New GW Instek GDS-2000E oscilloscope released

[PigSpark]

Try here:

https://www.elfa.se/en/manufacturer/gw-instek/man_gwi

[Carrington]

Just a short question that may not belong in this thread. Since there are Europeans here, where would buy the GDS-2000? Got my eyes on an GDS-2204E.

Also, maybe here?
http://www.tme.eu/en/details/gds-2204e/digital-oscilloscopes/gw-instek/

Usually, for other brands (rigol for example), I recommend this site (great guy): http://www.silcon.cz

[4cx10000]

Thanks all for the tips!  I will check all recommended, suggested distributors tomorrow. I don't know what price range to expect, but Elfa's wasn't exactly what I would say on the low side - I think...

[pascal_sweden]

The processing power of this scope seems really nice.
If I recall correctly it is built on the Xilinx Zynq 7000 series.

The software features look pretty nice too!

However there are 3 things holding me from making a purchase:

1) Design: The design is not inline with Rigol, Siglent, or even Owon.

It really looks like a toy. Come on GW-Instek... you can do much better!
Why don't you look at Rigol, Siglent or Owon?
Note that the new XDS series from Owon looks really nice!

2) PCB construction: Fabricated design using existing prototyping boards lying around

In the video Dave showed that it seems like they were in a hurry and fabricated some existing boards together. This seems very unserious design! Why not one single PCB board like the others are doing?
Even if you use several PCB boards, at least make it in a more structured way, and make sure
the boards are not flapping in the breeze, to use Dave's words.
In his video he showed several weaknesses in the design. Some bigger than others, but I have seen enough to not trust this construction whatsoever.

3) Shielding: They only shielded the front-end. Really?

While Siglent and Rigol have a very sturdy construction and shielding that covers the entire PCB board,
and the power supply, GW-Instek just shields the front-end.

Come on GW-Instek.. get your act together! Your scope is not cheap at all.. so people expect decent shielding. How can you expect such a huge sales price for something that is literally built like a toy.

Obviously GW-Instek has smart engineers in house, as they managed to pull of a very performant design, built on the Xilinx Zynq 7000 architecture. No doubt about that.

But why don't they get the other things correct? Decent PCB board construction, decent shielding.
And please, hire some industrial designer, that can make your scopes look like Rigol, Siglent or Owon.

GW-Instek.. do you really think yourself that this scope looks nice? Really?
How is it possible that taste can differ that much between Asia and Europe?
I don't think it differs that much.... at least all the others know how to do it.

Really... even if you are using a fancy Xilinx Zynq 7000 architecture, it does not weigh up against all the current drawbacks..... the scope looks like a toy! and it is not built to last... the clips that are used to keep those fabricated PCB boards together will certainly not last.. it almost looks like a prototype product, that was put together in a hurry to make it in time for a demo at a tradeshow.

I think it's a shame that in one and the same company, there can be extremely smart engineers who can pull a working scope together that is based on a state of the art Xilinx Zynq 7000 architecture, but that at the same time, the engineers have no clue whatsoever about basic construction, shielding, and design. Honestly, even the software and hardware engineers should recognize that their scope does look like a joke. One does not need expert industrial designer skills to recognize that

[wraper]

The processing power of this scope seems really nice.
If I recall correctly it is built on the Xilinx Zynq 7000 series.

The software seems pretty advanced too!

However there are 3 things holding me from making a purchase:

1) Design: The design is not inline with Rigol, Siglent, or even Owon.

It really looks like a toy. Come on GW-Instek... you can do much better!
Why don't you look at Rigol, Siglent or Owon?
Note that the new XDS series from Owon looks really nice!

, I'm full of that cramped Rigol front panel design that is a pain to control. Nothing to take an example from.

2) PCB construction: Fabricated design using existing prototyping boards lying around

They are NOT prototyping  boards  . They did similar thing in their old scopes too. CPU was a separate board.

While Siglent and Rigol have a very sturdy construction and shielding that covers the entire PCB board,
and the power supply, GW-Instek just shields the front-end.

Rigol  DS2000 case holding on 4 screws which are not even on the corners + not a single hook/clip top of that. Drop it once, and back cover will go flying separately.

3) Shielding: They only shielded the front-end. Really?

While Siglent and Rigol have a very sturdy construction and shielding that covers the entire PCB board,
and the power supply, GW-Instek just shields the front-end.

