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

[Muxa52]

Hello people!
Oscilloscope SDS7102 1143. The second batch, MB V1.1 2011/9/30.
Have a problem. After pressing the power button and display the logo screen, oscilloscope clicks relay and the screen goes blank and device reboot. Trouble arises with mains power and battery life. External inspection of PCB has not revealed any obvious damage. The power supply is apparently working well.
There are any ideas?

[TomC]

Have you contacted your dealer or Owon? What you describe doesn't sound good. The good news is that based on the MB date your scope is still under warranty.

[TomC]

The following is a summary of my interpretation of the attached PDF.

Peak Detect Mode is supposed to alter the original signal so that the higher and lower voltage values are emphasized when the signal is displayed. This is typically done by examining a given number of samples and selecting just the higher and lower voltage values. Only these selected values are stored in the display memory. One advantage of this technique is that it is possible to store a larger amount of relevant information (the highs and lows) in a shorter span of display memory. So if display memory is scarce, using Peak Detect Mode allows the capture of high and low voltage levels that otherwise would be missed. From this perspective Peak Detect Mode can be viewed as a form of compression.

To try to uncover some of the inner workings of the SDS7102, I captured some signals using both Sample Mode and Peak Detect Mode. On the first set of captures, #1 - #3a, I tried to set up the worst possible environment for Sample Mode to see how much of a difference using Peak Detect Mode would make. The signal generator was set to as close to 100MHz as I could get and the scope was set to 50µs/Div with a 100K acquire length. With these settings the sample rate is 100MS/s, exactly the same frequency as the input signal. This yields a capture window of: 100,000/100,000,000 = 0.001s or 1ms. At 50µs/Div this results in 0.001/0.000050 = 20Div of display. The sampling time at 100MS/s is 1/100,000,000 = 0.00000001s or 10ns. Ten nanoseconds is also the length of a cycle of  the 100MHz input signal. As a result, since the two are nearly synchronized, the samples should happen at the same point of each sine wave resulting in pretty much a straight line.

That's precisely what happens in Sample Mode. However, under the same conditions, Peak Detect Mode manages to detect a peak every 10ns. How can this be possible? The only answer that I can think of is that the sample rate is much higher than 100MS/s. I suspect that for Peak Detect Mode the SDS7102 uses the maximum sample rate, in this case 1GS/s. I also suspect that the memory is segmented into Display Memory, in this case 100K, and capture memory, where the 1GS/s are stored. From there I would think the firmware picks the high and low samples and stores them in the Display Memory.

The second set of captures, #4 - #6a, were obtained at 50µs/Div with a 1M acquire length. With these settings the sample rate is 1GS/s, plenty fast to properly display a 100MHz signal. So there aren't many surprises on the way Sample Mode and Peak Detect Mode behave in this case.

On the side by sides the signal on the left is Sample Mode and the signal on the right is Peak Detect Mode. The captures were obtained at 50µs/Div and the scope was stopped (#1 & #4). The remaining images were obtained by changing the time base while the scope was stopped.

[lemon]

Peak Detect Mode is supposed to alter the original signal so that the higher and lower voltage values are emphasized when the signal is displayed. This is typically done by examining a given number of samples and selecting just the higher and lower voltage values. Only these selected values are stored in the display memory. One advantage of this technique is that it is possible to store a larger amount of relevant information (the highs and lows) in a shorter span of display memory. So if display memory is scarce, using Peak Detect Mode allows the capture of high and low voltage levels that otherwise would be missed. From this perspective Peak Detect Mode can be viewed as a form of compression.

Very good analysis with perfect documentation for how the peak mode is working for SDS7102. Thanks for that, TomC!
For this reason at the Peak Detect Mode the intermediate information between max and min is missed and appears only with some inclines lines.

[rf-loop]

The only answer that I can think of is that the sample rate is much higher than 100MS/s. I suspect that for Peak Detect Mode the SDS7102 uses the maximum sample rate, in this case 1GS/s.

SDS series true ADC samplerate is allways 1GSa/s for single channel and 500MSa/s + 500MSa/s for two channels in use. This is what ever is displayed samplerate or samplerate in sampling buffer memory.  ADC itself samplerate is not changed, never.

