Help choosing DSO :) MSO5000 ?

[tautech]

https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/
Watch the Howardlong video in the first post.

[nubinstanley]

Does it make MSO5000 unusable and useless? Not by any chance. It is a scope that has many features and if you're not looking into low level signals and use it for digital  it is a tool that will let knowledgeable person do many things. It is not a bad scope, it's just not very good at certain, important figures of merit.
And this is all I already SAID in my post. For many people not doing precision stuff analog performance will be GOOD ENOUGH.

Do take a look at my reply (with a list of queries) to your earlier post.

[nubinstanley]

https://www.eevblog.com/forum/testgear/siglent-sds2000x-plus-coming/
Watch the Howardlong video in the first post.

I am assuming that you are talking about this video :

If yes, I had seen that. There is also a pinned post under the video (An update from 4 months ago), the last sentence reads :

Code: [Select]

The TL;DR: For a daily driver, I have changed my mind, and I now find I'm slightly more likely to go for the Rigol despite its several shortcomings, and this is mostly because I'm more comfortable with the workflow. The slightly smaller form factor of the Rigol means fits better on my bench, underneath the monitor without obscuring it. YMMV of course!
Which was why I kept the MSO5000 in contention even after going through the front-end noise thread. If this guy can use it as a daily driver, I definitely can - I'll most likely have 1/10th the use/knowledge starting off.

[Fungus]

Image you posted is not a representation of how MSO5000 shows trace and how much noise it has.

The text I wrote above that image was: "If a signal is repetitive then things like averaging mode can clean it up. Here's a screenshot taken from a MSO5000 with averaging mode on:"

But we cannot spread misinformation that it is EQUALLY GOOD as scopes that have SUPERIOR performance noise figures...

The text I wrote below it was: "I know they're not perfect solutions and the Siglent is better."

 

[Fungus]

I am assuming that you are talking about this video :

There have been firmware updates since that video.

Don't forget: There's a lot of happy MSO5000 users on EEVBLOG. Just go to any of the MSO5000 threads and see.

[nubinstanley]

I am assuming that you are talking about this video :

There have been firmware updates since that video.

Don't forget: There's a lot of happy MSO5000 users on EEVBLOG. Just go to any of the MSO5000 threads and see.

 

One more things I found : It seems to be easier to make DIY logic probes from the Siglent ?

[2N3055]

Despite some dubious things being said here, MSO5000 has much more noise than SDS2104X Plus/SDS1104X-E and that is important if you need to see low level signals.
No amount of wishful thinking and pushing a head in a sand will change that fact. Thick lines will hide the fact that in that noise there are other signals riding on large one...

You cannot replace low noise equipment with a high noise equipment. Higher noise equipment, by definition, hides more things from you.
You cannot replace bandwidth. If you need 1 GHz bandwidth, 200Mhz scope won't do. Period.
If you need to sample long periods with high sample rate, you need long memory. Scope with large memory can sample what a small memory scope can, but not vice versa.
If you need decoding, you need to have it.

All that being said, I personally like the scope to have lowest noise and best analog performance first. On top of that all other.

MSO5074, SDS2104X Plus and SDS1104X-E are really 3 different types of scopes.

I would say that MSO5074 (if equipped with MSO probes) is a decent mixed signal scope for looking at logic level signals. It has noisy front end, but it has lots of memory and MSO and decoding. I personally don't like their implementation, but is is a useful scope for amateur robotics, microcontroller arduino type stuff etc.. It has 2Ch of basic AWG. It has Bode plot but is not very good.

SDS2104X Plus is much better scope for analog stuff etc. It is more designed for that. Since I do that, I prefer it to MSO5000. I also, personally, like their software concept better. It is more logical to me. It has 1CH of basic AWG. It's Bode plot implementation is top notch.

But, If I were you, I would just go and get SDS1104X-E. It is a lot of scope for the money. Possibly all the scope you need for a long time. And for the rest of the money you can buy few other instruments....
And these  class of scopes actually hold their value. You will easily sell it later and upgrade if you have a need for it.

I do understand what you are saying. But my question is as follows :

I work a lot with microcontrollers and so I know that I want 4 channels, good protocol decoding and LA capabilities (cheap DIY probe possible with the MSO5000 ?). It's also unlikely that I'll get an acceptable resale value for a scope in my location - So, the one I buy now, will most likely be the one I'll be using 10 years from now!

