REVIEW - Owon SDS7102 - A look at the SDS series from Owon

[tinhead]

The first two are of my SDS7102 and Tek DPO20204 attached to the same signal generator using their supplied probes. 

with probes directly to signal gen output, right?

The Tek is being triggered off of the Owon's trigger out since the noise seems to be synchronized with the Owon's sweeps and this allows it to be better captured visually. 

ok, and how it looks when Tek is using own trigger and Owon is on? Is Tek having then clean picture and Owon noisy one?

The next is of the Tek connected to the same signal generator, but with the Owon turned off and it triggered internally.  The last is of the Owon connected to the signal generator using a coaxial cable and a 50 Ohm straight through terminator.

so here im principe both connected directly t signal gen and triggering with own internal trigger, they looks ok, but for sure coax cable is part of the answer here.

[bill2010]

with probes directly to signal gen output, right?

Yes -- directly connected using the pincher and ground lead clipped onto a BNC to banana adapter that is attached to the signal generator through a 50 Ohm passthrough terminator.

ok, and how it looks when Tek is using own trigger and Owon is on? Is Tek having then clean picture and Owon noisy one?

It is just as noisy, but the noise is "free running" so it gets somewhat washed out by the Tek's DPO processor.

so here im principe both connected directly to signal gen and triggering with own internal trigger, they looks ok, but for sure coax cable is part of the answer here.

Yes -- when either scope is directly connected it is fully shielded and the noise level drops significantly.  Forgot to mention that I placed a 10x attenuator in line on the cable so the amplitude would match the 10x probe.  As posted earlier, the noise seems to be comprised of 70 - 100 MHz bursts that are in sync with the Owon's sweep.

[rf-loop]

I have bought two brand new BNC terminators 50R and tried them. There is absolutely no difference between "fully inserted plug" or "just touching". I attach a picture which should count for all the four that you have asked me.

This your image looks ok. (It is nice if someone with example Tektronix do same. 1ms/div, 5mV/div, peak mode.)
It was important to find that 50ohm terminator fully in and "open but shield" make no difference. It was 2011 when there was manufacturing mistake and it make CH1 very noisy and there can see big difference if terminated or open. Shielding is important for block noise what is coming outside)
This peak - peak noise you show in this image is good if you compare for what ever this class of scope. You can also find some tables where this noise is littlebit compared to some others.
Scope analog front end  noise level looks ok.

The Owon people have come back to me and told me not to expect too much from the scope as it is a cheap scope. However I genuinely think mine is broken as there is all this noise which renders the scope almost useless for signals below 100mV RMS or thereabouts.

This need investigate more.
(I have some suspects and there may be something....)

But it is clear that if signal what oscilloscope see in its BNC input do not have noise also oscilloscope do not internally add this noise to signal. 

[akis]

But it is clear that if signal what oscilloscope see in its BNC input do not have noise also oscilloscope do not internally add this noise to signal.

In my "uneducated" view, the scope is useless for observing signals below a certain amplitude. The 20MHz internal filter makes things a bit better in these cases, but then you have a 20MHz (or less) scope, not a 100MHz scope, and it is still not clear enough for it to even lock on.

My old, 30+ year old, 20MHz scope displays the same curves with no problems.

[rf-loop]

My old, 30+ year old, 20MHz scope displays the same curves with no problems.

This is fun. But very very common wondering. Why my digital oscilloscope is so noisy if I compare it to my old conventional oscilloscope and what best.. 20Mhz oscilloscope. Try look 10ns pulse with it, say example it occurs 10 times per second. How about 100ns pulse what occurs example on time per second. You see something? Try same with digital oscilloscope. Why? What filters it out... and I can tell... it filters also noise by same way.

You can easu find example some Tektronix paper about this.

Then here more about Owon tests related to noise. Look also last table.

In this thread was bill2010 images with Owon and Tek
here in same scale (yes of course Tek is much more expensive, totally different class of tool. Also it have more BW but here sampling rate same and same horz speed and same vertical. Owon BW is around over 150MHz and Tek is "rated" as 200MHz but not know what is real BW.. perhaps lot of over 200MHz but it do not explain this noise... but also it need remember that in this picture Tek have collected lot of more data... if Owon is long time with infinite persistence it may change this situation some amount.. )

[akis]

Here are some more pictures to demonstrate the problem. Scope was away from any sources of interference and not connected to the mains.