Means that EMI is low enough that shielding is not essential.

but that at the same time, the engineers have no clue whatsoever about basic construction, shielding, and design

So some shielding automatically gives you a clue?  . You can build some freaking radiating "radio noise transmitter" and then just shield it to death. Read some treads here about how much clue had those DP832, DS1000Z and DS2000 designers. Like oscillating Vregs, improperly designed ADC clock oscillator, improper heatsink size and so on.

[nctnico]

@pascal_sweden:
Regarding the construction: my GDS2204E has a single board which holds all the components so no prototype style boards attached to it. I think Dave got an early engineering sample or prototype. Also some comments Dave made are totally unfounded. For example the flatcable for the PSU. Open up a Tektronix TDS500/600/700 series (around $20k when the dollar still was worth something) and you'll see a flatcable for power distribution as well from a 300W PSU (and not 30W like in the GDS2000E). When properly designed a flatcable is perfectly adequate for power distribution.

Tests with the noise levels have shown the shielding is more than adequate! So why waste the materials? The PSU GW Instek uses produces very little EMI noise so the scope doesn't need much shielding. When compared to a Siglent SDS1000X series the noise on the GDS2000E series is on par (perhaps slightly better but certainly not worse and at the same bandwidth).

About the looks... I rather have a pink hammer which works and gets the job done than a black one from which the head detaches itself every time.

IMHO the GDS2000E is a very good buy at the moment because it is packed with features which work as advertised and if you manage to find a bug it gets fixed quickly. What more do you want from a budget scope?

[pascal_sweden]

Can you make a "take it apart - 2.0" video?

[Vytautas]

The initial testing for a few days shows me that construction of the GDS-2000E brings no worries. They used there what is needed and did not use what would have been superfluous. For example, the back is screwed with metal threaded screws (Rigols, in comparison, - with self tappers).
The noise floor is excellent, under 400 micro volts, so shielding is a no issue.
One thing I do not like - is the fan implementation. When I turn on several devices that have fans, the total of those noises in my tiny lab becomes really annoying.
I paid 532 euros for my GDS-2072E. On the net, I have not found it for this price anywhere, the lowest is something like 680-700. But the company "Lokmita" somehow was able to get it for me that cheap. They have the GDS-2204E now for 1160 euros (plus x1.21 tax = 1403 euros).

[pxl]

, I'm full of that cramped Rigol front panel design that is a pain to control. Nothing to take an example from.

The Rigol desings are just perfect: made by Chinese warriors to Chinese warriors. No kiddin'. Anyone bought here - done by mistake.

[Piccolino]

 I have 2072E and that scope it is almost perfect. It is very fast, superb FFT , low noise , a lot of functions , superb design with no crowded buttons. The fan noise is very low. The display almost perfect too...The only major problem for me it is the power supply noise. That smps it is a really crap, very noiseing.

[Wuerstchenhund]

The processing power of this scope seems really nice.
If I recall correctly it is built on the Xilinx Zynq 7000 series.

The software features look pretty nice too!

Yes, it seems the GDS-2000E is a very solid offer in the entry-level range. Unlike most B-brand scopes it's obviously well balanced feature-wise, i.e. it doesn't just have large memory but can also make proper use of it. Also, it came out in very mature state, and it seems the few remaining bugs are fixed by GW Instek in a matter of weeks instead of months or years with Rigol and Siglent. The use of 'apps' is also interesting, as it opens the scope to the possibility of software expansions. The strong hardware platform should help here, too.

However there are 3 things holding me from making a purchase:

1) Design: The design is not inline with Rigol, Siglent, or even Owon.

Does it have to? If so, why?

It really looks like a toy. Come on GW-Instek... you can do much better!
Why don't you look at Rigol, Siglent or Owon?
Note that the new XDS series from Owon looks really nice!

You're kiddin', right? I mean, pretty much everything OWON made looks like a damn toy. Not saying it's not good gear (I don't know, I have no experience with OWON) but the looks of OWON gear is cheap and toy-ish, much more than anything else I've seen.

Siglent and Rigol have their own design which overall is pretty much slight variations of standard scope looks.

The GDS-2000E does look boring and uninspiring, no doubt about that, but it doesn't look 'toy-like', and from a Human Factors/usability perspective I really don't see a lot to critizize. Yes, the UI is bland as well and reminds me of Tek's awful DSO UI from the old TDSxxx days (*shudder*) but after having a somewhat closer look it seems logically structured and not overly cumbersome to use.