Coming to analog part - Right now, I don't even know a use-case where I would need to look at very low voltages. I am planning on learning more as I go and so I don't know for a fact that I'll not know such a use-case in future. So no, I cannot say with absolute certainty that the low noise floor will not be an issue for me. And as I mentioned earlier, I am not exactly able to comprehend how the front-end noise would effect the reproduction of higher voltage signals (1V+ ?) - the thicker traces - will it be too off-putting ?

So..

1) Assuming that for the next 3 years I'll mostly be in the digital domain with a little bit of analog stuff, is the MSO5000 a bad buy ?
2) The thicker traces - will it be too off-putting ? Will it be too difficult to work with analog signals - even if not very low voltages ?
3) How is the protocol decoding/LA capabilities of the SDS1104X-E/SDS2104X Plus (compared to the MSO5000)? Also, DIY LA probes are difficult to make ?
4) For the occasions where I do have to work with low voltages, can I use some work-arounds ? Pre-amps ?

As I did mention, I shortlisted 3 scopes : SDS1104X-E, MSO5104, SDS2104X Plus. The last one is off the list as I don't want to spend that much. Right now, after all the reading, I am a bit scared to go the MSO5000 way. But, to be frank, I don't know if that feeling is justified. Maybe if I get answers to the above questions, I can make a more informed decision.

I sure do hope if a few of the MSO5000 owners chime in. Anyhow, looks like I'll be saving some money (by buying the SDS1104X-E)

I'm sorry but I thought that enough was said in meantime that direct answer to this was not necessary.

What does the digital domain mean to you?
Scope use with digital signal is two fold: signal integrity and decoding protocols. With MSO scope you can also cross trigger between analog/digital with many channels. If you only do serial protocols no need for digital channels that much. 4ch scope will do.

Signal integrity on digital signals is analog work. Decoding is software works. Being able to do both on same scope   without switching cables and such is convenient.

If you do a lot of decoding and not much analog, and cross triggering is no important, a small USB protocol analyser and simpler scope will do better job. Just because decoding is better on PC.
So maybe you'l be better of with SDS1104X-E and a small logic analyser.

How are you going to look at digital signal? With passive 10x probe? That means that when you are at 1 V/div for 3.3V logic, you are actually using scope at 100 mV/div. What is noise then?

I'm not going to make a decision for you. I cannot.

As I said, if you want to just buy something because you want to give yourself a present of something nice, then buy whatever makes you happy. If you need something to do the job, just get simplest and cheapest thing that will do the job.
I think you have enough info. Now it is up to you to decide what you want/need.

[nubinstanley]

I'm sorry but I thought that enough was said in meantime that direct answer to this was not necessary.

What does the digital domain mean to you?
Scope use with digital signal is two fold: signal integrity and decoding protocols. With MSO scope you can also cross trigger between analog/digital with many channels. If you only do serial protocols no need for digital channels that much. 4ch scope will do.

Signal integrity on digital signals is analog work. Decoding is software works. Being able to do both on same scope   without switching cables and such is convenient.

If you do a lot of decoding and not much analog, and cross triggering is no important, a small USB protocol analyser and simpler scope will do better job. Just because decoding is better on PC.
So maybe you'l be better of with SDS1104X-E and a small logic analyser.

How are you going to look at digital signal? With passive 10x probe? That means that when you are at 1 V/div for 3.3V logic, you are actually using scope at 100 mV/div. What is noise then?

I'm not going to make a decision for you. I cannot.

As I said, if you want to just buy something because you want to give yourself a present of something nice, then buy whatever makes you happy. If you need something to do the job, just get simplest and cheapest thing that will do the job.
I think you have enough info. Now it is up to you to decide what you want/need.

 

[jacekowski]

I would strongly advise against MSO5000 (i've had 2 that both had serious issues with (presumably) ADC interleaving).

In all screenshots below it was fed with nice clean sine wave, artefacts were generated in the scope.

The issue is less obvious when scope is limited to 100MHz, but if you know where to look it is quite easy to find.

[normi]

The MSO5000 does have more noise than the Siglent, but so does the mso7000 and mso8000 which are way more expensive than the Siglent, so it is clear that it is not all about  noise. But to be objective, I have found that there are more Youtube videos from MSO5000 owners than I have found for the Siglent  SDS2104X Plus, so there are numerous active use cases that can be viewed online to get an independent review. Both are nice scopes and offer a lot of features but the US$500 savings gained by buying the MSO5000 could go a long way in buying other much needed equipment. If you want to work with low level signals you can buy cheap amplifiers or build one which will pickup  signals even lower than what the Siglent can. If you do need to work with low level analog signals then get the cheaper Siglent scope that has low noise floor.

