GDS-2000A series New economic oscilloscope by GW Instek comes to market

[marmad]

Ok - first video complete.  Be gently, it's my first time.

http://youtu.be/kbNKVD-ZoIY

I'll try to be a bit more organized and go point by point through some of the more interesting features the next time.

Thanks for posting, Greg - I'm sitting down with a snack to watch right now 

[marmad]

(Still watching...) One quick comment/question about the screen: do you think Instek is doing anti-aliasing during display (I mean pixels - not waveforms) but not for screen caps (i.e. better looking while running)?

[grego]

Definite possibility.  I will ask InstekUSA if they can get an engineer to confirm.

[marmad]

Ahhh... the sweet, sweet "where was that?" of early device ownership 

Nice, Greg - thanks for making and posting that. It was great to see it in action outside of the ridiculous robotic Instek engineers' videos (they really need to get these out to others for review instead of leaving it to their engineers)  I will definitely have requests for the next video (I think you already know one from my PM to you).

One question: it wasn't completely clear from the video - can you turn off interpolation completely (i.e. just sample dots)?

[grego]

Yes you can do dots or vectors.

[marmad]

Yes you can do dots or vectors.

Ahh... good to hear!

Another suggestion for the next video: (I'm really not trying to harp on this    but I really think Instek needs to be a little more clear in their promotional literature - it's kind of hidden*), is to explain in detail exactly how the Instek uses it's memory. Because with the 4-channel version, you can have 1M record lengths in AUTO/NORMAL modes by using channels from alternate banks - but it's impossible to do that with the 2-channel version, so prospective buyers should know that the 2-channel versions are basically a 1M / 500k DSO (except in single-shot mode).

Edit: * In the Features list of the User Manual it's written:
...
- All models feature a real-time sampling rate of 2GSa/s and an equivalent time sampling rate of 100GSa/s.
- Deep memory: 2M points record length.
- Waveform capture rate of 80,000 waveforms per second.
...
...which I think is a little misleading for 2-channel buyers.

[grego]

Ahhh... the sweet, sweet "where was that?" of early device ownership 

Nice, Greg - thanks for making and posting that. It was great to see it in action outside of the ridiculous robotic Instek engineers' videos (they really need to get these out to others for review instead of leaving it to their engineers)  I will definitely have requests for the next video (I think you already know one from my PM to you).

One question: it wasn't completely clear from the video - can you turn off interpolation completely (i.e. just sample dots)?

You can see the menu option on the left at 10:45 for example.

[marmad]

You can see the menu option on the left at 10:45 for example.

Doh!     Yeah, seems obvious now - I was looking elsewhere during that part and only thought of it when you mentioned sin(x)/x.

[Hydrawerk]

Thanks for the video. It has good quality.
Well, does the scope really crash when you press Measure and Display all soft button? It's strange.

[grego]

Thanks for the video. It has good quality.
Well, does the scope really crash when you press Measure and Display all soft button? It's strange.

Yep - it does - well, you hit display all and it crashes when you select your source.  Instek confirmed the bug and it's apparently already fixed in the next revision which should be available shortly.

[Hydrawerk]

When you set the Record Length to short, how long is it? Is it 8kpoints, 100kpoints per channel or what?
Anyway, the statistics feature looks great although I am not sure if i would need it as a hobbyist...

[grego]

When you set the Record Length to short, how long is it? Is it 8kpoints, 100kpoints per channel or what?
Anyway, the statistics feature looks great although I am not sure if i would need it as a hobbyist...

I mentioned it in the video - but it sets it to 1M.

[marmad]

I mentioned it in the video - but it sets it to 1M.

I would think it would be 500k or 1M, depending on the mode you're in and other channels being on, no?

[grego]

I mentioned it in the video - but it sets it to 1M.

I would think it would be 500k or 1M, depending on the mode you're in and other channels being on, no?

Yes - but the reference to 1M was the memory depth in short mode - not depth per channel/ADC.  Two separate things.

[marmad]

Hi Greg,

I'm very curious about the Instek's LA module. An in-depth demonstration of it's decode/analysis abilities would be greatly appreciated, if you have a circuit around to demo on.

[grego]

Hi Greg,

I'm very curious about the Instek's LA module. An in-depth demonstration of it's decode/analysis abilities would be greatly appreciated, if you have a circuit around to demo on.

That's in the works.  I need to get one of my eval boards set up to demo it properly.  Plus have some free time.   But I'll have that in the next week or so.

[Hydrawerk]

Grego, it seems that your Trig Out signal is 1.2 Volts peak-peak. That's low but still acceptable. What do you think about scope that was bought by Kiriakos A. Triantafillou? http://www.ittsb.eu/GDS-2102A%20Wfms%20measurments.html
He wrote:

c) The Trigger-Out has an voltage range not higher than 200 mV AC, which varies significantly depending the time base of the scope and usually ranges from 120 mV to 5 mV and even less than that.

