OWON VDS1022I Quick Teardown (versus the Hantek 6022be)

[Gyro]

Welcome to the forum Bigwizman. 

Indeed, I occasionally need to remind myself that it was 2015 when I posted the teardown and we're still talking about it. That seems a pretty long product lifetime, although looking back the Rigol 1054Z and even the (still flawed) Hantek 6022be were around back then, I guess products that find a solid niche survive. I wonder how many (free) sales the thread has generated for Owon in that time! I'm also retired and it's more or less my only DSO still - my goto is still my Tek 475A but I do have an old Pico ADC216 (333/166ksps!) 16 bit that I use for audio (mainly in FFT), I still hate the Pico UI though. As you say, the Owon H/W actually performs as advertised and Florent's S/W has certainly given it a new lease of life.

Interesting info on the external trigger input impedance, I've always assumed that it was a similar 1Meg like the channel inputs. I have to admit that I've only used it a handful of times and that has been a direct coax connection to the TTL trigger output on an HP function generator. I think somebody mentioned problems with external triggering a while back, so loading was probably the issue. You've prompted me to further investigation, I suppose it's possible that the pullup or whatever for the pass/fail output might still be in circuit, I'll see what it does on my early issue PCB version.

EDIT: Yes, the triggering in roll mode is one of Florent's biggest improvements - so much better than the little 'triggering not available' pop-up box in the original Owon S/W! I'm not sure what the problem is with enabling the channel roll off filters, I tried them myself a couple of days back and they worked for me. Yes, I think you're right on the peak detect.

[Bigwizman]

Wow! fast reply Chris and thanks for the welcome.  I suspected you might be one of life's more experienced members!  I like your diligence of repeating the dangers of not respecting circuits with high current capability and carelessness of ground connections.  I have a pair of lenses I was wearing, with embedded metal particles (I used them as training aides for graduates and apprentices) after believing the person who told me they had isolated the circuit I was investigating.  I had made several connections where my screwdriver had not touched the earthed housing but when aligning to a slightly more difficult connection I used the metalwork to steady the tool and after touching the electrical contact the end of the screwdriver vapourised into my face! It didn't even blow the fuse!  Had I not been wearing glasses, my eyes would have taken huge damage.  I was a youngster then but have never relied on others to confirm a safety situation since, without several confirmatory checks myself!

I will carry on investigating the external trigger but it worked perfectly as an output of the trigger (it does have an obvious delay), although to a high impedance 'scope input.  My assumption would be that it is a tristate pin, being TTL, so configurable as input or output, at best a line driver/transceiver.  It is interesting that you have used it as an external trigger, so I will try driving it from a signal generator rather than the unknown impedance of the probe calibrator or my micro-controllers.  It appeared to me that it was not just loading the circuit but clamping the lower signal state to a voltage level, so not quite a simple pull-up resistor, as that shouldn't have stopped my own circuit from triggering.  What fun!

The channel filter display I experience has a double hyphen against it and clicking/selecting has no effect.  I've tried several different settings to see if it becomes enabled but all have the same result so far.  I may be missing something but cannot find any other options or alternative references in the manual.

I'm about to look for the latest fault on the OS4000 digital section, which causes spurious steps to occur on top of the actual signal in a repeating cycle but the pattern is also dependent on amplitude and Y position control, so looks to be more than one fault.  The store is a set of eight 1 bit by 1K chips! Mainly 7400 logic based, so pushing the art of possibility for what was available then but all in individual sockets.  I actually toured the Gould/Advance factory (Hainault, Essex) when I was at university and got some development parts for my 1st analogue scope a few years before they started on digital storage 'scopes.

[Bigwizman]

A quick message regarding my issues yesterday.

1.  The channel filters are working fine today and each channel is independent.  I may have an explanation, as when I loaded the S/W yesterday, I double clicked.  I was unaware I had two versions overlaying, with the top indicating a sync issue.  On closing to try again, the lower version became visible and indicating sync. so I carried on.  Perhaps there was some interactive disturbance caused.  Put that down to my ineptitude.

