OWON VDS1022I Quick Teardown (versus the Hantek 6022be)

[SIBtronics]

have the owon vds 1022i comming in the mail, decided on that, since i dont think i will need 4 channels or 250mhz single channel bandwidth any time soon, the thing is there is no catch, of course it has a lower bandwidth and lower sampling rate but with owon its you get what it says. my college uses (desktop) scopes from owon and they are decent. as for the record lenght idk. will check when it arrives

[John Celo]

Too late, mine is already coming in the mail too!
With some ali coupons and sale got it for 26.88eur with 5day delivery. Not bad at all.

Was also considering HANTEK DSO2C10 for around 105eur.
HANMATEK DOS1102 (owon SD1102, 100mhz band, but 20mV/div) for around 85eur
HANMATEK DOS1104 (5mv/div, 100mhz, 4ch)  for around 125eur

Didn't want to deal with hantek bullshit even at 'low' price of 105eur.
Was seriously considering DOS1104, but 125eur is almost 1/3th-1/4th the price of Rigol DHO802/DHO804.

I just don't know how hard the limited record length and trigger options of DOS1104 and other low end owon-derivatives affect their usefulness.
Maybe I should have gotten DOS1104 for 125eur, because opporunity to buy it at such price won't come again anytime soon.
Ordering oscilloscopes from AE (effectively no warranty) seems like a risky endeavour too.

Worst case scenario I wasted 27eur. Other things I have gotten from ali, like AN8008 for 13eur have been perfect for a beginner like me.

[Gyro]

I think you'll be happy with your decision, especially at that price! Don't get too hung up on buffer length, memory is cheap these days on desktop scopes but the Tek TDS210, which still sells for healthy prices on ebay has precisely half as much and the 1022. The more expensive but lower bandwidth and sample rate Picoscope 2204 only has 8k too.

Higher USB bandwidth loading tends to go along with bigger buffer size. The Owon has particularly modest USB bandwidth requirements, hence it is possible to provide the highly desirable USB galvanic isolation VDS1022I, which none of the others have.

At this price range you can't have everything, but it's a good compromise.


P.S. For long buffer length on digital stuff you can't beat a <$10 8 bit Saleae clone and Sigrok Pulseview s/w (not real buffer - it's all in the PC).

[SIBtronics]

Have now just got the scope, it is good, doesnt feel cheap and doesnt work like a cheap scope. compared to the handyscope hs3s we have at college it behaves the same essentially and these are much more expensive (though also have 50mhz bandwidth, 16bit mode and awg). having also used a tds 220, i certainly prefer this to it, the tds 220/210s screen is awful, of on the owon the frequency and sampling rate is more limited but for me thats totally fine. I use florentebrs software, it works very nicely and i cant really complain about it, the lack of virtual dials is a plus for me tbh. i cant really see the complaints about the software, of course florentebr added some stuff, but the ui itself which is the same as the og one as far as i know is good. i havent tried out the recording capability in depth yet, it works with a single usb A, the dual usb a thing really is a thing of the past. did some lissajous figures of course, works aswell, although they do look nicer on an analog scope. the fft works too, not that i have too much use for it. the trigger is rock solid. edit: just checked the handyscopes specs and besides analog bandwidth, and  wavegen these are the same spec wise.

[John Celo]

I've also received the scope. Tested and it works well with florent software.
The only thing you have to do in addition to following instructions on florent github (install java JRE) is to get 'Zadig' to install USB-drivers for the device.

First impressions - It will definitely be useful (especially as educational tool), but I would have been pretty disappointed if I paid the full price for it (which is in 70-80eur range).
Right now I'm just hooking it up to everything and measuring things, such as looking at various LDOs and switching mode regulators on various MCUs I have.
5k record depth is just not good at all. Can't really zoom out or zoom in any captured signal to see more.

For full price it's really old & outdated and I would not recommend it. Actually any of the oscilloscopes with record length below atleast 1M is out of the question.

Oscilloscopes is just one of those things that are not possible to cheap out on.

I'll write a more indepth review some week(s) or month(s) later after I've had more time with it.

[cunningfellow]

Does anyone have a source of the BNC sockets on the VDS1022?

I have some that the plastic ring has perished.

[TizianoHV]

5k record depth is just not good at all. Can't really zoom out or zoom in any captured signal to see more.

For full price it's really old & outdated and I would not recommend it. Actually any of the oscilloscopes with record length below atleast 1M is out of the question.

I have one too, 5k memory is quite limited, 100k would have been much better, I don't know what they were thinking. Anyway, you have to remember to set a wide enough timebase, with a bit of practice you get used to it.