If scope samplerate is 100MSa/s in normal sampling mode it just take every 10'th sample and forward it to memory. If samplerate is 1MSa/s it just take every 1000's sample from ADC and forward it to memory. Other samples just dropped out.

In peak detect sampling mode it works some amount different.

If there is now 100MSa/s samplerate. (ADC works 1x 1GSa/s or 2x 500MSa/s. )
Very roughly and inaccurate principle: In case where it use 1GSa/s (one channel) it "listen" every sample what is coming from ADC every one ns. But it only select out min max pair (and becouse now 100MSa/s there is min-max result after every looked block (dependent of samplerate how long this block is.) 

If example samplerate is 1MSa/s (take one sample and drop out 999) and if it calculate  every 1000 samples block average, we have simple machines principle High-Resolution capture mode.
Of course perhaps it need also add noise to signal before ADC.  Pity High_res is not implemented in this scope.

[TomC]

In peak detect sampling mode it works some amount different.

If there is now 100MSa/s samplerate. (ADC works 1x 1GSa/s or 2x 500MSa/s. )
Very roughly and inaccurate principle: In case where it use 1GSa/s (one channel) it "listen" every sample what is coming from ADC every one ns. But it only select out min max pair (and becouse now 100MSa/s there is min-max result after every looked block (dependent of samplerate how long this block is.) 

Thanks rf-loop for clarifying some of this!

What I'm wondering is how does it proceed to select the min/max pairs. Since it can't look ahead of time, the blocks have to be stored someplace for processing after the fact. I was thinking segmented memory like I mentioned in my post, but this can be implemented in a number of ways. For all I know it could just be temporary memory that is discarded after each block is processed using alternating buffers. Anyway, if you have any information on how Owon does this it would be interesting to know!

[AndrejaKo]

I could be wrong, but I think that there's a good description of how that works in TI's datasheet for the ADC pattern used in Owon. I don't have the cloned and original part numbers right now, so I can't check.

[Muxa52]

Have you contacted your dealer or Owon? What you describe doesn't sound good. The good news is that based on the MB date your scope is still under warranty.

Thanks for your response!
No, I have not written the seller, and yes, the 3-year warranty is not over yet. I will try to take advantage of warranty, if I can not repair this device yourself
In fact there are only 2 options:
- Firmware is was destroyed, and I could not recover it.
- Reset/reboot gives the supervisor of power and I want to try to find the source of the fault dips on supply voltages.
Who can show where there is a reset circuit?

[lemon]

Have you tried to download it the firmware again to the scope or there is no any communication with this?

[Muxa52]

Have you tried to download it the firmware again to the scope or there is no any communication with this?

Thanks you for quickly replay!
Yes, I tried to download the update of FW, but the scope periodically reboot. The device driver appears and disappears, and there is no way to stop this process.
Now I'm learning the power supply circuit of scope, which is located on the "Adapter plate". 3.3V stabilizer works well and correctly. There are doubts about the work of the regulator to 5V. The voltage measured at the terminals 34 and 35 of the connector U1 is about 5.4-5.5V, which I think is not correct.
In an attachment of the circuit containing the 5V regulator circuit and a current source for LED backlight (connector U12).
Unfortunately, I now can not activate this card without LED backlight - because the risk of destruction of its stabilizer. As soon as I will prepare a dummy load, I will write here about the results.
Thank you all for participating and forgive my Googlish

[Muxa52]

I'm turned on "Adapter plate" without the motherboard. And yes, the regulator U13 gives 5.45V. And this is correct. In an attachment, a description of the chip. The reference voltage is 0.8V. Feedback resistors R18 = 10k, R19 + R21 = 56k +1.8 k. Thus, the calculated output voltage of stabilizer = 5.424V.
I believe that the "Adapter plate" is working properly.
Tomorrow I will try to explore on the motherboard. The problem apparently lies in it

[TomC]

Muxa52,

Thanks for reporting your findings!

Unfortunately I don't have any firm ideas of what your problem may be or how to fix it, just a vague suspicion that it may be firmware related.

Hope that you are successful in your efforts!

Please let us know what you find. 

[rf-loop]

Muxa52
"Yes, I tried to download the update of FW, but the scope periodically reboot. The device driver appears and disappears, and there is no way to stop this process."

