OWON Oscilloscopes

Owon Electronics and test equipments. Discussion area is now open!
It is currently Sat Jun 24, 2017 11:48 am

All times are UTC + 2 hours [ DST ]



Welcome
Welcome to owon

You are currently viewing our boards as a guest, which gives you limited access to view most discussions and access our other features. By joining our free community, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content, and access many other special features. In addition, registered members also see less advertisements. Registration is fast, simple, and absolutely free, so please, join our community today!


Post new topic Reply to topic  [ 13 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 2:54 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
This forum is mainly becouse our main forum do not show links and images to guests. There only members can look images and links.

Our main forum is here.



Here locally in Finland you are also welcome to visit our lab, test and look what is truth with your own eyes and with good test equipments.
-----------------------------

NOTE: Risetime and Frequency BW tests and numbers here are valid only with factory original 2.5.1 !

ADD: 30.06.2012 Owon have (now late 2Q2012) improved (agen) hardware and now frequency response -3dB and -6dB points have come to lower frequency but around -1dB point is still 100MHz. (look my comment before this later in this thread)

This have done for better signal quality specially for attenuate aliazing problems. Analog BW change and changes in front end circuits have dropped noise level (it have been good in this price class and now it is clearly better.)
Also front end temperature have decrased a lot.

So, thank you Owon!
--------------

In internet you may find many kind of reviews about many oscilloscopes.

Also from Owon.

It is markable that mostly these reviews have made very loosy and specially without measurements or with very poor measurements. Maybe due to lack of experience and skill to make tests and also many times total lack of any acceptable test equipments.

Good example about this kind of "rewiev" is here.

And more bad is that old rewiev seems to be updated due to OP but using still these old pictures without very clear note that all information is fully or partially obsolete. But very easy updating give feel that informatioon is also updated.

1. hardware have totally changed.
After this "review", hardware have changed, many minor updates and least one major update.

2. FW what have been in original review have been nearly first generation including lot of also serious bugs.
After original review FW have updated by Owon several times.

3. Pictures where can see some measurements are really piece of ********, it can see in some cases that there have used settings what really give "wanted" bad result. I can not find any other reason but bias. It is like mission to show "Owon is bad".


Most fun is that OP seems use Hantek 3x25 DDS box as signal generator for test 100MHz 1GSa/s oscilloscope. Oh well... really inferior. If this is situation it is maybe best to tell that sorry I can not do reliable measurements. Trolley takes the horse. It is like testing Keithley 6+ digit meter with UniToy 3+ digit toybox.

Maybe this whole review is highly biased for some other brand or with some other hidden reasons?


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:01 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
All tests have made with 1Q2012 produced V.2.5.1
This have 4Q2011 improved hardware.
Markable difference is only with frequency bandwidth in 100 - 500MHz area. More there is hidden things in littlebit better inside mechanical construction. Some service related things have made littlebit more ergonomy (less working time = cheaper) and maybe littlebit also for better reliability etc. In practice only dramatic change is BW and risetime.
FW updates can get also for older models.


Start with voltage accuracy test. (later other tests)

First. Oscilloscope have been >30min running and then defaulted and self cal.

Test signal from Tektronix CG5011 what have unofficially checked that accuracy is ok. (also looked with reserve unit CG5010 and between these can not see any difference in this accuracy class what is Owon)

Signal connected parallel to both channels using T split and 2pcs 1m 50ohm coaxial. (this method is enough good for this purpose)

Oscilloscope set for normal sampling, trigger fron CH1 and memory 1k and average 16. Both channels show also automatic measured "amplitude" levels (not peak). It is not specified how it measure related to Tektronix specification. Right way is look around middle part of pulse. There is specified overshoots in calibrator signal near edges, this is not error, it is Tektronix "feature". Note that Tektronix change risetime for higher Voltage levels.

Conclusion: With these measurements Owon is very good inside specified accuracy.

Image

2mV/div there can see small DC offset specially in CH2.


Image

Image

Image

Freq counter reacts to some enviroment electric noise peaks.
(it is not error in Owon or Tektronix)

Image

Image

Image

Image

Image

Image

Scope 2V/div, signal gen 1V/div mult 8


Image

Scope 2V/div, signal gen 2V/div mult 8
Here Tektronix change risetime (higher voltage area) and risetime is not important, only Voltage value in middle of top and bottom is meanigfull.



