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

[rf-loop]

Has anyone used the trigger holdoff on the SDS7102? I tried to use it today, but it seems that it only increments by 100ns at a time. I turned the knob for a minute and only got to 1ms. How in the world am I supposed to get to something like 100ms without turning the knob for an hour and a half? It's supposed to go up to 10 seconds.

Around 30 second.
If frequently need long holdoff times it is wise to store one (or two) setup profiles for this.
Example one for 10ms depending what area of holdoff need use in work.

Of course this is not best possible solution for working ergonomy.  This "multipurpose" knob speed up system is not best possible. Fast turning increments may be better selected for fast adjusting.

 
Here example  from expensive (old) professional class Tektronix.

First in menu there can find holdoff. It can adjust just from 0-100 on its own undimensioned scale. Nothing on the screen  tell what is holdoff time. Take user manual. It do not help. Exept that it tell it can change using control knob between 0-100. It is really "clever" user manual.

But user manual give one wink where can find more information.
Ok after time this very clear most deep imformation can find. This is most detailed information what can find.  Ok...  these are "limits".  It is fast to turn from 0-100.  How long time you need with HP48GX to solve what is holdoff time if you turn knob to example value 57 and you have 1us/div   but in Tektronix case it is of course M 1us (all know what is M 1us) horizontal speed.  (other question is that who need know exact holdoff time in practice)

But here in picture most deep information what can get from Tektronix TDS5xxA series.

Made in China?:

[rf-loop]

There is now available unofficial DIY version (it is preliminary version 2b) for this "Noisy GND" issue.
First published modify for Adapter card (PCB Version: SDS7102_CON_3.2).

Owon is now working for final version for production asap.
Of course there may be better result becouse they may (if need) also change PCB layout for example better handling for parasitic inductances and capasitances.

This preliminary version 2b modification include:

Adapter board modification:

- TFT backlight LED driver SMPS   (it was very bad and also this RF EMI noise radiate all around via to TFT panel)
-   +5.5V  SMPS (it was bad)
-   +3.3V  SMPS (it was slightly bad)

(Adapter board modification need good tools for SMD rework! Do not even try with cheap crap "noname" soldering(desoldering tools. It can only destroy components and board itself. It need real fast thermal transfer power without adjustable "overheating" iron. Also preheating helps.)

PSU board modification (DIY version 2b coming soon after some stress test):
(also version 2b is not yet enough good in situation where battery installed AND battery is charging  same time when use oscilloscope, still 8.4V produce too much RF noise.)

-  +8.4V main SMPS  (it was bad without battery and very very bad if battery is charging + oscilloscope in use))
-  -7.6V SMPS  (it was very bad) 

PSU board is much more easy for DIY.  It is single side board. (mixed with SMD and through-hole)

[AndrejaKo]

I just stumbled upon a new download on Owon's site. It's called Labview opearation cases for SDS series and is under driver download.

It contains the USB driver .dlls, some Labview files for interfacing scopes with it (I don't use Labview, so I don't know how useful they are) and a new and more detailed version of OWON Oscilloscope PC Guidance Manual (it's 1.3 now). At first look, it seems much more detailed than the previous version 1.1 which is currently circulating the Internet.  There are examples for USB and serial interfacing and a few mentions of LAN. Anyway, it looks like the main focus here was USB.

Hopefully this will allow creation of better PC side software than what we currently have.

As a side-note, the guidance manual file is actually called "application notes_en.doc", but is referred to as "OWON Oscilloscope PC Guidance Manual.doc" by the readme file.

[bill2010]

There is now available unofficial DIY version (it is preliminary version 2b) for this "Noisy GND" issue.
First published modify for Adapter card (PCB Version: SDS7102_CON_3.2).

Owon is now working for final version for production asap.
Of course there may be better result becouse they may (if need) also change PCB layout for example better handling for parasitic inductances and capasitances.

This preliminary version 2b modification include:

Adapter board modification:

- TFT backlight LED driver SMPS   (it was very bad and also this RF EMI noise radiate all around via to TFT panel)
-   +5.5V  SMPS (it was bad)
-   +3.3V  SMPS (it was slightly bad)

(Adapter board modification need good tools for SMD rework! Do not even try with cheap crap "noname" soldering(desoldering tools. It can only destroy components and board itself. It need real fast thermal transfer power without adjustable "overheating" iron. Also preheating helps.)

PSU board modification (DIY version 2b coming soon after some stress test):
(also version 2b is not yet enough good in situation where battery installed AND battery is charging  same time when use oscilloscope, still 8.4V produce too much RF noise.)