2. The External trigger issue has the following characteristics on my unit (VDS1022I):-
 
--  Multi switched to External input.
     The input sits at 4.4v when open circuit.  Maybe the pull-up mentioned yesterday.
     Connected to my 50 ohm source signal generator, a trigger can be obtained but on rising edges only.
     My signal generator is symmetrical output so I have to use the DC bias facility to provide a TTL compatible signal but this adds extra flexibility to the options.
     When connected to the ext trigger input, my source is pulled very slightly high but I adjusted back to TTL levels, varying sometimes during investigation.
     Investigating possible rise time restrictions a triangular wave was used, initially to define the slope but again triggering occurred even on slower signals but again rising only.
     Triggering occurs at a low voltage, so not a guaranteed high for TTL.  Changing the slope, consistently indicated the same low trigger voltage.
     Slightly negative transitions were allowed to see if it affected the triggering. Only rising edges obtained trigger.
     Using pulses, it appears to have a 290nS delay between trigger indicator and the triggering edge. Rise time indication is 14nS but that is likely the 'scope, although could also be my ancient HP3310A.
     Returning to the probe calibration output, a DC blocking capacitor did not allow any triggering and different values produced different distortions of the output.
     Signal distortion is mostly a sharp initial transition, followed by a time constant, so a pull-up is likely evident but there appears to be an active region too.

--  Multi switched to trigger output.
     Output sits at 0v, until triggered.
     The trigger output signal has approx. 1.5 microseconds delay from the trigger indicator.

--  Multi switched to Pass/Fail.
     Output sits at 0v.  I have no Pass/Fail criteria loaded.

Going forward, I may try an external TTL buffer between my controllers and the external input, to see if I can get the unit to trigger on my circuit, as it is more useful to have two independent signals, knowing they are referenced to a known point.
Still battling with the OS4000!
Best regards.

[Gyro]

You're welcome Bigwizman.

Yes, it's important to remember that even the isolated version isn't an automatic 'get out of jail free card'. By it's nature the Owon does appeal to beginners, so dire warnings are useful. I'm glad you sorted the filter issue, yes accidental multiple windows can be confusing (and non-functional!).

You have beaten me to it on the external trigger today (good investigation), but looking through the VDS1022 folder on my computer, I remembered that member camomille traced out the schematic back in January 2020 (he's been missing for a couple of years, I hope he's ok!), it's posted in Reply#526. I have attached a crop of the Trig in / out multi section for reference. The schematic was drawn for one of the previous PCB revisions that still used the Xilinx FPGA (before cost and availability forced changes to a sucession of different parts) but I don't see why that would affect the basic I/O circuit.*

As you can see, there is a permanent 1k pullup to +5V on the drain of the output Mosfet. Combined with the 470R series resistor, this will give around 1.5k pullup to +5V in Trigger-in mode - on the face of it, nothing that should worry a 50R push-pull output (or a TTL output). The Trigger input goes via a 470R series resistor and protection diodes into an FPGA input.

Even if the Trigger (or pass/fail) output Mosfet is turned on, this would only cause a 470R pulldown to 0V, but you are seeing around 5V anyway. so this is a little puzzling. The question I can't answer is whether some change has been introduced on a recent revision of the PCB (a wrong pullup resistor value for instance or a change to a push-pull output). I'm not sure how keen you would be to open up the scope and do a bit of investigation - the +5V in question is obviously the isolated one rather than the USB interface.

For reference, here's a link to a thread describing the Trigger out behavior... https://www.eevblog.com/forum/testgear/owon-vds1022i-trigger-out/


EDIT: * I suppose the input circuit may have changed a little, as it is less likely that more recent FPGAs would have 5V tolerant inputs**, but 3V3 logic thresholds should otherwise be an approximate match with TTL ones.

Another thought - It might be worth checking the resistance between the Trigger input and the Probe Cal output. If the circuit is still similar, you should see just under 3k.


EDIT1: ** Checking again, the positive input protection diode clamps to +3V3 so the Xilinx input wasn't 5V tolerant anyway.

[Bigwizman]

Hi Chris, thanks for the thoughts and the circuit availability had completely slipped my mind, although as you say changes may have occurred.  Your mastery of all the inputs is impressive!

I've continued battle with the OS4000 today and have used some of the VDS1022I's extra features to good effect.  The designers have used quite a few tricks to obtain the specification but in maximising bandwidth, it causes extra internal noise that is filtered late in the signal path.  I had to use both channel filtering and averaging to allow the erroneous signals to become visible above the noise.  I would have liked a proper signal integrator but it had the desired effect.  I have narrowed it down to a single data bit, so can now hunt the offending circuit.  The light diminished too much, so I will attack it another day.