I would like to remember how most of these chinese USB / battery oscilloscope can't work reliably with normal and single mode trigger. Probably due to software triggering or bad implementation.
The Owon on the other hand never misses an event and in general is always reliable.

Does anyone have a source of the BNC sockets on the VDS1022?
I have some that the plastic ring has perished.

I usually never screw them in because they seem flimsy. There's still enough grip to keep them from disconnecting.

[Gyro]

I've also received the scope. Tested and it works well with florent software.
The only thing you have to do in addition to following instructions on florent github (install java JRE) is to get 'Zadig' to install USB-drivers for the device.

First impressions - It will definitely be useful (especially as educational tool), but I would have been pretty disappointed if I paid the full price for it (which is in 70-80eur range).
Right now I'm just hooking it up to everything and measuring things, such as looking at various LDOs and switching mode regulators on various MCUs I have.
5k record depth is just not good at all. Can't really zoom out or zoom in any captured signal to see more.

For full price it's really old & outdated and I would not recommend it. Actually any of the oscilloscopes with record length below atleast 1M is out of the question.

Oscilloscopes is just one of those things that are not possible to cheap out on.

I'll write a more indepth review some week(s) or month(s) later after I've had more time with it.

I'm puzzled by the Zadig reference, It's not something I've ever had to do as both the stock S/W and Florent's S/W install the driver. You got me thinking that maybe I had always installed Florent's after the stock S/W, but then I remembered I had recently (only) installed his S/W on a Win7 VM under linux and it worked fine (the JRE is of course required). The Linux version works too of course.

Just looking back the start of the thread, my original review and teardown was in 2015 and the model is still available. It carved a niche at that price point that has pretty much survived ever since, based on price/performance. The only 'competitor' at the time was the Hantek USB range, that had no buffer, well maybe a couple of hundred bytes in the EZ-USB microcontroller (despite what the spec says) and relied on streaming everything in real time and triggering in PC S/W. There are certainly low end handheld and bench scopes coming down the line now but still not to the same price point (unless you get really lucky on Ali with some less known/unknown brand). The only real competitor in the USB space and H/W quality, as I mentioned previously, is the significantly more expensive, Pico 2204A. This arguably has much more flexible S/W (shared with the higher end models), but when you look at the fine print, it drops to 50Msps when you turn on both channels and it still has only a 8k samples buffer (I'm not clear whether this is per channel or shared, like the sample rate). I don't know of any low end USB scopes that have 1M buffers.

The 5k (per channel) buffer is certainly a limitation, it makes you go back to 'traditional' scope usage techniques - if you want to look at more detail of an edge, you increase the timebase and trigger on that edge rather than trying to zoom it up from a long capture (in your previous post, you mentioned several seconds). Unless you are massively oversampling, there is no such thing as the 'zoom and enhance' that you see in films. It's instructive to go into the display menu and switch from 'Vector' to 'Dots' to get an actual picture of how many samples the display of a particular waveform feature is being based on. If there are too few, you need to go to a faster timebase.

I was a unclear, from your previous post, about your requirement for capturing a couple of seconds of a fixed frequency (20-40-60kHz). If you are looking at one or more logic signals, a <$10 logic analyser, streaming to the PC is definitely the way to go. Again you need to fall back on traditional methods - the scope was used for evaluating the analogue domain integrity of a logic signal (rise and fall times, overshoot, crosstalk / ground bounce etc) and then the logic analyser for long captures and data decoding.

Looking forward to your more in-depth review anyway. Don't forget to pick up the keyboard shortcuts, they speed up the UI considerably.

[John Celo]

Unless you are massively oversampling, there is no such thing as the 'zoom and enhance' that you see in films.

Why wouldn't you massively oversample if your sample rate allows for it?
That seems like a default sane behavior if you have the sample memory for it.
100MS/s offers plenty of opporunities to oversample when you're zoomed out a bit (I find it's rare that I look at the signal at the very limits of the device's sample rate)

Even this scope oversamples teeny bit by default (two samples per pixel on 1080p screen), but limited memory lets you zoom in one step before there's less than a sample per pixel - thus zooming in more would be pointless.

Similarly to zoom out from captured signal, you are not limited by sample rate, purely by sample memory.

5k memory is barely enough to fill a 4k screen to have a sample per pixel.
Of course you can't zoom out. Neither can you zoom in. 5k memory simply does not allow for that.

You can't zoom out from a captured signal at all, because well.... that would require some memory, and this thing has barely any.

