A review of the GWInstek 1054B

[saturation]

There are few technical reviews of the 1054B available on the Internet or the forum, so this thread will supplement missing elements and avoid repeating what others have already mentioned.  For an executive summary, see only the bold items.

A key feature of this model as well as variants in the GWInstek lineup is the use of a Zynq SoC allows for faster data acquisition and processing at an entry level price range, $429 as listed.  However street prices vary and my unit was new < $330.

I received a firmware 1.09 1054B early May 2016 and confirmed nearly all items in its spec sheet as accurate  as published by GWInstek. I found one bug, a typo in the help file. 

Much of the GWInstek sales literature, videos and specifications target the market of the Rigol 1054Z.  It has been discussed elsewhere on the forum.  I compared the spec sheets for the 1000B and 2000E series against the 1054Z, see attached. Users can then differentiate both series together, assuming all data in the spec sheet of all models are accurate.

I used marmad's method for estimating wfms/s rates without using a trigger output port, and it returned at least 30,000 [ spec'd at 50,000 wfms/s] at its optimal setting for speed.

https://www.eevblog.com/forum/testgear/a-simple-technique-to-measure-waveform-update-rates-on-dsos-with-no-trigger-out/

I installed the latest firmware to date, 1.14, 3 weeks ago.  The firmware update instructions are vague and on first try it did not load correctly, but it did not brick the scope.  On second try, I used "safe mode" and it installed problem free.  I installed all available "apps" without issue, by simply opening the file on USB.  USB can be hot swapped anytime without locking the DSO.

There is no changelog from GWInstek.  Further testing did not reveal new bugs.

There are 2 marked changes compared to 1.09 not described elsewhere on the Internet, as of this post.  There is a noticeable increase in overall speed compared to 1.09, but not quantifiable.

• A trigger out in now available from the probe compensation port, just turn the adjust knob until option appears
  
  
• The wfms/s rate has been increased to ~ 120,000 wfms/s [ actually measured] from its specified 50,000 wfms/s, at optimal settings
  

A key feature of this DSO is most all processing can be done in real time or after acquisition, on the buffer.  The DSO can thus be configured with all processing turned off, adjusted for fastest wfms/s, a rough trigger set, data captured and analyzed in stop mode: filters, FFT, measurements etc., once the glitch is characterized, the trigger can be optimized for glitch capture, higher memory depth selected, and thus slower wfms/s, for better waveform fidelity. 

<sup>This model was released in middle 2015.  A mostly detailed hardware review was done by Dave on the 100MHz version on Nov 2015 in the forum.  Model variants were reviewed by others and linked below, including detailed teardowns.

https://www.eevblog.com/forum/blog/eevblog-824-gw-instek-gds-1000b-oscilloscope-teardown/msg809286/#msg809286

https://www.eevblog.com/forum/testgear/gw-instek-gds2204e-(200mhz-4-channel-dso)-review/

https://www.eevblog.com/forum/testgear/gw-instek-gds-2000e-teardown-short-review-comparison-with-rigol-ds2000(a)/25/

https://www.eevblog.com/forum/testgear/new-gw-instek-2000-series-digiscope-gds-2202e/

https://www.eevblog.com/forum/testgear/gw-instek-gds-2000e-released/</sup>


More, photos, added comments etc., to follow as time permits.

[SkyMaster]

Saturation, thank you for your review and comparison 

[nctnico]

I'm finding the PDFs difficult to read. Better put everything in a posting!

[saturation]

Thanks.  The color?  I have a B&W version with symbols replacing color.  If its still unreadable let me know; all pdf files I posted are open with security removed so you can change my notes as you please.  nctnico, with your good review I'm sure you know most of these already,  it helped make my decision to get a  1054B.

@SkyMaster your welcome.

I'm finding the PDFs difficult to read. Better put everything in a posting!

[saturation]

New firmware released V1.15 dated 8 31 2016 no change log from Instek.  Have not yet run it through it paces.

The firmware release has not been announced nor appears on the English website.  To find it you need to register as a "VIP" then go to the product of interest.  You can register with a bogus email address, they do not confirm the email address to enter the VIP area.

Problem free install from 1.14.  Ran tests for about 2 hours. Its appears mostly performance enhancements.

