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

[TomC]

AndrejaKo: This remind you to anything?

I've seen that noise, and all I had to do is plug the brick on the same AC outlet as the SDS7102 while doing the GND noise test. No loop probe required!

[Carrington]

I've seen that noise, and all I had to do is plug the brick on the same AC outlet as the SDS7102 while doing the GND noise test. No loop probe required!

Yep, we do not need a loop probe! 

[TomC]

Yep, we do not need a loop probe! 

At last!!

Owon beats Rigol!

Built in loop probe and noise synthesizer !

[lemon]

AndrejaKo, although the discussion is old, have in mind the Siri's measurements on MC34063.
https://www.eevblog.com/forum/testgear/review-of-owon-sds7102/930/

This measurement is from newer version of PSU: T115-N Rev 3.  (partially SMD components but still single layer board).

[TomC]

Lemon, any news on your AOZ1094 yet? I'm planning on opening my scope today and take some current measurements from the PSU. I need that to improvise some dummy loads for when I start scoping my old PSU.

[lemon]

Unfortunately, no news about this.
Usually the shipment from China delays enough.
If you are lucky =10-14 days
If you are normal guy = 18-25
If you are no lucky (gantemis like Greek says) = over 30 days.

It shipped 10th of August...

[AndrejaKo]

Unfortunately, no news about this.
Usually the shipment from China is long days ago.
If you are lucky =10-14 days
If you normal guy= 20-25
If you are no lucky (gantemis like Greek says) = over 30 days.

It shipped 10th of August...

If it makes you feel any better, I had a horrible experience with USPS. I ordered some stuff from Parallax and it usually gets here in two weeks. This time, it was sent on 11th June and came to my post office on 11th September.

Also thank you for linking to those measurements. Back when they were new, I was following the results, but in the meantime, I forgor about them completely. Lots of useful information is there!


By the way, the electrolytic capacitor on the adapter board is not low ESR, right? I could change it to a low ESR one without much problems. Perhaps that would help a little bit.

[lemon]

I had forget them, too!
Today I am looking at the old messages and found (again) all these usefull information. After so many pages, it is difficult to remember all of these. I bookmark some of these to easy recovery.

I had measure only the electrolytic capacitors of psu not adapter.
I have changed the C44 electrolytic with other with low esr but I had no improvement, but my trials stopped cause AOZ1094.
I am going to start again, when arrived them.

We have the same adapter (Ver. 3.0) and have some information from Siri's measurements. He had the same version with us, and his works will help us a lot.
I am thinking to come back all my mods to adapter board to the original and starts again slowly step by step with AOZ1094 topology first.
As I read today what the Siri found, he suggested that with the AOZ1094 there is no way for improvement, only with linear dc convertor.
I don't know we''ll see...

[TomC]

I measured the 8.4V and -7.6V supplies current today. While the scope was open I also captured some waveforms from pin 2 of the MC34063A IC. The results were as follows:

8.4V --- 2.2A - 2.22A
-7.6V --- 153.5mA after scope fully booted, 166mA while scope is booting.

The mains voltage during these tests was 125VAC. The scope does not have the battery option.

The 8.4V supply current was measured by obtaining the voltage drop between GND-B and GND-C. This is the voltage drop across parallel resistors R14/R14A/R14B with a total value of 0.0333 ohms. The voltage drop ranged between 73.5mV and 74mV. Using this method both the current used by the local -7.6V converter and the current used by the converters on the adapter board are taken into account.

The -7.6V supply current was measured by opening the circuit at the adapter cable and using an ammeter.

The attachments are the waveforms from pin 2 of the MC34063A IC.

[lemon]

Now it is explaining why Owon put small size inductor to 5.5V path of adapter board.
With 2.5-3A inductor we are OK.

It is very stange to me why the power suplly of -7.6V is too noise, with it has only 170mA of max. current!

TomC, thanks for all these measurement.

[rf-loop]

Now it is explaining why Owon put small size inductor to 5.5V path of adapter board.
With 2.5-3A inductor we are OK.

It is very stange to me why the power suplly of -7.6V is too noise, with it has only 170mA of max. current!

TomC, thanks for all these measurement.

It (-7.6V SMPS) is most crap design what is inside this scope, just this -7.6V SMPS. This can not understand - why. (I remember this - "perhaps designer have been drunk in this day")

[Carrington]

It is most crap design what is inside this scope, just this -7.6V SMPS. This can not understand - why. (I remember this - "perhaps designer have been drunk in this day")

You know what, if the "engineers/designer" was not drunk this scare me even more. 