There are others on the forum who have both scopes and are pleased with the MSO5000 because it does have a number of capabilities that the
 SDS2104X Plus does not offer.

On another note, it is far easier to dispose of a scope for a good price if you live in wealthy country, expensive scopes are harder to sell in countries with lower income and so you will have to give bigger discounts. If I were selling a scope I would have to ship it to the US to get a better price.


Youtuber  tomtektest does an extensive series on the scope, and his reviews are based on purchases he made and are not sponsored.

[rf-loop]

If you want to work with low level signals you can buy cheap amplifiers or build one which will pickup  signals even lower than what the Siglent can.

Fun argument. This can do also with Siglent. And circle close... back to starting line.

Btw, dc coupled low noise, 1Mohm impedance, 200 - 500MHz linear amplifier is not simplest case. Been there, not want replay.
 

[Fungus]

What does the digital domain mean to you?
Scope use with digital signal is two fold: signal integrity and decoding protocols. With MSO scope you can also cross trigger between analog/digital with many channels. If you only do serial protocols no need for digital channels that much. 4ch scope will do.

Yep. 4 channels can do a lot of "digital" work. I wouldn't get the logic breakout on day one, I'd wait and see how it goes.

Signal integrity on digital signals is analog work.

It doesn't need massively low noise though. An MSO5000 is perfectly capable of finding runt pulses and telling you if the pullups on your I2C bus are working.

If you do a lot of decoding and not much analog, and cross triggering is no important, a small USB protocol analyser and simpler scope will do better job. Just because decoding is better on PC.

Those logic analyzers are OK for examining large data exchanges but they aren't real time. You can't just look at the signal and fiddle with the controls. You have to press "record" on the PC, wait a while, press "stop", then go in and see what you got.

[nubinstanley]

I would strongly advise against MSO5000 (i've had 2 that both had serious issues with (presumably) ADC interleaving).

In all screenshots below it was fed with nice clean sine wave, artefacts were generated in the scope.

The issue is less obvious when scope is limited to 100MHz, but if you know where to look it is quite easy to find.

This is scary...

Where you able to find other folks facing the same problem ? Did you end up getting a different scope (if yes, which one ? ) ?

Has the issue been resolved in the latest firmware ?

[Fungus]

I would strongly advise against MSO5000 (i've had 2 that both had serious issues with (presumably) ADC interleaving).

In all screenshots below it was fed with nice clean sine wave, artefacts were generated in the scope.

The issue is less obvious when scope is limited to 100MHz, but if you know where to look it is quite easy to find.

Has anybody else had this issue? You'd think something so blindingly obvious would have been all over the forums since day one.

[nubinstanley]

The MSO5000 does have more noise than the Siglent, but so does the mso7000 and mso8000 which are way more expensive than the Siglent, so it is clear that it is not all about  noise. But to be objective, I have found that there are more Youtube videos from MSO5000 owners than I have found for the Siglent  SDS2104X Plus, so there are numerous active use cases that can be viewed online to get an independent review. Both are nice scopes and offer a lot of features but the US$500 savings gained by buying the MSO5000 could go a long way in buying other much needed equipment. If you want to work with low level signals you can buy cheap amplifiers or build one which will pickup  signals even lower than what the Siglent can. If you do need to work with low level analog signals then get the cheaper Siglent scope that has low noise floor.

There are others on the forum who have both scopes and are pleased with the MSO5000 because it does have a number of capabilities that the
 SDS2104X Plus does not offer.

On another note, it is far easier to dispose of a scope for a good price if you live in wealthy country, expensive scopes are harder to sell in countries with lower income and so you will have to give bigger discounts. If I were selling a scope I would have to ship it to the US to get a better price.

I am not an expert and I just want to buy a scope (at a particular price point) that can last quite a while. If I am being honest, the SDS1104X-E would probably be more than enough for me right now. But, I am not buying this scope just for "right now" - I am hoping to learn. I am from India, and I can confidently say that I'll not get a decent resale value for the scope. I am already buying at a higher price compared to someone from the US (MSO5074 : ~1300; SDS2104X Plus : ~1850 USD) So, whatever I buy now is the one I'll be using for a loooong time.

If the issues plaguing the MSO5074 can be "worked around", I'll of course prefer to get that compared to the Siglent. After going through a lot of threads on the forum, the primary issue seems to be the front-end noise in the MSO5000. I understand that working with low level signals can be difficult because of this. But I am not experienced enough to comprehend the effect it is going to have for other use-cases.