Increase FFT speed
Increase system speed
More contrast in intensity grading

[saturation]

Firmware 1.16 released 11 6 2016.  No changelog.  No announcement of its release.  I found it accidentally when another announced he located new firmware for the 2000 series.

Changes I noticed:

Further increase system response speed most notable in FFT
Rewritten HELP system
Band pass filter added

As an experiment, I rolled back firmware to version 1.14 prior to upgrade and it went without restriction or errors

Some details on the board layouts from Xilinx marketing:

https://www.linkedin.com/pulse/what-mystery-processor-resides-inside-new-gw-instek-dso-steve-leibson

[whited]

Saturation,
Have you come across additional information about the apps available for this series? I did a rudimentary search with nothing obvious as to where the app dev is taking place (in house/open) nor what sort of capability is available to app developers. I'm mainly interested as a consumer... programming is not my forte.
A serial decode app would be fantastic but I don't even know if that's feasible?

Just for fun - I see this list of apps available on the download tab of the gw instek site. Has some sort of function generator recently been added?

APP Software   1    en    581    2016/02/15   
GDB-03 Function Module for GDS-1000B Series. (It supports GDS-1000B V1.00 firmware or later.)
 
APP Software   1    en    327.7 KB    2015/09/18   
The DVM apps. of GDS-1000B Series.
 
APP Software   1    en    341.4 KB    2015/09/18   
The digital filter Apps of GDS-1000B Series.
 
APP Software   1    en    337.9 KB    2015/09/18   
The datalog App for GDS-1000B Series.

[saturation]

Yes, you don't really hear much, there are no announcements even with improvements or new firmware.  Even the marketing is very timid.

That said there is no external development of apps to date, not even an SDK.  There is a programmer's guide but that's for controlling the DSO via network connection; I have not tested the network connection.  Yes, a programmer could add apps including a serial decode and more, which would then compete with its more expensive models but how good it would be compared to more expensive models is anyone's guess.

But lets see what is available now to anyone if they chose this DSO.

No function generator, however the compensation port has a frequency adjust from 1k to 100k as a fixed amplitude and duty cycle square wave.  They also added a external sync on the same port, undocumented.

GDB-03 is a stand alone board that generates test signals for a how-to-use a DSO course.  The app simplifies using the DSO so a newbie needn't know how to use a DSO to see the demo waveforms.

The DVM app is a just a large display of the AV and DC voltages.  You can get the same thing in the measurement function.

The Digital filter is a very useful function missing in 1054B as shipped. IMHO, it a mandatory app and I use it a lot.  It comes as a basic function in many DSO that compete with the 1054B.  It does low, high and recently added, bandpass filters.  The cutoffs are very sharp [its a very high order filter], so its extremely useful and comparable to among the best DSO available.  For example, the same filters on the Rigol 1052e are barely usable as the rolloff is too wide, very low order filters.

The Datalog app is a programmable screen capture.  You can set the duration of each capture [ seconds to hours of data per capture] as well as the total run time.

I've used all the functionality except the GDS03 module, network connections, and the go-no go app shipped with the DSO.  To date, there are no bugs I can find that impact measurement quality.  For example, in the Datalog app using 'file function' screen caps cannot be viewed, but can be in the Utility app on the DSO.  Its a user interface  'bug' but causes no issues except UI, otherwise logged data is correct and at the right intervals.

While it appears the 1000B offers far less than competitors were is shines is the quality of the measurements.  Its hard to define since its qualitative, but in general there is no unexpected interaction between functions, measurements are stable as expected, and what is stated in the spec sheet is what is does at minimum, or better. This is what you'd expect only from top players in the DSO market as the video shows.

For example, FFT functions are a simple index of raw computing capability and eevblog has a video comparing this DSO line against others.

It shows what the 1000B can do and what the UI is not, jump to 9:00 for the 1000B review specifically.





 

Saturation,
Have you come across additional information about the apps available for this series? I did a rudimentary search with nothing obvious as to where the app dev is taking place (in house/open) nor what sort of capability is available to app developers. I'm mainly interested as a consumer... programming is not my forte.
A serial decode app would be fantastic but I don't even know if that's feasible?

Just for fun - I see this list of apps available on the download tab of the gw instek site. Has some sort of function generator recently been added?

