Improvements in Owon SDS7102 HW/FW and some tests

[aghp]

It is nice to see that Owon continue and highly focuses to develop HW. Signal quality is number one. Now it is even better than before although it was very good. (inside this price class)
I'm satisfied they do not hurry hurry run to add many more features to user interface. They focus to signal quality (first)  In my opinion They have selected right road.



Old rf-loop made table. I have add new measured data. (marked with color)
(measured with cursors, value is in () )  first value is oscilloscope automatic measured p-p value.

Scope settings. Normal acquire, 1M sampling buffer, 50us/div

Clear improvement! Look these 5 - 100mV/div values.
(resolution is too rough for measurement higher V/div values. (8 bit ADC and also display resolution)

Also I can tell that front end run now markable cooler.

Here you may find something more (all tests are not yet ready, just started)

-aghp

Note: HW V2.6.2   it means equipment version (including FW) It is functionally same as 2.5.1. If look with user eyes.
 
(It is maybe different if look  internal hardware interface)
Also it looks very stable.

[aghp]

Also this FFT tell something about noise level.
(there is infinite persistence for show peak levels of noise floor.
Scope settings 50mV/div   (50 ohm Tektronix feed thru terminator in scope input of course)
FFT zoom 10x  and upper cursor is 0db (maximum signal level but of cousrse 100MHz is littlebit attenuated.)

Signal level adjusted so that same level with 1MHz give just 10div p-p and with this level in signal gen frequency 100MHz (signal generator maximum error in flatness in this area is under +- 0.2dB)

Then next picture. Signal level from generator dropped 60dB
(in this case signal level is 172uV rms (-62.3dBm)

Not bad result!

[0xdeadbeef]

Well, some remarks...

Firstly it's rather annoying for a customer when Owon releases one "better" hardware after the other and customers of the now obsolete hw revisions feel like they're setting on a piece of outdated crap, wondering if they'll e.g. ever get a firmware update again.

Secondly this evolutionary development approach doesn't sound very professional. Usually you develop something first and then bring it to the market. Changing the hardware over and over again means nothing less than selling alpha and beta hardware to the customer in the first place. This might be an interesting financing model for Owon, but as said before, this is a bit frustrating for the customers who have an outdated hardware in no time.

Thirdly, it's not like the software was in a state where they should focus on the hardware. The hardware was always the best part about the Owons. So they should better invest a few Yuans in the SW finally before they scare away the last potential customer with the low end GUI, the lack of features and the crappy PC software. As long as you can't even measure a pulse in the 100µs range more exactly than 1µs at 1GS/s, who cares about improving their noise level a little bit more?

[muvideo]

Well, some remarks...

wondering if they'll e.g. ever get a firmware update again.

So they should better invest a few Yuans in the SW

As a (happy so far) Owon customer, you nailed pefectly my thought.
I hope they will follow old customers with firmware update and that they
can invest in more firmware features, but also battery management can be
improved (the unit discharges batteries when off    ).
Seeing the difference (thanks aghp) they lowered slightly the BW, aliasing problems?
As for the lower noise on the lower V/div settings, well this is welcome... to
future customers, to be honest this pisses me off a little bit

Fabio.

[aghp]

Well, some remarks...

Firstly it's rather annoying for a customer when Owon releases one "better" hardware after the other and customers of the now obsolete hw revisions feel like they're setting on a piece of outdated crap, wondering if they'll e.g. ever get a firmware update again.

Secondly this evolutionary development approach doesn't sound very professional. Usually you develop something first and then bring it to the market. Changing the hardware over and over again means nothing less than selling alpha and beta hardware to the customer in the first place. This might be an interesting financing model for Owon, but as said before, this is a bit frustrating for the customers who have an outdated hardware in no time.

Thirdly, it's not like the software was in a state where they should focus on the hardware. The hardware was always the best part about the Owons. So they should better invest a few Yuans in the SW finally before they scare away the last potential customer with the low end GUI, the lack of features and the crappy PC software. As long as you can't even measure a pulse in the 100µs range more exactly than 1µs at 1GS/s, who cares about improving their noise level a little bit more?

From start of SDS series  all versions meet specifications.

And yes, I know. Some time ago I buy TV. New nice TV and all ok...
Next week I go other reason to same store. There was new TV... lower price and BETTER and just same manufacturer (but slightly different model) 

In history I remember some TV. (yes tube)
Some months before my friend buy good TV and later I buy same model, exactly.
Oh no... there was (1980) this evolution... my unit was different.
First time it fails repaired and ok.. it fails later agen... (and agen and agen)
My friend TV, no any fail..  and my fails many times.

