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

[tinhead]

so far the highest frequency DIY DSO I've seen is this Polish school project

QurDSO Spring 2011 (look on ourdev.cn) have 2 x 125MSs, 30Mhz bw and nice 480x320 screen,
however my personal fav. is this one:

http://www.mikrocontroller.net/topic/228997

analog part designed for 150MHz, sampling with 50MSs only, but look at the self made enclosure
and the TDS2xxx "looks-like" firmware.

[marmad]

Display update rate is maximum around 25 frames per second.

Yes, that it is the absolute maximum - but it rarely goes this fast.  The average is probably closer to 12-16 FPS.

human eye can note around 10 per second (as simple  thumb rule only)

This is missing the point - the human eye can process 10 to 12 individual images per second - but the way we perceive speed vs. stuttering (slow, sporadic movement) is much more complex.  The Owon FEELS slow and stuttering at MANY settings - like it's not capturing and displaying waveforms as fast as it should - and that, IMO, is their biggest problem.

You can look what is waveform rate in hantek (maximum). maybe it is ten times more.

Testing the Hantek now, I think it's much more than 10x at certain settings.

But, it can capture full speed only to 4k memory. Under half speed  can go to 1M memory (max 400Msa/s and 200 for 2 channel max)).

Well, the Owon has similar problems.  For example, at 500ns/div capturing to 1K, the speed is 16 wfrms/s - but capturing to 10M, the speed drops to 12 wfrms/s - so 25% slower.

But example random spikes hunting it can use slow horz speed, high memory and so high samplerate and this way you can also try minimize blind time ratio. This is not of course same as can do with high wfrms/s but some cases these cind of "optimizaton" may give cheap way to solve some needs.

Yes, we all know these tricks for trying to reduce blind time on slow scopes.  Unfortunately, they don't help at all when hunting very short glitches - which won't be visually apparent at slower horizontal settings - and furthermore, you need to know that you're looking for them to even use these techniques.  What if you're just looking at a signal not expecting any glitches?  For example, when the Owon is set to 500ns/div, it's blind time is 99.98% (with 1K capture depth - even higher with more memory) - so it's very bad at seeing sporadic events (I've proven this over and over in many tests) - so you might just be looking at a signal thinking that everything looks fine (even though there are random glitches which the scope is blind to).

More I have look Owon - it feels better. (UI is not as good as it can do with these HW what scope have... there are unused resources lot of)

Now that I am testing the Hantek, I can say unequivocally that build quality and external design is better on the Owon - but the firmware is crap compared to the Hantek.

[rf-loop]

Well, the Owon has similar problems.  For example, at 500ns/div capturing to 1K, the speed is 16 wfrms/s - but capturing to 10M, the speed drops to 12 wfrms/s - so 25% slower.

1k 500ns/div ja 16wfrms/s
inside one second get 16k kaptured data and Owon use samplerate 100Msa/s
in this case ideal scope (full time active capture) give 100M
16000/100000000=0.00016

10M 500ns/div ja 12wfrms/s
inside one second get 120M kaptured data and owon use samplerate 1Gs/s
in this case ideal scope (full time active capture)  give 1G
120000000/1000000000=0.12

So -  what is problem with deep full speed memory. (this "question" do not need answer)
Hantek have not at all this deep full speed memory.

[tinhead]

Hantek have not at all this deep full speed memory.

right, unfortunately like all these other scopes with pipelined SRAM (Rigol, ATTEN, Instek, HanTekway)
only 400MSs possible, when you patch firmware to clock faster and solder faster SRAM (org. SRAM can 500MSs too,
but that's 25% overclocking, so might not work for long time) like i did then 500MSs ...
so whatever you do - still only 50% of what Owon can do.

[hilltop]

Wow, that Selbstbau scope is a really well done job.  Very neat!

Ryan

so far the highest frequency DIY DSO I've seen is this Polish school project

QurDSO Spring 2011 (look on ourdev.cn) have 2 x 125MSs, 30Mhz bw and nice 480x320 screen,
however my personal fav. is this one:

http://www.mikrocontroller.net/topic/228997

analog part designed for 150MHz, sampling with 50MSs only, but look at the self made enclosure
and the TDS2xxx "looks-like" firmware.