I did a quick check of the resistance between the two connectors suggested and it was just under 3k as you and the circuit originally uncovered by Camomille predicted, so looks to be unchanged.  I don't think there is anything in the circuitry that contradicts the results I noted on my unit, apart from possibly the ac coupled trigger test results but even that is likley to have incomplete charging, so would exhibit some fast apparent edges.  I think my only confusion is that my 50 ohm source could trigger the external input even when using a relatively slow triangular waveform, albeit on rising signals only but a fast transition on the probe calibrator failed to obtain a trigger, along with my own controller circuit using 3.3 v logic?

As I commented at the time, Ext triggers at a very low voltage, which may be a clue.  Perhaps the circuits failing to trigger the system do not quite get to a low enough level but I'd be surprised if that is the case.  My second confusion is that I failed to get any negative transitions to trigger the unit?  I wonder what others have experienced using the external trigger?

That previous link to the trigger out behaviour has very similar findings to my own.  The delay then, was slightly shorter at around 1.2'ish microseconds, whereas mine was closer to 1.5 but I suspect it is dependent on what version of code has been put in the FPGA.  It is at least deterministic when using FPGAs, so despite having a delay, it is easy to compensate.

Returning to the circuit diagram aspect, I had already downloaded it but had yet to study the level of detail, so it probably wouldn't have tweaked my deliberations for some time.  Looking, it does have all the details and a very good effort indeed, so I will use it in future.  I am possibly one of the few, along with those long time members who witnessed the comments as they were made originally, who has read every one of the comments from start to finish.  I did it over a relatively short period and possibly overloaded what I could remember, as my main aim was to see if some features that initially appeared good had distinct limitations that only long term use identifies (a bit like external triggering  ) and largely speaking there are very few limitations associated with the features actually incorporated.  Yes, others have more features but at considerably more cost and many add them to the feature list and either fail to work (or unreliable) or worse are actually not incorporated but designed to capture the search engine traffic.  I've seen so many claiming the ubiquitous 1 GS/s, that I disbelieve it initially unless it is a well known manufacturer.  Usable sensitivity is another claim or avoiding the fact it has no low signal capability.  Rant over!

Best regards.

[Bigwizman]

Hi All,

I've been looking at my EXT trigger input issue a bit more and sent Owon a few of my observations, asking for an opinion, but mainly to provide them with feedback.  The observations I sent were obtained using their latest S/W 1.1.5 but I did the same tests with Florent's cf19 and there were no differences.

I can now get usable EXT triggering on my 3.3v logic but it is temperamental.  Taking the EXT input negative by 0.5v obtains very reliable triggering but as I noted before, the triggering voltage is very low.  I suspect that many will have used 50 ohm source generators without termination, as most of the time we are just after something that will clock a circuit but the ringing produced (normally hidden at the lower scan speeds) is likely to provide the overshoot that mine needs to trigger more reliably, and probably others similarly if it is common to all VDS1022 units.  I also noted that, although the 50% trigger option is available, it has no effect on EXT trigger inputs.

I cannot get any EXT triggering when set for negative edges, however much I push the biasing on the triggering signal, without exceeding the TTL spec too much or taxing the internal clamp diodes of course.

Looking at the Camomille circuit from Gyro's reminder, it is very obvious that connecting the Ext trigger input to the probe calibration output will never get close to the actual EXT trigger level with the potential dividers created by the isolation resistors and pull-ups incorporated.

In my mind, even the positive EXT edge triggering is easy to live with and a simple external circuit with a schmitt input inverter for triggering on negative edges, or adding more sophistication with variable trigger levels, is possible.  I've used a schmitt XOR with a short delay on one input to provide a positive pulse (effectively both edges, although the positive edge is the actual sync) to trigger on positive and negative transitions, as I needed change of state then.  The schmitt improves reliability, even on slow edges and cleans up noisy signals.

These are all small foibles and part of getting to know one's equipment.  It has been a useful exercise, as I have now used the Owon S/W and Florent's.  Florent's version has more fluent results with that all important triggering at all scan rates and I like it the most for all the extras, which I'm still exploring but I do like the visibility of the Owon S/W during maths functions, as it keeps the CH1 & 2 signals visible during the display of the maths results.  The fact the two versions can be alternated between so easily removes any restrictions but Florent, along with some of the others, have driven what is essentially a good product to a higher level of sophistication.   Interestingly, on another area of these forums, a Perytech unit gets mentioned that has similar hardware, perhaps with slightly more capability.  The software suggests it can synchronise up to four units on the same display, using the external trigger outputs.  Now that would be some facility, two of these units, providing a four channel input, displayed on the same synchronised GUI....