[Carrington]

Now it is explaining why Owon put small size inductor to 5.5V path of adapter board.
With 2.5-3A inductor we are OK.

It is very stange to me why the power suplly of -7.6V is too noise, with it has only 170mA of max. current!

TomC, thanks for all these measurement.

Do whatever you want, I'm just saying that mohm x A = mV.

[AndrejaKo]

Does anyone know a Pi filter calculator which lets user enter component values and then get output based on that? I've been thinking more and more about adding a Pi filter to the -7.6 rail on the adapter board and would like to play with some values. So far I've only found MATLAB's rfckt.lclowpasspi, but figuring out how it works is taking a while.

[lemon]

http://www.calculatoredge.com/electronics/ch%20pi%20low%20pass.htm

http://www.designers-guide.org/Design/bypassing.pdf

[TomC]

Does anyone know a Pi filter calculator which lets user enter component values and then get output based on that? I've been thinking more and more about adding a Pi filter to the -7.6 rail on the adapter board and would like to play with some values. So far I've only found MATLAB's rfckt.lclowpasspi, but figuring out how it works is taking a while.

They usually work the other way around, you input the specs and it gives you the components values. Here is a link to the latter in case you don't have one. Also, there is a pi filter already designed and recommended for that circuit, see datasheet in second link. Owon uses this filter on later version PSU boards.

http://www.calculatoredge.com/electronics/ch%20pi%20low%20pass.htm

https://www.eevblog.com/forum/testgear/review-of-owon-sds7102/msg278699/#msg278699

[Carrington]

Does anyone know a Pi filter calculator which lets user enter component values and then get output based on that? I've been thinking more and more about adding a Pi filter to the -7.6 rail on the adapter board and would like to play with some values. So far I've only found MATLAB's rfckt.lclowpasspi, but figuring out how it works is taking a while.

See at PM. 

[lemon]

Now it is explaining why Owon put small size inductor to 5.5V path of adapter board.
With 2.5-3A inductor we are OK.

It is very stange to me why the power suplly of -7.6V is too noise, with it has only 170mA of max. current!

TomC, thanks for all these measurement.

Do whatever you want, I'm just saying that mohm x A = mV.

Carrington could you explaining this better? Why the law of Ohm is identical with the ripple of V?

[AndrejaKo]

Thanks for the Pi filter help! What I actually wanted to do was to take some component values from stuff I have in my junkbox and then compare results. I'm aware of the filter in the MC datasheet  and the component values from the later PSU.


It turns out that the version of MATLAB I have access to has GUI for that class which provides interesting output.

[Carrington]

Carrington could you explaining this better? Why the law of Ohm is identical with the ripple of V?

This obviously is not as simple. Decoupling capacitors are everywhere mitigating the voltage drop that occurs in the transient.
All the transients will not reach to this point*, but a current passing through a resistor will generate a voltage right? I think if less better. what do you think?
 
* The point is where we want put "the PI filter".

[TomC]

Just some ramblings to help keep my brain focused!

- DC/DC converters produce fast changing signals.

- Fast changing signals have high frequency components.

- High frequencies are easily radiated and electromagnetically coupled or conductively coupled through stray capacitance.

- When this is allowed to happen common mode signals are produced.

- Fast changing signals can be rendered harmless when their dissipation is confined to the circuit producing them.

- Striving for this goal is the best defense against common mode noise.

- Existing circuits can be altered sometimes to mitigate blunders in this area.

- Providing short low impedance paths to complete the circuit discourages radiated and capacitive coupling as alternative paths to complete the circuit.

- Keeping other circuits at a safe prudent physical distance can also help discourage alternative paths.

- When all else fails wholesale mitigation techniques such as decoupling capacitors, ferrite cores, and similar strategies are commonly used.

- Since at this point the common mode noise is usually widespread these strategies are seldom completely effective.

Any other thoughts, additions, or deletions?

