GW Instek MDO-2072EX MSO / MDO Oscilloscope - Review

Table of contents

RoadTest: GW Instek MDO-2072EX MSO / MDO Oscilloscope

Author: nermash

Creation date:

Evaluation Type: Test Equipment

Did you receive all parts the manufacturer stated would be included in the package?: True

What other parts do you consider comparable to this product?:

What were the biggest problems encountered?:

Detailed Review:

First of all, I would like to thanks Element14 and GW Instek for the privilege to roadtest this very impressive piece of test gear.


My experience in DSOs is so far focused on budget standalone models from Owon and Siglent (and Welec if anyone remembers), and high-end Picoscope 5444D USB based MSO.


I will just mention here specs that were interesting to me when applying to this roadtest:

  • 1 GS/s capture rate per channel with 10 MB max memory
  • 120 000 waveforms/s
  • Spectrum analyser with 500 MHz bandwidth on all models
  • 8” TFT screen 800x480
  • Arbitrary waveform generator, dual channel, 25 MHz, 14bits@200 MS/s, 16 kpoint memory
  • Serial decoding out of the box
  • True RMS DMM
  • Power supply 5V max
  • Frequency response analysis




In this part of reviews I usually go over the unboxing, initial overview of overall look and feel, weight, robustness, build quality, supplied accessories.

And also other impressions, calibration certification, quick guides, manuals, and of course setup and SW installation process.


What is in the box?

You get the scope of course, 2x 70 MHz 1/10x switchable probes, multimeter probes, power supply connecting cable and 2x BNC terminated cables.

Unlike budget models, this MDO comes with Declaration of traceable calibration, confirming that device meets or exceed published specifications.


Here is it compared to my 7” screen Siglent SDS1202X-E scope. With only 1” smaller screen and being more compact unit, it does look significantly smaller.


Scope itself is tall and nice looking unit, with great build quality and with nicely sized buttons and knobs.

Rotary encoders feel like quality units, not like those typically found on lower cost scopes.


Measured boot time: 29 seconds


All the buttons are logically placed, and UI experience is great, scope is fast and responsive.


First nuisance for me here is the lack of push functionality on Variable and Scale knobs (vertical and horizontal).

On horizontal position knob, push serves as zoom functionality.


I would also personally prefer having push on knob for selections instead of using dedicated Select button bellow Variable knob, as well as having them switch between fine and coarse vertical sensitivity adjustment (which this scope does not have).


UI is however highly polished, with many options not found on budget scopes, such as:


  • Adjustable Rise/Fall time values
  • Variable probe calibration frequency
  • Properly implemented velocity function on knobs
  • Probe deskew


Regarding other accessories, DSO probes feel like quality probes, however they do not come equipped with grounding attachments for high speed signals, nor any BNC adapter.

Probes are definitely not usual probes that come from China, with 2 features that I particularly like:


* plastic locking ring on the BNC connector side that turns 90° instead of bare metal

* white nylon inserts in probe hook grabber attachments that allow for firm and positive mount onto probe itself


Multimeter probes are nice and sharp, with fairly flexible  leads, same as power supply connecting cable with alligator clips. Although they feel like silicone leads, I am not 100% convinced that they are actually silicone.





In the following sections I will present examples and experiments I used to roadtest this MDO, in order to try to show it's strengths and weaknesses.

Some experiments and setups are basic demonstration of capability and some are more elaborate, depending on my own personal preference and my interests.


AM modulation


Like many others, I like to use AM modulated signal to assess intensity grading properties of DSOs.


Signal source is my Agilent 33622A generator, 1 kHZ AM modulation on 1 MHz carrier, with 100% modulation depth.

NOTE: I would have used inbuilt AWG for this, however it's max AM frequency is 100 kHz.




While I could not find information on exact number of intensity grading levels for MDO2072EX, I guess that this is around 64.

Datasheet for MDO2000EX series does mention 256 color gradients at one place, but I could not find color gradation option anywhere.


Here it is compared with Siglent, both at 70% and 50% intensity settings.


{gallery} AM modulation


GW Instek 70% intensity


Siglent 70% intensity


GW Instek 50% intensity


Siglent 50% intensity

Setting display persistence to 16 ms in order to "force" VPO mode showed no perceivable difference.


I have a feeling that Siglent's 256 levels paints a prettier picture.