Best regards.

[Gyro]

Hi Bigwizman,

I'm sorry, I haven't had time to check my early unit yet (I'm literally up to my elbows in an audio project). One thing occurs to me, just a vague possibility. I don't know what FPGA is in your unit, but presumably it's one of the recent variants. If the internal pullup on the input pin happened to be enabled, it might act as a potential divider with the series 470R resistor and artificially lower the input low threshold. As Florent uses the same Owon FPGA stream files, this would be the same for both softwares. This is pure conjecture on my part but it's the only way I can see of getting a voltage drop across the series 470R resistor (the resistive 1k+470R pullup is acting directly in the input connector), the FPGA logic thresholds being pretty well defined. A quick thought anyway - I don't think anyone has reverse engineered the FPGA streams so no help available there.

I will check the behaviour on my Xilinx FPGA equipped early model as soon as I get a moment.

[Bigwizman]

Thanks for the reply Chris, at some point, past when I'm comfortable that my unit doesn't have any functional faults that would require warranty claims, I will get inside, as I too was thinking it would be FPGA input thresholds, although again, as the FPGA history is vague on later units and knowing most of the devices used have separate supplies for banks of I/O, internals and perhaps a few other internal blocks, which could actually be lower again than the 3.3v main supply, the actual trigger point could be relatively low anyway.  I note Camomille's circuit shows a number of lower voltage supplies, that could come into play.  Once I get past that checking period, I will endeavor to see what the actual level is at the FPGA input.  A nice thought about an internal pull-up possibly being enabled, as it would indeed influence the reaction to the external signals.

The only real item I had hoped somebody else could check, is if their units accept negative edge EXT triggers (on this round of software), purely from the functional aspect. Is mine typical or the odd man out?  Please don't see that as anything pressing, it isn't, I can happily live with what I have found but if the unit has a fault, I'm in the window to get it replaced.  Owon may even get back to me!

I have now resolved the Gould/Advance OS4000 DSO issue.  It was one of the 1bit, 1K memory chips, from Intel (2102 static RAM).  The data sequence coming out, isn't the same as that going in.  I cannot source any of these, so have moved it to the lsb position, where it is hardly distinguishable.  It had been a bit of a gut feeling for some time and I've contemplated putting a block of more modern memory in, even though most of it won't get addressed.  I noted these old chips have very variable levels between them too, some marginal on TTL transition levels.  I have some around 2.7v.  The "stuck" bits make an attempt but never reach the threshold.  The VDS1022i is pretty much superior in every respect to the OS4000, except nostalgia and my own appreciation of how it helped me produce, what were pretty unique designs at the time!

Good to hear you are also into audio!  I have an omni-transmission speaker project running, that will eventually have 5 drivers per stereo channel, each driven separately to maintain damping levels.  The lower end is taking some time to design, as it is inverse tapered transmission line loaded and using concentric tubes to reduce panel radiation.  Top end is the Dayton air motion transformer, mid range is covered by two 3" metal coned units and the bottom end is a composite coned unit.  The extended bass, when required, is likely a 10 or 12" but has yet to be bought.  All but the very top end are forced to radiate in an omni pattern but forcing the tweeter becomes a bit pointless (and rather difficult, with the short wavelengths involved, so easy to produce lobes), although it sits in a spherical housing to minimize diffraction effects.  Except the extended bass, which would be guilding the lilly, they are point sources.  There are reflectors involved but well within a wavelength at their crossover points.  Sorry, not intending to hijack the subject of these discussions!

Best regards!

EDIT:  I had an acknowledgement from the UK Owon office, that my observations were sent to the development engineers and will respond when a reply is received.  Owon certainly have a professional approach to customer liaison that should be commended, responding the 1st working day after I sent my observations.  I'm certainly impressed.

[magic_smoke]

Hello, newbie here.
I just ordered this scope (isolated version) for US$77.99 shipped, original price US$92.99.I had a $15 coupon code. Did I score a good deal? I'm on a tight budget and I'm just starting, so I didn't go for the higher end stuff.

My intention would be to probe hobby-related ac or dc electric motors and their driver circuits, low voltage stuff.
Would I need to isolate the unit under test for safety?