-  +8.4V main SMPS  (it was bad without battery and very very bad if battery is charging + oscilloscope in use))
-  -7.6V SMPS  (it was very bad) 

PSU board is much more easy for DIY.  It is single side board. (mixed with SMD and through-hole)

Thanks again for all the investigation. So what's the plan for Owon to fix scopes under warranty? What should I be telling my distributor?  I could DIY, but would rather have the redesigned boards from the factory and not risk voiding my warranty.

[akis]

I also do not have the necessary tools or experience to carry out SMD repairs. It would be easier to receive a replacement card from Owon. Do we have any news from them?

[AndrejaKo]

I'm also interested in this!

Is there any response from Owon? Is there even information about no response? Any new hardware revisions?

[chromesphere]

Hi guys,
I'm thinking about getting my first scope (the owon 6062) and I was about to pull the trigger on it, but I just wanted to ask if anyone knows if the 6062 suffers from the same ground noise problem as the 7102?  I'm very much a beginner with electronics but ground noise sounds like something serious and best to avoid?  Im also not very confident reworking smd
Thanks for your help!
Paul

[akis]

In my very simple opinion and comparing with an old analogue scope, the OWON SDS 7102V is incapable of displaying a trace of a small amplitude, it even has troubles with +/-100mV - locking on will be problematic and there will be noise spikes almost obliterating the signal.

There may be a "fix" on the way, not least due to the contribution of a few members of this board, but it will require DIY skills even if OWON decide to ship replacement "adapter" boards.

[chromesphere]

Thanks Akis,  so does the problem include the 6062 as well?
Cheers,
Paul

[Hydrawerk]

Well, wait a few months, or if you want your scope today, go and buy a cheap GW Instek or Rigol. They will not have the 2×10MB acquisition memory, but who cares. 

[onlooker]

2 comments/questions

1). I do not remember  seeing it mentioned, but, is the Owon "noisy GND" issue similar in nature as Rigol's as mentioned somewhere in eevblog. Maybe all scopes of the same class have this problem? Can anyone verify?

2). If the DUT is some distance away from the Owon, the noise is way less, right? If true, I do not really see it is a big problem.

[AndrejaKo]

My speculation could be wrong, but I think that the issue currently discussed would affect most of the SDS series, since it appears to be caused by components that would be most likely common to the whole series.

Since the problem is caused by EM radiation from components, it would be reasonable to expect the "noise" to drop with distance, but that leaves the other side of the issue:

 How are we going to measure precisely what's happening? For long distance measurement we'd need probes with long cables and the cables themselves are going to affect measurement. I'm looking right now at manual for Testec's TT-LF series probes and for example the probe with 1.2 m cable has 150 MHz bandwidth, the 1.6 m cable probe has 130 MHz bandwidth and the 2.0 m probe has 100 MHz bandwidth.

The other option would be to go through the trouble of placing dedicated BNC connectors or probe testpoints on the PCB in order to avoid the radiation, but I suspect that such measures could easily be overkill for users of scope such as SDS7102.

This leaves us with the springy ground lead accessory which ships with the scope probes, but it's a bit awkward to use and isn't all that durable. On the other hand, they are relatively easy to make.

So perhaps it would be best to wait until next revision ships and get a unit of that revision.

[marmad]

1). I do not remember  seeing it mentioned, but, is the Owon "noisy GND" issue similar in nature as Rigol's as mentioned somewhere in eevblog. Maybe all scopes of the same class have this problem? Can anyone verify?

I think you're referring to a thread that was started because of a particular problem that one member (A Hellene) was having with his Rigol DS1052E. But I think there are tens (or hundreds) of thousands of Rigol users without significant noise problems.

[kuson]

NEWBIE and POTENTIAL OWON CUSTOMER

Hello -- I'm very interested in the Owon SDS7102 and was about to buy it on Ebay, and after seeing the problem, I decided to wait until the problem is fixed (Thanks to everyone in the forum especially the "Improvers and Tinkerers".   

A couple of questions:
1. Is there a list of outstanding issues posted somewhere for tracking and
2. For each issue, what kind of functions will it adversely affect, or is it an overall showstopper?

Something on the lines of a vehicle -- I want to buy a vehicle, but if there are 2 issues  P1) Leaking Roof P2) Problematic 4WD mode,  I could understand if I wanted to buy the vehicle NOW, I could P1) Avoid driving on rainy season and P2) Do not use 4WD mode.  That, or should I wait for the complete fixes from manufacturer.


Thanks in advance.

[akis]

Well I am no expert oscilloscope reviewer, and this is what I have found out so far:

1) Low voltage signals are affected by all the noise sources inside the scope as mentioned in the pages above, anything below +/-0.5V pk is problematic.