1) During the (possibly rare) occasions when I need to work on low level signals, are there work-arounds available ? (Seems like I can possibly use cheap/DIY pre-amps ?)
2) The thicker traces - will it be too off-putting ? Will it be too difficult to work with analog signals - even if not very low voltages ? Using an example that @2N3055 pointed out :

How are you going to look at digital signal? With passive 10x probe? That means that when you are at 1 V/div for 3.3V logic, you are actually using scope at 100 mV/div. What is noise then?

What is the answer here ? With the front-end noise that the MSO5000 has, is it too difficult to work with 3.3V logic with a 10X probe at 100mV/div ?

Any idea about the issue @jacekowski mentioned in the previous post ?

Youtuber  tomtektest does an extensive series on the scope, and his reviews are based on purchases he made and are not sponsored.

Yes, found the list yesterday. Going through those videos now.

[nubinstanley]

If you want to work with low level signals you can buy cheap amplifiers or build one which will pickup  signals even lower than what the Siglent can.

Fun argument. This can do also with Siglent. And circle close... back to starting line.

Hey, if I didn't have to spend ~570 USD extra, I would have already bought the SDS2104X Plus.

Btw, dc coupled low noise, 1Mohm impedance, 200 - 500MHz linear amplifier is not simplest case. Been there, not want replay.
 

Forgive my ignorance... Could you give me a situation where this is applicable ? Maybe then, I'll be able to make a judgement - whether or not such a use-cases would ever be one that I'll need to consider. (I do agree that I cannot say with absolute certainty that I'll never be in such a situation. )

[nctnico]

2) The thicker traces - will it be too off-putting ? Will it be too difficult to work with analog signals - even if not very low voltages ? Using an example that @2N3055 pointed out :

This. Even with digital signals you will want to check whether they are within the threshold limits so having the thinnest trace possible helps to make a quick assessment without guessing what the signal level actually. Sure you can enable averaging or hi-res but those are just crutches if your actual problem is a noisy oscilloscope frontend.

And don't get fooled by 'low level signals'. An oscilloscope scales (amplifies / attenuates) the input voltage to match the ADC's input range so whatever noise is in the frontend, it will affect every V/div setting. So even a 50Vpp signal is affected.

[nubinstanley]

2) The thicker traces - will it be too off-putting ? Will it be too difficult to work with analog signals - even if not very low voltages ? Using an example that @2N3055 pointed out :

This. Even with digital signals you will want to check whether they are within the threshold limits so having the thinnest trace possible helps to make a quick assessment without guessing what the signal level actually. Sure you can enable averaging or hi-res but those are just crutches if your actual problem is a noisy oscilloscope frontend.

Say I use averaging/hi-res mode while working with 3.3V digital logic as mentioned in the example. Can there be something that gets hidden due to it ?

And don't get fooled by 'low level signals'. An oscilloscope scales (amplifies / attenuates) the input voltage to match the ADC's input range so whatever noise is in the frontend, it will affect every V/div setting. So even a 50Vpp signal is affected.

Mmmm. Let's keep aside the low level signals. Let's say I am working on stuff in the 1Vpp - 50Vpp range. How thick a trace are we talking about ? Is it bad enough to justify me spending around 570 USD more ? (Not saying that this is the only thing that differentiates an SDS2104X+ from an MSO5000 - Display size, a "bit" sluggish UI etc are things I am willing to overlook for the money... Am I overlooking something else that is important ? )

[rf-loop]

Btw, dc coupled low noise, 1Mohm impedance, 200 - 500MHz linear amplifier is not simplest case. Been there, not want replay.
 

Forgive my ignorance... Could you give me a situation where this is applicable ? Maybe then, I'll be able to make a judgement - whether or not such a use-cases would ever be one that I'll need to consider. (I do agree that I cannot say with absolute certainty that I'll never be in such a situation. )

Sorry this my comment was not at all for you.

Based on what I have read from this chain, I still believe that you will never find yourself in such a situation. This  not to absolutely exclude that it still couldn’t be possible.

If I am in a situation as you have explained and from which I have formed an imaginary reality of yours, then my choice would be SDS1104X-E. If my image of your reality were the same as your real reality.

[Fungus]

1) During the (possibly rare) occasions when I need to work on low level signals, are there work-arounds available ? (Seems like I can possibly use cheap/DIY pre-amps ?)

Yes.

From DC up to about 1Mhz you can use a $2 opamp to preamplify the signal.
Above that there's all sorts of amplifiers available.

A single box that does all frequencies would be awfully expensive but if you have a specific need then it can be quite cheap.

Realistically though, you probably won't need it. The MSO5000 will probably be OK without them.