Image

Image


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:02 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Here tested risetimes.

First single channel in use using samplerate 1GSa/s
Next 2channel in use so that samplerate is 500MSa/s simultneously.

Signal come form Tektronix CG5011 calibrator "Edge" mode. Signal is teminated in oscilloscope input with Tektronix 011-0049-01 50ohm terminator. Calibrator edge risetime is around or under 1.3ns

With 1GSa/s sampling corner wobbling is minimal.
With 500MSa/s there is more corner wobbling and mainly this affect to automatic measured risetime difference between CH1 and CH2.
(number vary randomly around and in picture time it was as it was)

This corner wobbling is same as "aliazing" what is better known.

In pulse edges it affect corners wobbling just becouse there are harmonics over Nyquist frequency.

Oscilloscope analog BW before ADC is always compromize.
With 1GSa/s all over 500MHz pruce aliazing. Over 500MHz signals are highly attenuated but some small amount is there still.

Situation is different with 2 channel use. Sampling is now 500MSa/s and Nyquist mirror is 250MHz.
As you can see 250MHz is not so much attenuated (around only -3dB)
So it is natural that it produce aliazing and it can see with pulse edges as corners wobbling.

These things related to signal harmonics, Nyquist mirrors etc are specially important to understand. In serious FFT use I recommend special filtering related to signal under test. I will make FFT tests later but now first lot of other tests after they are ready. (it is slow becouse lot of checkings, most of tests dobled or tripled so that I can be sure I have not make mistakes and so that it is repeatable. Also equipments used need check somehow that all accuracy etc are ok.

I do not "tests" like in some "review". First thing and study, and thenh publishe so that truth is not totally loosed.

Example there are one review in EEVblog where nearly all tests have done without any understanding how to test. Without any self criticism they all published what ever ******** and crap.

With repetitive waveforms it can filter with averaging and or there can use persistence so that human eye can see wobbling limits and make intelligent opinion about signal under test.

In pictures here used averaging with setting 4.

In this price class pulse response is extremely good.
Best what I have ever seen and I have seen lot of.


Image

Image



Image

Image


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:03 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Trigger have also ALT (Alternate) what can not find every oscilloscope.
Owon works well and also both channels have separate adjusted trig level.

Here is one simple test. Totally different leval and frequency to both channels. CH1 high loevel ~4.9Hz and CH2 low level ~14.1MHz (not accurate). Trig mode ALT. CH1 and CH2 trig level edjusted just as randomly want for picture. Trigger is very stabile.

Also both channels have now separate frequency counter. (it is trigger counter)

Original signals are not very pure so you can see noise etc.. this test was only for show this ALT feature.

Note: Also these signals rise and fall times are slower than oscilloscope itself rise and fall time. So this slow risetime is in signal under test, not in scope.


Image


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:04 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Frequency response.

Owon SDS7102 freq response is very flat. It can say, it is straight line from 0 - 50MHz. After then it very smoothly drops down. I have not seen so pure flatness and smooth curve so many times in oscilloscopes frequency response. Specially in this kind of price class.
Owon analog front end is really good design if think its price class.
One minor think I miss there. Real 50ohm selectable input imbedance. I mean real 50ohm imbedance over freq BW. (I do not mean this ******** what have seen in somewhere, relay connect just 50 ohm resistor parallel to input on the circuit board behind BNC. THis ******** have not any use.
It is not so important feature in 100MHz oscilloscope where input capacitance is low but becouse Owon freq response is much more than just 100MHz it may be nice feature)

Test configuration:

Signal source: HP8644B
1m good quality coaxial cable.
Oscilloscope input 50ohm Tektronix terminator as before.
Signal level over used frequency area checked also with Agilent/HP RF power meter. It is enough flat to this kind of simple measurement.
(absolute level accuracy not meaningful becouse adjusted so that 1MHz is reference point for 8 div p-p level.)

And signal purity is so good that here we do not measure signal source quality with Owon. Here we measure Owon as DUT.
(this have totally forget in some published kinderkarten level "reviews" and this maybe due to total lack of understanding how to test or maybe even biased for show Owon as bad as can do so that always can explain that "not me but this signal"... sorry... or just as forget tell that Rigol picture was made using averaging and Owon made without. Just random forget?)