Yes becouse we both are electronic peoples we open these... oh well... nearly all was different. (outside exactly same, same model... also inside first looks like same but lot of different components and some sub units different but in this case, later version was more bad... not better)



Why it is bad if Owon do it but ok if  AR (Agilent-Rigol), Hantek, Tektronix, Mercedes Benz, Fluke, Keithley, Rockwell-Collins and Rohde&Schwartz  etc do it.
Have you seen these nearly hand made Tektronix "things" what are nearly every unit different...

I have example looked many Tektronix 2445 and 2465. Just same model what outside can look, model same, specs same. Inside lot of differencies. Rev this and Rev that.   

They are not professionals???

I have looked these things more than 35-40 years. I do not see any "issue" in this developing. Even every manufacturing lot may be different in many kind of manufacturing!

(also I have been working in manufacturing industry (not for this main product but with machines and old times "robots" in production so that they work), do you know we together product developing engineers and doctors  did tens of changes sometime nearly weekly or monthly without any word to customers level. How I know it... becouse many times changes need different program or other adjustments in machines. )

Do you know your car? What version and what subversion subversion subversion it is and what product lot... all different but model name is just same. My version have different many things, than my neigbourg have in just exatly same model and year but base number is just different... Some weeks different. This evolution is everywhere. Do I know if I buy today Agilent analyzer it is not just tomorrow improved (or opposite) somehow. I go buy to spare parts, I need my base number so that they find if I have this very good air condition blower motor or if I have just this piece of crap motor.  (yes I'm happy I have this better... but how I can buy car becouse I do not know what version it really is and if next week is better? )

Manufacturers do always changes, one day some find that too many warranty time fails.. need change something and continue.

Owon have published specifications and description and from start to this day all is same and all they meet these specifications.

Did you know that if I do deep tests, every single unit is different!
Different noise, different BW, different vertical accuracy etc...

But Owon new small "improvement" means also littlebit less BW (what is also very good!)

All who have buyed Owon, they have very good signal quality oscilloscope in this price class.
Do you know if they have less noise if they have buyed Hantek or Rigol.


It is difficult... evolution and revolution is difficult:

I can not buy camera becouse if I buy now, tomorrow I can get somehow better?
This is reason why I do not have any pictures. But my friend take extremely good pictures with 10 years old camera and he have lot of nice pictures.... and he is happy with his pictures (and with his camera)

[0xdeadbeef]

Sure, also the big brands have a new revision here and then. But not every 3 months. And most of the times only since they need to 2nd source some component, not because they still develop their hardware in the field. That's why devices of the bigger brands usually report their hardware revision. For Owon, they tell you the production week and looking at the remarks for the firmware release you get the impression that they build other scopes every 2nd week since the update is only for a new production dates, but not nearly for all.

And still I wished they would finally invest a little money in a few decent SW developers to improve the GUI and features of the
scope and to completely rewrite that excuse for a PC software. Well really, look at that PC software. If they spent more than 1k€ for developing this, they have been cheated. It's ridiculous.

[SeanB]

Sounds like KIC, or, as it was better known in the repair trade Kak in Case. You could buy 2 TV sets, made on the same line on the same day, with serial numbers consecutive, and the boards inside would be totally different, with even the identical remote control units being not interchangeable.

[aghp]

Still waiting someone make comparable tests with example Rigol...


Owon SDS7102 - Tektronix TDS520A
Noise

Setup is made so that scopes are as close comparable as possible.
"apples to apples"

Note: Tektronix 100MHz BW limit is more low than Owon full BW (100MHz)
Signal from HP function genrator using 50ohm feed thru terminator. (Tek have built in 50ohm but here used same as with Owon)


Owon  set here  10k memory and Tek 5k memory and "fit to window"
With this setting they are more comparable becouse then Owon have 10k 20div Tektronix 5k 10div.

Both normal sample, realtime, no any averaging... just as single shot realtime.


Owon 100MHz full BW just as it is (-3dB far over 150MHz)



Tektronix with 100MHz rejected BW



Tektronix with full BW







Noise without signal.

Oscilloscope 1M input with 50ohm terminator.




(cursors show Owon OLD version noise peak level.)


NOTE: Owon have used 10M memory. Tektronix  have only tiny sampling memory so also trace looks different.
(if drop Owon memory it looks same but, these tests was done different time and I did not use Owon small memory becouse with it noise looks "too good".