Also, there is a visible difference in 10 vs 14 horizontal divisions.


Bandwidth sweep


While having the 120 MHz AWG connected, I wanted to test -3 dB bandwidth on this MDO.

AWG is set in linear sweep mode, with sine signal going from 20 MHz to 120 MHz in 10 seconds.

Amplitude is set for 1.6 Vpp.

MDO is set to 200 mV/div with 1s/div horizontal scale, triggered by sync pulse and CH1 terminated externally with 50 Ohm. Unfortunately MDO does not have internal 50 Ohm termination.



MDO is set to single shot mode, and trigger point moved to far left of the screen on -5s mark, so that 20 MHz lands on first division, and 120 MHz completes on the last division.


We can see that it has -3 dB point at around 110 MHz, with corresponding voltage of 0.8 V x 0.707= 565 mV on positive and negative period. I expect that frontend is the same for all MDO versions, and just limited in software for each version. Interesting thing is that same frontend is used for SA functionality with 500 MHz bandwidth, although SA specs are saying “uncalibrated” and do not give out any linearity numbers for the entire span.

Unfortunately my signal generation capabilities are limited at 120 MHz, so I could not verify this theory.

For additional information, frequency flatness of 33622A generator is specified at     ± 0.50 dB.


Spectrum analyzer overview - 10 MHz sine signal


Next test was the first experience with Spectrum analyzer part of MDO. Switching to SA mode is very easy, pressing Option button while in DSO mode brings up selection of measurement options: AWG, Spectrum Analyzer, DMM, Power Supply.


Signal source is again Agilent 33622A generator, 10 MHz sine with 0 dBm amplitude into 50 Ohms, and 0 V DC offset.


Using Freq&Span menu option I selected 1 MHz span, 10 MHz center, and selected an option to put a marker to center. RBW mode set to Auto, and Span:RBW set to 5000. Hanning window mode wa used throughout the entire testing.



Everything is correctly displayed and all settings are easily visible for quick check.

Marker shows measurement of -0.8 dB of signal power, which can be accepted as true due to cable and connectors losses.

Noise floor at this sensitivity setting (200 mV/div) sits at arround -80 dBm, which is pretty good result for this class of instruments.


Next I turned on MAX HOLD ON option in spectrum traces, and again resulting trace is nicely contrasted to changing signal.



So, let's try to reduce further span and test several more traces option, seen in gallery bellow.



{gallery} SA traces


10 MHz with 100 kHz span, MAX HOLD, amplitude change from -10 dBm to 0 dBm


10 MHz with 100 kHz span, MIN HOLD, amplitude change from 0 dBm to -10 dBm


10 MHz with 100 kHz span, 256 Averages


Again, nice presentation and performance, no objections.


Time to check cursors, pressing Cursor button on the front panel brings up H cursors, and pressing again brings also V cursors.


Cursors are easily manageable by using H cursor menu option to select one or both cursors, and then Variable know moves them.


Here I set the first one to arbitrary position 20 kHz away from the main signal, and leaving the default S (seconds) for H units.

Windows shows useful information such as signal power at both cursors and frequency difference between them.


Once again, I am impressed by ease of use and intuitive nature of selecting various options and performing basic SA operations and measurements.

Of course, there is a learning curve involved, especially for beginners, however with some enthusiasm and effort even beginners will quickly come to speed with this.


Spectrum analyzer - RF signals


This time I wanted to see some real world RF signals introduced to MDO, and again I used Agilent 33622A as source.

During my review of this AVG, I managed to get help from the creator of ArbIQ software, which was kind enough to provide me with waveform data to create 4 QAM64 DVB-C signals with PRBS payload.


This is the screenshot from the app, 4 carriers at 22-31-40-49 MHz, 5120 ksym/s symbol rate, alpha factor of 0.25 and raised cosine (Nyguist) filter applied.



Here is the SA view on MDO, with some cursors on.



Due to 0.25 Alpha coefficient, actual channel occupied bandwidth is 6.4 MHz.


With 256 averages on, we get a much clearer picture of the captured signals.



I also turned on Search functionality, Max Peak method. Number of peaks selected is 4.

Of course, due to flat top signals, automatic peak search will have some difficulties guessing the correct peak frequencies.


Peak table menu button gives an overview of identified peaks with frequency and magnitude.