[bertchai]

Hi,

I hope someone can give me a hand on this.   

Compensation of the probes.  I followed the instruction on P13-14 of the manual to compensate the probe at 10x,  then I switched everything to 1x to check but I don't quite get the square waves I got in 10x.   I tried both probes and channels.  Is this normal?  What results did you guys get?  Is there something wrong with my Owon?

Many Thanks

[rf-loop]

This is normal and not only Owon if scope input is AC coupled.

It depends also AC coupling low freq response how much it do this. Very slow square wave top or bottom you can nearly think as they are DC. (If square mother frequency is dropped example to 100 or 10Hz you can not see that it is square) 1kHz 50/50% square: "top and bottomm of signal you can think they are 0.5ms duration DC levels.

Normally probes what have 1:1 and 10:1 compensation is only for 10:1

In this kind of double probes 1:1 is also highly reduced frequency bandwith upper end. It reads in all specs.

You can see it if you look this rising edge with higher speeds (turn 1us/div and compare. Also 10:1 looks slow rising (becouse 1kHz output rising time is slow)  but 1:1 probe looks lot of more slow. 1:1 is slow. Scope itself 1x 10x etc settings do not affect anything to freq response. Only how scope display level numbers.

 

But what is wrong with your scope settings. Wrong or "wrong" but:

Set CH for DC, not for AC.  and probe to 10:1  (1:1 is not adjustable. Expensive probes may have inside termination box many "service" settings for calibration and freq response for freq response flatness)

use 100us/div

If need more accurate use one step lower V/div value and shift then signal so that upper (or lower) part is more than 1-2  div from top (or bottom)

[marmad]

10M 500ns/div ja 12wfrms/s
inside one second get 120M kaptured data and owon use samplerate 1Gs/s
in this case ideal scope (full time active capture)  give 1G
120000000/1000000000=0.12

This comparison is not valid.  At the 500ns/div setting, only 1/1000 of the captured data is visible on the screen - unless you stop it, which defeats the purpose of real-time monitoring.  It is NOT as useful or the same as a scope which might capture 1K @ 500ns/div @ 1200 wfrms/s - because in this case, most of the data captured would be visible on the screen, not just a tiny piece.

The 10M of memory is really only useful in single-shot mode.  But unfortunately, even when using it this way, Owon has not included ANY tools for looking at these 10M records of captured data: no dual-window mode, no position indicator, no mark or search functions, and worst - no auto-scrolling mode.  Have you tried to look through a single-shot 10M record scrolling with the little knob?  It is TEDIOUS in the extreme, and you lose your position very quickly.

Simply adding more memory to a scope, if it's not implemented well, is not the answer to solving problems.  There is a reason that more expensive scopes have less memory but faster waveform update rates.

Again, if the Owon captured 1000x (or even 100x) more waveforms per second when it used 1K of memory than when it used 10M of memory, it would be a much more useful tool.

[rf-loop]

  Have you tried to look through a single-shot 10M record scrolling with the little knob?  It is TEDIOUS in the extreme, and you lose your position very quickly.

In stop mode scrolling around 10M is easy and fast. (if example 1us/div 1Gs/s 10M).
Stop > zoom out with horz speed > shift wanted position by horz pos > zoom in. Very fast can go to every place there in 10M capture.  It is not good but better than nothing and same with hantek but sampling is now more slow and capture is more short.
But as before, Owon is conventional DSO with long full speed memory.
Hantek is "tiny" DPO with short memory and faster wfrms/s but really not fast. Try 20ns/div and 1M memory. You can calculate with your eyes how many waveforms/s. (be sure that you know when scope use long memory. Setting in menu do not tell always truth (it is one bug in Hantek - and there is tens of these bugs - and as talk Hantek, need know what FW becouse they are with different features and differrent bugs. (Ht. is so that some old bug go and new bugs come in every FW update. No one know exactly how they have make all this FW spaghetti.). Setting may be 1M but truth may be totally different becouse if you accidentally turn faster, scope drop 1M out and use 4k (or 3.2k depends...who knows)
After one week you find many "interesting" things. But still, Hantek is faster. Pity that this zoom is what it is. Toy zoom.