2) MATHS and MEASUREMENT seems to work from what is displayed on the screen. It seems that the scope has two memory areas: one is what it captures (raw data) and he second memory area is what it displays on the screen. Maths and measurements are derived from the screen data, it seems, which is a very bad idea. Quick example: take two sinusoidal signals, one is 50V pk, the other is 48V pk. The difference is 2V pk, and this is what the MATHS will display. However this 2V curve is almost a flat line when you are also displaying a 50V, ie the division is 10V. There is nothing you can do to "zoom in" on the MATHs curve, because this will take the other two curves out of range and you will get garbage on the screen.

3) fan makes constant and annoying noise - wondering if it could be variable speed like on my laptop :-)

[marmad]

Maths and measurements are derived from the screen data, it seems, which is a very bad idea.

Perhaps - but from what I understand, it's standard operating procedure for virtually all low to mid-range DSOs (e.g. the Agilent InfiniVision X series does the same).

[akis]

The example I have above was not theoretical, I was really trying to measure the voltage drop across a known resistor and it was 1-2V out of 60V. It is very hard to read the MATH curve under those conditions.

In my opinion there should be an option to hide ch1 and ch2 when in Maths so the screen does not get cluttered, and also to measure and zoom on the Maths curve, which would necessitate new software I suppose.

[rf-loop]

Well I am no expert oscilloscope reviewer, and this is what I have found out so far:

1) Low voltage signals are affected by all the noise sources inside the scope as mentioned in the pages above,

anything below +/-0.5V pk is problematic.

This problem is in your Owon becouse you are one who have this most bad noisy GND issue.

Here is some pictures from factory new Owons and mesurments made so that it really show worst. Oscilloscope run with mains power (not with battery)
Probe connected to signal with its own GND wire with alligator. Oscilloscope is in normal enviroment and signal source is also connected to mains power.
Scope and signal source are so that common mode noise picked up from equipments itself and from enviroment is as bad as it is typically and also connected to laptop so that situation is as bad as can be if think what kind of "antenna" this whole test setup is.
20mVpp do not look very problematic. There is connection with normal GND wire and then with short GND spring what is very normal practice if people have any real experience about measurements and for avoid problems.


(also it is very good to look this  youtube from Dave where he hanging around  Rigol DS2202 and common mode noise.

Look what happend there related to Rigol TFT. (yes Owon do same)

Look carefully. There is some very "fun" details.
Specially starting after 13.30.  (TFT)
Also 17.38 is short very interesting point. Coaxial cable near Rigol TFT.


But these pictures attached show how it is here and related problems under +/-0.5Vpp signals as you tell. My opinion is that it is not very problematic. And methods and settings I use for these are just these "entry level" practices where all conncted as "magpie nest".
Some reader may see more and some may see less problems.

2) MATHS and MEASUREMENT seems to work from what is displayed on the screen. It seems that the scope has two

memory areas: one is what it captures (raw data) and he second memory area is what it displays on the screen.

Maths and measurements are derived from the screen data, it seems, which is a very bad idea. Quick example: take

two sinusoidal signals, one is 50V pk, the other is 48V pk. The difference is 2V pk, and this is what the MATHS

will display. However this 2V curve is almost a flat line when you are also displaying a 50V, ie the division is

10V. There is nothing you can do to "zoom in" on the MATHs curve, because this will take the other two curves out

of range and you will get garbage on the screen.

There is 2pcs 8bit (*1) ADC's
((*1)   and then note all things in practice what reduce conversion result quality)

It is good to understand this simple math.

Data for math is 8bit. What happend if in one point result is 120-120 and then next place 122 - 120. 

Difference between these is 2. There is 4 pixels vertically. Yes you can zoom it.  Where is now more information if you have nearly half div for one resolution step. Zoom can not add resolution!
For my eyes some times zoom factor 2 or 4 is enough for better visibility if difference is quite small.  (btw. how it differ in vertical direction from raw vector data in practice?)
In picture I have zoomed by 10x  (inputs 2V/div and math 200mV/div)

It is different case if we have more resolution than 8bit and then we have lot of less resolution in display. Now our display resolution is more (double) than data resolution.

Look these example pictures.

3) fan makes constant and annoying noise - wondering if it could be variable speed like on my laptop :-)

Do you know what this price class oscilloscope have more silent fan? If there is fan.
I know one. Siglent but difference is very marginal.
But in other hand, time ago Owon have some extra problems with fans. Many fans was assembled wrong and / or damaged in assembly in factory and after short time or just new from package make very very bad noise.
These fans need change.



Pictures  (picture names give some explanation also)

First 2 pictures for this Math "zoom"
and then some pictures related to: "anything below +/-0.5V pk is problematic."
In some pictures these is Acquire mode PEAK. If filename read acqN there is Normal mode.

[akis]

1) I do not have those two extra options on my maths functions - vertical position and V/div. I updated to the latest software 4-5 months ago. I am glad that they added that feature though, and I hope it is software and not hardware changes, so I can get it too.