As I said, 4 flat top signals are not the best way to demonstrate peak search, but I wanted to make an example of peak search functionality, and I think that it still does a decent job.



Let's compare SA view with FFT representation of the same signals, using the same timebase and vertical scale settings.

This MDO has max 1 Mpoint of FFT memory, which should provide enough resolution for decent representation of signals in frequency domain.



Even with 1 Mpoint of FFT memory selected, signal representation is far worse than using SA part of this MDO.


Serial decoding and triggering


This MDO comes equipped with nice range of serial decoding and triggering options that cover most of educational or hobbyist needs:  I2C, SPI, UART serial bus and CAN/LIN bus.


For i2c decoding and triggering test, I used Mbed LPC1768, programmed it to read data from 24C02 serial eeprom.

Eeprom is filled with values that translate to ASCI characters "GW INSTEK MDO2072EX ROADTEST". On the end of eeprom memory space I placed values that translate to ASCI characters "TRIG".


Bus feature on the MDO is enabled, set to I2C, threshold levels for SDA and SCL set to CMOS recommended levels 1.65V.

Inputs are set SCL on CH2, and SDA on CH1.

Triggering is also set for i2c, trigger event being Start frame.

Here is the picture of captured and decoded data:




Here I have one gripe with this MDO, Bus display only provides options for HEX and BINARY. ASCI would be nice to have also.


Event table gives a ordered representation of decoded captured data.

Sometimes random value is added in front of data itself that increments with each data point.

I haven't been able to figure why this happens.




And Data Detail provides a matrix view address/value:



Last 4 bytes of eeprom are set with characters "TRIG", or 54 52 49 47 HEX.

I tried triggering on that word since MDO i2c trigger allows up 5 bytes of data to trigger on.


I have not been able to get a capture on that word. I am not sure if is something in my setup, or because sometimes MDO adds random value in front of data itself...


So I tried trigger on single byte, ASCI *, HEX 2A.




Let's suppose that we now want to see how many trigger evens there are in this capture.

Search functionality can do this very easily. Enabling search and selecting Copy Trigger Settings to Search MDO locates all search events.




All identified search evens can be seen in Zoom mode also with white opaque markers above them.

Moving through events is also easy with left and right arrow buttons.




Waveform update rate and VPO


This MDO claims maximum waveform update rate of 120 000 wfm/s, which is a great figure for this class of instruments.

High update rate allows for more "natural" look of captured signal that resembles that on analog scopes.

It is useful for spotting rare transients that can ruin your day.


VPO stands for Visual Persistence Oscilloscope, which is GW Instek's technology for visual enhancement of displayed waveform.

This VPO technology takes in consideration signal strength beside time and amplitude to provide more genuine representation of signal, again close to analog CRT display.


I have noticed that VPO is enabled only with display persistence enabled, which is somewhat vaguely mentioned in the manual:

"The persistence function allows the MDO 2000EG/2000EX to mimic the trace of a traditional analog oscilloscope"


Since there is no Trig Out output on this MDO, I could not verify the maximum waveform update rate numbers.

However, I tried this with my Mbed LCP1768 outputting 50 kHZ square wave signal with 50% duty cycle. Once in 50k cycles, a glitch pulse is created that drops to 0 after 1 us and returns to Vcc.


I tried to capture this using Run/Stop key (not triggering on runt pulse), with 16 ms display persistence.

With 1000 points of capture memory, this glitch pulse is easily visible 2-3 times per second.




Segmented memory capture and statistics, runt pulse triggering


Nice feature of this MDO is capability of Segmented memory acquisition, ability to place one triggered signal capture into one segment of available memory.

This is useful when trigger occurrences are far between, and continuous capture after first trigger would waste memory.

With segmented memory enabled, available memory can hold much more triggered captures.


Signal source is same as before, Mbed LCP1768 outputting 50 kHZ square wave signal with 50% duty cycle. Once in 50k cycles, a glitch pulse is created that drops to 0 after 1 us and returns to Vcc.


First I turned on advanced triggering capability of this MDO, Pulse Runt triggering. Acquisition memory is set to 100k points, which allows for total of 290 segments.



Segments are then enabled in the Acquisition menu.

After capture of 290 segments, each of them can be replayed and examined individually.




Segmented capture aslo offer nice statistics analysis of captured segments, but measurement must be enabled before segments are captured.