In this kind of situation DO NOT upgrade FW. (also 99% for that this is not FW problem. This is not windows PC what OS need suddenly reisntall or update after update becouse system destroy itself)  If there is FW inside and check sum pass, it is very very rare it have some wrong bits.
If it reboot at wrong time during FW upgrade procedure is running it may destroy your scope to state where any kind of FW loading is impossible.
During FW upgrade procedure, oscilloscope need run rock steady until procedure itself reboot scope for finalize FW upgrade.
It is serusly writed: Do not shut off oscilloscope during FW upgrade procedure!
And it is same if oscilloscope have HW failure what shut system down (and perhaps reboots).

Follow service manual simply diagnose (troubleshoot) instructions and take contact to Owon after you can exatly explain what is symptom.

One question, (I suppose you have battery).

If you take battery out and run without battery, do this reboot still happend just same as with battery.   If you have battery, do it reboot just same if you run scope with battery and mains power disconnected.

Btw, you name it 5V regulator on the Adapter board. It is (nominal) 5.5V supply what can do it from 6 - 8.4V. (Main PSU nominal 8.4V is in practice 6V to 8.4V)
Scope need (adapter board connection to main board) -7.6V, 5.5V, 3.3V  and 8.4V (it can vary around 6 to 8.4V(max). 
You can run it without adapter board and PSU with these voltages and look if it still reboots. It do not need display.

Also on the main board there is some regulators and SMPS circuits.

Reason may be what ever. Not only power lines. There is databus things, there is BGA chips where may be hidden bad soldering, there may be bad trace on the board some layer. Example I have seen one bad trace on the adapter board. Before solder mask there have been small damage on the copper foil. It have worked one year but in random situation, temperature or some mechanical stress change and this one trace is ok or broken. Under microscope I find this very very small break. In rare case reason may be what ever.

If failure can isolate so that you know if it is PSU, Adapter board, Main board, front banel board or sidepanel board or TFT module it is enough.  This is not diy scope, exept if you change it to diy scope. It is product what have normal 3 year factory warranty.

First contact your seller. If seller is real seller who also have after sales customer care, he will help you. If you have purchased it from "box shipper" seller who only reship boxes, perhaps he can not help you.
If seller can not help  then contact Owon for instructions how to do and perhaps for free shipping to manufacturer repair center using normal Owon RMA procedure or other instructions how to do.

Exept if you have destroyed warranty or purchased it from markets what do not cover international warranty. Also in this case, you can always get (free or with payment)  spare parts or quite cheap repair service from Owon. 

But, in what situation exactly it reboots.

Is it possible to see any dropout in voltages. Drop outs may be so short that normal DVM can not see it at all.

Btw, I can see metal pillar on the main bord near CH1 BNC. Did you have extra noise specially in CH1. Later this pillar have changed to plastic pillar. (it was other case before this later "noisy GND" issue) (only pillar what is very near CH1. Owon have send instructions and these pillars to many end users and also for Owon services  if example  seller have Owon service)

First, with this kind of failure, try get this scope to Owon RMA procedure. It may be also that Owon is interest to see what have happend becouse of course they also want know failures better for develop better)

[lemon]

Have you tried to download it the firmware again to the scope or there is no any communication with this?

Thanks you for quickly replay!
Yes, I tried to download the update of FW, but the scope periodically reboot. The device driver appears and disappears, and there is no way to stop this process.
...

Before you examined that all said rf-loop do a check to your usb cable...

[Muxa52]

Muxa52
"Yes, I tried to download the update of FW, but the scope periodically reboot. The device driver appears and disappears, and there is no way to stop this process."

In this kind of situation DO NOT upgrade FW. (also 99% for that this is not FW problem. This is not windows PC what OS need suddenly reisntall or update after update becouse system destroy itself)  If there is FW inside and check sum pass, it is very very rare it have some wrong bits.
If it reboot at wrong time during FW upgrade procedure is running it may destroy your scope to state where any kind of FW loading is impossible.
During FW upgrade procedure, oscilloscope need run rock steady until procedure itself reboot scope for finalize FW upgrade.
It is serusly writed: Do not shut off oscilloscope during FW upgrade procedure!
And it is same if oscilloscope have HW failure what shut system down (and perhaps reboots).

Follow service manual simply diagnose (troubleshoot) instructions and take contact to Owon after you can exatly explain what is symptom.

