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

[marmad]

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.

It's not a bug - all DSO's set their sample rate based on the sec/div setting and the memory depth - even if they don't show the rate on the screen.  At 100ms/div, a full screen of waveform data = 2 seconds of real time - how is the scope supposed to do that precisely?  It has to drop the sample rate in order to have the time and memory space to capture the data.

[rf-loop]

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.

Yes but only IF you are working with 2 channels and need nearly risetime limits or need look 2 channel signal nearly or even over 100MHz. (200MHz do not look nice if samplerate is 500Ms/s)
-------------------

to mzz:

Yes FW 2.1.1 is latest.

100ms/div and you still get 10Ms/s with 0.5k$ oscilloscope? And you are disappoint or think bug...hmm
And you get same with 2 channel. 10M+10M memory.

Try same with Rigol, Atten, Hantek, Tektronix, Agilent, LeCroy.....

Btw HantekDSO5000B with 1M memory, (only possible with 1 channel), 80ms/div (becouse next is 200ms/div)
it use 500ks/s.

Simplified: 1M memory, 1Ms/s. How long time there need before memory overflow. Yes, exactly 1 second. IF this memory is whole your display and display is divided (for simple) 10 divs. Then, every div is 1/10s and it is 100ms/div.
Now, how about if your memory is 10M (and same other things) now you can capture 10Ms/s 1 second before memory overflow.  If looking samplerate this Owon beats all other new <1k$ "1Gs/s" oscilloscopes what I know.

Oscilloscope XYZ is very good to read so there is not some many thinking that scope is fail or bug after you see all other things what scope can and what can not. This you find here:


http://socrates.berkeley.edu/~phylabs/bsc/Supplementary/Lab1/xyz_scopes.pdf

And what type of digital oscilloscope Owon is:

Figure 16. The serial-processing architecture of a digital storage oscilloscope (DSO).

It is not as: Figure 18. The parallel-processing architecture of a digital phospor oscilloscope (DPO).

This is perhaps more clear for read. Also some principle draws are maybe more clear, also this is Tektronix what have been clear leader in oscilloscopes 50 years but not anymore.  http://ecee.colorado.edu/~mcclurel/txyzscopes.pdf

[Mechatrommer]

the low end...
SDS8102 100MHz 2GS/s
SDS9302 300MHz 3GS/s

seem tempting esp SDS8102 if someone can hack to 200MHz. SDS9302 maybe pricey (cant find in ebay), if it can be hacked to 1GHz scope, that can be cool.
http://cgi.ebay.com.my/New-OWON-Digital-Storage-Oscilloscope-SDS8102-1G-s-2CH-/280714920525?pt=BI_Oscilloscopes&hash=item415be9c24d

[rf-loop]

the low end...
SDS8102 100MHz 2GS/s
SDS9302 300MHz 3GS/s

seem tempting esp SDS8102 if someone can hack to 200MHz. SDS9302 maybe pricey (cant find in ebay), if it can be hacked to 1GHz scope, that can be cool.
http://cgi.ebay.com.my/New-OWON-Digital-Storage-Oscilloscope-SDS8102-1G-s-2CH-/280714920525?pt=BI_Oscilloscopes&hash=item415be9c24d

AFAIK  SDS9302 is not yet ready (this info 2wk ago)

Mod:
Example 7102 have amazing good frequency response. -3dB point is typically over 140MHz. But it do not tell whole truth. -6dB point is far away... >>200MHz. (nearly 250MHz but this is not reliable result, it may have errors becouse matching..)
With its own probe nearly same! )

Rising -3dB point to >200MHz is maybe easy possible with HW mod in front end. BUT... it need carefully think is it wise.
Maybe better and more useful modification is reject BW! 2 channel mode it samples 500Ms/s. There leaks lot of signal over Nyquist freq to ADC. This is wrong! Also if think 1 channel mode. It is interleaved 2x500Ms/s. (Hantek is interleaved 8x 125Ms/s what is even more problematic)

it have pros and cons  if think that oh... I want 100MHz scope mod to 200MHz. Well it can do but result may be more bad than good. If mod BW (-1.5 - -3dB point)  to example 200MHz it need also do nearly brickwall filter after 250MHz. Without this 100 --> 200MHz mod imho it is also maybe clever  think how can modify this high frequency end of BW.  Front end BW is too wide... even 500MHz can leak to ADC input. It may be nice as long as you know your signal but then need no oscillocope. Oscilloscope need with unknown signal things and if there is some higher freequency components in signal it may destroy your measurements very easy. This is not so big problem with Hantek becouse its unmodified BW shape is different. It decay more fast as freq is rising after some 200MHz. Owon have more wide  analog front-end  and signal decay slowly. (or maybe Hantek have also but then it means that it is filtered after ADC what is possible but very danger if think that scope need "nice picture" builded from garbage.

[Alexey]

That's what I said.

1PCS SDS7102 Unit price @ USD400 The shipping cost from here to your country is USD75by EMS.

1PCS SDS8102 Unit price @ USD580 The shipping cost from here to your country is USD75by EMS.

Should I take SDS8102 Unit with the difference in price with SDS7102 Unit = USD180, or stay on the price / quality on SDS7102 Unit?