NOTE: All these pictures are as single shot sampled without any averaging.
Just raw capture as is. (image pick-up via USB to computer and scope in RUN state.)

( live running picture looks even littlebit better due to feature of the eye and the brain's perception.)


Explanations for pictures more later (editing text)

Image
One channel in use. Adjusted signal generator level so that 1MHz sinewave level is around 8div peak to peak. (around 400mV) and with this same level all next measurements. This is reference level for measurements.
Cursors position is 3dB under reference level.

Image
20MHz

Image
50MHz

Image
100MHz

Image
150MHz

Image
200MHz (still over -3dB level)

Image
250MHz around -3dB still can not see any markable wobbling.

Image
300MHz. Here can see small wobbling if you look peak values and how these peaks levels are not exatly same but wobbling is still very small.


Image
350MHz Wobbling can see clearly. Every peak is littlebit different level.
Next picture show it fully.

Image
350MHz. Display have now infinite perisist so it can remember all values where signal have been. (of course for repetitive signals can use averaging and hide all wobbling)



Image
400MHz. Only test with infinite persistence. Wobbling is terrible but still if use averaging this wobbling nearly disappear.


Image
450MHz. Signal is still somehow visible. Wobbling is really terrible. But still example if use high number average for continuous repetitive signal it can hide lot of this wobbling. Trigger still works if you look trig point!


Image
1MHz. Other things same as before but now two channel in use. (CH2 signal is not any meaning, just meaningful is that it is on and samplerate drops to 500MSa/s

No need test 20 and 50MHz becouse just same as before.

Image
100MHz

Image
150MHz.

Image
150MHz 10Msample memory. (you see long memory do not drop samperate.) And if look very carefully it can see there is littlebit wobbling.

Image
150MHz. Infinite persist for show total full wobbling. (as before collecting time is enough for collect all worst cases and it is not much)

Image
175MHz. Wobbling is clear and high.

Image
200MHz. Wobbling is terrible. (yes with averaging it still can hide)


Image
200MHz and 10Mpoints memory without persist. Wobbling can see very high and clear. If signal is repetitive there can use averaging, it can still hide this. (but averaging also hide all fast changes in signal so it can use in very limited situations)





Some kind of conclusion after risetime and sinewave bandwidth test is that it is well over 200MHz oscilloscope if single channel in use and over 150MHz oscilloscope if 2 channel is in use. Not becouse BW but becouse samplerate with 2 channel use is maximum 500MSa/s.

If use 5 sample criteria (referenced to Agilent publications) 1 channel is 200MHz and two channel is 100MHz.

Frequency / BW, Samperate / risetime, BW shape need think carefully.
It depends also signals what need measure. You can not measure 200MHz square wave risetime! But with < 1.4ns risetime signal uoy can measure this oscilloscope risetime.

Other hand risetime is markable, in other hand sinewave BW is markable but also very markable is sample time interval if think resolution. WE do not mesasure and look only square waves and sinewaves. There is never ideal square wave available. For normal use it is many times enough you see 5th harmonics but for real accurate measurements it is goot to see 7th, 9th and even 11th harmonics. It means in thinb rule that for 10MHz fast rising square wave you need least 100MHz oscilloscope.

(oscilloscopes BW is traditionally so that BW sahape is gaussian and BW limit is -3dB point compared to some reference frequancy. (some use 50kHz some 100kHz reference... not so important as long as low frequancy flatness is ok)

Owon label this as 100MHz BW oscilloscope. It is right, if use tight standards and also keep care that it is with two channels. (example Hantek label they DSO5202B to 200MHz. It is not acceptable. It can not do it with 2 channel in use if 4k memory is in use. And it can not do at all if 2 channel and deep memory (>4k) is in use. It can just do 40-60MHz. (Becouse with 2 channel and deep memory Hantek maximum samplerate is 200MSa/s. In HW1005 e8.)


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:05 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Related to pictures before with real measurements here is random example how this unfair game go.

Just think whole this topic name. "do not believe all you see"

These 2 pictures are some random examples from some Owon "review" and here marked with big blue cross. Least with these pictures and comments in original review every people who have real experience and skills to make tests may think - why this kind of total junk.