I apologize I lost my temper a bit, that "in the films" remark really got me LOL. You can do everything like in the movies and more, including time travel to past if you have the memory for it darn it(or bandwidth to stream it over USB at full speed to a computer that has bountiful memory)!

One of the signals I'm looking at has noise spikes roughly every 500ms/1s (those periods can't be captured by edge trigger since the amplitude of the noise is the same, just way more of it). Precise timing of which I can't tell...... because it doesn't have enough samples.

For other signals and LDOs I can see that 3.3v LDO greatly amplifies noise spikes coming from USB VBUS at ~90kHz and no other aperiodic spikes. So no issues analyzing signals of that type.

[TizianoHV]

I don't know if someone has already mentioned this but if you enable the peak detect mode the adc will run always at 100MHz and it will display the min and max. I found out that it's very convenient when looking for casual events.

[Gyro]

Unless you are massively oversampling, there is no such thing as the 'zoom and enhance' that you see in films.

Why wouldn't you massively oversample if your sample rate allows for it?
That seems like a default sane behavior if you have the sample memory for it.
100MS/s offers plenty of opporunities to oversample when you're zoomed out a bit (I find it's rare that I look at the signal at the very limits of the device's sample rate)

Even this scope oversamples teeny bit by default (two samples per pixel on 1080p screen), but limited memory lets you zoom in one step before there's less than a sample per pixel - thus zooming in more would be pointless.

Similarly to zoom out from captured signal, you are not limited by sample rate, purely by sample memory.

5k memory is barely enough to fill a 4k screen to have a sample per pixel.
Of course you can't zoom out. Neither can you zoom in. 5k memory simply does not allow for that.

You can't zoom out from a captured signal at all, because well.... that would require some memory, and this thing has barely any.

I apologize I lost my temper a bit, that "in the films" remark really got me LOL. You can do everything like in the movies and more, including time travel to past if you have the memory for it darn it(or bandwidth to stream it over USB at full speed to a computer that has bountiful memory)!

One of the signals I'm looking at has noise spikes roughly every 500ms/1s (those periods can't be captured by edge trigger since the amplitude of the noise is the same, just way more of it). Precise timing of which I can't tell...... because it doesn't have enough samples.

For other signals and LDOs I can see that 3.3v LDO greatly amplifies noise spikes coming from USB VBUS at ~90kHz and no other aperiodic spikes. So no issues analyzing signals of that type.

I'm sorry that the 'zoom and enhance' got to you, but in this situation it really is quite apt. As you said yourself of course you can massively oversample, as long as you have the memory to do it. The trouble here is that you can't afford that amount of memory and still meet a low price point in a USB scope.

You need to read more of the thread to understand the architecture of the 1022. Unlike competitors of the time (and many since) it implements all of the core functionality in an FPGA. This allows the solid hardware triggering and other features. The 5k buffers are implemented inside the FPGA. Moving to an external RAM would require a larger pin count FPGA that would blow the budget. As it is, Owon have changed the FPGA something like five times over the life of the 1022 in order to cope with supply and pricing changes (check out the firmware folder in the S/W). Added to that, it would increase power consumption, requiring an external PSU or USB power bodging, there would be a lot more data flying about so you would lose the galvanic isolation that protects you from blowing up your PC, you would also lose the fast screen update rate and snappy response that you don't get with something like a Hantek USB scope.

I've designed loads of high volume consumer products and can tell you that at this end of the market, every cent in the BOM cost matters. If you miss your niche, where you please 90% of your customers for the price, then you don't get to sell.

Owon do produce a range of higher spec USB scopes with 5/10M with higher sample rates, 4 channels etc. but you are talking several multiples of the 1022 price. mikeselectricstuff (member) did a video review of one of the 4 channel models (3104) and liked it. It's ideal for sticking in a slim bag when traveling (which he does a lot) but price wise, they don't stack up well against equivalent desktop scopes, so never achieved the same degree of traction. As I said, the 1022 dates back to 2015 (it was an unknown back then and I think I was the first to risk a teardown and review). Time moves on, but it's still a solid performer for a lot of people on low budgets / portability / space constraints.

Sorry, but I'm beginning to feel like I'm doing your research for you. If you want something with a combination of features that matches what you want then you are going to need to develop deeper pockets or take risks (I'm assuming the reason you even looked at the 1022 instead of one of the new range Siglent or Rigol benchtop is shallow pockets).