APP Software   1    en    581    2016/02/15   
GDB-03 Function Module for GDS-1000B Series. (It supports GDS-1000B V1.00 firmware or later.)
 
APP Software   1    en    327.7 KB    2015/09/18   
The DVM apps. of GDS-1000B Series.
 
APP Software   1    en    341.4 KB    2015/09/18   
The digital filter Apps of GDS-1000B Series.
 
APP Software   1    en    337.9 KB    2015/09/18   
The datalog App for GDS-1000B Series.

[whited]

Thanks Saturation - great info. I looked at all the info.
One thing I am very curious about is the capability to analyze and compute from scope memory (on frozen waveform). Can all of the 1000B computations access the full memory of stopped waveforms? Or are only subset of computations available on waveforms in memory? I am not sure how this capability compares across competing scopes like the 1054Z, and I have only used analogue scopes - no DSO.

While it appears the 1000B offers far less than competitors were is shines is the quality of the measurements.  Its hard to define since its qualitative, but in general there is no unexpected interaction between functions, measurements are stable as expected, and what is stated in the spec sheet is what is does at minimum, or better.

Yes, this is another aspect I have been having trouble evaluating compared to competitive DSOs. It seems with GDS-1054b Zynq SoC there are some architectural improvements underlying some of the usability aspects of the user interface, like general responsiveness, but also the ability of processor to access memory directly due to SoC platform. E.g. RAM controller, FPGA, and display driver included in Zynq SoC architecture rather than separate.
I do not have experience with system architecture of competitive DSOs, so it is hard for me to pick out what is significant about the GDS-1000b series implementing Zynq.

Here Dave mentioned the 1000B series uses Z-7010 chip, which is dual core 866MHz cpu version of Zynq 7000 family. You can see family comparison here: https://www.xilinx.com/support/documentation/selection-guides/zynq-7000-product-selection-guide.pdf
Also, interestingly, XILINX case study on GW Instek GDS-2000E and Zynq SoC here: https://www.xilinx.com/publications/prod_mktg/zynq7000/goodwill-casestudy.pdf

GW Instek seems to be terrible at marketing, yet produce a good product... very hard for me to understand apples to apples against other DSOs. App SDK maybe is a wildcard, but would be great if GW Instek can leverage open dev community.

Edit: I think all the effort went into product dev, and nowhere else I like the nod to Star Wars @ 55s
https://youtu.be/_uVC7eDQ7fo

[saturation]

Welcome.  More later.  Yes, that is one great strength, all processing [ so far I've discovered, since I do this by default ] can be done on "frozen waveforms", with acquisition stopped . 

Thus, you can dedicate all the horsepower to CPU intensive issues like FFT, complex math  [ you can build your own function rather than use preset as in the Rigol 1054z], filtering etc., making the full 10M of memory [ were applicable, in FFT its only 1M max] practical and minimize or eliminate the slow grind as Dave/eevblog shows in the video.   For most all tasks then, processing can be 'instant' with acquisition turned off. 

Thanks Saturation - great info. I looked at all the info.
One thing I am very curious about is the capability to analyze and compute from scope memory (on frozen waveform). Can all of the 1000B computations access the full memory of stopped waveforms? Or are only subset of computations available on waveforms in memory? I am not sure how this capability compares across competing scopes like the 1054Z, and I have only used analogue scopes - no DSO.

[whited]

I am also trying to compare against GDS-2000E series. It is interesting because 120k wfm/s added for GDS-1000B is matching the 2000E. They both use zynq 7010 SoC.

Seems 2000E advantages over 1000B are

• I2C/SPI/UART/CAN/LIN Serial Bus Trigger and Decoding
• bigger 8" screen, but same 840x480 resolution
• 2 ADCs so 1Gs/s can be on two channels simultaneously
• updated FFT with some sort of search and peak detection plus no overlap display of time and frequency domain (can 1000B do this?)
• expanded time base to 1ns/div
• VPO (visual persistence oscilloscope)
• Memory segmentation

I think I missed some. But I like the Serial decode and trigger in 2000E. But what if they add it via app to 1000B??
I will email Instek.

[saturation]

Yes, the 2000E is older than the 1000B.  The 1000B is more refined learning from the 2000E growing pains, you can read the thread about bugs found by early adopters on V1.x release, which did not exist in the 1000B V1.x release.