[rsjsouza]

I personally think you got a great deal, especially because it is the isolated version.

I don't know how you define low voltage, but if it is 12 or 24V you don't need to have concerns about the isolation of the unit and its controlling PC.

Only keep in mind the channel inputs are not isolated between them, so you need to be cautious if you clip the ground probe in different parts of the circuit.

[jasonRF]

Hello, newbie here.
I just ordered this scope (isolated version) for US$77.99 shipped, original price US$92.99.I had a $15 coupon code. Did I score a good deal? I'm on a tight budget and I'm just starting, so I didn't go for the higher end stuff.

My intention would be to probe hobby-related ac or dc electric motors and their driver circuits, low voltage stuff.
Would I need to isolate the unit under test for safety?

I agree with rsjsouza - you got a great deal.  For many hobbyists, especially those just starting out, it has plenty of capability.  My first scope was a picoscope 2204a that cost twice as much and has inferior basic specs (10 MHz bandwidth, 10 mV/div sensitivity, 50 MS/s x 2 sampling, 8 kSample shared memory but usually you get less); it served me well for a handful of years before I decided to upgrade.  While the picoscope has a few extra features that are great for niche applications, I find the vds1022 to be a much better value and is quite fun to use. 

EDIT: I recommend using florentbr's software, which many of us find superior to Owon's. 
https://github.com/florentbr/OWON-VDS1022

cheers,

jason

[Bigwizman]

That is a good deal MagicSmoke and from what you say, will be a good learning platform, as it is a good stable unit you can rely on.  Not knowing the exact control circuits of course means the advice given can only be based on assumptions.  In this instance though, even if you are working with H bridge controllers, the isolated version of the DSO will limit the possibilities of shorting one side of the motor output to a common ground, as neither of the two motor outputs stay at 0v.  This would happen if you connected a normal grounded type scope input directly across the motor.  If you are really worried, supply the motors from a floating psu but shouldn't be necessary with the isolated scope.  Once you start using the DSO, many of the pitfalls will almost explain themselves, as you visualise the operation of the circuits under test.  However, do stay withing the voltage rating of the scope inputs, so switch the probes to the 10x setting if you are uncertain of what voltages under test might be and then once confirmed, set them appropriately.  There will also be inductive spikes due to interrupting currents through the motor windings but there should be diode clamps as part of the circuit to ensure they do not destroy the driving transistors and if they are missing, those spikes will damage the DSO inputs, so those clamps are important for operation and circuit reliability.  Otherwise have fun!

[indman]

I use Windows 10 64x at work and I want to talk about a small problem with installing florentbr's software on this system.
I'm used to using the 32-bit version of the popular TotalCmd file manager for file operations. If you run the install.win file in this version of the manager, then there will be no errors or messages and the program will install correctly, but only in the \Windows\Program Files (x86) directory. In this case, the oscilloscope will not work because it will not connection with USB. Therefore, on a 64-bit system, the install.win file must be run either through explorer or in the 64-bit version of TotalCmd.
In this case, the program will be installed in the desired directory \Windows\Program Files as well as the driver file libusb0 along the path Windows\SysWOW64. Now the oscilloscope will work correctly. What is also interesting is that the official software of the latest version OWON_VDS_C2_1.1.5 is installed correctly, regardless of the environment in which the installer is launched. It is possible that my hint will be useful to someone.

[lotas]

A new version "OWON_VDS_C2_1.1.6" has been released on the official website.
https://files.owon.com.cn/software/pc/OWON_VDS_C2_Setup.zip

[jasonRF]

A new version "OWON_VDS_C2_1.1.6" has been released on the official website.
https://files.owon.com.cn/software/pc/OWON_VDS_C2_Setup.zip

Thanks- I just installed it.  Unfortunately it doesn't fix the problem I have been having with corruption of waveform samples sent over SCPI when sampling slower than 100 MS/s.     Was worth a try, though!

jason

[jasonRF]

For folks in the US, newark.com has the Multicomp Pro mp720017  (rebadged vds1022i) for $72.23 + shipping (which I think is probably around $10). 
https://www.newark.com/multicomp-pro/mp720017-us/pc-based-osc-2-ch-25mhz-100msps/dp/10AH2400

Not a bad deal. 

jason

[lotas]

A new version "OWON_VDS_C2_1.1.7" has been released on the official website.
https://files.owon.com.cn/software/pc/OWON_VDS_S2_Setup.zip

[Vict20]

Hi All

I've had my 1022 about a year now and not used it much. I decided to update it to a 1022i first so recently ordered the components detailed earlier in this thread.