The signal is 120mV pk-pk at 1KHz. This is not a "small" signal. I repeat : it is 120mV pk-pk.

The situation gets worse, much worse, as the signal gets weaker. I even wonder why bother have a "5mV" or a "10mV" scale when in reality they are so useless. Also the only way to get a clear picture is to "limit" the scope to 20MHz, in which case why advertise it as a "100MHz".


Picture 1 shows the amount of random noise :

Picture 2 shows the same but with the scope set to 20MHz.

Picture 3 shows the same but with the addition of "sample" rather than "peak". Had I put the screen persistence on (like old phosphor displays) it would also be much worse.

[firehopper]

I'm looking to buy a owon sds7102 sometime within the next week or so, and was wondering, what is the newest firmware for it. It will hopefully be the version with vga and ethernet port. I also hope to obtain a battery for it as well. dont know what serial number I will be getting till it gets here.

[rf-loop]

Here are some more pictures to demonstrate the problem. Scope was away from any sources of interference and not connected to the mains.

The signal is 120mV pk-pk at 1KHz. This is not a "small" signal. I repeat : it is 120mV pk-pk.

The situation gets worse, much worse, as the signal gets weaker. I even wonder why bother have a "5mV" or a "10mV" scale when in reality they are so useless. Also the only way to get a clear picture is to "limit" the scope to 20MHz, in which case why advertise it as a "100MHz".


Picture 1 shows the amount of random noise :

Picture 2 shows the same but with the scope set to 20MHz.

Picture 3 shows the same but with the addition of "sample" rather than "peak". Had I put the screen persistence on (like old phosphor displays) it would also be much worse.

There is something wrong.


Here some images.

Now here used Owon probe what is connected to 50ohm terminated signal generators output. (50ohm source need 50ohm termination for correct level.)
All shots are real time single captures and of course without averaging. Normal acquire.
for 300kHz used HP8644B  signal generator (it have very clean signal)
for 1kHz used HP3314A "noise gener....oops... function generator"
Owon used for test:
SDS7102V  series: 1246xxx 
Version 2.8.2
Probe: "T5100"  (Oscilloscope normal accessory)

[akis]

Therefore I will email the company I bought it from and ask them to replace it, obviously the scope is capable of better quality, I must have been given a bad one.

[bill2010]

Here are some more pictures to demonstrate the problem. Scope was away from any sources of interference and not connected to the mains.

The signal is 120mV pk-pk at 1KHz. This is not a "small" signal. I repeat : it is 120mV pk-pk.

The situation gets worse, much worse, as the signal gets weaker. I even wonder why bother have a "5mV" or a "10mV" scale when in reality they are so useless. Also the only way to get a clear picture is to "limit" the scope to 20MHz, in which case why advertise it as a "100MHz".


Picture 1 shows the amount of random noise :

Picture 2 shows the same but with the scope set to 20MHz.

Picture 3 shows the same but with the addition of "sample" rather than "peak". Had I put the screen persistence on (like old phosphor displays) it would also be much worse.

There is something wrong.


Here some images.

Now here used Owon probe what is connected to 50ohm terminated signal generators output. (50ohm source need 50ohm termination for correct level.)
All shots are real time single captures and of course without averaging. Normal acquire.
for 300kHz used HP8644B  signal generator (it have very clean signal)
for 1kHz used HP3314A "noise gener....oops... function generator"
Owon used for test:
SDS7102V  series: 1246xxx 
Version 2.8.2
Probe: "T5100"  (Oscilloscope normal accessory)

Here is a screen capture from mine mirroring the setup you used for your first picture.  My signal generator isn't as clean as the HP, but I've verified the output on my Tek DPO (with the Owon turned off) and it is much cleaner then the signal the Owon is seeing. Serial 71021246xxx, firmware 2.8.2.  Other than this noise problem, I really like the scope.

Looks like mine is heading back as well...

[bill2010]

Here are some more pictures to demonstrate the problem. Scope was away from any sources of interference and not connected to the mains.

The signal is 120mV pk-pk at 1KHz. This is not a "small" signal. I repeat : it is 120mV pk-pk.