[rf-loop]

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

Correction:
Also this correct my answer before.

I just remember chek this. (before I did not have posibility to live check)
100ms/div and 10M selected DSO7102 display 5Ms/s with 1 or 2 channel use.
Scope memory is 20div lenght (with this setting).  (display is 15div)
with 5Ms/s 10M memory is full after 2 seconds. 20div 100ms/div is 2 second. This is logical.

[marmad]

Should I take SDS8102 Unit with the difference in price with SDS7102 Unit = USD180, or stay on the price / quality on SDS7102 Unit?

@Alexey:

No, I wouldn't get the SDS8102 - I don't think it's worth the amount of extra money they want, unless signal fidelity in the 50MHz to 100MHz range is VERY critical to you.  You don't get any added hardware features (as you do when you go from the SDS6062 to SDS7102), and the SDS7102 works well enough in that range (you can just use a single channel as needed for those speeds).

But make sure you specify you want the VGA output on the SDS7102 - it appears they've turned this into an option now - and it IS worth the extra money.

[Alexey]

Thanks for the advice.
Outcome: SDS7102 and VGA output - it would be a wise decision.

[bertchai]

Yes, buy SDS7102.   I regret a bit after I got the SDS8102.  I know it may be stretching the budget but the Agilent 100 Mhz  2 channels 2000 series is USD1400 or so and it does a lot more....   To me it should have been either SDS7102 or go for Agilent.

[mzzj]

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.

It's not a bug - all DSO's set their sample rate based on the sec/div setting and the memory depth - even if they don't show the rate on the screen.  At 100ms/div, a full screen of waveform data = 2 seconds of real time - how is the scope supposed to do that precisely?  It has to drop the sample rate in order to have the time and memory space to capture the data.

IF 50ms/div gives me 20MS/s and 100ms/div gives me only 2 kilosamples/s I call it a BUG.
See the difference?

[mzzj]

just checked out exact numbers:
using 1 channel and 10M sample memory:

50ms/div scope displays sample rate as 10MS/s, time between samples 100ns ==10MS/s
100ms/div scope displays sample rate as 5MS/s, time between samples 500us(!) ==2KS/s

Anyone else see problem with that?

[marmad]

IF 50ms/div gives me 20MS/s and 100ms/div gives me only 2 kilosamples/s I call it a BUG.
See the difference?

Either you aren't understanding how the scope operates - or you have a defective scope.  At 50ms/div (single channel - 10M depth), the scope is sampling at 10MSa/s and capturing 1 second of data (50ms * 20 divisions).  That is exactly correct:  10M / 1 second = 10MSa/s.  At 100ms/div, the scope is sampling at 5MSa/s and capturing 2 seconds of data (100ms * 20 divisions), which is exactly correct: 10Meg / 2 seconds = 5MSa/s.   At 5MSa/s, each byte in the 10Meg memory is equal to exactly 200ns of time (2 seconds / 10M) - which, again, is exactly correct - and the same thing that every DSO does.

Are you saying your scope does not capture down to 200ns at 100ms/div?  If so, your scope is defective.

[mzzj]

Are you saying your scope does not capture down to 200ns at 100ms/div?  If so, your scope is defective.

Exactly. My scope captures down to 500us (instead of 200ns) at 100ms/div.
So sample rate is 1/2500 of what it should be.

Now, can anyone check their scope and confirm if this is bug in firmware?

Or is it  only on my scope? and/or wich firmware versions are affected?

1. Set scope to measure 1khz kalibration signal, 100ms/div and 10M sample memory.
2. Hit STOP-button.
3. switch timebase to 500us/s and it should be obvious if your scope was sampling at 2.5KS/s or 5MS/s 
(4. set scope display to dots instead of vectors to check time between samples, adjust timebase to see invidual dots and determine time between samples.)

My unit works perfectly fine until i switch to 100ms/div or slower.   

[rf-loop]

mzzj:

What SDS model you have?

One time you tell that scope display 10Msa/s if setting is 10M memory and 100ms/div
other time you tell that scope display 20Msa/s if setting is 10M memory and 50ms/div

I have look with 2.1.1. version SDS7102.
same settings and 100ms/div it display 5Msa/div
50ms/div it display 10Ms/div.
I have not yet looked carefully what is real samplerate becouse just now too busy.

[mzzj]

mzzj:

What SDS model you have?

One time you tell that scope display 10Msa/s if setting is 10M memory and 100ms/div
other time you tell that scope display 20Msa/s if setting is 10M memory and 50ms/div

I have look with 2.1.1. version SDS7102.
same settings and 100ms/div it display 5Msa/div
50ms/div it display 10Ms/div.
I have not yet looked carefully what is real samplerate becouse just now too busy.

SDS7102 like you should guess
Sorry, I had wrong samplerate numbers on my first post.

[mzzj]

OK.

I just got e-mail confirmation from my scope importer/seller (rf-loop) that this bug affects probably all SDS7102 FW 2.1.1 versions.
(rf-loop, please correct me if you are not the person I bought my scope.)

No idea about other FW/HW revisions. Is there any other FW version out?