[TomC]

Here is a complete set of images for the MC34063A that I just captured. The subject PSU is installed in the same scope used to capture the waveforms and is the new PSU released by Owon. Because of the cramped conditions I had to use the long ground leads, and for some captures and additional 2" wire to connect the probe tip to some of the IC's pins. However, the probe cables are equipped with ferrite cores which should counteract some of the effect of the long ground leads. On all captures CH1 is attached to pin 2 of the MC (Switch Emitter) and is used as the trigger source. CH2 was used to view the signals on all the other pins. In most cases there is a capture using DC coupling, so that the voltage levels can be read, and one or more captures using AC coupling to show a more detailed image of noise and spikes. There are also captures with CH2 connected to pin 4 of the adapter cable (-7.6V to adapter).

#1 - Shows the relationship between pin 2 and the timing capacitor.

#2 - Shows the voltage at the IC end of the current sense resistor. The voltage during pin 2 low is 8.4V, and ramps down about 200mV when pin 2 is high. If this voltage reaches 330mV it shortens the duty cycle to prevent damage. However, as can be seen, this doesn't happen during normal operation.

#3 & 4 - Shows the 8.4V as it appears on pin 6 of the MC.

#5 thru 7 - Shows the 7.6V as it appears on pin 4 of the MC. This is before it's filtered by L3 and EC9.

#8 & 9 - Shows the 7.6V as it appears on pin 4 of the PSU to adapter cable, adapter side of the cable.

#10 thru 13 - Shows the relationship between pin 2 and the COMP input. This pin is very sensitive and as can be seen, probing it alters the operation of the MC IC.

#14 thru 17 - Shows that the operation of the MC IC is altered even when just a 2" piece of wire is attached to the COMP pin. I fear that this extreme sensitivity can spell trouble if there is a high level of EMI nearby even if nothing is attached to this pin.

[TomC]

AndrejaKo,
in the set of captures I just posted I obtained some images that look similar to what your MC looks like. It turns out that pin 5 of that IC is very sensitive and just a short piece of wire attached to it causes havoc with the duty cycle. I observed the frequency go down to about 48KHz at times, which is similar to what you have. Looking at images of PSU boards similar to yours it seems that the resistors attached to that pin are thru hole labeled R40 & R37 located right next to the power switch housing. I would undo anything you have done in that vicinity, including the cutting of the ground traces that connected the switch housing to ground. Also, looking at a picture of your board's printed circuit side, I see some kind of residues in that vicinity, see attachment, I would thoroughly clean that up and undo any mods you may have done in that vicinity.

Good Luck!

[rf-loop]

There is 08.March.2013 published modification prototype version 2b for PSU version PCB-T115-N REV3.
This is -7.6V circuit.  There is not this problem.

Of course for other versions it is only principle and need carefully look how to apply it for other versions (before Owon made new supply).

One main question was, why this circuit disturb itself and of course this common mode radiated/conducted noise what was one part of total noise sum. 
Also note that this version works well with original electrolytics in this circuit.
It need note that also some added capacitors positions are critical. One example can see in picture where I have removed one added capacitor and moved it right around 10mm. (C in orange box)
In this mod 2b all is not optimal. But it works. Of course it need some experiments with other board versions but I think it give good start point what to do. I think, in these now showed test results there is not used any these modifications. If there is, then some capacitor is broken. With these modifications can not get these results what I see now there in signal pictures. 

[AndrejaKo]

AndrejaKo,
in the set of captures I just posted I obtained some images that look similar to what your MC looks like. It turns out that pin 5 of that IC is very sensitive and just a short piece of wire attached to it causes havoc with the duty cycle. I observed the frequency go down to about 48KHz at times, which is similar to what you have. Looking at images of PSU boards similar to yours it seems that the resistors attached to that pin are thru hole labeled R40 & R37 located right next to the power switch housing. I would undo anything you have done in that vicinity, including the cutting of the ground traces that connected the switch housing to ground. Also, looking at a picture of your board's printed circuit side, I see some kind of residues in that vicinity, see attachment, I would thoroughly clean that up and undo any mods you may have done in that vicinity.

Good Luck!

At this moment, the PCB itself is clean from contamination (it was completely covered in flux residue when I first opened the scope) and there are no modifications close to the comparator pin of the MC right now. Also instead of cutting ground, I disconnected the switch's cover from the GND.

In any case, I don't believe that any modifications  affected the MC duty cycle issue. It was erratic even before any of the modifications, so I'm convinced that they're mostly influenced by bad layout. I'll see if I can post a picture of the PSU with current modifications.

Unfortunately, I have to stop my experiments for a while. I have some exams in the next two weeks for which I need to prepare and fixing the noise is currently draining too much of my resources.