Owon looks like this with 75MHz test signal:

Image




Rigol looks like this with same 75MHz test signal. (Rigol real size is ok)

Image





Only for just for fun I resize Rigol image to around same size what is Owon.

Image


After some questions then just informed that "forget" set Owon to averaging. Is it fun? But still this same picture show to readers.

--------


I am biased, I am seller. But I want still show true measurements and not do nearly lie "trick". But all who have Owon can repeat my measurements.

Personally I have been nearly 50 years hobbyist (electronic, radio etc) and I have also walk long road in the professional field of industrial electric and electronic and also long walk experience with professional test equipments use and repair.

I really respect all peoples and hobbyist so much that I want be as honest as possible and least if I show some test measurement, they are made well and carefully. Specially in entry level phase it is important that get some kind of trusted information. In worst case it may destroy whole nice just starting hobby or make other kind of harm and also money loose.
And it is also same for second hand test equipments what I sell and have sell lot of.

It may decay or make harm for sales but maybe in my world my faces are more important than sales volume. Finland is small country, here we nearly konow all each others and faces can loose just onöly one time and thats it.

In China situation is littlebit different,.. whole culture is different. Buyer is totally responsible about all. Specaially about quality. Seller is good businessman if he can sell ******** to stupid with most high price... but this is just other world.

It is not easy stay between chinese business and you.

But these kind of global shared and advertised "reviews" read just more these peoples who do NOT have Owon for check. This is just most bad type of peoples manipulation.

Then some claims that reviw is so good that many seller want also use it as ad. Do this tell that review is ok or do it tell that there are lot of sellers who do not really understand anything what they are sell. They sell just "nice box". If seller understan really anything he understand that this kind of review can only damage brand. But in other hand there are also lot of buyers who do not have any experience.. example just starting his hobby. Just entry level. It is very pity and bad if these peoples get wrong information just becouse they can not select what information is reliable and what is not.


If you see some error in my measuments or question about it, do not hesitate to contact and ask.
I will also correct if I have made accidentally some error.
I'm very thankful if someone find some fail in my tests or information and I'm happy to correct it. If find error it is always win-win in final game.

But remember. Owon is not Rolls-Royce, Owon is not Rohde&Schwartz.

Owon is very good in its price class! Speacially in signal and capturing quality. But remember, there is not at all any scope what is just best. It depends what are things you need, want and how you weight these.


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:05 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Testing with video signal.
Test setting: Signal from PM5518TXYC video test generator.
Signal simple T splitted to both oscilloscope inputs. (this split method is really bad and can use only becouse both cables are equal and both scope inputs are enough equal and frequencies inside video signal is quite low with also low risetimes of course). But still something is of course wrong - I know, this kind of real mistake can not do in serious lab work but this is only scope demonstration and becouse know this setting it can take count.
It is not meaningful here becouse here we do not measure exactly this signal, it is only for demonstrate oscilloscope. (Video output is 75ohm. It is T splitted to 2 1meter lenght 50ohm coaxial what are ended directly to oscilloscope CH1 and 2 inputs. So nearly all is wrong! But acceptable for this test.)

I can later also do something with line sync and using line number adjustment.


Note: Owon display show around 15 horizontal time div but whole capture lenght is 20div (with this time and memory setting)


Special NOTE:


I will littlebit explanate what you see in samplerate information on the screen. It is FW bug or maybe wrong thinked "feature".
But after stop and then change timebase it give wrong information. It tell what is samplerate IF scope is running with these settings.
True samplerate in capture is what is was as scope stopped to "STOP" status.
I hope Owon repair or clarify this information later. Now it is just like "wrong" information.
Yes, it tell what is samplerate if you push RUN with this setting but... with this way it do lot of confusion. These are marked in pictures with orange line over samplerate ____


Image
2 channel same videosignal, captured with 20ms/div, 25MSa/s (10M memory)
Whole capture is visible 15 whole field and whole capture is 20 field.
(Signal is PAL, 625, color, greyscale)
After capture with RUN mode scope set to STOP for looking more details in capture.
(next 4 pictures). This is equal if whole capture have done as single shot.
With this horizontal speed and 2 channel in use, realtime samplerate is 25MSa/s for both channels simultaneously.




