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

[EU1]

Yesterday I wrote that the CH1 works good with DC coupling but there is a problem with AC.
Today it works good with AC coupling, but cannot switch to DC.
There is a +10...20mV offset on CH1 in AC coupling when both channels are connected to probe compensation output, probably it is caused by optocoupler leakage. When I switch CH1 to DC, waveform shifts about 20...40mV up.

So I guess that the optocoupler itself is a root cause. Also it is possible that there is a problem with corresponding digital output with controls the optocoupler or with related components (probably there are at least a transistor and resistors which also are under a cloud).

I communicating with OWON, truly hope they will propose some adequate solution for my problem.


Also I've notes some low frequency noise caused by LCD backlight supply.
It increases when I decreased the display brightness.
Noise level on both channels is the same. On the screenshot below CH1 is open (probe is not connected) and CH2 input is connected to the ground with a short wire near the BNC connector.
Thus a noise is present only if an input is open.
Probably it is not an issue, because noise peak-to-peak amplitude it less than 1mV. Just want to ask is such noise typical for other SDS7102?

[EU1]

NEC relay what you see through bottom hole near right down corner is nothing to do with CH1 and CH 2 channels signal pathway. This relay what you see can be normal relay, And also, what is real problem with this NEC relay. Nothing. Why it is not problem in Hantek or Rigol but if it is inside Owon and out from primary signal pathway it is just now problem?
Fujitsu relay is just littlebit better than this NEC model. Difference is not big. It not any problem if this NEC is just used also later in new HW version - just becouse it is not problem.
I'm very interesting to see real data about "problem" with NEC relays. Some real data, where.

It was only opinion that Fujitsu relay is littlebit better but it was just mutua. Example becouse this datasheet really read that it is for RF purposes. But I have not see any true measured data so that can compare what is exactly difference if only variable is this relay.

Looking trough this hole you can not see these CH1 and CH2 band  relays. These band relays are under front end RF shields and they are solder closed.

Of course this is a trigger input relay.
It would be great if somebody know which relays they used now in channel front-ends.

[EU1]

So I have written to OWON email from new (.hk) site.
Then I have written again..
Then again..

So, I have no answer from OWON for a month. Not even "this is your problem, go away". Just nothing.

I defenitely don't want to send oscilloscope so far wawy, because it means two or three monthes without it.
What should I do to get new FAN and plastic detail to repair to be sent to me via post mail (as it was with another people here)?

Is email "info -at -owon.com.hk" working?

I've contacted them from my gmail account and they answered pretty fast. Thus "info -at -owon.com.hk" is definitely working .

Actually I had a very good impression of the Owon support. I will describe this when my problem with CH1 will be resolved (now I have a reason to think that it will be resolved ).

[EU1]

As I've stated before, I've contacted OWON and they offered a very nice solution for my problem with SDS7102.
At Monday midday they sent me a shipment by DHL and today I've already received a box (just 3 day to ship from China to Ukraine - it is amazing) with the perfectly packed mainboard with pillars installed and the FAN .
Thus I very pleased with the OWON support. They always answered professionally and fast, even during weekend. Only a week has passed since I've contacted OWON regarding HW problem and today the problem is almost resolved (I've not replaced the board yet ).

I just made a cursory look and there are some changes comparing to my old board - it seems that a board they sent me is of newest HW revision.
I've noted that relays are again Fujitsu and TI's amplifier is used in the trigger channel (it seems that on they used some AD amp before, not sure because I was not disassembled the oscilloscope but just looked at the board through small holes on the case). And it seems that the trigger channel is significantly reworked.

Will see how it works when I will replace a board .

[saturation]

EU1, its good to hear you are getting results.  The continuous revisions to their boards can get confusing, but at least if you can do your own service, it makes for a better scope in the end.

[aghp]

Good.

If you do not have this it may help littlebit.

In Fan assembly look it is well centered. Other way later fan blades may hit this metal in some cases and make bad noise and soon destroy also bearings. (in Owon there is some workers who are loosy sometimes and fan may be not assembled correct. First time they start oscilloscope it may just be nice... later this noise may start specially if moving or turning scope when it is powered.