The situation gets worse, much worse, as the signal gets weaker. I even wonder why bother have a "5mV" or a "10mV" scale when in reality they are so useless. Also the only way to get a clear picture is to "limit" the scope to 20MHz, in which case why advertise it as a "100MHz".


Picture 1 shows the amount of random noise :

Picture 2 shows the same but with the scope set to 20MHz.

Picture 3 shows the same but with the addition of "sample" rather than "peak". Had I put the screen persistence on (like old phosphor displays) it would also be much worse.

Did you use the supplied probes to connect to the signal source in picture 1?  If so, was the probe set to 10x or 1x?  I am trying to reproduce on my own SDS7102 as I believe that mine has a similar defect. Do you have a coaxial cable you can use?  On mine, the noise seems to be external (but generated by the Owon itself) to the actual signal path as it is greatly attenuated when using a fully shielded connection between the scope and the signal generator.

[akis]

Here are some more pictures to demonstrate the problem. Scope was away from any sources of interference and not connected to the mains.

The signal is 120mV pk-pk at 1KHz. This is not a "small" signal. I repeat : it is 120mV pk-pk.

The situation gets worse, much worse, as the signal gets weaker. I even wonder why bother have a "5mV" or a "10mV" scale when in reality they are so useless. Also the only way to get a clear picture is to "limit" the scope to 20MHz, in which case why advertise it as a "100MHz".


Picture 1 shows the amount of random noise :

Picture 2 shows the same but with the scope set to 20MHz.

Picture 3 shows the same but with the addition of "sample" rather than "peak". Had I put the screen persistence on (like old phosphor displays) it would also be much worse.

Did you use the supplied probes to connect to the signal source in picture 1?  If so, was the probe set to 10x or 1x?  I am trying to reproduce on my own SDS7102 as I believe that mine has a similar defect. Do you have a coaxial cable you can use?  On mine, the noise seems to be external (but generated by the Owon itself) to the actual signal path as it is greatly attenuated when using a fully shielded connection between the scope and the signal generator.

Yes, I used the supplied probes for this test. Set to 1X. Source resistance about 600R.  Signal 120mV pk-pk.

I forgot to say, the ONLY way to get a steady curve, is with trigger set to "HF".  If you are looking at DC forget it.

[bill2010]

Here are some more pictures to demonstrate the problem. Scope was away from any sources of interference and not connected to the mains.

The signal is 120mV pk-pk at 1KHz. This is not a "small" signal. I repeat : it is 120mV pk-pk.

The situation gets worse, much worse, as the signal gets weaker. I even wonder why bother have a "5mV" or a "10mV" scale when in reality they are so useless. Also the only way to get a clear picture is to "limit" the scope to 20MHz, in which case why advertise it as a "100MHz".


Picture 1 shows the amount of random noise :

Picture 2 shows the same but with the scope set to 20MHz.

Picture 3 shows the same but with the addition of "sample" rather than "peak". Had I put the screen persistence on (like old phosphor displays) it would also be much worse.

Did you use the supplied probes to connect to the signal source in picture 1?  If so, was the probe set to 10x or 1x?  I am trying to reproduce on my own SDS7102 as I believe that mine has a similar defect. Do you have a coaxial cable you can use?  On mine, the noise seems to be external (but generated by the Owon itself) to the actual signal path as it is greatly attenuated when using a fully shielded connection between the scope and the signal generator.

Yes, I used the supplied probes for this test. Set to 1X. Source resistance about 600R.  Signal 120mV pk-pk.

I forgot to say, the ONLY way to get a steady curve, is with trigger set to "HF".  If you are looking at DC forget it.

Replicated your setup.  1 KHz sine wave @ 120 mV pk-pk feed into channel 1 through the probe set on 1x.  Peak mode, full bandwidth.  Looks similar, but yours seem to have even more noise. Output impedance of my signal source is 50 Ohms, so that may account for some of the difference.  I've contacted my distributor for a replacement as well.

[AndrejaKo]

@firehopper

Newest firmware is 2.8.2 and as far as I can see, newest hardware revision is also 2.8.2. Do read the resources rf-loop posts on his forum. They're good if you want to be up to date.

