complete and total lock of the UI....only a power on/off cycle fixed it....that has stopped, since latest FW.  I must say the Rigol triggers are def it's weakest point....

Most SMPS are DC/DC....maybe I should further clarify and say POL, rather than SMPS.  The differential probe, is most certainly the correct way to do things, but I also have to assume most people buying a Rigol DS1000/2000 can't afford a proper differential probe, and are doing the poor mans method....which is another problem in the Rigol scopes....the math functions are NOT real time....since the operations are done in software. 

I can't imagine buying an E series Rigol over a Z series....and unless you need the low noise front end on the 2000A srries I can't imagine buying it over the Z series.....the 1000Z series is really a no brainer....nothing comes close for the value represented.....

[andrija]

You are right, it can be seen as a trigger bug and we know Rigol have serious trigger issues. I only had one hard freeze where it stopped responding to any button so in that respect it is reliable. There is another thread that seems to imply problems when you have it plugged into LAN so that may be a factor too.

I agree that new Z Rigol seems like a no brainer for a new scope but I'm no expert.

This whole tread made me take a look at used LeCroy scopes. I was looking previously for Agilent ones, and the main reason is to get an active probe with it. There's just too many available in a wide range of prices and with no knowledge which model to look for nor what their real value is, I am sure someone would fleece me. They sure look impressive even just from their photos, compared to Agilents.

[TunerSandwich]

You are right, it can be seen as a trigger bug and we know Rigol have serious trigger issues. I only had one hard freeze where it stopped responding to any button so in that respect it is reliable. There is another thread that seems to imply problems when you have it plugged into LAN so that may be a factor too.

I agree that new Z Rigol seems like a no brainer for a new scope but I'm no expert.

This whole tread made me take a look at used LeCroy scopes. I was looking previously for Agilent ones, and the main reason is to get an active probe with it. There's just too many available in a wide range of prices and with no knowledge which model to look for nor what their real value is, I am sure someone would fleece me. They sure look impressive even just from their photos, compared to Agilents.

I recently picked up a WR64MXi and I couldn't be happier.....it easily outclasses the other scopes I have on hand, and have used....honestly I would take it over 9000 series Agilent.  The probe options are vast.  I picked up two (essentially new) AP020 FET probes for $130 each.  The differential LV probes are under $500 each, and most current probes can be had for under a grand each....obviously there are higher end probes and what not, but used most don't surpass $1500. 

The way I made my value judgement is the following

1. WR64MXi (fully loaded with options and keys) $5000 usd (upgrade CPU RAM and HDD for another $200 ish)
2. AP020 active FET probes x 2 $130 each
3. Lecroy DA1855A-rm2 (2channel) $3100
4. assortment of passive differential leads (about $900 each)
5. Lecroy AP015 current probes x2 $1000-$1500 each
6. Lecroy ms500 18 ch logic probe ($2k ish) (500mHZ)

that is a full analysis grade system for around $12-$15k.....you just can't touch that value....especially with the lecroy triggers and analysis functions....buy a 9000 or even 4000 series Agilent, and load it up with probes and licenses, and you will be at easily double that price point....and won't get the same quality of triggering or math/analysis functions....

Obviously some folks aren't going to agree with that opinion, but for my specific needs, nothing else is going to do what that system can do.....the only thing I can imagine that would trump that level of function, would be an HDO8000.....the 12 bit resolution is useful for me. 

i have never been the type to gush about one brand or another....but owning a lecroy has completely changed my opinions on that subject.....I had played with and demo'd units before, but until you work with one day to day for a few hours a sitting, you don't really get a fair overview, of just how powerful their upper end systems are....it's hard to believe this unit i have is a 2009 vintage. 

I am sure some of the grumpy folks, who have never really spent time on a higher end lecroy system will call me a fanboi....or throw out the usual "lecrap" statement....but those are based on lower end rebadges....I highly doubt anyone who has lived with an MXi would call it "crap"....