Relays.
In this scope NEC and Fujitsu relays are nearly as same value. Freq response and risetime and BW shape is nearly equal with both relays.
 
Sometimes availability just on right time for use in  product assembly line may be also question...

Personally I like littlebit more these Fujitsu but without real data. Just becouse I like more F than N letter.. In Owon service and tests before sell, I have not yet meet any fail related to relays if they have been NEC or Fujitsu. (Note that in this Fujitsu relay serie is also RF version... but afaik, used versions are basic versions without extra "RF" mark. Also I know some oscilloscopes are continuously in heavy professional "everyday" use - no problems in relays, no problems in fan IF fan have been OK after our extra burn in tests.
Also we have tightened this fan quality control and we do not accept for sell any scope if we find signs about problem real problem. 

[EU1]

Good.

If you do not have this it may help littlebit.

Thank you. Nice manual.
I have two questions:
1. You warned to do not open display connector if it is not exactly necessary. I've worked with many types of ZIF flex cable connectors and never had problems with them. Is OWON's connector really not tough enough or it is just a warning for people who newer worked with such kind of connectors?
2. It seems that SDS has thought-out mechanical design, with holes for screwdriver in boards and Z-plate. Looks like I can just unscrew two screws on Z-plate, then unscrew three screws from mainboard through holes in Z-plate, and then remove Z-plate with display, PSU, adapter and main boards. Then I can replace mainboard without unscrewing adapter board and disconnecting the display. Have I missed something?

In Fan assembly look it is well centered. Other way later fan blades may hit this metal in some cases and make bad noise and soon destroy also bearings. (in Owon there is some workers who are loosy sometimes and fan may be not assembled correct. First time they start oscilloscope it may just be nice... later this noise may start specially if moving or turning scope when it is powered.

Looks like exactly my case. My oscilloscope worked about 15-30 hours, thus even if FAN bearing was initially in good condition, probably now it is damaged. Anyway, now I have a new one which I will center precisely .

Relays.
In this scope NEC and Fujitsu relays are nearly as same value. Freq response and risetime and BW shape is nearly equal with both relays.
 
Sometimes availability just on right time for use in  product assembly line may be also question...

Personally I like littlebit more these Fujitsu but without real data. Just becouse I like more F than N letter.. In Owon service and tests before sell, I have not yet meet any fail related to relays if they have been NEC or Fujitsu. (Note that in this Fujitsu relay serie is also RF version... but afaik, used versions are basic versions without extra "RF" mark. Also I know some oscilloscopes are continuously in heavy professional "everyday" use - no problems in relays, no problems in fan IF fan have been OK after our extra burn in tests.
Also we have tightened this fan quality control and we do not accept for sell any scope if we find signs about problem real problem.

I will check this when get home.

[voidptr]

 
i got my new 7102 yesterday firmware seems to be 2.6
i did a quick calibration and function testings,  it seems to be really great, around 0.5mv of noise on channel ...

but i just found this this morning ....   
darn ! i have 1 stuck on green pixel !
is this warranty ? 

is this common problem ?

i'm going to write to my ebay seller ...


EDIT:  Added pic of stuckOn pixel, left bottom screen, just below the graticule

[aghp]

I have two questions:
1. You warned to do not open display connector if it is not exactly necessary. I've worked with many types of ZIF flex cable connectors and never had problems with them. Is OWON's connector really not tough enough or it is just a warning for people who newer worked with such kind of connectors? 

Exactly this. If write some help text it need take care about many different level peoples (also becouse privately I know some peoples who I was helping with this and this also was in mind when I write littlebit "over" carefully. If people have not knowledge about this kind of connectors and never handled these connectors,  it can easy break.

2. It seems that SDS has thought-out mechanical design, with holes for screwdriver in boards and Z-plate. Looks like I can just unscrew two screws on Z-plate, then unscrew three screws from mainboard through holes in Z-plate, and then remove Z-plate with display, PSU, adapter and main boards. Then I can replace mainboard without unscrewing adapter board and disconnecting the display. Have I missed something?

I think you have not missed anything, but also if unscrew adapter board it do not need disconnect display.