On opening the unit up I discovered it's the later version (VDS1022 V2.2 2021.08.08) with the different layout for the DC-DC converter instead of the DCP020505.

Anyway, with a bit of head scratching and studying of datasheets I made it fit!

Note: I didn't remove R79, It's part of a chain of 2 and now does nothing.

Vic

[kds]

Hi All,
Many thanks to all contributors, this teardown stream is amazing.

I have just purchased the isolated version VDS1022i and… I realized that the differential HV probe (micsig) I am planning to buy next, requires a grounded oscilloscope (I need to control the output of 2 solar inverters).

What would be your recommandation to ground the VDS1022i?
I am considering to ground it to PE via CH2 or EXT for instance (since I won’t invest in 2 differentials probes  I won’t use CH2).

Note that I will operate the S/W on a headless Raspberry 4, remotely via VNC, if that matters.

[Gyro]

Hi All

I've had my 1022 about a year now and not used it much. I decided to update it to a 1022i first so recently ordered the components detailed earlier in this thread.

On opening the unit up I discovered it's the later version (VDS1022 V2.2 2021.08.08) with the different layout for the DC-DC converter instead of the DCP020505.

Anyway, with a bit of head scratching and studying of datasheets I made it fit!

Note: I didn't remove R79, It's part of a chain of 2 and now does nothing.

Vic

Thanks for posting, it's useful to know that the original DC-DC can still be made to fit reasonably easily.

[Gyro]

Hi All,
Many thanks to all contributors, this teardown stream is amazing.

I have just purchased the isolated version VDS1022i and… I realized that the differential HV probe (micsig) I am planning to buy next, requires a grounded oscilloscope (I need to control the output of 2 solar inverters).

What would be your recommandation to ground the VDS1022i?
I am considering to ground it to PE via CH2 or EXT for instance (since I won’t invest in 2 differentials probes  I won’t use CH2).

Note that I will operate the S/W on a headless Raspberry 4, remotely via VNC, if that matters.

Hi, welcome to the forum.

Yes, the shields of CH1, CH2 and EXT are solidly [grounded connected] to each other on the PCB so you could use either. The tag adjacent to the 1kHz Probe Cal output tag could also be used (though soldering a wire to this would be better left for a permanent installation, where you want to avoid accidental disconnection).

[SIBtronics]

Looking to buy my first scope, of course as a Student my budget is smaller, so i was looking at cheaper scopes, do you guys think the owon vds 1022 is good? i dont want to buy something i will regret in a few months. might be able to save up a bit more and get the 60mhz vds 2062 version, but i also want it to be compatible with the hscope app where only the 1022 is listed, have heard the higher end ones use the same protocoll so maybe these also work in it, if anyone has a 2062 can you please check?

[Anthocyanina]

Looking to buy my first scope, of course as a Student my budget is smaller, so i was looking at cheaper scopes, do you guys think the owon vds 1022 is good? i dont want to buy something i will regret in a few months. might be able to save up a bit more and get the 60mhz vds 2062 version, but i also want it to be compatible with the hscope app where only the 1022 is listed, have heard the higher end ones use the same protocoll so maybe these also work in it, if anyone has a 2062 can you please check?

for 100$ i think there isn't a better option new. maybe for a bit more you can get one of the handheld scopes they make, like the HDS242s which has similar oscilloscope functionality to the VDS, except for math and FFT, and a more limited set of measurements, but you also get an ok multimeter and function generator. the VDS of course has the advantage of using a computer screen, plus the extra measurements and math and trigger options.

i don't know how good the hscope app has gotten. i tried it once, the free version, and was not impressed at all.

as a student, i think you might be better off finding a used analog discovery 2, it's basically a small lab in a single package, its software is amazing, and the hardware itself is also very capable. you get a 2 channel scope, 2 channel waveform generator, 16 logic I/O channels, and positive and negative power supplies. the software has features like frequency response analysis which will definitely come in handy when you're studying filters and amplifiers and whatnot, and with an add-on board, or a basic circuit of your own making, it can also do impedance analysis. you can find the analog discovery 2 on ebay used for about 150usd, which isn't much more than what the VDS1022i cost