You really need to do some detailed research on some of the options you listed in reply #801 and see whether the feature combinations, s/w quality (presence or absence of bugs) meets your requirements. Maybe take some risks on quality versus price (there are bound to be compromises!), and pass the 1022 onto somebody else, you won't struggle to make a profit on the price you paid for it.


P.S. Just looking at the options you listed, the Hanmatek / Owon ones only have 10k record length, hardly an improvement on the VDS1022. Only the Hantek has buffers in the Megasamples range, 8M?


EDIT: You piqued my interest with the devices you listed (in fact I spent most of the day looking). I don't know where you got your pricing, but in the UK, both the Hantek DSO2C10 and the Owon SDS1102 are currently around 210eur (making the price differential with the VDS1022I considerable, >2x).

Looking at the options, the Hantek has some good reviews and seems to be the winner on features, e.g. protocol decoding (although that seems to have issues like buffer length shrinking to 4k when decoding and not being able to decode stopped waveforms), It doesn't come close to what you can get with a <10eur LA clone and Sigrok anyway, unless you need higher frequencies. There seems to be an issue with low level measurement accuracy too.

I wouldn't touch the Hamatek DOS1102 over the Owon SDS1102, 20mV/div means minimum of 200mV/div with a x10 probe - not sensitive enough!

As far as Owon VDS1022 vs SDS1102, depending on max frequency requirement it's not as clear cut as I assumed for USB vs bench. If you need 100MHz bandwidth, then the SDS is the only option, if 25MHz is sufficient then it's not as clear cut. Stuff I observed from the specs and video reviews... The SDS lacks the number of trigger options of the VDS, it only has 2 BNC inputs (no external trigger input / output or pass/fail functionality which can be very useful on a 2 channel scope). The 10k buffer vs 5k seems fairly academic when comparing with multi Msample scopes. The SDS does have a PC app, but it is so slow as to be unusable, disappointing. Lastly, you need to store waveforms to a USB stick for transfer to the PC rather than just clicking the save button to save a waveform to the PC. I asked myself whether I would buy the SDS while owning the VDS (I've been wondering about a new scope) and I honestly don't think I would. Not unless I really needed the bandwidth and could live with the compromises. I don't have a UK price for the SDS1104, but I might buy one of those if I needed 4 channels and it wasn't much more expensive than the 1102. This comparison surprised me, I assumed the SDS would at least have everything that the VDS does. I haven't even touched on display size and UI.

Would I personally buy the Hantek? Maybe, but I would worry about bugs and functionality issues (I don't know what, if anything has been fixed in F/W since the reviews I saw. They look a bit fundamental. At least you get the massive buffer for non-decode stuff and get back the external trigger. Again, bandwidth requirements might make it a no brainer.

I think with both of them, I would have a nagging regret that I hadn't bought something Siglent, the SDS1102CML or higher, but those are in an entirely different price bracket still.

Just my thoughts anyway, Ymmv.

[John Celo]

You're correct.

My conclusion in the end was that these low end scopes are not worth even considering.

They might seem be cheap, but even on sale they offer poor value.
Some combination of outdated, miniscule memory (5k/10k/20k or so), 8bit ADC, low sensitivity, low samplerate, only 2c etc.
Buying them only offsets a purchase of something decent.

Rigol DSO804/Siglent SDS804X are the entrypoint worth considering as far as I can tell.

VDS1022I is a usable functional "better than nothing" type scope I was lucky to get for less than 30eur basically.

[Gyro]

I don't know if someone has already mentioned this but if you enable the peak detect mode the adc will run always at 100MHz and it will display the min and max. I found out that it's very convenient when looking for casual events.

I have a vague memory that it may have come up sometime in the past, but I had completely forgotten and others probably have too, or haven't spotted it, thanks. An excellent example of getting the most out of what you have.


P.S. I'm impressed by your website, you have a flair for the dramatic!

[Gyro]

...
Rigol DSO804/Siglent SDS804X are the entrypoint worth considering as far as I can tell.

The recommended 'entry' point consensus on the forum does change over time. for many years it was the Rigol DS1054Z. It is only within the last year that things have been pushed towards 12 bit by the more affordable Siglent SDS800 series. 12 bit is clearly an improvement, but whether it is a necessary entry point is another matter. 12 bit is only useful if accompanied by a similarly low noise floor. In some ways, depending on application, it is neither one thing nor the other. For instance I keep a very old but very low noise 16 bit Picoscope that I use for low level audio and distortion measurements. Something that a modern entry level high bandwidth 12 bit scope FFT could never match.

Anyway, I have deviated far too far OT for this thread. Good luck with your search.