[andrija]

That sounds like a great deal since I am seeing the same scope - with unknown options and no probes - for $14k on Ebay (never mind shipping and taxes). I won't even ask how you got that price. Anyhow that complete system is way beyond price range of most mortals unless it's a business expense, especially since it depreciates. Thanks for the probe info, it may come in handy if I start looking into other LeCroy scopes in the lower price bracket (is it safe to assume all 2000+ built  LeCroy scopes use the same probe interface?).

[TunerSandwich]

That sounds like a great deal since I am seeing the same scope - with unknown options and no probes - for $14k on Ebay (never mind shipping and taxes). I won't even ask how you got that price. Anyhow that complete system is way beyond price range of most mortals unless it's a business expense, especially since it depreciates. Thanks for the probe info, it may come in handy if I start looking into other LeCroy scopes in the lower price bracket (is it safe to assume all 2000+ built  LeCroy scopes use the same probe interface?).

ALL lecroy scopes using the probus interface are cross compatible with the majority of probus probes....

granted, this is not a tool for hobby or personal lab use (unless you are wealthy and have the need). 

I doubt the depreciation curve will be that steep, over the next 2-4 years.  I speculated a loss of less than 20%.  Who knows though 

For my business the specific tools and functions were necessary, so it does represent an excellent value for my business.  Just in the last week the tool has saved me immense amounts of time.  Easily will pay for itself in a couple months....

don't let some of the artificially high ebay prices scare you off....I did find one unit 104MXI-A that sold with probes and some other goodies for $4700.....obviously the licenses weren't there, but that is still a fantastic value, considering the alternative options....

[J-D-H]

Hacking....

Several contributors here have warned against trying to double the BW of a DSO via the hacking which some others have mentioned.  Presumably hacking would void the warranty, so on that basis it might be best to avoid.  However, suppose the DSO is no longer under warranty, then what?  Unless hacking can cause a DSO to lock up in a permanent way, why not try it?  Since DSO are largely computers, is it always possible to back up the firmware so there is always a means to get back to an earlier, reliable configuration?  Or can a lock-up be so bad that even this isn't possible?

John

[David Hess]

If I wanted to do SPMS I would buy a differential probe or two (which would be mandatory for mains devices anyway, don't they produce 600-800V, well beyond oscilloscope input rating?

Even worse is that add and invert mode requires each channel to see the entire input common mode range.  True differential inputs remove the common mode input early allowing orders of magnitude better sensitivity.  Analog oscilloscopes and some early DSOs have an advantage over later DSOs because they do the subtraction before digitizing.

4 channels is a must, for measuring even the simplest SMPS.  Assuming you are doing the "poor mans differential pair".

I cheated to get this capability.  One reason I like the Tektronix 7000 series mainframes is that 4 single ended inputs can be replaced with 2 high performance differential inputs while keeping the separate external trigger inputs although getting good performance requires matched probes.  I figured this was cheaper than buying dedicated differential probes for a DSO.

Most SMPS are DC/DC....maybe I should further clarify and say POL, rather than SMPS.  The differential probe, is most certainly the correct way to do things, but I also have to assume most people buying a Rigol DS1000/2000 can't afford a proper differential probe, and are doing the poor mans method....which is another problem in the Rigol scopes....the math functions are NOT real time....since the operations are done in software.

How fine is the vertical scale adjustment on the Rigols?  One thing I like about my analog oscilloscopes (and a couple of early DSOs) is that the analog input variable gain adjustment allows for relatively good DC and low frequency common mode rejection ratio.  If the Rigols do the math before the fine vertical scale adjustment is applied, then it does not matter and they will suffer from terrible common mode rejection but I have never seen this tested.

I can't imagine buying an E series Rigol over a Z series....and unless you need the low noise front end on the 2000A srries I can't imagine buying it over the Z series.....the 1000Z series is really a no brainer....nothing comes close for the value represented.....