[rf-loop]

... as a scope which might capture 1K @ 500ns/div @ 1200 wfrms/s - because in this case, most of the data captured would be visible on the screen, not just a tiny piece.

Can you name this kind of scope?
It is not Hantek DSO5000B series, this I know,  but what it is?

[rf-loop]

....no mark or search functions, and worst - no auto-scrolling mode. 

Also it is nice if you can name this scope?

Hantek?

[marmad]

It is not good but better than nothing...

This was exactly my point.

...same with hantek but sampling is now more slow and capture is more short.

Yes, but with the dual-window mode of the Hantek, you can at least see where you are in the waveform.  As I mentioned before, it's VERY easy to lose your position on the Owon.

You can calculate with your eyes how many waveforms/s.

This is not necessarily true - from what I can tell, the Hantek wfrms/s is not linked to the refresh rate of the display - so it might refresh the display less often, but be capturing 100x more waveforms - you would have to test to be sure.  But, unfortunately, it is true with the Owon - and the sad thing is that you see this effect sometimes when only using 1K of memory!

... as a scope which might capture 1K @ 500ns/div @ 1200 wfrms/s - because in this case, most of the data captured would be visible on the screen, not just a tiny piece.

Can you name this kind of scope?
It is not Hantek DSO5000B series, this I know,  but what it is?

I was writing hypothetically - it's obvious that the makers of the cheap lower-end scopes are not being forthcoming with their waveforms per second specs - so you would need to test them thoroughly to find answers.  But even 120 wfrms/s would decrease the blind time by a factor of 10.

....no mark or search functions, and worst - no auto-scrolling mode.

Also it is nice if you can name this scope?
Hantek?

Yes.  The Hantek allows you to drop marks, jump to those marks (although it doesn't search for trigger events, etc), and auto-scroll through a waveform record.  I can't believe that the Owon doesn't even do the auto-scroll, which is SO easy to implement in firmware.

I was hoping, when I bought the Owon, that it would represent a new breakthrough in price/performance at this level - and perhaps it would have if they had hired software engineers who were as clever as their product designers.  I do think it represents a better level of design, build, and included hardware for cheap Chinese scopes - but it's performance is stifled by bad choices at the firmware level.

Now that the new Agilent 2-channel 70MHz is available for 2.5x more money than the Owon (but, IMO, much more than 2.5x more powerful), you have to wonder if it isn't false savings to buy one of these $500 - $600 scopes - unless, of course, you absolutely can't afford more money - or - you only do the simplest kind of work with the scope - or - you are planning to upgrade to a better scope later.

[saturation]

Yes, thanks for the update and I stand corrected.  those scopes are now the best I've seen for DIY.  Steffen says that the actual parts cost isn't more than Euro 100, but I'm not so sure.  Its interesting the discussion does suggest the cost is also ~ Euro 300 as an estimate for get it fully assembled.

Wow, that Selbstbau scope is a really well done job.  Very neat!

Ryan

so far the highest frequency DIY DSO I've seen is this Polish school project

QurDSO Spring 2011 (look on ourdev.cn) have 2 x 125MSs, 30Mhz bw and nice 480x320 screen,
however my personal fav. is this one:

http://www.mikrocontroller.net/topic/228997

analog part designed for 150MHz, sampling with 50MSs only, but look at the self made enclosure
and the TDS2xxx "looks-like" firmware.

[bertchai]

Set CH for DC, not for AC.  and probe to 10:1  (1:1 is not adjustable. Expensive probes may have inside termination box many "service" settings for calibration and freq response for freq response flatness)

use 100us/div

If need more accurate use one step lower V/div value and shift then signal so that upper (or lower) part is more than 1-2  div from top (or bottom)

It is fine now.  Many Thanks

[SnakeBite]

Hello guys,

i've sent owon yesterday the second review video and they respond very quickly. here's a copy of what they say:

Hello,

 

i want to buy this oscilloscope and some of the problem in the video are annoying please fix them in the next firmware update.