What functions both support are nearly identical in specification and speed suggesting they have very similar cores but the 1000 is missing hardware present in the 2000, that would make it thus of limited hackability and secure Instek's product differentiation and pricing value.

Originally the 1000 had 50,000 wfm/s but the first firmware update quitely 'opened' it so maybe 120k is its maximum, as well as engaged the sync out capacity again suggesting this was just crippling the core in firmware.

Yes, the detailed review of nctnico is best for the 2000E.  Your list is correct but comments on bolded items :

Yes, you can fudge it.  Since the FFT can be analyzed on the buffer, it doesn't need the time domain waveform viewed, however you cannot turn the channel its using off.  What I do is turn the acquisition off, RUN->STOP then shrink the waveform until it disappears or roll it out of the way! See photos later in these replies.

Yes, expanded timebases needed to properly view waveforms at 200+ MHz bandwidth.  This is much bigger than it appears.  The usuable bandwidth is easily >= 500 360 MHz.  However, with 1GS/s you'd have to keep an eye out for aliasing, you can use the built in filters as a DIY anti=aliases frequencies as you please, they have very high order cutoffs. 

Keep in mind 200 MHz probes no-name Chinese probes are < $15 each, < 100 MHz $6 each, but a single Chinese branded 500 MHz probe is about $100 each, and at 500 MHz we are at the limit for passive probes and true quality matters in its performance.

https://www.eevblog.com/forum/testgear/gw-instek-gds2204e-(200mhz-4-channel-dso)-review/msg855862/#msg855862



I do not know if 1000B has enough onboard memory to implement serial decode, even if an app could be written for it.

1000B do not have segmented memory, I cannot recall if this a hardware issue.

I'm guessing the 1054B is a firmware crippled 1104B as its entirely identical except for the bandwidth.  All that's need is to release the bandwidth filters.

Both uses VPO, but Rigol's is far superior as far as CRT like view.  I think Instek can improve on this easily, they need to increase the contrast options to the variable persistence, what exists barely changes the appearance.

I am also trying to compare against GDS-2000E series. It is interesting because 120k wfm/s added for GDS-1000B is matching the 2000E. They both use zynq 7010 SoC.

Seems 2000E advantages over 1000B are

• I2C/SPI/UART/CAN/LIN Serial Bus Trigger and Decoding
• bigger 8" screen, but same 840x480 resolution
• 2 ADCs so 1Gs/s can be on two channels simultaneously
• updated FFT with some sort of search and peak detection plus no overlap display of time and frequency domain (can 1000B do this?)
• expanded time base to 1ns/div
• VPO (visual persistence oscilloscope)

I think I missed some. But I like the Serial decode and trigger in 2000E. But what if they add it via app to 1000B??
I will email Instek.

[wraper]

Both uses VPO, but Rigol's is far superior as far as CRT like view.  I think Instek can improve on this easily, they need to increase the contrast options to the variable persistence, what exists barely changes the appearance.

Dunno, at least on 2000E, you set persistence to minimum setting (16ms) and get pretty much Rigol like "analog view". Except zero intensity is not really zero, so won't completely hide "low intensity" waveforms.

[saturation]

eevblog screen cap of FFT setting of 1104B




My copy of the same fo= 1 MHz , FM at 5kHz and 500 Hz dev
a = RUN STOP
b = 1Mpts on FFT, 50MS/s sampling
c = 20dB vertical scale, 5kHz horz. div.
d= spur from my FM generator
e = suppressed source waveform
f = CH1 still active but RUN STOP and waveform 'rolled' away



fo = 50 MHz  FM 5kHz and 500 Hz dev
span to 10 kHz for clarity
a =fo suppressed [ not sure why]
b = source waveform suppressed





fo = 75 MHz,  5kHz and 500 Hz dev; gain to 500mV and span to 20 kHz for clarity, now past the 1054B rated bandwidth

[whited]

Glad you are sharing screenshots, saturation. Is source waveform amplitude the same for all 3 carrier frequencies? It looks like -40dbv attenuation from 1MHz to 75MHz? That's a lot. Or maybe it's happening before the scope or I am missing something?

Instek has not yet replied regarding serial decode app for 1000B. It has only been 1 day though. Perhaps there is not enough memory as mentioned.