Original old version was lot of different. Main board different, front panel big board and small boards different and adapter board different and even power supply littlebit different.
Also old version did not have ribbon cable between main board and front panel.
Also it was originally for front end  (service) adjustment with minimum work time using service cable between adapter and main board. (original old version did not  have ribbon cable between front panel and main board...

My way is only one way, there are others maybe better or just parallel ways.

[aghp]

i got my new 7102 yesterday firmware seems to be 2.6
i did a quick calibration and function testings,  it seems to be really great, around 0.5mv of noise on channel ...

but i just found this this morning ....   
darn ! i have 1 stuck on green pixel !
is this warranty ? 

is this common problem ?

i'm going to write to my ebay seller ...

This is first time I hear pixel problem (in Owon SDS)
These Chimei-Innolux panels have been very good quality.
It is NOT common problem in Owon oscilloscopes.

(btw. Tekway - Hantek use these same TFT panels but different size)

[T4P]

Pixels die sometimes really . I had a Samsung LCD screen die completely within 2 years 
The samsung LCD panel i am typing had a dead green pixel from the start , but since it's a high reso one i usually can't see it 

That said , contact your ebay seller .

[voidptr]

ZAP
darn ! i have 1 stuck on green pixel !

This is first time I hear pixel problem.
These Chimei-Innolux panels have been very good quality.
It is NOT common problem.

I put a pic of it in the original message i'm gonna take a better pic tomorrow !

[voidptr]

i put my scope to some test tonight and i found 2 things   

bad first i discover i have a faulty probe,  when the probe lays flat no problem, it also bends in 3 directions, but a really small bend in the other and i got full noise ....
here a video of that ...
http://s1114.photobucket.com/albums/k539/n45w73/dev%20stuff/?action=view&current=MVI_9868.mp4

good new  it is fun to have a scope ! 
here a small video of PWM of an hbridge in red  and voltage of the neg pole of the motor in yellow.
it is a small low standard no torque 50mA motor
http://s1114.photobucket.com/albums/k539/n45w73/dev%20stuff/?action=view&current=MVI_9862.mp4

strange traces  ,  a bit because there was no load on my cheap motor so i choose the worse frequency  of pwm and also because of the "fast" refresh of the scope who stacks many trace together on the screen ...
that bridge will need a lot more study for sure 

[saturation]

Stuck pixels maybe a mechanical issue, rub around it or press on it with your finger and see what happens.

See more here:

http://en.wikipedia.org/wiki/Stuck_pixels#Stuck_versus_dead_pixels

ZAP
darn ! i have 1 stuck on green pixel !

This is first time I hear pixel problem.
These Chimei-Innolux panels have been very good quality.
It is NOT common problem.

I put a pic of it in the original message i'm gonna take a better pic tomorrow !

[EU1]

I've replaced the board but had faced with another issue: there ia a large interferences caused by backlight LED DC-DC converter. Earlier on old board I've discovered some weak interference from backlight (https://www.eevblog.com/forum/product-reviews-photos-and-discussion/review-of-owon-sds7102/645/), but it disappeared when I connected a 50 Ohm BNC terminator or probe, even when I connected a 10kOhm resistor or any signal source with any output impedance there was no interference, so it did not affected measurements because it was visible only when nothing is connected to inputs.
I've compared peak level of noise of my oscilloscope with screenshots which other people shared on the Internet and found it is about the same. Thus I've considered that it is OK, now I have reasons to gues that it is not OK .

With new mainboard interference is much larger and they are enlarged when I connect terminator to input. I've installed the old mainboard again and see the same very small interference visible only on open input. Thus it is not due to incorrect installation - because I swapped boards few times and always I get negligible interference with old board and large interference with the new one.

The problem appears when powered from battery or from AC mains. So it is definitelly not a PSU problem.


I've made screenshots with new boards. X10 probe configuration is selected for both channels. I just forgot to switch it to X1, thus V/Div readings on the screen should be divided by 10 to get the actual input sensitivity.
An interference on both channels is the same. I leave CH1 open (no probe connected) and connect 50 Ohm BNC terminator to CH2 to show that 50 Ohm terminator increases an interference.

NewBoard_1.png: CH1 is open, CH2 shorted to ground through 50 Ohm BNC terminator, brightness is 0%.
There RF noise extremely low for 2mV/Div sensitivity. Seems that it is even much lower than a noise on the old board.
But there is a terrible 30kHz interference from backlight LED driver.