100MHz BW




Full BW


Need realize that Tektronix is very old but totally different price class oscilloscope and still  good equipment for many kind of use. In second hand markets still far over Owon price if condition is somehow demonstrably good condition.

My opinion is that Owon result in this tiny test is very good.

Still Owon have noise, of course (all have), and it also affect to triggering stability. But also so do this Tektronix.

[aghp]

Also here new 2.6 HW bandwidth.

(these small waves in shape are propaply from impedance mischmatch (feed thru terminator in scope input is not perfect match))



Risetime:



100mV/div


Same with infinite persist (for show possible corners wobbling (one form of aliasing) and also trigger stability!

Test signal risetime is around 0.8ns.
Owon risetime with 200mV/div or higher is typically littlebit better.


Then littlebit more with face to face with Tektronix 520A.

Scopes get exactly same signal.

Settings so that highly comparable.
Tektronix 100MHz BW becouse Owon is 100MHz oscilloscope.


















Looked more deep this triangle corner becouse it is real "case" in real life. (Equipment what need service).  Just after voltage range change over 100mV p-p level there appear this ringing. (proportionally most high ringing with this level).

It is problem in this function generator - need service adjust (or repair) using oscilloscope. 
This is just borderline for Owon. (note that Tek is set for 100M BW)) If oscilloscope have not this signal quality or better this work need forget.


















Owon front end signal quality is amazing good as we look 0.5kEur price class. Maybe becouse they have designed it themselves, not try make cheap copy from some other.
If do copy it also may mean that copycat do not understand fully this circuit what he is copying. Then they may do some changes for more cheap without understanding how it affect and then we see "Siglent" type result.

If design and test in own lab, it need understand something and result may be this what Owon show here.



Why this noise is important.
Bad noisy oscilloscope (or example spectrum analyzer) is just piece of bullshit independent of how many features it have in UI. 
But then it is also pity that if there is good signal quality oscilloscope have poor UI. This is somehow "waste of resources".  But poor UI do not still destroy signal. Fine and rich UI do not rescue if signal is grap.

Of course it mean how reliable and with what accuracy you can tell something about signal under test.  (if you take example single shot, how you know what is oscilloscope made and what is really in signal under test - I think it is good to ask)
Then, noise affect also to Trigger and this is very important. Part of trigger instability come just from noise - naturally.



It is also good to imagine that these pictures are still pictures.
In practice with live screen human eye-brain system make some effects and it may mostly look better.  This effect is extremely clear if you run video with fast movements and then you stop it. Did you really know from how bad image your eye-brain complex produce image what you "see". We have high amount of image processing. This help also with oscilloscope in many cases.
If you have long experience with analog oscilloscope it may need some time before you learn how to look digital oscilloscope screen and what you look and what you "filter" away (you see but with intelligence you do not care)



Add:
Here one image what may tell something how BW affect.
It is just with same signal as before but now with full 500MHz front end BW. This is why I use 100MHz BW in Tek (becouse also Owon is 100MHz. But Owon BW is higher than Tek 100MHz filter. (Owon BW is around over 50% higher)

[electroguy]

hi,

those screenshots you have look like a really clear signal!

i have SDS7102V (SDS710212191xx) firmware 2.7.0
i try to measure the 5V probe square wave on the PROBE COMP pin.
Can someone please compare to their SDS7102 and tell me if this is normal?
There seems to be a lot of noise if i zoom in... why?
first picture is with nothing connected at all, which to me looks good, nothing wrong with it.

But the others are zoom in of the 5V square wave from the PROBE COMP signal. The last photo is the GND of the probe connected to the tip. Something seems wrong? 800mV noise? maybe the grounding is not connected inside CRO?

I tried both inputs, i tried different probes, same result

ideas?

Thanks.

[electroguy]

is there anything going on inside this scope at about 14.5khz (69us) ? backlighting? anything else? just wondering why it is happening at exactly 69us every single time. weird.

[arekm]

On mine

[electroguy]

Thanks.
I found that if i try grounding the tip using the ground terminal near the tip of the probe, it is much better than using the aligator clip connected to the ground half way down the probe. Maybe the probe is faulty? But both probes do the same thing, so i don't know...
Something strange is it also picks up the signal and the noise when the probe is about 10mm away from the pin (without touching the pin). Is that normal? You can clearly see the 14.5khz spikes.