It's no beauty but as an engineer I'l always prefer substance over style. My Siglent SDS2204 may have had better optics but that doesn't help when two years after release the firmware suffered from major bugs. The ugly GDS2kE on the other side seems to have come out in pretty mature state.

3) Shielding: They only shielded the front-end. Really?

Yes, really.

While Siglent and Rigol have a very sturdy construction and shielding that covers the entire PCB board,
and the power supply, GW-Instek just shields the front-end.

Come on GW-Instek.. get your act together! Your scope is not cheap at all.. so people expect decent shielding. How can you expect such a huge sales price for something that is literally built like a toy.

I don't know if the GDS-2000E is built like a toy but shielding should be appropriate to the emissions caused by the circuitry. If the emissions are low then there's no need to stuff the case with shielding that at the end of the day has no impact.

Obviously GW-Instek has smart engineers in house, as they managed to pull of a very performant design, built on the Xilinx Zynq 7000 architecture. No doubt about that.

But why don't they get the other things correct? Decent PCB board construction, decent shielding.

Who says their design is incorrect? You? With what qualification? Just asking because judging a board design on the amount of shielding is pretty silly.

Also, GW Instek has been around long enough to probably know a bit or two about proper board design.

And please, hire some industrial designer, that can make your scopes look like Rigol, Siglent or Owon.

Please, don't. Just don't. OWON's optics are a lot worse and would be two steps back.

GW-Instek.. do you really think yourself that this scope looks nice? Really?
How is it possible that taste can differ that much between Asia and Europe?
I don't think it differs that much.... at least all the others know how to do it.

Really... even if you are using a fancy Xilinx Zynq 7000 architecture, it does not weigh up against all the current drawbacks..... the scope looks like a toy! and it is not built to last... the clips that are used to keep those fabricated PCB boards together will certainly not last.. it almost looks like a prototype product, that was put together in a hurry to make it in time for a demo at a tradeshow.

What makes you think the plastics clips that keep the boards together won't last? Clips like these are used in lots of electronics products, and have been so for decades. They last because they are in a position that doesn't see much use (boards are usually left alone, they're not removed and replaced on a daily basis), and if they crack under force then the impact the device must have suffered from this to occur would have been so strong that the clips are the least of your worries.

I think it's a shame that in one and the same company, there can be extremely smart engineers who can pull a working scope together that is based on a state of the art Xilinx Zynq 7000 architecture, but that at the same time, the engineers have no clue whatsoever about basic construction, shielding, and design.

Again, you say this based on what qualification? Because I'd think you should be a bit more careful about whom you call 'clueless' especially when one of your shining examples include Rigol who managed to screw up the most basic ciruitry:

https://www.eevblog.com/forum/projects/project-yaigol-fixing-rigol-scope-design-problems/


Can you point out specifics where you think more shielding is required in the GDS-2000E, and why you believe that is the case (and no, saying 'because Rigol and Siglent have more shielding' isn't a valid argument)?

Honestly, even the software and hardware engineers should recognize that their scope does look like a joke. One does not need expert industrial designer skills to recognize that

You could say the same about Rigol who obviously can't do hardware properly, or Siglent who has a very long track record of utter incompetence in everything software.

I've owned Rigol and Siglent products, and I have further examples of products from both in my environment. I never had GW Instek, but frankly, if I needed a cheap low-end scope now then GW Instek would be the only one on my short-list, based on its well balanced features and its apparent maturity.

Looks are for posers.

[Vytautas]

Speaking shortly, IMHO:
Design and layout is excellent, nobs well spaced, easy to access without bumping around.
Love the separate SELECT button under the variable nob (much better than the pushable option).
Menus and functions are easy to use, extremely intuitive.
Scope very responsive.
Goes to 1nS per division.
1Gs per ch; 10Mp per ch; 120000Wfm/s
Gated + segmented + event search + 1M fft...
Brilliant!

[rf-loop]

120000Wfm/s
 segmented

I have not seen any true data about wfm/s speed with different settings and not seen any true data about segmented memory acquistion speed. Not seen any real speed tests with pass/fail mask test true speed.
Yes, I know there is not trigger out but it can test also with different methods. Also segmented acquistion maximum trigger/acquistion speed is extremely easy to check. Just calculate from segment time stamps over example 1000 or 10000 segments.  Rigols, Owons, Siglents etc there is lot of data available.  Nice if someone who can do reliable tests can do these also with this new GW.