NewBoard_2.png: CH1 is open, CH2 shorted to ground through 50 Ohm BNC terminator, brightness is 100%.
30kHz noise is much lower comparing to an interference with 0% brightness, but it is still present and spikes occasionally appears.

NewBoard_3.png: FFT for CH2 which is shorted to ground through 50 Ohm BNC terminator, brightness is 0%.
30kHz interference with harmonics is clearly visible.

NewBoard_4A.png, NewBoard_4B.png: Even with low sensitivity 30kHz interference is clearly visible. With 2V/Div interference is higher than with 1V/Div. Thus interference comes rather through power line than through input.


It doesnt look like a mainboard issue, because there are a lot of DC-DC converters on the adapter board and on the mainboard itself and only an LCD backlight step-up causes interference. I guess that a mainboard of new revision is just more sensetive to low-frequency ripple on supply voltages rails and a real problem is a LCD backlight step-up on the adapter board or new mainboard has some compatibility issues with old adapter board (I'm not sure that OWON changed adapter board, this is only an assumption). There are a lot of changes in analog and power management parts on the new board comparing to my old one.


Thus I would like to ask people who have SDS7102 - have you any signs of 30kHz noise with visible low display brightness with open input (no probes connected)?
Because I have this noise on two mainboards of different revisions - on old 1209*** it is wery weak and produces no visible impact on measurements, but it is still visible when nothing is connected to input BNC, especially on FFT. On new board (1211****) interference level raises when I connect a probe. I've recorded a video, you can see that on 2V/Div with an input is connected to probe compansation output an interference just makes any precission measurements impossible (while on 1V/Div it is much lower):



And the second strange thing - I've noted some strange behavior of external synchronization input when it is configured to trigger on falling signal edge, on old mainboard the trigger works fine, so owners of 1211**** devices, please check check is this issue appears on your devices. On the video both ext trigger input and a channel input are connected to probe compensation output and an AC coupling is selected for trigger and channel input itself:

[EU1]

I think you have not missed anything, but also if unscrew adapter board it do not need disconnect display.

When I disassembled the unit I've discovered that there is a thing that I've missed - the BNC connector on the real panel prevents removing the mainboard without unscrewing the adapter board. So the way described in your manual probably is a simplest possible. At least I did not found any better way .

[aghp]

This disassembly is possible also other way but it need small trick with Z-plate.

About backlight control.

It is described here (old) but sorry, most txt is with finnish lang.
After this there is small change in PWM frequencies and later they also have changed current control reistor and typical now is 1R0.  No need anymore change this resistor. (older some systems have been too dim becouse current was too low due to wrong resistor and due to wrong thinking in design = it need be current, not voltage what need adjust ok so that it meets Chimei-Innolux specifications for driving back light LED's.

Here this old (obsolete) pic
( it is explained there with finnish language what is 2nd difficult after chinese what is number 1.))

[EU1]

This disassembly is possible also other way but it need small trick with Z-plate.

About backlight control.

It is described here (old) but sorry, most txt is with finnish lang.
After this there is small change in PWM frequencies and later they also have changed current control reistor and typical now is 1R0.  No need anymore change this resistor. (older some systems have been too dim becouse current was too low due to wrong resistor and due to wrong thinking in design = it need be current, not voltage what need adjust ok so that it meets Chimei-Innolux specifications for driving back light LED's.

Here this old (obsolete) pic
( it is explained there with finnish language what is 2nd difficult after chinese what is number 1.))

On my board 1 Ohm resistor is soldered and the display is rather too bright for me even with 0% brightness.
Brightness adjustment works. The problem is a 30kHz noise which comes from the backlight step-up to input channels. I know that you worked with many SDS7102 units, have you seen any 30kHz noise with brightness less than 100%?