[saturation]

Welcome, I've had some for a while never posted. 

I'm not sure if my FM generator is good past 25 MHz; Vo likely drops with higher frequency.

Here's a crude frequency response curve I generated some time back and never posted.

This is the raw dBVpp  and dBV using the DSO's native RMS calculator.  Its consistent with a typical -6dB/octave low pass filter with 66 MHz at -3dB point.

Response from a Tek 284 pulse generator.  Between 6ns and 3 ns, called it 4.5ns or 0.35/4.5 = 78 MHz, consistent with the frequency test.

Glad you are sharing screenshots, saturation. Is source waveform amplitude the same for all 3 carrier frequencies? It looks like -40dbv attenuation from 1MHz to 75MHz? That's a lot. Or maybe it's happening before the scope or I am missing something?

Instek has not yet replied regarding serial decode app for 1000B. It has only been 1 day though. Perhaps there is not enough memory as mentioned.

[saturation]

Yes, the VPO is a major improvement over its prior models, but it still could be improved further with a firmware tweak.  Some waveforms looks better than others, but my view is against my old Trio CRT.  The fast update rate makes it much easier to see transients.  I think if zero intensity were made truly zero it will help a lot, it would give the adjust for 'contrast' a greater span and act much more like a true CRT intensity adjust.

Both uses VPO, but Rigol's is far superior as far as CRT like view.  I think Instek can improve on this easily, they need to increase the contrast options to the variable persistence, what exists barely changes the appearance.

Dunno, at least on 2000E, you set persistence to minimum setting (16ms) and get pretty much Rigol like "analog view". Except zero intensity is not really zero, so won't completely hide "low intensity" waveforms.

[wraper]

I think if zero intensity were made truly zero it will help a lot, it would give the adjust for 'contrast' a greater span and act much more like a true CRT intensity adjust.

And what is practical use for it?

[nctnico]

I think if zero intensity were made truly zero it will help a lot, it would give the adjust for 'contrast' a greater span and act much more like a true CRT intensity adjust.

And what is practical use for it?

None. Actually it is rather counter productive to want to have a dim (invisible) signal on a DSO! Fortunately GW Instek didn't hop on the 'emulate a CRO display' band wagon but went the practical way instead by starting at (approx.) 50% intensity and go from there. Ofcourse you can turn the trace intensity down to invisible.

[whited]

Yes this makes more sense.

Above I missed memory segmentation in my 2000E diff list. You mentioned it too. I am adding to my list (since it's there, so I don't forget).

Welcome, I've had some for a while never posted. 

I'm not sure if my FM generator is good past 25 MHz; Vo likely drops with higher frequency.

Here's a crude frequency response curve I generated some time back and never posted.

This is the raw dBVpp  and dBV using the DSO's native RMS calculator.  Its consistent with a typical -6dB/octave low pass filter with 66 MHz at -3dB point.

Response from a Tek 284 pulse generator.  Between 6ns and 3 ns, called it 4.5ns or 0.35/4.5 = 78 MHz, consistent with the frequency test.

Glad you are sharing screenshots, saturation. Is source waveform amplitude the same for all 3 carrier frequencies? It looks like -40dbv attenuation from 1MHz to 75MHz? That's a lot. Or maybe it's happening before the scope or I am missing something?

Instek has not yet replied regarding serial decode app for 1000B. It has only been 1 day though. Perhaps there is not enough memory as mentioned.

[tautech]

A screenshot captured on a USB stick is much more helpful and clear for those following than something like this:



 

[saturation]

Sorry, mea culpa, I have them but didn't have time to take it out; will replace that photo with the screen grab.

A screenshot captured on a USB stick is much more helpful and clear for those following than something like this:
 

[saturation]

I agree with nctnico as is any instability of the waveform can be seen with the defaults as it is today, its just less starkly contrasted.   So its practically not an issue, its more usability and a way to see such events better.

The spec sheet states that the screen supports 256 shades, it looks less than half that so it can easily be improved.  The  Rigol 1054z traces are only 64 grey shades. IIRC. 

I think over time some improvements in usability would help, but they have put a lot of good work into its measurements fidelity and being bug free.