2) The thicker traces - will it be too off-putting ? Will it be too difficult to work with analog signals - even if not very low voltages ? Using an example that @2N3055 pointed out :

How are you going to look at digital signal? With passive 10x probe? That means that when you are at 1 V/div for 3.3V logic, you are actually using scope at 100 mV/div. What is noise then?

What is the answer here ? With the front-end noise that the MSO5000 has, is it too difficult to work with 3.3V logic with a 10X probe at 100mV/div ?

No.

a) Nobody's forcing you to use 1V/div. Feel free to fill the screen when/if you're interested in "noise".
2) Noise will be what it is - a percentage of the on-screen divisions. Here's a typical screenshot:



You'd think people who post questions like the above would show what happens instead of just invoking FUD/bogeymen, but... noooo.

If the signal is repetitive like in that screenshot then you can turn on averaging mode or hires mode and get much thinner traces. Most of the time though, you'll be interested in the frequency, amplitude, rise times and squareness of the corners. Noise does not hide any of those things, see image above.

(images taken from noise discussion here)

[nctnico]

Now show the same signal with 2 or 3 divisions like you would normally do to fit more traces on the screen and you'll see that the amount of noise versus amount of signal gets quite large.

[nubinstanley]

Btw, dc coupled low noise, 1Mohm impedance, 200 - 500MHz linear amplifier is not simplest case. Been there, not want replay.
 

Forgive my ignorance... Could you give me a situation where this is applicable ? Maybe then, I'll be able to make a judgement - whether or not such a use-cases would ever be one that I'll need to consider. (I do agree that I cannot say with absolute certainty that I'll never be in such a situation. )

Sorry this my comment was not at all for you.

Based on what I have read from this chain, I still believe that you will never find yourself in such a situation. This  not to absolutely exclude that it still couldn’t be possible.

If I am in a situation as you have explained and from which I have formed an imaginary reality of yours, then my choice would be SDS1104X-E. If my image of your reality were the same as your real reality.

Even so, just out of curiosity, could you give me an example use-case ?

Also, maybe I'll ask a question from a different perspective : In your experience, what are a few use-cases where you think the SDS1104X-E can't cut it ?

[nubinstanley]

1) During the (possibly rare) occasions when I need to work on low level signals, are there work-arounds available ? (Seems like I can possibly use cheap/DIY pre-amps ?)

Yes.

From DC up to about 1Mhz you can use a $2 opamp to preamplify the signal.
Above that there's all sorts of amplifiers available.

A single box that does all frequencies would be awfully expensive but if you have a specific need then it can be quite cheap.

Realistically though, you probably won't need it. The MSO5000 will probably be OK without them.

2) The thicker traces - will it be too off-putting ? Will it be too difficult to work with analog signals - even if not very low voltages ? Using an example that @2N3055 pointed out :

How are you going to look at digital signal? With passive 10x probe? That means that when you are at 1 V/div for 3.3V logic, you are actually using scope at 100 mV/div. What is noise then?

What is the answer here ? With the front-end noise that the MSO5000 has, is it too difficult to work with 3.3V logic with a 10X probe at 100mV/div ?

No.

a) Nobody's forcing you to use 1V/div. Feel free to fill the screen when/if you're interested in "noise".
2) Noise will be what it is - a percentage of the on-screen divisions. Here's a typical screenshot:



You'd think people who post questions like the above would show what happens instead of just invoking FUD/bogeymen, but... noooo.

If the signal is repetitive like in that screenshot then you can turn on averaging mode or hires mode and get much thinner traces. Most of the time though, you'll be interested in the frequency, amplitude, rise times and squareness of the corners. Noise does not hide any of those things, see image above.

(images taken from noise discussion here)

Ok....

[nubinstanley]

Now show the same signal with 2 or 3 divisions like you would normally do to fit more traces on the screen and you'll see that the amount of noise versus amount signal gets quite large.

I can understand (I think). But (an excerpt from an earlier post),

"Is it bad enough to justify me spending around 570 USD more ? (Not saying that this is the only thing that differentiates an SDS2104X+ from an MSO5000 - Display size, a "bit" sluggish UI etc are things I am willing to overlook for the money... Am I overlooking something else that is important ? )"

[Fungus]

Now show the same signal with 2 or 3 divisions like you would normally do to fit more traces on the screen and you'll see that the amount of noise versus amount signal gets quite large.

When you have multiple traces on screen you're usually more interested in the timing and positions of the signals relative to each other.

Noise doesn't hide that either. eg. can you see if the data signal changes only when clock is low?