I enabled Rise time measurement, and here is the representation of 290 segment's rise time divided in 10 bins. Numbers of bins is also user selectable.




Arbitrary waveform generator


This MDO is equipped with dual channel AWG, with 14 bits of vertical resolution at 200 Msa/s.

Maximum frequency is 25 MHz which is on a lower end for AWGs, but still plenty enough for vast majority of projects.

Built in waveforms are the usual ones: Sine, Square, Pulse, Ramp, DC, Noise, Sinc, Gaston, Lorentz, Exponential Rise, Exponential Fall, Haversine, Cardiac.

There are also usual modulation options: AM, FM, FSK as well as Sweep option.

Voltage output range is 20 mVpp up to 5Vpp into High Z, with possible DC offset of ±2.5 V into High Z (with 5V being absolute maximum output level).


Here we have both channels enabled, CH1 is noise and CH2 is cardiac signal. Both outputs can be controlled independently, and with an option to select High Z or 50 Ohm load.




State Disp displays all settings for both channel in a way usually found on standalone units.



UTIL menu offers some nice features like coupling of CH1 and CH2 in freqeuncy and/or amplitude, or tracking one waveform to another.


Here is an example of frequency coupled sine signals with 1:10 ratio



And setting arbitrary phase of one signal to another is also easily set.




Having an AWG, important part is the ability to create custom waveforms.

First I checked how easy it to create something on the MDO itself.

I selected Arbitrary waveform and then Create New.




Here you can select how many points you want to define (2-16384 points) and initial function to base new waveform on.

I selected 128 points and Ramp function, clicked Create New and voila, new arbitrary waveform is ready.


Of course, you can edit it further, and MDO offer 2 edit methods: Normal and Functon edit.

Normal edit allows you to adjust amplitude of each point to get the desired waveform shape.


On the picture above, point 72 is adjusted to 1V amplitude to create a spike in the ramp.


Second edit method is Function edit which is a little bit more versatile and allows editing is several ways:


Point/Line: Insert a point or horizontal line into the ARB waveform

Diagonal: Insert a diagonal line

Scale: Scales the ARB waveform vertically

Copy/Paste: Copy or paste a section of the ARB waveform

Clear: Clears a section of the ARB waveform and replaces it with a 0V DC waveform


Here I used Scale to easily convert Ramp to sort of double flat top Ramp function:



Saving it internally to one of four ARB positions and here is the final result on the DSO screen:



Of course, proper arbitrary waveform creation is a job for software running on a PC.

GW Instek has a free download on their site called OpenWave V1.05.

Download size is 24 MB in a ZIP archive which is surprisingly small in this age. Zip contains only 2 files, one txt license file and one exe file.

Clicking on exe brings up Microsoft Defender warning popup. Not good signs so far...


Final result looks like this:



You can not actually draw waveform, only load it from CSV/LSF file.




Connecting MDO to PC via USB goes smoothly, MDO is recognized as Serial Com port.

I failed to establish connection using Putty as terminal, even with GW Instek drivers.


OpenWave managed to connect and is able to capture the signals connected to MDO on CH1 or CH2:




You can Zoom, Pan or save the captured waveform, however no direct editing is possible.

Also, there is no option to upload waveform directly to MDO.


Ethernet connection works fine after setting the necessary networking details, however OpenWave is unable to use it.

There is also no web server in MDO, ethernet  is used only for NI Measurement and automation explorer which I haven't tested.




MDO 2000EX series has 2 distinctive functionalities not found on other MDOs/MSOs/DSOs: DMM and power supply.

At first I thought that is software type DMM that uses DSO data to calculate values, but this DMM is fully separate unit inside MDO.

Specs are decent for this class of instruments: 5000 count DMM with 0.1% DC accuracy +5 digits.


DMM is enabled with Option key, and it is visible along side primary screen, for example main DSO view.

This means that you can use DSO and DMM functions at the same time, which could be very useful in some troubleshooting sessions.




All the usual ranges are present, DC and AC voltage, current, mA measurement, Ohms, continuity, diode test and temp.

As expected, this is autoranging DMM, but user has an option to manually select desired range.

Scree update rate is not specified, but I am prettu sure it is 3 readings/s, which is for me somewhat on the slow side.

Continuity speed is also pretty slow for my liking, but it is latched.