 



 

tell me if you working on it

 

Ido

Hello Ido,

Thank you for your message on our website of www.owon.com.cn

After checking the video, we welcome to have good advice to our products, and will improve them in the future or take them into mind in developing new products.
But as you know, different brand of DSO has its own features and design, such as the menu design, some customers like it, others like that. And we also have our below features like video said:

8 inch big LCD
10M record length
VGA port
Battery support, etc.

If one day, we will have F/W update, we will send email to you to realize it
By the way, which country are you from?

Thanks,

Mr. Colin Lee
Xiamen Lilliput Technology Co.,Ltd
Fujian Lilliput Optoelectronics Technology Co.,Ltd
www.owon.com.cn
Tel:             86 592 2575666       ext 1047
Fax: 86 592 2575669

 

 Ido
 

 

[RCMR]

Translation:

Thank you for your email.  As the video says - our product is great.  If you want any changes then tuff-cheese.

Please provide more information so we can forward it to the relevant marketing people for later follow-ups.
 

[rf-loop]

You or someone else have never try this with  Agilent or Tektronix?

[alm]

I wouldn't expect a more helpful reply for them either, but I would expect a more polished product with less to complain about, so it's not as much of an issue.

[hilltop]

OK I returned my Owon SDS7102... they sent the wrong model.  Apparently there are two models, one with serial port and one with VGA port.  I ordered a VGA-enabled model, and apparently the sales person didn't realize there are two different variations.  So they (Saelig) gave me an RMA.   They won't even have any of the VGA models for at least six weeks. 

BTW: Really liked the scope in the few hours I had to tinker with it.   However it seemed like there was a lot of noise being picked up, even without a probe connected.  Tested the scope in two different locations, same thing.  About 5mV p-p noise and from my limited knowledge, looked like about 200khz frequency.  Seemed like a lot to me.  The tech rep at Saelig turned his on, and only saw about 0.8mV noise.  Maybe it has to do with my setup.  I did not connect the "ground screw" but I assumed that would be handled by the grounded 3-prong AC cable?  Am I wrong on that?

I could definitely see 60Hz noise with floating probes, but I guess that's normal?

Anyway, for the price, I was very pleased.  Just sorry it didn't have the VGA port!


Ryan

[Bored@Work]

Translation:

Thank you for your email.  As the video says - our product is great.  If you want any changes then tuff-cheese.

Let me translate your translation

We don't want to invest more money in the development / bug fixing. We don't care if the product is really good or not, we just think it is good enough to make money for us. Take it as it is or leave it.

[rf-loop]

Translation:

Thank you for your email.  As the video says - our product is great.  If you want any changes then tuff-cheese.

Let me translate your translation

We don't want to invest more money in the development / bug fixing. We don't care if the product is really good or not, we just think it is good enough to make money for us. Take it as it is or leave it.

What is problem with this. How you thing if you are buying car, camera, TV, Phone, washing machine, etc.
What is big difference if buy test euipment.

How many of you have send letter to car factory and ask that blinking light switch need different timing, please what time you update it. Or how is washing machine who always forget timer if you have also selected extra rinse. have you send letter to factory and ask FW update - what is answer. "Do not use this setting".  How about thousands of letters arrive from peoples who do not know what is oscilloscope or and how to use it and still also they want better FW. Or some other changes.

I do not like my camera, shutter speed selection menu is bad and in wrong place, it is more easy to handle if I can go directly after I push... ok I send letter to factory. What they answer. Something that: thank your interest about our products, we allways are happy we get feedback from our customers for developing our production. (and answer come from automatized answer machine)

I have reported two small bugs in Owon FW. I have show exactly what is bug, how it works. Documented it. Sen sinfo to Owon and get very professional answer and talking. It is really so that in factory level there need be some order what to do, when to do and who do. This is not as mixewd level of hobbyist club and hobby chat box if you start talking with engineers about product designing, bugs repair etc. I well undertand that first factory need "filter" peoples answering letters. If you btal professionally and also undertand even littlebit all this complex system what is making product then you get also real answers. They have not time to play "blinking led" kidding.