Hopefully it is something easy like wrong prefix with units and OWON is able to fix this
sooner or later.

[rf-loop]

Ok.

SDS7102 display bottom of display:

(10M memory selected)

50ms/div 10MS/s 
100ms/div 5MS/s 

with 50ms/div scope use "normal" capturing display. (capture - display - capture - etc)
with 100ms/div scop use "scroll mode" (just as plotter with moving paper.)

50ms/s and faster, all goes right.
100ms/div ans slower.  Something is wrong in FW. Looks like FW bug and IMHO bad bug. It do not use 10M memory at all in scroll mode.

If some people read this who have FW before 2.1.1  and/or Serial number below SDS71021127000
Please test this!

[rf-loop]

OK.

I just got e-mail confirmation from my scope importer/seller (rf-loop) that this bug affects probably all SDS7102 FW 2.1.1 versions.
(rf-loop, please correct me if you are not the person I bought my scope.)

No idea about other FW/HW revisions. Is there any other FW version out?

Hopefully it is something easy like wrong prefix with units and OWON is able to fix this
sooner or later.

AFAIK you buy your scope from user "aghp" Finland

[marmad]

If some people read this who have FW before 2.1.1  and/or Serial number below SDS71021127000
Please test this!

I don't have the Owon anymore - and I don't remember testing this in continuous mode.  I DO remember testing in single-shot mode at slower div settings (> 500ms) - and it seemed to work fine with the entire memory.

[duke3k]

Hey... Been Lurking and Reading this thread over the past week...  Thanks everyone for the interesting reading

(0) I just bought a new SDS7102 last week (from mortoncontrols.com  here in Denver,CO)

(1) marmad:  great review you did on the SDS7102 - thanks for all your effort even though you don't have it any more.  Why did you get rid of it ?  what did you end up with?

(2) My firmware rev: 2.1.1    Serial#SDS71021127100

(3) And yes mine is exhibiting the incorrect sample rate when doing the 100ms/div test.

duke

[mzzj]

If some people read this who have FW before 2.1.1  and/or Serial number below SDS71021127000
Please test this!

I don't have the Owon anymore - and I don't remember testing this in continuous mode.  I DO remember testing in single-shot mode at slower div settings (> 500ms) - and it seemed to work fine with the entire memory.

THANKS!
I checked mine on single-shot mode and it works perfectly indeed.
That made me think that this behaviour might have something to do with trigger.
Next obvious thing was to check "normal" trigger --> works perfect!
This "bug" needs  100ms/div or slower timebase and "auto" trigger to be seen.

So this is not nearly as bad bug as it seemed. Maybe it is intented feature instead of bug?

Just remember to use "normal" trigger with 100ms/div and slower timebases

BIG thanks to "marmad" pointing me to dig in right direction and  "aghp" who have been superb fast delivering my scope and ansvering all of my questions.

[rf-loop]

If some people read this who have FW before 2.1.1  and/or Serial number below SDS71021127000
Please test this!

I don't have the Owon anymore - and I don't remember testing this in continuous mode.  I DO remember testing in single-shot mode at slower div settings (> 500ms) - and it seemed to work fine with the entire memory.

In single shot mode, 10M memory, 100ms/dif or more slow. It capture just right speed what is display bottom of screen.
Slow speed single shot mode is not very nice to use becouse scope do not show anything when it is capturing. Then after it have captured entire memory it push captured waveform to display and stop. After it stop, signal can zoom and there is just this speed sampled data what is displayed as samplerate. 100ms/div 10M mem is now captured 5MS/s. As long as keep it stopped it can zoom in and out and samplerate is ok. But if try store single shot to USB it do not store any capture. Only picture can store. (it can push store wfrm but there is no captured data in bin file) If do store in continuous scroll mode it store only around 6kb captured data.

So it seems that continuous scroll mode and single shot also with 100ms/div and more slow have some bugs.
Together with USB store findings and these others it looks now more as just FW bug. 
But in continuous scroll, something go wrong (or then it is feature but if it is feature it need clearly tell in specs and AD's.
More it looks now like FW bug after  I tested also with USB trying store wfrm and reading this binary.

[marmad]

in and out and samplerate is ok. But if try store single shot to USB it do not store any capture. Only picture can store. (it can push store wfrm but there is no captured data in bin file) If do store in continuous scroll mode it store only around 6kb captured data.

I remember doing this when I had the Owon - it saves the data to a USB stick fine if you have the trigger set to Normal (not Auto), as mentioned by mzzj.

But, as I pointed out in my review - it's ridiculously slow saving 10M files to the USB stick.  It's much better reading them from the scope using the PC - it only takes a few seconds.  The USB host (memory stick) uses USB 1.1 protocol, but the USB device (connection to PC) used USB 2.0 - so it's many many times faster.

[House91320]

Is there a part 3 of the review coming?

[marmad]

Is there a part 3 of the review coming?

Yes - I was just swamped with work (and a deadline) the last couple of weeks - plus I had a Hantek scope to test as well - so I couldn't find the time to edit the video that I shot together.  But I think I can finally get it done today - and hopefully will have it posted by this evening.