[pxl]

no trigger out? probably it could be resolved by firmware update

[rf-loop]

no trigger out? probably it could be resolved by firmware update

Segmented acquisition have segment time stamps.
Pass/fail have output. It can count pass and fail. Stopwatch and look example 100s time how many pass (signal so that all test pass)  and then look same but how many fail, using signal what sure give result fail.

Things are not so difficult.

Also two channel pulse generator and human eyes (idea is detect how long pulse delay need that (2nd) pulse exist in next acquisition, and so on)   can use for define wfm/s speed if there is no trig out. (there is @marmad long time ago explained how to do it. I do not try with my chingfinnglish.

[nctnico]

You have to realise that the maximum waveforms/s occur at the shortest memory length, a short timebase setting and probably some other settings which could impact the screen update speed. That is not the general use case of a scope with deep memory. All in all you can go through a lot of trouble for producing these numbers but I don't think they have a lot of meaning for using an oscilloscope in a typical usage scenario. If GW Instek claims it can do 120 000 waveforms/s I'm pretty sure it can do that for one obscure setting just like any other oscilloscope.

Speed of segmented recording depends on what GW Instek aimed to achieve: update the screen after each segment or record first and show later (like my Agilent DSO7104A does IIRC). The latter is faster but makes it impossible to view the progress of the acquisition. I did some testing on the GDS2204E and it seems it updates the display after each segment so this function is definitely aimed at capturing rare events and allowing the user to see what is going on.

[rf-loop]

All in all you can go through a lot of trouble for producing these numbers but I don't think they have a lot of meaning for using an oscilloscope in a typical usage scenario.

What kind of trouble?   

I have not meet any real troubles  Exept if  manufacturer marketing peoples do not like but - so what. It is not my head ache.

And in this case, they claim 75kwfm/s
It is "true" but it is highest peak value what do not include processing and display refresh cycle pause in acquisition.
True (or least my true) value (average over several screen update-processing period) is around or bit under 10% lower.

Examples (First rable have typemistake and it is only Ch 1 on)
Other proplem is that it is with old first FW release. So some numbers may be bit low but meaningless.





Where is trouble?

[rf-loop]

Speed of segmented recording depends on what GW Instek aimed to achieve: update the screen after each segment or record first and show later (like my Agilent DSO7104A does IIRC). The latter is faster but makes it impossible to view the progress of the acquisition. I did some testing on the GDS2204E and it seems it updates the display after each segment so this function is definitely aimed at capturing rare events and allowing the user to see what is going on.

GW GDS2000E  manual:

Also notice that the scope doesn’t need to
rearm the trigger between each segment, this makes the segmented
memory function especially useful for high speed signals. The time
between each segment is also recorded so that accurate signal
timing can also be measured.

Is it copypasted from some other scope manual or how it feel that this story is not about this scope... 
I have nothing to say. Just "unbelievable". Nearly like Rigol joke about segmented memory.

[nctnico]

Also notice that the scope doesn’t need to
rearm the trigger between each segment, this makes the segmented
memory function especially useful for high speed signals. The time
between each segment is also recorded so that accurate signal
timing can also be measured.

Is it copypasted from some other scope manual or how it feel that this story is not about this scope... 
I have nothing to say. Just "unbelievable". Nearly like Rigol joke about segmented memory.

Your 'if the number ain't high it is definitely bad' attitude is getting boring. There are pros and cons to each design choice. Respect that and don't go jumping up & down because you can't mount a roof rack on a Ferrari.

[rf-loop]

Also notice that the scope doesn’t need to
rearm the trigger between each segment, this makes the segmented
memory function especially useful for high speed signals. The time
between each segment is also recorded so that accurate signal
timing can also be measured.

Is it copypasted from some other scope manual or how it feel that this story is not about this scope... 
I have nothing to say. Just "unbelievable". Nearly like Rigol joke about segmented memory.

Your 'if the number ain't high it is definitely bad' attitude is getting boring. There are pros and cons to each design choice. Respect that and don't go jumping up & down because you can't mount a roof rack on a Ferrari.

Just data? Dataa can not change with opinions.  Speeds are just numbers. Where are they?