As example of added peculiar usability decisions in 1054B, the statistics function begins at n = 2 samples.    Rotate the encoder to increase or decrease up to n= 1000 where it stops.  If they let it simply cycle back to 2 it would save 1/2 the time to get n = 500 and it would be instant to get n = 1000, for the best analysis possible, code wise that is faster to implement than adding a coarse and fine, which would make it even speedier.

That said and aside, the numbers it generates are superb.

I think if zero intensity were made truly zero it will help a lot, it would give the adjust for 'contrast' a greater span and act much more like a true CRT intensity adjust.

And what is practical use for it?

None. Actually it is rather counter productive to want to have a dim (invisible) signal on a DSO! Fortunately GW Instek didn't hop on the 'emulate a CRO display' band wagon but went the practical way instead by starting at (approx.) 50% intensity and go from there. Ofcourse you can turn the trace intensity down to invisible.

[whited]

Saturation, I have a little more time now to see your numbers; seems 1050B has ~70 MHz bandwidth instead of its rated 50 MHz. It is not too surprising but nice to see in a budget scope. I wonder if all GDS 1000B and 2000E have similar conservative bandwidth rating?

Maybe ntcnico can give another datapoint on frequency response? ntcnico I read your review thread; I hope I didn't miss it. Nevermind I found it here on post #67 https://www.eevblog.com/forum/testgear/gw-instek-gds2204e-(200mhz-4-channel-dso)-review/?all
It's a good measurement showing 2204E roll off (rated 200MHz bandwidth)
-3dB @ 290MHz with 50mV/div, more steep curve
-3dB @ 210MHz with 10mV/div, more gradual curve

I am personally thinking 2070E is a sweet spot @ $920... but 2204E is so close @ $1256. Can I convince myself?
I know this is a 1054B thread - I try to keep it relevant.
Thanks to you all.

Yes this makes more sense.

Above I missed memory segmentation in my 2000E diff list. You mentioned it too. I am adding to my list (since it's there, so I don't forget).

Welcome, I've had some for a while never posted. 

I'm not sure if my FM generator is good past 25 MHz; Vo likely drops with higher frequency.

Here's a crude frequency response curve I generated some time back and never posted.

This is the raw dBVpp  and dBV using the DSO's native RMS calculator.  Its consistent with a typical -6dB/octave low pass filter with 66 MHz at -3dB point.

Response from a Tek 284 pulse generator.  Between 6ns and 3 ns, called it 4.5ns or 0.35/4.5 = 78 MHz, consistent with the frequency test.

Glad you are sharing screenshots, saturation. Is source waveform amplitude the same for all 3 carrier frequencies? It looks like -40dbv attenuation from 1MHz to 75MHz? That's a lot. Or maybe it's happening before the scope or I am missing something?

Instek has not yet replied regarding serial decode app for 1000B. It has only been 1 day though. Perhaps there is not enough memory as mentioned.

[saturation]

The 1054B is <$366 street price via Tequipment not including the eevblog discounts and possibly other promotions.

At the published bandwidth and list price they are about the same proportion, 50 MHz vs 200 MHz is still 4x more at 4x the price as well as serial decodes and segmented memory.

However, its in the extreme end, the usable bandwidth were you can get more.  Its likely to still go past 500 MHz based on the curve slope at its more sensitive vertical amp setting.

So the big question is do you need that much bandwidth to pay more for it?  Or in the end its all part of happy holidays?

Saturation, I have a little more time now to see your numbers; seems 1050B has ~70 MHz bandwidth instead of its rated 50 MHz. It is not too surprising but nice to see in a budget scope. I wonder if all GDS 1000B and 2000E have similar conservative bandwidth rating?

Maybe ntcnico can give another datapoint on frequency response? ntcnico I read your review thread; I hope I didn't miss it. Nevermind I found it here on post #67 https://www.eevblog.com/forum/testgear/gw-instek-gds2204e-(200mhz-4-channel-dso)-review/?all
It's a good measurement showing 2204E roll off (rated 200MHz bandwidth)
-3dB @ 290MHz with 50mV/div, more steep curve
-3dB @ 210MHz with 10mV/div, more gradual curve

I am personally thinking 2070E is a sweet spot @ $920... but 2204E is so close @ $1256. Can I convince myself?
I know this is a 1054B thread - I try to keep it relevant.
Thanks to you all.