EEVblog #1057 - Digilent Open Scope MZ Review

[G0MJW]

Seems to be neither fish nor fowl. And using JS doesn't give me any confidence in the product either. I'm a hobbyist and I think the 90 bucks for the Open Scope MZ could be spent on something more useful.

$90 woudl almost buy a nice dinner.

Or a couple of rounds of drinks in a London pub. Perspective.

[Brumby]

Given it's open source I can't help but wonder how many weeks before the Chinese are pumping them out at a fraction of the price with links to Digilent's website for the software.

This is presumably why it is open source in the first place. To allow anyone to copy, improve on and develop the project.

From what I saw, I think they are relying on that.

[hans]

This looks like a product made by some intern that wanted to use some "nice components", like the PIC32MZ, Wifi module and USB comms, and made this project given those parts.
Looking at the review and schematic it looks like a shockingly bad value:

NO input overvoltage protection
NO appreciable ESD protection on any I/O
NO EMI filters on any I/O
NO mechanical enclosure
What update rate does it have? (intermittent by the looks of it, else at best 5Hz?)
Only 32K sample depth even if you turn of all other features of the device? (PIC32MZ should have lots more available..)
Why does it use a webbrowser as an interface?

Okay, the product is open and people can take it to improve upon. But I don't see the point if the software is a turd as it is. I would only invest my time to build a compatible or improved hardware module if the software was any good, after all I think this is where the power of "open source" lays.
Given that they used a PIC32MZ in BGA, doesn't help that fact to respin the board with a better analog front-end and/or ADC.


(small note: this topic should be called episode 1056)

[Brumby]

Why does it use a webbrowser as an interface?

For the same reason we see web browser interfaces in a lot of stuff... the display engine is already built (and maintained by someone else) and the probability that there is one already installed is deliciously high.

[NiHaoMike]

Seems to be neither fish nor fowl. And using JS doesn't give me any confidence in the product either. I'm a hobbyist and I think the 90 bucks for the Open Scope MZ could be spent on something more useful.

$90 woudl almost buy a nice dinner.

Almost? For me, $90USD would easily cover 2 weeks of groceries from Whole Foods (a place apparently known for high prices) with room to spare.

From what I can see, the most value the device provides to us is inspiration to make a better version.

[Brumby]

From what I can see, the most value the device provides to us is inspiration to make a better version.

It's almost as if the only purpose of the thing is to get the chips out in the world.

Make some chips that are probably half-decent and get them made up into a half-baked board.

The response from many being "I can do better than that".  So they go out, get the chips and make better versions of the hardware and software - and others build on that.

I suppose that's one way for a chip manufacturer to get some differentiation in the marketplace.

[imidis]

It's gotta be. I had stopped watching halfway through the video, but decided to continue watching the other half just to make sure he wasn't impressed by it.   

[NiHaoMike]

It's almost as if the only purpose of the thing is to get the chips out in the world.

Make some chips that are probably half-decent and get them made up into a half-baked board.

The response from many being "I can do better than that".  So they go out, get the chips and make better versions of the hardware and software - and others build on that.

I suppose that's one way for a chip manufacturer to get some differentiation in the marketplace.

But how many would use the same chips if they're tasked with making a better version from scratch? At the least, the 802.11g module is way outdated.

I would take the Pocketbeagle, which already has some coprocessors that can do 100MHz logic capture, and add an ADC, AFE, and mux to give it mixed A/D capability. There's plenty of RAM to get a decent memory depth and a good amount of bandwidth for real time streaming. If you want wireless, add an off the shelf USB 802.11ac adapter.

[Brumby]

Oh, I agree - absolutely ... but then they wouldn't be "improving" something.  They'd be doing something completely different.

Some people don't think past the parts that are presented to them - so they'll use those chips.
Other people might wonder what these chips might actually be capable of - so they'll use those chips.
... and some people will just like to be able to say they could do something better than a known brand name.

[Howardlong]

One of these arrived on my desk a few days ago.

I can only express my reaction has been similar to that in the video. At first I was worried it wouldn't work without an internet connection, but once you've configured it it can be run without the internet. Sadly though though I can't seem to figure out if you can get the WiFi run as an AP, which means that in the field it'll need to be physically tethered to the laptop unless I mess about setting up my own mobile AP.

I really, really, don't like the numerous compromises on the UI compared to the full fat Waveforms. Things like not being able to change all setting without stopping the scope, slow response (particularly when running on a tablet), and fundamentally being tied to run through a browser rather than a proper app has clipped its wings badly. I'd much rather they'd got it to work with the full Waveforms software than this Frankenstein mix of browser, agent and dependency on an AP of some sort for WiFi.

This'll just be thrown in the drawer, until (if) the software and whole user experience is improved very significantly: just because you can run stuff through a browser doesn't mean you always should, and this is a stark reminder of why. The good news is that it's proved to me that last dwindling thoughts I had of moving to a browser UI for one of my products (which is also very interactive and video-rich) can be laid to rest, and I'll stick to my tried and tested thick app method.