[Carrington]

Look what I found. Another "all in one" in the near future, perhaps? 
Source: http://www.stantronic.co.uk/GWINSTEK-GW-GDS-2204E.html

[4cx10000]

What?!  It doesn't look like that in the data sheet!

[pascal_sweden]

You're kiddin', right? I mean, pretty much everything OWON made looks like a damn toy. Not saying it's not good gear (I don't know, I have no experience with OWON) but the looks of OWON gear is cheap and toy-ish, much more than anything else I've seen.

I know very well, that OWON scopes in the past used to look like toys.

But I explicitly referred to the XDS series from OWON, which you have either overlooked,
or you did not bother to take the time and google on it.

Anyhow, to make things easy for you, I hereby provide you a photo of the new XDS series and a link

Professional design look of new XDS series from OWON



Link to XDS series from OWON, with details on product specifications

http://www.owon.com.hk/products_info.asp?ProID=182

What makes you think the plastics clips that keep the boards together won't last? Clips like these are used in lots of electronics products, and have been so for decades. They last because they are in a position that doesn't see much use (boards are usually left alone, they're not removed and replaced on a daily basis), and if they crack under force then the impact the device must have suffered from this to occur would have been so strong that the clips are the least of your worries.

Have you seen the video from Dave Jones about the GW Instek GDS-1000B oscilloscope?
Then you will know what I mean about fabricated prototyping boards that were put together in a hurry,
almost like they had to be ready in time to show off their product on a tradeshow

Link to EEVblog video about GW Instek GDS-1000B Oscilloscope Teardown

https://www.youtube.com/watch?v=9SXZk2tsQ4g?t=902

I have to admit that this was on their entry model, the GDS-1000B series, but it could very well be,
that they use the same design practise on the GDS-2000E series, as both scopes were released in the same delivery time window (a few months ago), and both scopes are cosmetically the same from the outside.

Non-mainstream design look of GDS-1000B series from GW Instek



Non-mainstream design look of GDS-2000E series from GW Instek

[nctnico]

I have to admit that this was on their entry model, the GDS-1000B series, but it could very well be,
that they use the same design practise on the GDS-2000E series, as both scopes were released in the same delivery time window (a few months ago), and both scopes are cosmetically the same from the outside.

I already wrote that the GDS-2000E uses a single board (seen it with my own eyes) so why keep insisting it isn't so? Don't forget that building a product from seperate boards is more expensive due to higher manufacturing (labour) costs. Besides that the 'Panda look' with the dark bezel to make the screen look bigger is common for many oscilloscopes nowadays:

I still don't get why you keep going on about how an instrument looks. It's a friggin instrument and I'd be more worried about the knobs being in the right place and the UI easy to navigate than how it looks. Philips used to have this doo-doo-brown color scheme:

Do you want the mechanic who works on your car to use red spanners and wear a blue jumpsuit?

[pxl]

I am just watching again Dave's video, and here is my personal opinion:
- I don't like that double GEN output on the front, probably I would not buy 2 options for that, and just seeing that two dead outputs, a bit sad
- no trigger out
- design is not my fav, but I could live with that without problem, looks practical
- horizontal pos. knob is just not that responsive, need to turn a lot to move the signal
- not real 1mV/div, just zoomed, why they need to fake that?
- definitely slows down the screen update when moving the signals up and down, not a real problem, but weird a bit it was because the persistence was set to 240ms, even more wierder
- a bit laggy / freezes from time to time
- usually the screen update rate seems fluctuate all the time and generally very slow
- I really like the "app" approach, should the API open and documented, would be a real killer
- when the func. gen was connected but off (at 27.51) and Dave moved his finger close to the scope the noise came up
https://youtu.be/cRe9b7iV6Kc?t=1671
- at the same time the trigger pos was at 0 V, but the waveform was not triggered at all, yet it seemed a well formed sine wave
- at the same time there was a sign at the top of the screen [Trig'd], hmmm
- there is no sign on the screen of the memory mode / sample rate, should be important to see whether a signal could be aliased or not
- short/auto memory? okay, but exactly?
- measurements +
- search +
- segmented mode +
- the "color" type waveform looks really bad and not that useful
- Dave left the filter on (10 kHz) (35:05) but is not written on the screen that we have the filter on!
- no dedicated intensity knob
- strong fan noise
- no HiRes mode

So all in all it looks generally capable, but there are things here and there to be polished later.