TEMP range allows for a wide selection of thermocouple types: B E, J, K, N, R, S, T


DMM also has MIN/MAX and HOLD feauture. And unfortunately HOLD is not the good one touch Hold, just the regular value hold feature.


Of course, no DMM test can be full without DC accuracy check image

As reference I used my trusty unit based on AD584LH with 4 voltages.

Reference DMM is my Fluke 287 and results are in the table below:


Reference voltage GW INSTEK MDO 2072EXFluke 287
2.499887 V2.500 V2.4990 V
5.00187 V5.004 V5.0021 V
7.50026 V7.50 V7.501 V
10.00280 V10.00 V10.003 V


Power supply


Second distinctive feature of the MDO2000EX is built in dual channel power supply.

It is fully integrated in the UI of the MDO, with voltage levels adjustable between 1V and 5V in 0.1V increments, and max 1A of current.

It will not replace your bench power supply, but again, for majority of smaller projects it will be enough,for example it can provide 3.3V and 5V rails

NOTE: Power supply's ground is connected to scope ground, and of course to mains earth.


PSU is enabled with Option key, and on first start it takes few seconds for scope to configure it, I don't know exactly why.

After that, usage is pretty straight forward, with dual Output setups controlling Output ON/OFF and voltage. There is also quick access to 3.3V and 5V presets.

Reconfigure button will reset PSU in case Over Current Protection is tripped, ie current exceeded 1A.


Here is a picture of PSU supplying 3.3V, with DMM connected and scope capture of power up characteristic on CH1.

No negative surprises there, ramp up to desired voltage is linear with no overshoots.



In case output current exceeds 1A output is disabled, and you have to click Reconfigure to enable output again.




Unfortunately, there is no current limiting option.


Frequency Response Analysis application


This MDO has another interesting feature which is being able to add/run/remove applications, and it has dedicated APP button for this on the front panel.

It comes with some preinstalled:


GO/NO GO - sort of mask testing with user definable signal boundaries and and with an option to output result to another device via open collector output on the back of MDO

DVM - standard "software" based DVM which uses sampled signal to calculate and displays values of voltage and frequency

Data Log - logging the current waveform data or screenshot at set intervals for a set duration of time

Digital filter - applying digital high/low/bandpass filter on the analog channel capture

Mask - standard mask testing


One of the features that interested me the most, and was one of the main reasons I applied to roadtest this was the freely available FRA application.

This application is able to create graphs of gain and phase response or Bode plots of  DUTs using the MDOs inbuilt AWG and math functions.


FRA app is easily downloaded from GW Instek web site, and it comes with nice manual. Installation on MDO is also pretty straightforward.


Of course before FRA is started, there are 4 thing that need to be setup first:

INPUT source

OUTPUT source

AWG setup - Start/Stop frequncy and amplitude of sine signal

Number of points per decade


These is also nice reference circuit available in the app for easy reference:



For my experiment I built a simple class A single transistor amplifier on the breadboard, powered by MDO PSU set to 5V.




Sweep was set from 100 Hz all the way up to 25 MHz with 200 mV of amplitude.


After I connected everything and ran the analysis with FRA Run button, here is the result I got:




Let's do some basic measurements and check the -3 dB point of this amplifier.

It only takes few quick operations to set measurements and scroll to desired point and now we can see that at 1.2 MHz my amplifier has lost 3 dB of gain and that output phase is now at 130 degrees.






In short, I am impressed!


This device offers a huge array of tools packed into one single unit at relatively affordable price. It is aimed primarily for educational and hobbyist market, and GW Instek has made a great job here.

You get dual channel DSO, SA, AWG, DMM, PSU, serial decoding, fast update rate, statistical analysis options, logging, etc... all in one device, and you are all set to start learning, tinkering, making stuff and measuring.

FRA application is a great example of what is achievable with this highly integrated measurement station.


If I had to point out some areas for improvement it would be:


* Fine/Coarse vertical scale option

* Enhanced vertical resolution option via oversampling

* ASCI i2c decoding

* AWG software with better capabilities


I believe that we will see more smaller improvements with future firmware releases.


Once again, thanks goes to Element14 and GW Instek for privilege to roadtest this very nice piece of test equipment.


At the end, all I can say is: Wish I had this when I started learning about electronics image

Parents Comment Children
No Data