[voidptr]

Thus I would like to ask people who have SDS7102 - have you any signs of 30kHz noise with visible low display brightness with open input (no probes connected)?
Because I have this noise on two mainboards of different revisions - on old 1209*** it is wery weak and produces no visible impact on measurements, but it is still visible when nothing is connected to input BNC, especially on FFT. On new board (1211****) interference level raises when I connect a probe. I've recorded a video, you can see that on 2V/Div with an input is connected to probe compansation output an interference just makes any precission measurements impossible (while on 1V/Div it is much lower)

i got Vp 0.640mV of that 30kHzon both channels with backlight 0%
a little bit less noise at 80%
and i got Vp 0.400mV at 100%

with ch1 probe grounded on itself and put away (didn't test for ch2)
i got Vp 1.280mV on ch1  30kHzon with backlight 0%
i got Vp 0.880mV on ch1  30kHzon with backlight 100%

i pick noise, with probe grounded , from the top of the screen  around Vp10mV

[rf-loop]

i pick noise, with prove grounded , from the top of the screen  around Vp10mV

Owon backlight produce littlebit noise specially if other than 100% brightness.
Also after some peoples talk about LCD backlight audible high pitched noise what can listen with human ear Owon made small changes.
It was really pity they did this change. (change was made in FW changing EN signal (PWM))  Now it can dim only littlebit. 0% is really bright now. 0-100% is now nearly as 80% - 100% before this change.

Right solution is in HW but then it do not help these peoples who allready can listen this "noise" becouse some very rare unit produce it in some component as some kind of interference/resonance.  (propably one inductor).

With open CH input I can find around <1mV around 30kHz noise.  But who measure anything with open input.  Do it really affect real measuremets if you use the normal working practices, and good measurement practices. Of course, it require also some experience. In the laboratory, one of the most difficult things are to know error sources and be able to avoid them. It helps the most extensive experience and adequate knowledge. It must also be able to assess what kind of things are really important and what not.But yes, it need littlebit care if do some very sensitive measuremets that do not pollute tests with enviroment RFI/EMC

Also oscilloscope itself is some source and it also radiate noise field. But lot of care need if lab have schwitch mode small things as example energy saving lamps, laptop, laptop power supply, and many other this kind of things in today electronic equipments.  Example if take normal Tektronix analog 2235 oscilloscope (what is not at all bad). It is still terrible emi noise source and still it is all inside aluminium case. Its power supply pollute oscilloscope inside and lab room outside.  Also it radiate lot of to enviroment.  Becouse this, they have made also MIL version about same scope with better EMI shielding, even CRT front have metal grid. 

Probe is as antenna. If connect GND to center tip this is nice  near field loop antenna. It is grounded ONLY if think pure DC. It is not grounded at all if look it with RF eyes.

It is "poor man" easy made near field RFI probe.

Connect probe GND wire to probe tip it make D shape loop. Turn probe to 1:1.
Try example with laptop, try find example backlight inverter noise. I move it around laptop and Lenovo T series near right bottom corner of display it give around 2000mV peak to peak around 58kHz.

For many purpose better loop antenna can easy do.  50 ohm  loop antennas for near field proping of noise is easy made using semirigid coaxial (example from dismantled equipment "waste").

These are very handy for example spectrum analyzers.
(these 50 ohm loops are not very good for oscilloscope what have 1M input)

Here is random example picture collected from internet.

[EU1]

voidptr, rf-loop: thanks for your investigations, so the interference presence seems to be a normal case.
It is really strange, because it can be easily avoided by reducing a dimming frequency to be less than 2kHz.
30kHz is a too high dimming frequency for G5126 step-up converter and this is even explicitly defined in the datasheet (http://www.gofotech.com/upfiles/201026105539.pdf):

LED dimming control is pos sible by forcing PWM signal to EN pin.  When the power-on or start up with EN pin, softstart function works, however, after that, if the EN pin is set as  “L” and set EN pin  “H” again during the shutdown delay time, softstart function is disabled and starts up fast to normal mode, therefore 200Hz to 2kHz PWM signal is standard. By the EN pin input, LED turns on and off. Average LED current varies depending on the duty cycle of EN input. Too high frequency PWM signal is not ffective because of its delay.

Of cource, lower frequency may cause acoustic noise to be audible, but but this is only 0.2A DC-DC converter, such low power converters usually stay silent on any frequency if a type of inductor is chosen properly.