I think that some of the GDS-2000A scopes have a hardware problem. But your instrument is OK, it seems. As seen here:

[grego]

Grego, it seems that your Trig Out signal is 1.2 Volts peak-peak. That's low but still acceptable. What do you think about scope that was bought by Kiriakos A. Triantafillou? http://www.ittsb.eu/GDS-2102A%20Wfms%20measurments.html
He wrote:

c) The Trigger-Out has an voltage range not higher than 200 mV AC, which varies significantly depending the time base of the scope and usually ranges from 120 mV to 5 mV and even less than that.

I think that some of the GDS-2000A scopes have a hardware problem. But your instrument is OK, it seems. As seen here:

Mine is actually higher than that as you can see in:

https://www.eevblog.com/forum/testgear/gds-2000a-new-economic-oscilloscope-by-gw-instek-comes-to-market/msg217468/#msg217468

More like 1.5-1.6V which is fine.

I don't think it's a hardware problem - I can't say without being there but I haven't seen anything to indicate any issues.

I'm not going to generalize with "some of the GDS-2000A scopes have a hardware problem".  Especially since I appreciate what Kiriakos has done but I don't think he's measuring things 100% correctly -- I wish he had a second scope at his place to verify his data with.

[marmad]

Hey Greg - Another thing I'd love to see you measure - since Rigol / Agilent X / GW-Instek don't provide this info in their specs (as used to be common with analog scopes) - is the Trigger Out delay. This is easily accomplished by feeding the Trigger Out into one of your other channels. Here is an image I just made of the Rigol's delay (which appears to be ~220ns).

[Hydrawerk]

I'm not going to generalize with "some of the GDS-2000A scopes have a hardware problem".  Especially since I appreciate what Kiriakos has done but I don't think he's measuring things 100% correctly -- I wish he had a second scope at his place to verify his data with.

Well, it seems that Kiriakos measured the 80 kHz spikes with his multimeter. Multimeter never shows you Vpp, it shows Vrms.
But he also wrote that most multimeters cannot measure the output frequency. That's bad. My DSOX-2002A has Trig Out 5Vpp, so it can be measured by any multimeter...

[jahonen]

For me it looked like that the waveform capture is not continuous, picoscope screenshots showed long periods with no trigger out pulse at all. I think I saw this in Instek marketing video too. Perhaps this could be further checked?

Regards,
Janne

[grego]

For me it looked like that the waveform capture is not continuous, picoscope screenshots showed long periods with no trigger out pulse at all. I think I saw this in Instek marketing video too. Perhaps this could be further checked?

Regards,
Janne

Sure - but remember the 80k wfm/s is the average peak so you would expect things to be faster/slower depending on what the scope is doing.

I'll check it again but I don't want us all to get wrapped up in it.  Personally I think it's a display issue with the pico because at a lower timebase you don't see any delays.

[marmad]

Playing around, I just found out that I can make the Rigol count it's own wfrm/s rate fairly accurately by just feeding it's Trigger Out into one of it's own channels (the other channel is turned off), turning on the frequency counter for that channel, and then sending a 1MHz square wave to the External Trigger In. The displayed frequency is very close to the numbers I recorded using my Fluke for counting. Maybe this works for the Instek as well?

[marmad]

While looking at this Agilent doc about waveform update rates, I noticed that the Tek DPO3000 series listed in their tables (and also the Tek DPO2000/4000 to a lesser degree) has a very similar wfrm/s output graph to the GDS-2000A series. And by output graph, I don't mean the absolute rates (the Instek is clearly faster at many timebase settings) - I mean that both of their wfrm/s rates ramp up - and then stay high - to the 50ns/div setting (or thereabouts), and then drop way down for the final few settings. It makes me wonder if there is some design similarities in the strategies both Tek / Instek used for intensity-grading, etc, that makes the graph the way it is (as opposed to the Agilent or Rigol). I'd love to know more about particular design choices being made in this area - although I guess they're all guarding their secrets closely 

[Hydrawerk]

The GDS-2000A screen design is quite similar to Tektronix DP04000. Even both scopes have no pushable knobs...

I noticed that the Tek DPO3000 series listed in their tables (and also the Tek DPO2000/4000 to a lesser degree) has a very similar wfrm/s output graph to the GDS-2000A series. And by output graph, I don't mean the absolute rates (the Instek is clearly faster at many timebase settings) - I mean that both of their wfrm/s rates ramp up - and then stay high - to the 50ns/div setting (or thereabouts), and then drop way down for the final few settings.

Well, really? Then it means that measurements by Kiriakos are wrong. http://www.ittsb.eu/GDS-2102A%20Wfms%20measurments.html
EDIT: Once he measured 154870 waveforms per second. That's strange.