There's is no need to hope for VGA out or not. The version with VGA out is called SDS7102V and is a bit more expensive than the version without.  From what I've seen (but I could be wrong), in some markets, the available version is with V and in some is without V. In some markets, both versions are available, so check which version exactly you're getting.

As far as I can see, all scopes from 2012 have Ethernet.

[rf-loop]

As far as I can see, all scopes from 2012 have Ethernet.

Ethernet (LAN) Yes, today it is factory default.
More than one year ago it was factory option.

VGA is factory option (need buy from factory with extra charge). If scope have V it have VGA.
(SDS8302 and 9302 have VGA + LAN as factory default without letter V.)
----------------------

Other:
This suspected  noise problem in some units need study more.
It looks like if it come from oscilloscope it do not pollute internally signal pathway. But if there is strong noise generated by oscilloscope itself it may connect externally to signal.
Later I will take one oscilloscope open and do some tests.

[bill2010]

Other:
This suspected  noise problem in some units need study more.
It looks like if it come from oscilloscope it do not pollute internally signal pathway. But if there is strong noise generated by oscilloscope itself it may connect externally to signal.
Later I will take one oscilloscope open and do some tests.

My testing definitely seems to indicate that it is generated by the oscilloscope, but does not pollute the internal signal path.  The noise appears to be somewhat broadband in nature ranging from around 70 - 100 MHz, which explains why when you apply the 20 MHz bandwidth limiter it reduces it.  It almost looks like some kind of ringing. You can see this in the attached, with the greatest amplitude right at the trigger point.  If you speed up the sweep it is very visible.  Will try to post another picture showing this later.

[rf-loop]

My testing definitely seems to indicate that it is generated by the oscilloscope, but does not pollute the internal signal path.  The noise appears to be somewhat broadband in nature ranging from around 70 - 100 MHz, which explains why when you apply the 20 MHz bandwidth limiter it reduces it.  It almost looks like some kind of ringing.

Yes I know this. Of course fast rising/falling edge include lot of frequency components.
Also there may be internal ringing due to these fast edges.

Your message Reply #880 at 10:20:51 AM
And there image: 20130130_264121.bmp
It tells to me "lot of" things.

Now is time to do lab work for this.  I will dismantle one SDS and look more deep things related to this.

Information of a potential problem now is enough here. Now need  starting  study. What is needed now is only to work in peace and arrange time slot for do it. I will also contact with Owon  after I have more than suspects and nearly empty hand. (I have a couple of suspects, it should be examined.) 
It is fully possible there is only some units if problem is example in some circuits for filtering these switch mode noise... perhaps lack of some component or "whatever"...

[akis]

My testing definitely seems to indicate that it is generated by the oscilloscope, but does not pollute the internal signal path.  The noise appears to be somewhat broadband in nature ranging from around 70 - 100 MHz, which explains why when you apply the 20 MHz bandwidth limiter it reduces it.  It almost looks like some kind of ringing.

Yes I know this. Of course fast rising/falling edge include lot of frequency components.
Also there may be internal ringing due to these fast edges.

Your message Reply #880 at 10:20:51 AM
And there image: 20130130_264121.bmp
It tells to me "lot of" things.

Now is time to do lab work for this.  I will dismantle one SDS and look more deep things related to this.

Information of a potential problem now is enough here. Now need  starting  study. What is needed now is only to work in peace and arrange time slot for do it. I will also contact with Owon  after I have more than suspects and nearly empty hand. (I have a couple of suspects, it should be examined.) 
It is fully possible there is only some units if problem is example in some circuits for filtering these switch mode noise... perhaps lack of some component or "whatever"...

What is your hunch? Is it maybe a manufacturing line problem (eg dirt on the tracks), a failing component, or a flaw in the design ?

[rf-loop]

What is your hunch? Is it maybe a manufacturing line problem (eg dirt on the tracks), a failing component, or a flaw in the design ?

Now is the time to examine first,  not tell the dozens of statements about what I believe or suspect. I will report back once I know something concrete. The study first, and then the speech is the wiser.

But it takes now some time.
Also soon now in China is  starting "new year" time and it make lot of delays in communication with Owon.

[bill2010]

My testing definitely seems to indicate that it is generated by the oscilloscope, but does not pollute the internal signal path.  The noise appears to be somewhat broadband in nature ranging from around 70 - 100 MHz, which explains why when you apply the 20 MHz bandwidth limiter it reduces it.  It almost looks like some kind of ringing.