[ogden]

From what I can see, the most value the device provides to us is inspiration to make a better version.

It's almost as if the only purpose of the thing is to get the chips out in the world.
Make some chips that are probably half-decent and get them made up into a half-baked board.

Digilent academic price list suggests that they are targeting universities. Obviously manufacturers want to push their products to be on university bench, not competitor's.

Regarding "better than that" - ADI is preparing much better product, at least in my opinion. Most likely I/O protection is not "top notch" as well, but such kind of products rarely are properly hardened for rough use/environments. They are happy that user blows I/O - can sell replacement

[Howardlong]

From what I can see, the most value the device provides to us is inspiration to make a better version.

It's almost as if the only purpose of the thing is to get the chips out in the world.
Make some chips that are probably half-decent and get them made up into a half-baked board.

Digilent academic price list suggests that they are targeting universities. Obviously manufacturers want to push their products to be on university bench, not competitor's.

Regarding "better than that" - ADI is preparing much better product, at least in my opinion. Most likely I/O protection is not "top notch" as well, but such kind of products rarely are properly hardened for rough use/environments. They are happy that user blows I/O - can sell replacement

On paper, the AD device looks very similar to the Analog Discovery 2. Its success or failure will be down to the software, and whether they actually deliver. I’ve had an ADALM-PLUTO on order for six months from Mouser. Seems like that intro price of $99 has crept up to $149 if you buy direct from AD, who do have stock.

[inventhouse]

While we're down in the <$100 performance range, I've been playing with the Xprotolab Plain from Gabotronics:

http://www.gabotronics.com/oscilloscopes/xprotolab-plain.htm
https://www.amazon.com/Xprotolab-breadboard-Oscilloscope-Waveform-generator/dp/B00HWZSAPI/ref=sr_1_1?ie=UTF8&qid=1519965766&sr=8-1&keywords=xprotolab

2 analog + 8 digital channels, 8-bit 2MS/s, 200KHz, MSO for $20

I'd definitely recommend it as a starter / ultra-low budget, but good for some basic audio or arduino stuff.

Open source enough I was able to hack the app to run on a RasPi, too.

[ogden]

2 analog + 8 digital channels, 8-bit 2MS/s, 200KHz, MSO for $20
Open source enough I was able to hack the app to run on a RasPi, too.

2 MSps logic analyzer? To do what? - Look at LED blinker signal? 

Add few dollars and get much better, 24MSps/3MHz analyzer/scope. There's sigrok support as well.

[inventhouse]

OOoo nifty!! I hadn't seen that, might (who am I kidding?  ) have to pick up one of those...

[chris_leyson]

Damn, shame about the crap GUI. I had hoped it would solve an instrumentation problem I'm having. If it will run off a small external USB power bank and doesn't need a USB connection then I might give it a try despite the GUI. I'm trying to debug parts of a PCB whilst testing immunity to fast transients, up to 7kV, and that is before I've done surge testing ! There is no way you can use a conventional scope for this sort of application so the open scope MZ might be worth a shot. Looks like it might need some protection diodes though. Just exactly how do you measure signal integrity with several kV of fast transients all over the place ?

[Crazy_Pete]

> Why is the Analog Discovery so well received?

Good Question.  Especially since  the Analog Discovery and Analog Discovery2 use the same hideously crappy software.   (I honestly am not sure if the problems you demo'd in your video are all hardware, all software or a mix.)      I spoke to the chief engineer responsible for the waveforms software for ALL three devices (OpenscopeMZ , AD and AD2)   I was flabbergasted.    I wrote offering to help them fix the bugs in the AD2 software since it is very very well received.     The standalone software crashes on startup if you don't give it any arguments.   I explained to the engineer what the problem was and his answer was "You need to look further into this, argument processing is handled by the QT Classes".    Yes the QT framework handles the arguments for you, but if you don't tell the QT classes how many arguments to expect it can't "handle" them.  :-)   IOW these guys (as i told them) don't know enough to pass a first semester "intro to programming" class, and yet they insist that their software is so precious that only they can look at it.     I explained a path to open source this based on some work others interested in the AD2 have already done and heard only mocking responses.

Basically this is the old "Hardware Genius, Software Idiot" syndrome.    My honest opinion is that you hardware gurus are 1000x smarter than us software gurus.   Hardware design is more difficult on many levels and the consequences are more severe.    (I let the smoke out of a $100 chip.  Now i have to find the money and buy another!!!! AGGH! Versus "I segfaulted a $100 tablet, now i have to wait 20 seconds to reboot!  AGGHHH!"   :-)

For some reason  (maybe for THAT reason :-) ) hardware gurus simply do not trust software engineers and over and over again we see excellent hardware designs literally bricked by the engineer refusing to let anyone but himself write software for the device.  I have offered to help Digilent fix this mess and they simply shrug it off.   (I am sorry for the lack of modesty but i honestly believe i am as good at software as you guys are at hardware) :-)

Watching these guys scream and cry that they will go broke if they open source their useless software reminds me of Gollum (sp?)  in Lord of the Rings "The Precious!   You are only here to steal my precious"  Uh Gollum we were just passing through and your 'precious' is just a stupid rock.     Except that these Hardware Gurus are Gandolf not Gollum!     Hey if anyone wants to design a REAL open source scope, there are thousands like me waiting to write the software for it.