One example in short time "history" was 3'rd party Linux driver developing for Owon scopes.
There factory professionals also was helping this.  But as all this kind of "open" systems, they never go ready or developing stop and then nobody update anything for new models.

Last time I tell factory I have problem to read  waveform binary files. Inside 48 hours and both send 2 emails... now I know.

Some small bug I have find and reported these to Owon. They promise later  repair it after they now inspect and look it deeper and then add to "need to do" list.

Try send letter to Apple factory and teach them how to do. I'm iteresting what kind of answer they give. Or try same with Nikon, or maybe with Tektronix...

Main stream is that there is product. You look this product pros and cons, then you look others, also what you like or nt, also what are you feels about product. Then you select.  That's it.  Buy what you have inspected what is ok for you.

If you have find really bug or fail then please document it so that this simpliest first phase "filter" people do not put it to garbage and give some trivial answer.  We can help factory do better or we can only make load to factory peoples who read thousands of emails. Also if you report something. Do not write as here in forum. If you want they really take you serious. make simply and clear document about what you see problem. Take good care that chinese culture and specially they language and laguage based thinking is littlebit different. Do NOT never go as wild westerns to saloon... show gun and tell that YOU are wrong, this YOU have make wrong.  It is better to littlebit adapt also to cultural things. What is normal here is not always normal in China. If you can do so that chinese EE or sales person feel loose his faces... you just have loose this game. Do not show your finger to hima that YOU do wrong.  You can ask that you want do cowork and help develop together better machine and it do not cost anything and also you can win more selling and more money if we do these small developments.. 

One problem is also that there is just one product, example SDS7102VGA.
One like orange one like apple. It is impossible to grow appleorange.  I go to store and I do not buy apple if I like orange better. If I still buy apple and then go to for appeal and whimper with farmer that: I do not like this becouse it do not taste as orange, YOU have made something wrong. Clever?

If I do this... where I find one stupid.  It is easy, I go and look mirror, then I find.

I have never get stupid answer from Owon after first they have understand what we are talking and why.

[Alexey]

All the good times.
Please choose from two things, which is better?
I want to purchase OWON SDS7102, sent an inquiry about the price of the plant, sent a reply that 100 MHz is also SDS8102, except that the Sample rate (Real time):
for SDS7102 = 1GS / s
for SDS8102 = 2GS / s

[rf-loop]

They are all same. Really only difference is ADC sampling speed. This is my quess but there is different ADC, propably MTX2001 (quess). 7102 use MTX2002 (this is known)
But all other things are same.

If I look 2 channel use. 7102 have maximum 500Ms/s per channel.

I have measured visible rise time with one channel use and two channel use.
One channel give 2.6ns/div (averaged16) Samplerate 1Gs/s (so samples every 1ns)
Two channel give 3.2ns/div (averaged16) Samplerate 500Ms/s (so samples every 2ns)

Now if think what 8102 can give.
Maybe one channel do not rise faster becouse analog parts is limiting.
maybe two channel use give better result. maybe just like 7102 with one channel.
Also it reduce aliasing.

If need lot of 2 channel use for both 100MHz... well, think 8102. 500Ms/s is not very nice for 100MHz. (it is enough but signal do not look so  nice.)

[Alexey]

Thank you very much.
As I understand it:
- If you often work one channel, then 7102;
- If you often work with two channels, then 8102.

[mzzj]

New user popping in..

I just go my new OWON and I am quite satisfied BUT found one very annoying software bug:

Scope set to 100ms/div or slower and REAL samplerate drops down to nothing.
settings:
mem set to 10M, 100ms/div.
scope shows sample rate as 10MS/s
try measuring 1kHz  calibration signal, hit stop-button and "zoom in" turning timebase to 500us/div.

1kHz square wave has become total garbage, turns out sample rate is less than one-thousand of what it should be.

Can anyone check this behaviour on their own scope, is it the same in all units and firmware versions?

My scope has firmware 2.1.1