[aghp]

On mine

This looks like well known CH1 noise issue in units what are first generation "improved HW"  late 2011.
It was some first manufacturing lots after Owon change original front end mechanical construction.
There is simple solution for this. There is also available instructions for this if you littlebit use search in forum.
Also Owon send needed small part to you if problem is this known and if you have enough experience and sklls to do it. (it need disassemble whole unit before this part can change and it need do carefully so that do not damage anything inside. But overall it is simple work. No need soldering anything.

What is this unit manufacture lot? (leave 3 last numbers out from whole serial number)

in this your picture:

[arekm]

It's sds71021149, so it's buggy right?

[aghp]

It's sds71021149, so it's buggy right?

Yes together your image and this 1149..
There is small change need in harware (mechanical).
There need cut one noise road from Z-plate to CH1 input through GND.
Z-plate itself is very noisy.. it collects lot of EMI inside machine and there are noise current trawelling through it.
(one metal pillar from Z-plate go to inside CH1 analog part shield and there it pollute very sensitive area of GND) This metal pillar need change to nylon pillar. It solve this problem.

(later this whole construction have changed very different. )

You can also find detailed Illustrated instructions how to do it.

You can ask Owon this small part, I think they will ship it to you if you can do it yourself... or you can negotiete with Owon some other solution.

[asbokid]

What is the latest news on the Owon SDS7102 and its LAN protocol for communicating trace data to a PC?

In the past Owon was generous and supportive of open source code. It released the protocol specifications for the USB interface of its PDS scope family.
The Owon USB protocol supports two types of trace data - vector-based and bitmap.  Vector-based is excellent for scientific analysis, especially long-term trends.

Is the 7102 trace data format and its IP protocol publicly documented for us, please?

This is important for people who need to use Owon oscilloscopes with Unix/Linux PCs.

cheers, asbokid

[aghp]

Here is latest available version. (V1.3)
Not good but better than nothing and more new than original V1.0 what was first version after PDS series.

[asbokid]

Thank you, aghp.

So the scope trace "datagram" format for the Owons is documented, and the USB specs (endpoints, etc) are there. There's a brief description of the libusb functions used: from the libusb-0.1 synchronous API.    Nothing as yet, though, documenting the TCP/IP network protocol(s) used by these scopes.  e.g. the network ports used, handshaking commands, data transfer commands, etc.   It would make sense to find re-use of the USB data transfer protocol in the firmware?   Would be good to know though.

Are owners generally happy with this generation of scope from Owon?

cheers, a

[arekm]

Could anyone check what probes are attached to SDS7102?

I've got replacement osciloscope with "T5060 60MHz/5MHz 16pF/115pF" probes. I guess they send me wrong probes since this is 100MHz osciloscope.

Edit: SDS7102 comes with T5100 probes (100MHz).

[aghp]

Could anyone check what probes are attached to SDS7102?

I've got replacement osciloscope with "T5060 60MHz/5MHz 16pF/115pF" probes. I guess they send me wrong probes since this is 100MHz osciloscope.

Edit: SDS7102 comes with T5100 probes (100MHz).

With SDS7102 T5100 probes or equal. 100MHz of course. (also these have low frequency BW with 1x setting, of course. It is typical for all probes what ahve 1x/10x selection)

T5060 probes are for SDS6062. So if you have get SDS7102 with T5060 probes they are wrong.

[jstan]

Hello

I just bought Owon sds 7102V version: 2.6.2. It's my first scope and i wanted to check it.So I took it to work to test it with agilent 33220A generator. I checked 8 Mhz and 20 Mhz square wave and I thought it would be be sharper( I heard that DSO can measure square wave which has 5 times less frequent than BW ). Maybe I'm doing something wrong with oscilloscope settings or it's just normal? I think i checked all the settings. I also connect generator with BNC cable without probes but the effect was the same, maybe this generator is old and have some faults( they use it in another department to generate only kHz signals ).

Thanks for any help.

[KTP]

The 33220A has a rise time of around 12ns...so for your 20mhz wave there just isn't a lot of time for the pulse to get to the high point before it has to turn around and go back low.

This is why I had the idea of making a little cheap add-on board for function generators such that faster rise time pulses and faster noise glitches could be experimented using existing slower rise time waveform generators.   But nobody has responded to that thread/idea, so maybe only you and I are interested 

[jstan]

Thank for pointing me out. I didn't realize that this awg have such long rise time. I don't have any awg so i won't be interested in your ad-on.

[mihlit]

Hi,
I'm considering buying this oscilloscope in local store. They have it on sale, so I'm afraid it can be some old version. How can I find out what HW version they have? Is there some sticker on the oscilloscope? What is the latest hw version? Are there some version I should not buy?
Thank you