2) Both Maths and Measurement display garbage if the signal exceeds the screen size, ie if you have selected a smaller V/div so that CH1 or CH2 (or both) curves do not fit. I understand what you mean about the ADC being 8 bit and the screen vertical resolution being almost twice as large.

3) Since these changes seem to be software related, could you please ask them to add a "Maths measurement" function too? Currently you can measure CH1 and CH2 but not the resulting Maths curve, and it is easier to read a number off the screen than try to count divisions etc.

4) My Owon may be more "noisy" than yours because as I have said earlier on in this thread it has problems with anything less than 0.5V pk. I was the one that started this discussion if you recall. Are you using your "DIY fixed" Owon ? Can we try some of the fixes you and others have attempted to reduce the noise or should we wait for an official reply from Owon? Last you had said it was Chinese New Year and the Owon factory was closed.

[akis]

Maths and measurements are derived from the screen data, it seems, which is a very bad idea.

Perhaps - but from what I understand, it's standard operating procedure for virtually all low to mid-range DSOs (e.g. the Agilent InfiniVision X series does the same).

Based on another comment on another thread where I was informed a the 500 UK pound "Agilent" and "Tektronix" were simply re-badged cheap Chinese scopes, how do we know which "Agilent" is actually the real thing?

[rf-loop]

4) My Owon may be more "noisy" than yours because as I have said earlier on in this thread it has problems with anything less than 0.5V pk. I was the one that started this discussion if you recall. Are you using your "DIY fixed" Owon ? Can we try some of the fixes you and others have attempted to reduce the noise or should we wait for an official reply from Owon? Last you had said it was Chinese New Year and the Owon factory was closed.

Yes, this is case. You have more noise as told earlier. Of course I remember this becouse after this start heavy hunting to find problems and solution.

I take two random factory new owon from original new factory package just to my table and look they are just same related to this noise issue. Then I randomly take other one and do these tiny tests and images.

This is NOT from my prototype DIY.
From my prototype DIY I have made documentation for Owon. (This proto is totally disassembled on my workshop table for more deep experimentals)


I have talked with OWON after the holiday period.

Owon answer  about "DIY" was that they accept it IF I do it  to my customers, in this case it do not void warranty (as described in other place).

Owon have also told (time ago) that they transfer this to production asap (they take something from me and something how they do. What is asap: They need do they own tests and becouse they need also made  new PCB's and so they can also do it littlebit different (I hope they do better than my DIY). In mass production it is not so fast to make changes you know. Design, protoboards, tests, possible changes, protoboards, tests... iteration as many times as need.. then order PCB's and components to production line..  it take absolute minimum least one month. In practice perhaps 2. 

(It is fun, I have tested very very old Owon and there is not at all this extra "noisy GND" problem.
And more fun is that circuits are very very same.)

[marmad]

Based on another comment on another thread where I was informed a the 500 UK pound "Agilent" and "Tektronix" were simply re-badged cheap Chinese scopes, how do we know which "Agilent" is actually the real thing?

There is tons of information around about the completely in-house designed Agilent InfiniiVision X series - including Dave's reviews and teardowns. Suffice it to say, if you spent $15,825 for the top of the line Agilent MSOX3104A 1 GHz Oscilloscope with 4 Analog and 16 Digital Channels - you would still be making measurements on the displayed data - and not the actual sample data. I believe (higher-end?) Tektronix DSOs (and perhaps Hameg?) make measurements in sample memory.

[rf-loop]

2) Both Maths and Measurement display garbage if the signal exceeds the screen size, ie if you have selected a smaller V/div so that CH1 or CH2 (or both) curves do not fit.

Yes becouse in Owon, ADC whole vertical data range is nearly equal  to display 10 divs (500 vs 512)

In my old tek workhorse it display 8 div and if I remember right (not so exactly) it show something like 200 ADC values on the screen. So signal can go maximum 56 over screen border.

About this measurement fronm display. What make it less accurate as long as display resolution is much more than raw data.

[AndrejaKo]

1) I do not have those two extra options on my maths functions - vertical position and V/div. I updated to the latest software 4-5 months ago. I am glad that they added that feature though, and I hope it is software and not hardware changes, so I can get it too.

Currently newest software versions are 2.8.2 from 2012-12-30 and 2.8.1.6 from 2012-11-05, depending on serial number. Also it appears that there was firmware version 2.8.1 which was removed from their site for some reason.

If you can, try the newest firmware and report back. I think I didn't see (or at least notice) the math divisions option on my scope with 2.6.2 hardware and 2.8.2 firmware. I could be wrong but I can't check since the scope is currently being repaired.

[rf-loop]

2.8.3 is newest. (and this is first verion where this scale and position is added for math menu )

Perhaps soon it come also available for free download.