Yes, I understand that. I even not tried to run the update utility.

One question, (I suppose you have battery).

If you take battery out and run without battery, do this reboot still happend just same as with battery.   If you have battery, do it reboot just same if you run scope with battery and mains power disconnected.

Yes, scope reboots as on battery power only, and powered by the PSU (with/without battery)

Btw, you name it 5V regulator on the Adapter board. It is (nominal) 5.5V supply what can do it from 6 - 8.4V. (Main PSU nominal 8.4V is in practice 6V to 8.4V)
Scope need (adapter board connection to main board) -7.6V, 5.5V, 3.3V  and 8.4V (it can vary around 6 to 8.4V(max). 
You can run it without adapter board and PSU with these voltages and look if it still reboots. It do not need display.

By calculation I have determined that this source should produce 5.422V at a current of up to 5A (measured value of 5.45V). AOZ1094 chip has a minimum current limit of 6A.
Thank you for your comment about voltages . It adopted me into thinking that PSU and Adapter plate and are working properly and well

Also on the main board there is some regulators and SMPS circuits.

Yes, I plan to gather documentation on chips, located on the MB. At a minimum, the chip of analog supply TPS76850 has a PG (Power good) output. This signal can be used to reset the CPU.
I'm working remotely and do not have in home any appliances, except for the multimeter. I have to go to the ofice and take an oscilloscope for monitoring this and other signals.

Reason may be what ever. Not only power lines. There is databus things, there is BGA chips where may be hidden bad soldering, there may be bad trace on the board some layer. Example I have seen one bad trace on the adapter board. Before solder mask there have been small damage on the copper foil. It have worked one year but in random situation, temperature or some mechanical stress change and this one trace is ok or broken. Under microscope I find this very very small break. In rare case reason may be what ever.

If failure can isolate so that you know if it is PSU, Adapter board, Main board, front banel board or sidepanel board or TFT module it is enough.  This is not diy scope, exept if you change it to diy scope. It is product what have normal 3 year factory warranty.

First contact your seller. If seller is real seller who also have after sales customer care, he will help you. If you have purchased it from "box shipper" seller who only reship boxes, perhaps he can not help you.
If seller can not help  then contact Owon for instructions how to do and perhaps for free shipping to manufacturer repair center using normal Owon RMA procedure or other instructions how to do.

First contact your seller. If seller is real seller who also have after sales customer care, he will help you. If you have purchased it from "box shipper" seller who only reship boxes, perhaps he can not help you.
If seller can not help  then contact Owon for instructions how to do and perhaps for free shipping to manufacturer repair center using normal Owon RMA procedure or other instructions how to do.
But, in what situation exactly it reboots.

Is it possible to see any dropout in voltages. Drop outs may be so short that normal DVM can not see it at all.

Unfortunately I can not use the standard free procedure, since the device was purchased as a consumer device, and do not miss the free parcel customs. In any case, the cost of delivery will not stop me. I was stopped by a great time of delivery parcel. Now I plan to try to repair the unit yourself.

Btw, I can see metal pillar on the main bord near CH1 BNC. Did you have extra noise specially in CH1. Later this pillar have changed to plastic pillar. (it was other case before this later "noisy GND" issue) (only pillar what is very near CH1. Owon have send instructions and these pillars to many end users and also for Owon services  if example  seller have Owon service)

Yes, I read about the noise in channel 1, but they are not great, about 5 mV. I did not change the pillar, because I do not want to measure very weak signals. The measurement of digital signals, this noise does not affect

First, with this kind of failure, try get this scope to Owon RMA procedure. It may be also that Owon is interest to see what have happend becouse of course they also want know failures better for develop better)

Yes, I will try to contact the Owon, but it requires some effort from me. I am well read and understand technical English, but absolutely do not have the practice of spoken language, and do not feel the style of a written text. For this reason, I am embarrassed of me English texts and The Google translator does not help me.
On the other hand, I, as a developer of industrial electronics, I get a lot of experience studying other people's creations. Regarding this device can tell you that at least PSU developed highly economically, but it is possible that the battery charging circuit designed for extremely basic level. This does not help extend battery life. Uses a simple current limiter, but the battery like in the beginning of charge the high-current charge, and only after the main charge of high current (with temperature control of battery) is possible charge current limit.
In the scheme of supply LED backlight, stabilizer G5126 possible breakdown of high voltage (greater than 28V) in the case of a breaking LED  backlight and open-circuit load. Developers forgot one resistor in parallel load circuit. Never turn on "Adapter plate", if not connected LED backlight. There is a probability of failure of the G5126.