Is the DS2000A series actually lower noise?  How is this measured?  The images I have seen show a noise level orders of magnitude lower than can be expected from a 1 Mohm 15 pf input so I assume they use some type of DSP noise reduction.  The trace when ground input coupling is selected does not look real.

The old Tektronix 2232 DSO has DSP noise reduction when using peak detection which returns a result very much like the DS2000A but it can be disabled for a more real view.  It does make for great if misleading photos though.

[David Hess]

Several contributors here have warned against trying to double the BW of a DSO via the hacking which some others have mentioned.  Presumably hacking would void the warranty, so on that basis it might be best to avoid.  However, suppose the DSO is no longer under warranty, then what?  Unless hacking can cause a DSO to lock up in a permanent way, why not try it?  Since DSO are largely computers, is it always possible to back up the firmware so there is always a means to get back to an earlier, reliable configuration?  Or can a lock-up be so bad that even this isn't possible?

There is another issue.  The transient response calibration is done digitally and may not have been calibrated at alternate input bandwidths.  If this is the case, then raising the bandwidth limit can result in poor transient response.  On one hand, most users are not going to be equipped to test for this problem but on the other hand, the same users are also unlikely to notice it.

This is also an issue if the calibration data is lost somehow which is a common problem on old DSOs which use battery backed up memory to store the calibration data.

[Wuerstchenhund]

Thanks for the probe info, it may come in handy if I start looking into other LeCroy scopes in the lower price bracket (is it safe to assume all 2000+ built  LeCroy scopes use the same probe interface?).

All LeCroy midrange and highend scopes since around 1994 (when the 9300 Series was introduced) use the same probe interface (called 'ProBus'). Their low end scopes (ScopeStation, LiteRunner, WaveAce, WaveJet) however don't, these scopes have plain standard BNC connectors. Some of the ultra high end scopes (i.e. WaveMaster 8000) use a common interface called 'ProLink' instead of ProBus where the BNC connector is replaced by an SMA type connector. However, for the common LeCroy scopes you want to look for ProBus compatible probes.

The common interface makes buying active probes a bit simpler, however don't forget that an active probe also needs to be supported by the scope software. This is generally not much of a problem with older probes (i.e. AP-034) and newer scopes, but for example the current ZD1000 active probe won't work with an old 9300 Series scope as the scope software won't know it. But then I guess only very few people would buy a brand new and very expensive high end active probe for using it with some 20 year old scopes. The older probes like the AP-033/034 are often sold for very low prices (much lower than the Tek and Agilent equivalents) and work fine with older and newer LeCroy scopes.

In terms of scopes, TunerSandwich's 64MXi is a great scope (and without doubt did get a great deal there) but I guess for hobby use spending this money is probably not an option for you. Depending on what you need have a look at the WaveRunner LT and WaveRunner2 LT Series. These scopes don't run Windows but VxWorks as OS, and still offer most of the options that can be found on the newer Windows scopes at a much lower price.

[TunerSandwich]

Thanks for the probe info, it may come in handy if I start looking into other LeCroy scopes in the lower price bracket (is it safe to assume all 2000+ built  LeCroy scopes use the same probe interface?).

All LeCroy midrange and highend scopes since around 1994 (when the 9300 Series was introduced) use the same probe interface (called 'ProBus'). Their low end scopes (ScopeStation, LiteRunner, WaveAce, WaveJet) however don't, these scopes have plain standard BNC connectors. Some of the ultra high end scopes (i.e. WaveMaster 8000) use a common interface called 'ProLink' instead of ProBus where the BNC connector is replaced by an SMA type connector. However, for the common LeCroy scopes you want to look for ProBus compatible probes.

