VOLTCRAFT DOV series

[goaty]

Ok, so the OpAmp is a SGMOP17C. https://www.mikrocontroller.net/attachment/700811/SGMOP17C.pdf
Probably regulates the TPS7A92 ?
So some kind of feedback loop leading to the 40kHz ?

If it is just a buffer for the ref, this IC is a 'REF3425' 2.5V reference. ("19ED" REF3425IDBVR from TI)

https://www.mikrocontroller.net/topic/586101

Edit: The newer Datasheet (https://www.sg-micro.com/rect/assets/e93e49f3-6052-44b4-a620-dd9b0780a842/SGMOP17C.pdf)
states 2,8MHz instead of 1,5MHz of the Datasheet on top.
Also only the new datasheet states "unity gain stable"...

Can we replace it with another part ? Maby the new one with marking "GE3xx" ?

I have some OPA340 in the bin (worse input offset). Supply seems to be 5V on Pin 5. So maby I give it a try if it behaves differently.

One more thing. In the picture https://www.eevblog.com/forum/testgear/voltcraft-dov-series/?action=dlattach;attach=2844874;image
I think I see a different resistor (R1782) on the top left side of the TPS7A92 (U623). One is marked 12B (1,3k), the new board looks like something different. More like 682 or so.
Hope there´s not more than one modification depending on another.

[SL75]

Hi, I'm new here – so far I've only been reading along.
From my point of view, Owon made a classic beginner mistake with his reference voltage buffer. The designers didn't consider that directly feedbacked unity gain amplifiers have a particularly small phase margin and are therefore especially sensitive to capacitive loads at the amplifier's output. So it's practically guaranteed that this amplifier design, in combination with a large capacitive load, will oscillate. The oscillating amplifier is only not that obvious because the AC component at the output is partially knocked down by the large capacitors (10 µF, 22 µF?), which keeps it relatively small.
Using a different op-amp usually doesn't help much in this case because the circuit determines the behavior. Besides, the chosen SGMOP17C has a particularly low input offset. So Owon made a good choice there...

[goaty]

So can I measure the capacitance at the output (Pin 4) of the buffer using a LCR meter ?
Is it feasible to add compensation components to avoid oscillation ?

The datasheet suggests an isolation resistor and feedback capacitor.
"Driving Capacitive Loads:
The SGMOP17C is unity-gain stable with heavy capacitive load. If greater capacitive load must be driven..."
What is a "heavy capacitive load" ? 10pF ? 100pF ? 1nF ?

Could a modification like drawn be a solution ?

But then again, why a 10mA OpAmp if the Ref can drive 10mA ?
I probably also remove the buffer.

[SL75]

Owon's solution for bridging the (poorly dimensioned) op-amp circuit, as introduced in PCB Rev 1.1.1, is one option, since the REF3425 can drive up to 10 µF according to the datasheet (and maybe a bit more). In addition, Owon saves the cost of the op-amp this way... 
However, it's questionable whether the 10 mA from the REF3425 is sufficient. 
I would therefore properly modify the op-amp buffer so that it meets the requirements. 
In the circuit shown, the capacitor on the pads of R2437 ensures stability in combination with the increased decoupling resistor R2434 (1 Ohm --> 10 Ohm). The additionally required resistor R DC FB closes the loop for DC and thus ensures that the circuit maintains absolute DC precision. R DC FB can, for example, be soldered in 0402 package from the lower pad of R2434 to the lower pad of R2437, since Owon kindly placed them right next to each other :-).
R DC FB and the sum of R1840 and R2435 are chosen to be almost the same value so that the (assumed equal) input currents of the op-amp drop the same amount of voltage across them, and thus do not worsen the offset behavior of the op-amp circuit. Additionally, the filtering effect is improved due to the increased time constant (larger R1840 and R2435) which better suppresses noise components from the REF3425. With this dimensioning, it's easy to drive output capacitances larger than 22 µF without running into stability issues.

[goaty]

Great. Seems someone knows a bit on electronics design ?
I will try that out the next days.

[Martin72]

Ok, so the OpAmp is a SGMOP17C

That's a great find—it changes things quite a bit, at least for me.

[SL75]

One more thing. In the picture https://www.eevblog.com/forum/testgear/voltcraft-dov-series/?action=dlattach;attach=2844874;image
I think I see a different resistor (R1782) on the top left side of the TPS7A92 (U623). One is marked 12B (1,3k), the new board looks like something different. More like 682 or so.
Hope there´s not more than one modification depending on another.

I think it's not that critical since the TPS7A92 is just a simple adjustable LDO from TI. As long as the stabilized 5 V (V5.0A) is there, everything is fine. However, it should be present since V5.0A provides the supply voltage for the reference buffer op-amp...

[Martin72]

With the realization of what “ST3SK” actually is, I redrew the buffer circuit.
I actually desoldered the two capacitors at the IC's output and measured them.

[CeeJay]

Can someone confirm that this is how the improved op-amp buffer (thanks to @SL75) would look like on the pcb?