Image
Zoomed 10x (right samplerate when scope stopped is 25MSa/s)

Image
Zoomed 4000x (right samplerate when scope stopped is 25MSa/s)


Image
Zoomed out to zoom factor 2000x and shifted to looking color burst.
(right samplerate when scope stopped is 25MSa/s)


Image
Zoomed to 40000x ! (right samplerate when scope stopped is 25MSa/s)


Note, single shot both channels 20 video frames, zoomed to single video line color burst!






Image
2 channel same videosignal, captured with 2ms/div, 250MSa/s (10M memory)
Whole capture is visible 1.5 whole field and whole capture is 2 field.
(Signal is PAL, 625, color, greyscale)
After capture with RUN mode scope set to STOP for looking more details in capture.
(next 5 pictures). This is equal if whole capture have done as single shot.
With this horizontal speed and 2 channel in use, realtime samplerate is 250MSa/s for both channels simultaneously.



Image
Zoomed 200x (right samplerate when scope stopped is 250MSa/s)


Image
zoomed 4000x (right samplerate when scope stopped is 250MSa/s)


Image
zoomed 10000x to color burst (right samplerate when scope stopped is 250MSa/s)


Image
zoomed 40000x to color burst and measured frequency with cursors.
(right samplerate when scope stopped is 250MSa/s)


Note that BOTH channels have same Video signal and you can barely see it. (yellow CH2 is color that it is "over" red.)

Image
Just only zoomed back and adjusted one line center and measured time with cursors.
(right samplerate when scope stopped is 250MSa/s)



With this small example you may littlebit also imagine what full speed 10M deep memory for both channels is really meaning.
This affect dramatically to samplerate what can use with slow horizontal speed.
It can very easy understand that with half speed 1M shared "deep" memory this whole test fails totally.

Also it is good to know that Owon have hardware video suncronization (trigger) and it is more realiable and faster that software based Video synch.
With 2ms/div 10M memory owon can capture 16 times per second. (two full video frame in one capture)
In this price class it is really good - amazing good.

Of course there can also use line sync, and adjust what line number want, and also odd and even fields sync.
Bictures have ALL field sync.


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:06 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Here is something about noise.

Inputs terminated with 50ohm terminator.
In this price class this is good. There are many others what have more! You can also esy find how is Agilent or Tektronix noise.

Also it is good to note that example analog oscilloscope looks like "no noise".
This is just wrong. For random noise analog oscilloscope phosphor is good filter!

With repetitive signals random noise can clean easy with averaging.
But manytimes peoples forget try digital oscilloscopes adjustable persistence. This may be sometimes very useful. But still there is not this attack delay what is important part of analog scope. If there is "zero" delay phosphor, then you see lot of noise and "fat" line on analog oscilloscope. What is this noise. Amplifiers have noise, every resistor have thermal noise, ADC have noise, digitizing etc etc.

Before you make any opinion about noise it may be good to know some basic fundamentals about noise.

Here is tiny answer from Tektronix:
http://www.tek.com/support/faqs/why-doe ... cilloscope

Here littlebit more and also more eye opening from Agilent:
http://cp.literature.agilent.com/litweb ... 3020EN.pdf
(specially look side 5 tables)

Keep it mind that RMS noise is not what you see. You see total peak to peak noise example in these Owon pictures.
But many times noise is specified and RMS.
And analog oscilloscope CRT phosphor is very good filter for random noise.

Also note that more wide band more noise. This scope front end is wide and even 500MHz is not so much attenuated. (this wide onpen front end have pros and cons. It is compromize)

More about noise also with some later tests. (but before this, look BW measurement sinewaves. There is not averaging. Just as real time single shots.

Here is only 2 - 20mV. Other V/div show visually less.



Image

Image

Image

Image
NOTE: 2mV/div have always BW limit on. It can not shut off.


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:07 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Here is one test where have been one random Hantek DSO5102B and one random SDS7102V. Also I have look many same models and it is like typical.

Owon have totally different ADC than Hantek.