Have you tried to download it the firmware again to the scope or there is no any communication with this?

Thanks you for quickly replay!
Yes, I tried to download the update of FW, but the scope periodically reboot. The device driver appears and disappears, and there is no way to stop this process.
...

Before you examined that all said rf-loop do a check to your usb cable...

USB cable in order and the driver is installed. I finally came to the conclusion that there is a hardware reset signal, which is produced by an over-current analog circuits of the motherboard, you turn them on. On MB I counted on one pulse and 3 linear voltage regulator. It is possible that the board has still some sort of a scheme of testing analog voltages, which provides a hardware reset signal. Most likely out of order one of the decoupling capacitors or (worse) chip ADC (MT2002). Now I plan to explore this part of the scheme.
Thank you all for participating in the discussion. I will write about their results.

[Muxa52]

Muxa52,
Thanks for reporting your findings!
Unfortunately I don't have any firm ideas of what your problem may be or how to fix it, just a vague suspicion that it may be firmware related.
Hope that you are successful in your efforts!
Please let us know what you find. 

Thank you for your participation. I read online about a case where such fails after saving the measurement results. But just as I was reading about a few cases where this effect was associated with a malfunction of the power supply. I regard the probability of FW / HW failure, as 50/50%.
Here's a video of such a failure (from 0.40 '), but I have no picture interference.
http://youtu.be/e4iRH-DENBw
Yes, I'll write about the results of my research.

[BBAAHHOO]

Muxa52     
   -       
 And trying to debug via JTAG? Or upgrade FW?

http://radiokot.ru/forum/viewtopic.php?f=10&t=70052&sid=c1c18115b263a82e84503efc6c119f61&start=60

[Muxa52]

Muxa52     
   -       
 And trying to debug via JTAG? Or upgrade FW?

Thank you for your suggestion!
I do not have the appropriate JTAG adapter. In addition, Samsung provides access to documentation and developer tools of processor on a separate individual agreement. Finally, for the firmware via JTAG need to have the image of ROM, which can not be obtained from the update FW. Owon will not give it to me, because this is the subject of her property.
I'll send the oscilloscope on the warranty, if I am sure that the problem is in the software.

[Carrington]

Muxa52, BBAAHHOO thanks guys, I had not seen the page 17 of this topic, this has helped me to resolve some of the questions that I had.
Muxa52 I hope you get lucky.
"tinhead" You say" I don't want to be human!"  I say you are a machine.
Did you find the BSDL model for S3C2416?

[BBAAHHOO]

BSDL model for S3C2416? - As long as there is no experience.
The matter was only CPU RT 3052
  And you have some experience with S3C2416?

[Carrington]

BSDL model for S3C2416? - As long as there is no experience.
The matter was only CPU RT 3052
  And you have some experience with S3C2416?

No, I have no experience with the S3C2416.
I do not understand. Why the RT3052? Is a Ralink SoC with a MIPS32 24KEc core.

[BBAAHHOO]

RT3052 - it is used in routers ... he dead firmware,  i had  restores. something like this ...

[Carrington]

RT3052 - it is used in routers ... he dead firmware,  i had  restores. something like this ...

Ok.
By the way this has something to do with the SDS series?


http://radiokot.ru/forum/viewtopic.php?f=10&t=70052&sid=f0c6b067c1522483e9cfb701355c613d&start=20

[tinhead]

"tinhead" You say" I don't want to be human!"  I say you are a machine.
Did you find the BSDL model for S3C2416?

sure, i've posted one some time ago here in forum, but take better the attache version (in org. Samsung BSDL there was an error)

[Carrington]

"tinhead" You say" I don't want to be human!"  I say you are a machine.
Did you find the BSDL model for S3C2416?

sure, i've posted one some time ago here in forum, but take better the attache version (in org. Samsung BSDL there was an error)

Thank you very much, but what do you mean with "in org. Samsung BSDL there was an error"?