The common interface makes buying active probes a bit simpler, however don't forget that an active probe also needs to be supported by the scope software. This is generally not much of a problem with older probes (i.e. AP-034) and newer scopes, but for example the current ZD1000 active probe won't work with an old 9300 Series scope as the scope software won't know it. But then I guess only very few people would buy a brand new and very expensive high end active probe for using it with some 20 year old scopes. The older probes like the AP-033/034 are often sold for very low prices (much lower than the Tek and Agilent equivalents) and work fine with older and newer LeCroy scopes.

In terms of scopes, TunerSandwich's 64MXi is a great scope (and without doubt did get a great deal there) but I guess for hobby use spending this money is probably not an option for you. Depending on what you need have a look at the WaveRunner LT and WaveRunner2 LT Series. These scopes don't run Windows but VxWorks as OS, and still offer most of the options that can be found on the newer Windows scopes at a much lower price.

I can't imagine NOT running windows as being a severe limitation....especially on a hobby bench, where the need to mount iSCSI or join a domain is not necessary.  If all the math and functionality is the same, I think the 64MXi or other XI/MXi/6100/7000 etc is overkill for most applications? 

[Wuerstchenhund]

I can't imagine NOT running windows as being a severe limitation....especially on a hobby bench, where the need to mount iSCSI or join a domain is not necessary.  If all the math and functionality is the same, I think the 64MXi or other XI/MXi/6100/7000 etc is overkill for most applications?

Well, of course the Windows scopes offer more advanced features, plus a few nice things like touch interface, but the majority of features found in the Windows scopes can also be found in the older VxWorks-based devices.

The biggest difference is in serial decode which is absent from the older non-Windows scopes, but in most cases a cheap USB logic analyzer is probably the better choice for a hobby lab anyways. Especially when considering how limited the MSO functionality in scopes like the Rigols, Siglents or even Agilent's DSO-X 2k/3k is.

I'd say a WaveRunner2 scope plus a good USB logic analyzer (i.e. Salae) should be a good and economical choice for a hobbyist lab.

[J-D-H]

Several contributors here have warned against trying to double the BW of a DSO via the hacking which some others have mentioned.  Presumably hacking would void the warranty, so on that basis it might be best to avoid.  However, suppose the DSO is no longer under warranty, then what?  Unless hacking can cause a DSO to lock up in a permanent way, why not try it?  Since DSO are largely computers, is it always possible to back up the firmware so there is always a means to get back to an earlier, reliable configuration?  Or can a lock-up be so bad that even this isn't possible?

There is another issue.  The transient response calibration is done digitally and may not have been calibrated at alternate input bandwidths.  If this is the case, then raising the bandwidth limit can result in poor transient response.  On one hand, most users are not going to be equipped to test for this problem but on the other hand, the same users are also unlikely to notice it.

This is also an issue if the calibration data is lost somehow which is a common problem on old DSOs which use battery backed up memory to store the calibration data.

To say that cal data can be lost implies that backup wasn't possible?  What about modern DSOs?  If they allow firmware updates via the USB port, I certainly hope they also allow a full backup of all internal data prior to attempting the update....  Yes?

[Wuerstchenhund]

To say that cal data can be lost implies that backup wasn't possible?  What about modern DSOs?  If they allow firmware updates via the USB port, I certainly hope they also allow a full backup of all internal data prior to attempting the update....  Yes?

I can't speak for other scopes but on LeCroy Windows scopes the cal data is on the hard drive and can easily be backed up to USB stick/CD/DVD/cloud storage/wherever.

On the older scopes which run VxWorks instead of Windows (WaveRunner/WaveRunner2/WavePro 900) the cal data won't be lost when the lithium backup battery dies (the scope will just refuse to boot until the battery has been replaced).

[rstoer]

I rarely need 2 channels, let alone 4, for analog use. It all depends on what do you need and how do you use it...

I have a four channel scope but I agree, I almost never need more than two. There's is a side benefit to four channels though - On most two channel scopes using both channels cuts your sample rate in half. This can usually be avoided on a four channel scope by using channels 1&3, 2&4, etc.