Hantek have 4 pcs 2x100MSa/s ADC (8 ADC in one channel use)
Every ADC have own clock outside from ADC chip. There need be 8 separate clock phase shifted so that it make as perfect interleave as possible. Every chip and ADC is littlebit different and every clock have jitter, more or less. These clock signals are derived inside FPGA from one 100MHz oscillator. But it is not enough. It can do 800MSa/s. For 1GSa/s system clock signal need rise to 125MHz. (It is sone inside FPGA)

ADC's need be well balanced with timing and level.
It need large correcting data tables in system and more complex becouse ADC's use two different clock speed. Normal 100MHz and overclocked 125MHz with full 1GSa/s samplerate. Data tables are produced to every machine in factory. How accurate it is depends many things.
Self calibration do not touch these data tables.


Owon use totally different lot of better ADC system.
It have one chip 2x500MSa/s ADC.
ADC chip have also build in special calibration system. (oscilloscope selfcal is not this).
If one channel use.. ADC chip can set to 1x1000MSa/s mode.
It is one kind of interleaved also, there is inside chip now 2 ADC. But it do not have all disadvantages becouse still both ADC use exactly same clock outside from Chip. So there is not timing jitter due to separate clock signals. Inside chip this one clock what normally simultaneous clock both ADC's is only just inverted! If there is clock jitter, both get same.
Chip is ADC08D500 clone made by RuiFeng.
Clone specification differ only very little. It is just very little less perfomance as original. Difference is nearly marginal in this purpose.

Both scopes get exactly same signal, all is same. Both scopes have warmed enough (more than hour and then selfcalibrated.)

Hantek noise is very clear. Also Owon trig is more accurate.

Both oscilloscope have normal sampling without averaging and just as realtime single shot.

(but I do not want so much put these scopes competite becouse both have also things better than other. Pros and cons list is totally different so what is winner depends highly user needs.

Here is typical result.

Hantek is not bad if start looking more this price class members.


Image


Top
 Profile  
 
 Post subject: Re: SDS7102 "reviews" = do not believe all you see
PostPosted: Wed May 16, 2012 5:08 pm 
Offline
Site Admin

Joined: Wed May 16, 2012 7:54 am
Posts: 187
Before next tests my some personal opinions about frequency bandwidth.

It is too high! Yes it is nice to have but in turn it have also some negative effects.







More about this:

Of course always need think that many thigns are compromises. Many things have pros and cons.

If the analog front end frequency band is increased but samplerate stay same there need think aliasing. Also it need littlebit think noise.

It need keep tightly in mind that oscilloscope is for analyze (partially) uncnown signal. Not for use known signal for analyze oscilloscope. (but testing oscilloscope may also be fun)

The starting point is an unknown signal and it may have all levels and all frequency components.
Of course mostly we know "something" about signal or we have some idea what it need be. Sometimes we know nearly exactly what it need be and this we can check with oscilloscope (example if we have failed equipment and service manua with oscilloscope pictures about test points). This helps of course some amount. With real totally undefined signal it is sometimes more or less difficult (depends how accurate and sure we need analyze "truth" - in analog world there is never something exactly in practice. And all know if practice differ from theory - we have wrong theory.)

We need oscilloscopes for many kind of different works.
Some ppl repair old analog equipment, some may develop new equipments with lot of critical and fast digital signals etc. Finding rare clitches in fast digital signal is one narrow sector in common oscilloscope use.

Extra BW without real need and specially without enough samplerate is more bad than good.

Many people want "wide BW" -- oh my scope have 200 or 300 or what ever.
If there is possible to hack and modify many peoples just want more BW - without deep thinking what all it means. This is fun. Of course if people do it for himself and he know what are limits and he have good experience for use it for some special purpose it may be ok for him. BUT manufacturer perspective is different, it need be different.

If talk about real time oscilloscopes there need be enough samplerate related to maximum signal frequancy (upper limit of freq BW).
If there is oscilloscope what have repetitive sampling case is littlebit different. Good example is old HP54501A what have 100MHz BW but only 10MSa/s maximum samplerate. It can use as realtime scope upt to 2-3MHz.

Owon SDS7102 is labeled "100MHz" oscilloscope.
And it is this. And it is this as real time oscilloscope.
It do not have repetitive sampling mode at all.
It do not need it becouse 100MHz (and more) can solve with 1GSa/s (and 500MSa/s) single shot.

It is it also with deep memory and with 2 channel use!

It do it also better than most of this price class 1GSa/s competitors if look signal fidelity. Sampling guality is good (amazing good). Front end BW flatness and pulse response is extremely good in this 0 - 100MHz area.