[Martin72]

Yes, that should be correct.

[Martin72]

So...
I read through and simulated SL75's approach to improving the buffer circuit instead of just throwing it out.
I think this is the better solution, precisely because it's not clear how much current the reference would have to drive on its own.
Based on my measurements of the capacitors, it is now clear that there is at least 10 µF present—plus an additional amount X related to the buffer capacitors for this 2.5 V in the front end.
I wanted to measure the actual current being drawn, but to do that, I need a very fine soldering tip—or a soldering iron—so I can solder test leads to an 0805 resistor.
In the end, I ordered a soldering iron for my Weller station, but it still hasn’t arrived.
That also means I can’t implement the “SL75 solution” right now.
But maybe the soldering iron will arrive this week after all...

Owon's solution for bridging the (poorly dimensioned) op-amp circuit, as introduced in PCB Rev 1.1.1, is one option, since the REF3425 can drive up to 10 µF according to the datasheet (and maybe a bit more). In addition, Owon saves the cost of the op-amp this way...

I assume that no “fully qualified” engineers were hired for the design; this is also evident from the fact that some circuits followed the recommendations in the respective IC data sheets exactly.
Removing the buffer circuit—which not only eliminated ripple but also helped optimize costs—probably earned the employee responsible the “Employee of the Year” award.

[evava]

Gentlemen, you’re amazing!

But the question arises: what about all the other circuits in this oscilloscope? Are they designed correctly, or are they also suboptimal?

[Martin72]

You can never rule that out—neither here nor with other manufacturers.
I wouldn't worry about it too much; we found a clear error here.
Others might never be found because they don't stand out in practice.
And in that case, it doesn't really matter anyway.

[Olex]

I have a small gift for everyone 
The very latest firmware version for Owon ASD800/900 with some nice improvements and hopefully fixes.

https://drive.google.com/drive/folders/1rE0NNm-Won3Vsb7VVnA8pbmk0wyUDhND

[Aldo22]

I have a small gift for everyone 
The very latest firmware version for Owon ASD800/900 with some nice improvements and hopefully fixes.

Thanks. 
How did you come across it?
Was it sent to you or someone else to solve a specific problem?

[SL75]

Can someone confirm that this is how the improved op-amp buffer (thanks to @SL75) would look like on the pcb?

Yes, that's exactly what I had in mind (but Martin72 was faster :-)).
By the way, with the modification you can proceed step by step. The components circled below are the most important. Once you have success, you can also adjust the resistors of the integration network at the input (R1840; R2435).
Do you have the possibility to measure the ripple on the reference voltage with an AC-coupled external oscilloscope before and after the modification (test point: C995 or C2329)? That would also be interesting.

[Olex]

I have a small gift for everyone 
The very latest firmware version for Owon ASD800/900 with some nice improvements and hopefully fixes.

Thanks. 
How did you come across it?
Was it sent to you or someone else to solve a specific problem?

The firmware was obtained from the official Lilliput/Owon support.
To be honest, I don't understand why they don't upload updates to the website ( they understand its weak security?)

[Martin72]

Do you have the possibility to measure the ripple on the reference voltage with an AC-coupled external oscilloscope before and after the modification (test point: C995 or C2329)? That would also be interesting.

I measured the 2.5V ripple—but I can't remember if it was right at the capacitors or at the logic IC in the front end:
https://www.eevblog.com/forum/testgear/voltcraft-dov-series/msg6290902/#msg6290902

EDIT:
That can't have been directly, because the circuit is covered by the heat sink—so I must have measured it in the front end.

[Martin72]

The very latest firmware version for Owon ASD800/900 with some nice improvements and hopefully fixes.

So what has been improved or fixed?
Unfortunately, as usual, there's no revision description included...
But there's one thing I've already figured out: the axis labels are there....
“Of course,” they excluded the FFT from this....
This should definitely be improved as well.

[Olex]

But there's one thing I've already figured out: the axis labels are there....

I knew you'd appreciate that 

[Aldo22]

 
Are you sure that's right?
From ...0.17 straight to ...8.2?
Either there are a lot of changes in there, or they've messed up the version numbers. 

[Martin72]

As far as I'm concerned, they can even call it XYZ—as long as things keep moving forward.

[Olex]

Are you sure that's right?
From ...0.17 straight to ...8.2?
Either there are a lot of changes in there, or they've messed up the version numbers. 

I hope there are a lot of changes/fixes there, since the difference between the firmware versions is about 1 year.

[Aldo22]

Are you sure that's right?
From ...0.17 straight to ...8.2?
Either there are a lot of changes in there, or they've messed up the version numbers. 

I hope there are a lot of changes/fixes there, since the difference between the firmware versions is about 1 year.

Yes, a few changes.
What I noticed right away:
- The fan is quieter again. Very nice! Better than the original version.
- You can now select German as the language. As expected, this is pretty confusing (Screenshot).
- Kernel version from May 2026

[Martin72]

Here are the two most noticeable changes: the axis labels in the display menu, and significantly more entries under “More.”