Yes I know this. Of course fast rising/falling edge include lot of frequency components.
Also there may be internal ringing due to these fast edges.

Your message Reply #880 at 10:20:51 AM
And there image: 20130130_264121.bmp
It tells to me "lot of" things.

Now is time to do lab work for this.  I will dismantle one SDS and look more deep things related to this.

Information of a potential problem now is enough here. Now need  starting  study. What is needed now is only to work in peace and arrange time slot for do it. I will also contact with Owon  after I have more than suspects and nearly empty hand. (I have a couple of suspects, it should be examined.) 
It is fully possible there is only some units if problem is example in some circuits for filtering these switch mode noise... perhaps lack of some component or "whatever"...

What is your hunch? Is it maybe a manufacturing line problem (eg dirt on the tracks), a failing component, or a flaw in the design ?

akis, rf-loop -- my supplier pulled another scope from stock and was able to reproduce the noise problem, so it appears that this affecting more than just a few units.  Hopefully we can get a fix from Owon (perhaps through rf-loop's investigations), since the only alternative seems to be to return the unit for a refund.  That would be a real shame as it is hard to find a portable unit with similar specifications that is even close to the price of the SDS7102.
 

[GEuser]

My testing definitely seems to indicate that it is generated by the oscilloscope, but does not pollute the internal signal path.  The noise appears to be somewhat broadband in nature ranging from around 70 - 100 MHz, which explains why when you apply the 20 MHz bandwidth limiter it reduces it.  It almost looks like some kind of ringing.

Yes I know this. Of course fast rising/falling edge include lot of frequency components.
Also there may be internal ringing due to these fast edges.

Your message Reply #880 at 10:20:51 AM
And there image: 20130130_264121.bmp
It tells to me "lot of" things.

Now is time to do lab work for this.  I will dismantle one SDS and look more deep things related to this.

Information of a potential problem now is enough here. Now need  starting  study. What is needed now is only to work in peace and arrange time slot for do it. I will also contact with Owon  after I have more than suspects and nearly empty hand. (I have a couple of suspects, it should be examined.) 
It is fully possible there is only some units if problem is example in some circuits for filtering these switch mode noise... perhaps lack of some component or "whatever"...

What is your hunch? Is it maybe a manufacturing line problem (eg dirt on the tracks), a failing component, or a flaw in the design ?

akis, rf-loop -- my supplier pulled another scope from stock and was able to reproduce the noise problem, so it appears that this affecting more than just a few units. Hopefully we can get a fix from Owon (perhaps through rf-loop's investigations), since the only alternative seems to be to return the unit for a refund.  That would be a real shame as it is hard to find a portable unit with similar specifications that is even close to the price of the SDS7102.

Ha HA Ha HA  , that was funny! .
And even more so in brackets after 

[StubbornGreek]

Well now, you can add me to the list of 'noisy' scopes. Can anyone comment on the warranty service from Owon as of late? Best avenue of approach?

I have to say that for what it offered at the price point it offered, it's been an ok little device but this is an unreasonable defect that seems to be occurring to more than a few owners of the scope. I knew of its limitations when I purchased it and they were acceptable for what I paid/expected but one would expect for hardware to last a bit longer than this.

I'm cringing just thinking about what its going to be like getting a repair/replacement. I would even be willing to do the repair myself if they've already identified the problem and want to pass along the information/parts (obviously not going to happen - just part of my rant). 

[bill2010]

My testing definitely seems to indicate that it is generated by the oscilloscope, but does not pollute the internal signal path.  The noise appears to be somewhat broadband in nature ranging from around 70 - 100 MHz, which explains why when you apply the 20 MHz bandwidth limiter it reduces it.  It almost looks like some kind of ringing.

Yes I know this. Of course fast rising/falling edge include lot of frequency components.
Also there may be internal ringing due to these fast edges.

Your message Reply #880 at 10:20:51 AM
And there image: 20130130_264121.bmp
It tells to me "lot of" things.

Now is time to do lab work for this.  I will dismantle one SDS and look more deep things related to this.