It need think that one channel use oscilloscope can use 1GSa/s samplerate but 2 channel use it can use maximum 500MSa/s.
And it is 2 channel 100MHz oscilloscope.

With this 500MSa/s sampling speed Nyquist limit is 250MHz.

All frequancy components what ADC see over Nyquist frequency produce aliazing!

Now is good to read least this:

There is expensive oscilloscopes but theory is same.
http://cp.literature.agilent.com/litweb ... 5732EN.pdf

Also this Agilent paper give nice information.
http://cp.literature.agilent.com/litweb ... 8008EN.pdf


It need note that Owon analog BW is same if you use 1 channel or 2 channel and also if you reduce horizontal speed and or memory so that samplerate drop (This is common in all this kind of oscilloscopes.)
Also Nyquist freq drop. But still analog front end is wide open. There is not front end filter what reduce BW as samplerate drops.



Now my personal opinion.

Owon analog front end is littlebit too wide (in 4Q2011 improved versions) and roll off is too slow. It have TOO "GOOD" bandwidth.
(turn it 1 channel mode and use 350MHz signal. It is some amount attenuated (6dB or something nearly this). 200MHz is just fine. Less than -3dB. (quess what happend if you now turn oscilloscope to 2 channel mode. Do you regognize this signal now. If it was 350MHz absolutely not. What you see is just alias and nothing els. You can not tell anything about signal in time axis. Peak-Peak value you can tell if you wait and note all highest point you can collect (using example persistence).


With experienced user it is not so bad becouse he know and he can be alarmed about it and also select right tool to analyze signal and he can do also some tricks to simply detect if there is alias things in some cases.
But it is danger and leads easy to wrong diagnoses about device under test with unknown signals.

With many kind of compromise and pros and cons weighting my opinion is that Owon need reduce this BW upper limits and also maybe try do roll off more fast that it is now. (yes I know it also change FFT frequency response.)

Some maybe do not like as I say that it is better to reduce BW than try make it more wide as many "hack and modify" peoples seems think with other brand oscilloscopes. (good examples Hantek and Rigol)

My recommendation is that Owon reduce -3dB point to around 150-170MHz This is a compromise between several respects.
If there is possible to adjust BW different for single channel use and two channel use (and for FFT) then my opinion is different.
If there is possible adjust analog front end filter then it may be 250MHz/125MHz. But as long as this is not possible it need keep in some compromise value.

In this price class oscilloscope need be ok also for entry level peoples who maybe just have firsta oscilloscope and or also started just whole hobby.

I have made some tests and example with fast risetime signals corner wobbling is not very heavy with 160MHz BW using 500MSa/s After this, more wide BW start looks "not so nice". With this BW it still give very good value for user and very clean nice signal using one channel.

My opinion is that it is better balanced compromise if BW -3dB is around 160MHz than 250MHz what is maybe ok for single channel but ONLY for single channel.

If Owon reduce this frequency response so that -3dB point is around 160MHz +-10MHz it is better than now. (now -3dB point is around 250MHz +-25MHz) You may think that higher BW is better. It is NOT so simple.

(as told 250MHz is acceptable for one channel with samplerate 1GSa/s)
But this is 2 channel oscilloscope and it affect very easy bad effect if 2 channel is in use.

If people do not have good knowledge and experience he get very easy wrong measurements due to high aliasing if 2 channel is in use.

If people have good knowledge and lot of experience he maybe can use this higher BW if he is always cereful with possible aliasing. For avoid mistakes and difficulties it is better so that BW corner and roll off is well blanced with sampling speed. These linked Agilent two papers give some knowledge for understanding.


If need more, there is SDS8102 and SDS8202.
But they give only more amplerate with 1k and 10k settings becouse deep memory (>10k per channel) speed limit is 1GSa/s.


Later more tests and also some look for user interface and features.


-aghp


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 13 posts ]  Go to page 1, 2  Next

All times are UTC + 2 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
 cron
Donate Now
Donate Now



Hosted by © 2017 FreeForums.org | Create a free forum | Powered by phpBB
About FreeForums | Legal | Advertise Here | Investors | Contact FreeForums.org
Report Violation

subSilver+ theme by Canver Software, sponsor Sanal Modifiye
suspicion-preferred