[EU1]

Connect probe GND wire to probe tip it make D shape loop. Turn probe to 1:1.
Try example with laptop, try find example backlight inverter noise. I move it around laptop and Lenovo T series near right bottom corner of display it give around 2000mV peak to peak around 58kHz.

Of cource for all measurements with grounded input I've used a 50 Ohm BNC terminators or just a short piece of uninsulated wire inserted into the BNC connector.

As for inverters, I placed a probe near my laptop's backlight inverter and get something about 20V RMS AC voltage . Because there is even no EMI shild on inverter board, just a piece of plastic covers switch transistors and HV transformer.

Owon backlight produce littlebit noise specially if other than 100% brightness.
Also after some peoples talk about LCD backlight audible high pitched noise what can listen with human ear Owon made small changes.
It was really pity they did this change. (change was made in FW changing EN signal (PWM))  Now it can dim only littlebit. 0% is really bright now. 0-100% is now nearly as 80% - 100% before this change.

The PWM duty cycle for 0% brightness is about 35%. I guess that it is not easy to get a low brightness with 30kHz PWM.
If the frequency will be reduced to 2kHz or less the brightness will be much more controllable and a soft-start function will operate making a G5126 much more "silent".
Really hope that OWON will correct this in new FW.

And, guys who have newer HW (probably all boards with FW 2.6.2), please check the external trigger channel operation:
Just attach probes to CH1 and EXT TRIG inputs (ensure that the switch on the trigger probe is in the X1 position) and connect both probes to the PROBE COMP output. Then in CH1 menu select AC coupling, in trigger menu select AC coupling, trigger source EXT/5 and triggering by falling edge (H4 button - "Slope"). Then select 1us/div sweep and by adjusting trigger level with the knob from -3V to +3V see is the trigger fires exactly at signals edge.
On my the following video you can clearly see that the trigger fires with a delay from hundreds of ns to tens of us:

I just want to know is it a specificity of my new board (1211****) or all new boards work with falling edge trigger with such delay. On old board (1209***) I was not noted this.

[rf-loop]

The PWM duty cycle for 0% brightness is about 35%. I guess that it is not easy to get a low brightness with 30kHz PWM.
If the frequency will be reduced to 2kHz or less the brightness will be much more controllable and a soft-start function will operate making a G5126 much more "silent".
Really hope that OWON will correct this in new FW.

In early version, this LCD backlight control works just fine!

It need return back. Why they make change to good. This is interesting question.
It is good that manufacturer listen customers. They of course need collect opinions and reports about bad things, bugs etc.
 
But there it go wrong if too fast make changes just as in "panic".  Many peoples want so many things. There need be calm in manufacturer side and do ONLY changes what are really examined and tested and studied with experienced professionals before change. Also need carefully think if change is ok. This LCD backlight is just good example about wrong reaction after customers feedback. It need study in manufacturer laboratory first and really think what is good solution (or do it need anything) but control better example inductors quality or some other component.

So, I hope Owon return this LCD control just back so that it works like it works in early "not improved" HW version but using right (1R0 - 1R2) current monitor resistor value and original part of FW for control.

Second. Why they have make changes to HW now? 1149xxx - 1209xxx is best HW and it works well. Why change good?

It is nearly as "repairing" (or many times adjusting)  unit what have not any fail or problem.

[voidptr]

here an update with my Owon scope's problems

bad stuck on green pixel :

i got update from my ebay seller and also from Owon's service,  1 bad pixel is NORMAL thing and is not a defect...
my seller said i need a bunch of them to be considered has a defect...

so people be aware of that  you may end up with a bunch of them 

bad probe:
for my probe defect, they said they will ship me a replacement ...

[T4P]

here an update with my Owon scope's problems

bad stuck on green pixel :

i got update from my ebay seller and also from Owon's service,  1 bad pixel is NORMAL thing and is not a defect...
my seller said i need a bunch of them to be considered has a defect...

so people be aware of that  you may end up with a bunch of them 

bad probe:
for my probe defect, they said they will ship me a replacement ...

Yup . Samsung does the same thing , they wouldn't change my panel when i asked them about a few dead pixels on a full hd 37" set , only until the whole screen died . Maybe your lcd was manufactured quite a long time ago , some new LCD's have some dead pixels and it takes time for them to appear  eventually .