Information of a potential problem now is enough here. Now need  starting  study. What is needed now is only to work in peace and arrange time slot for do it. I will also contact with Owon  after I have more than suspects and nearly empty hand. (I have a couple of suspects, it should be examined.) 
It is fully possible there is only some units if problem is example in some circuits for filtering these switch mode noise... perhaps lack of some component or "whatever"...

What is your hunch? Is it maybe a manufacturing line problem (eg dirt on the tracks), a failing component, or a flaw in the design ?

akis, rf-loop -- my supplier pulled another scope from stock and was able to reproduce the noise problem, so it appears that this affecting more than just a few units. Hopefully we can get a fix from Owon (perhaps through rf-loop's investigations), since the only alternative seems to be to return the unit for a refund.  That would be a real shame as it is hard to find a portable unit with similar specifications that is even close to the price of the SDS7102.

Ha HA Ha HA  , that was funny! .
And even more so in brackets after 

Well, Owon did produce a fix for the channel 1 noise issue that cropped up after a hardware change and rf-loop seems like a fairly knowledgeable guy who had the added incentive as a distributor to want his products to be competitive, so this statement didn't seem like nearly enough of a stretch for someone to be LOL over it...

Regardless, I bought the scope as an "experiment" given the cost and I am seriously thinking of just debugging and fixing it myself vs. returning it for a refund.  If it was a signal path issue I'd be more skeptical, but it looks like more of a general shielding / bypass problem which should be far easier to solve.

[rf-loop]

Here two images.
This new power supply in picture have  also SMD components on the solder side (not in picture)

These both need now more deep investigations.

(these have lot of small changes after older (many) versions.
PCB layout, routing and some componets are changed.
They have try do it better - but something is now wrong in my opinion.
It is even possible that secret is in some components specifications.

Parasitic osc......

EDIT/ADD:
-----------

Owon have holiday 7th to 17th Feb due to Spring Festiwal time.
(whole China is more or less on the holiday)  and after holiday, no one know who workers return back (hehe) from holiday and so or so but after holiday all services are heavy loaded, rush and queyed....


Disclaimer: This next all is now my unofficial opinion!
After some  talking with Owon about some findings (in scope internals) I believe there is coming some, perhaps very simple, solution but not just immediately. There is not just one problem... it is perhaps group of problems perhaps related to some components parameters and perhaps also PCB design  (This power supply and adapter card is very new version if looks also used components)   With very fast rising edges there need handle tens and hudreds of megaherz freq components in single simple 50kHz switch mode regulator.  Now something  have changed and...

This is only example, NOT related to this case.
I remember once in history some fun (and expensive)... there was one case... all works in lab very nice... in proto.
One proto PCB inductor was take from "lab test and evaluating component box" but it have wanted inductance etc.. all was fully tested and exactly ok.
Then later someone buy these "looks like" inductors, same inductance and so on...
All was then wrong... extreme heat... components burn... it was like spread spectrum VHF transmitter for EW... one simple power supply.

Before solution and repair it need take extreme care about probing methods so that this partially common mode noise do not pollute signal under test or it do not disturb device under test. Inside scope it do not (significantly) pollute internal signal pathway from BNC input to ADC.
--------------------


Also note: If you have OLD version service manual or other OLD material how to open oscilloscope.
It have changed. If you try pull power button out before opening case...  DO NOT. It do not come out without damage.
Push it fully inside same time carefully opening bac cover.

[bill2010]

Owon have holiday 7th to 17th Feb due to Spring Festiwal time.
(whole China is more or less on the holiday)  and after holiday, no one know who workers return back (hehe) from holiday and so or so but after holiday all services are heavy loaded, rush and queyed....


Disclaimer: This next all is now my unofficial opinion!
After some  talking with Owon about some findings (in scope internals) I believe there is coming some, perhaps very simple, solution but not just immediately. There is not just one problem... it is perhaps group of problems perhaps related to some components parameters and perhaps also PCB design  (This power supply and adapter card is very new version if looks also used components)   With very fast rising edges there need handle tens and hudreds of megaherz freq components in single simple 50kHz switch mode regulator.  Now something  have changed and...

Thanks for putting some time in on this and working with your contacts at Owon. Seems like a reasonable line of thinking to explore. One thing I should add is that the noise is the same when running off the line or battery, so it would have to be related to the power supply components that are common to both modes of operation.