Thanks
Crazy Pete

[Crazy_Pete]

Some final thoughts on all this:

I want to apologise to Siglent for the comment:  "Hey if anyone wants to design a REAL open source [software] scope, there are thousands like me waiting to write the software for it."  It seems Siglent has done just that with their 1000x series of scopes.   They run Linux internally and are SCPI and LXI compliant.

I did send a PM to the Analog Discovery 2 software developer and offered to help fix these software problems.    He never bothered to respond.   I sent him a second email pointing out that the demo program crashed on an error that COULD have been caught in the 1 line shell script that calls the program.    Forgive me for repeating myself but although the Digilent Hardware is EXCELLENT, the programmer who wrote the closed source software is so skilled at software development that he cannot even write a 1 line shell script to call his own program!    And that's their CUSTOMER DEMO!   Good Lord "Script Kiddie" used to be an insult, the Digilent programmers can't even rise to that level!

To make matters worse, he claims the OTHER reason they can't open source it is because it uses a Xilinx FPGA for the main MCU.   Funny thing is, the Siglent uses a Xilinx FGPA as well.   Funny that Siglent could do it and doesn't even think it is important enough to put it in the market literature, but  Digilent says it is "Technically impossible" :-)   But what i do know about hardware?   On the other hand, i don't market that i can write a 1 line shell script that doesn't crash and Digilent seems to think that is "Technically impossible" as well.  :-)  Interestingly the OpenScope will work with waveforms live (the software Dave ridiculed) a javascript front end.   I can't even get javascript to operate email correctly from Browser to Browser and they write their TEST INSTRUMENTATION IN IT?   Any thoughts?

Thanks

Crazy Pete

[ogden]

I want to apologise to Siglent for the comment:  "Hey if anyone wants to design a REAL open source [software] scope, there are thousands like me waiting to write the software for it."  It seems Siglent has done just that with their 1000x series of scopes.   They run Linux internally and are SCPI and LXI compliant.

Could you please provide URL to source code of Siglent 1000x series of scope? Note that Linux inside some device does not automagically make it open source [software] device. There are loads of Linux applications that are proprietary, thus not open source. The same with application software of most commercial Linux-based embedded devices.

[Crazy_Pete]

>Could you please provide URL to source code of Siglent 1000x series of scope?

"They run Linux internally and are SCPI and LXI compliant."   That is honestly all i asked for.     You want a url for the source for linux:

http://www.kernel.org   :-)

A siglent rep kindly showed me how to get into the Linux internals and poke around.   Ms. Amanda Harris at Saelig also called Siglent and obtained the same information for me.   (Which is why i turned into such a Saelig fanboy.   She went out and GOT that information for me.)

There is no simple way (AFAICT) to write your own PC software for this Digilent stuff.    You can do it with a siglent right from the box.  (Unless the siglent rep is lying, which i really doubt.)  So whats on the Siglent that you can't get to?    Besides open source or not, if you can't write a one line shell script without crashing a machine, (Digilent) i don't need to get involved with your software.

thanks
crazy pete

[ogden]

"They run Linux internally and are SCPI and LXI compliant."   That is honestly all i asked for.
[..]
A siglent rep kindly showed me how to get into the Linux internals and poke around.

Fact that you can compile Tetris and run it on Siglent scope does not make it open source [software] scope If I need Linux, I run it on a normal computer anyway. OK, OK. Kidding. Obviously I agree that this is more than nothing and by some reverse engineering you can do lots of fun stuff and not only, with it. What are your plans regarding this scope?

[Crazy_Pete]

What are your plans regarding this scope?>

Well like any self respecting Schizoid, i intend to breed an army of nanobots and take over the world! :-)   Seriously i have 2 plans.   

1)  I want to use it to play with Arduino, ESP32 boards and the like.    (The fact that i won't be looking at signals above 10MHz anytime soon motivated me to get a 100MHz Scope.)

2)  I hear so much about crappy OScope software that i am hoping to write a really cool (tm) open source PC based OScope application.    As i said in my "newbee intro" this entire foray into electronics is sort of "stave off senility" move on my part.   Electronics was always my weakest point in science so i that is where i am focusing in my old age.

IOW, this scope, this forum, and electronics in general is my new Jack Shack!    (I think down in Oz they call them Man Sheds?)  Since i have the experience already in writing numerical and scientific software, things like FFTs and Bode Plotting software are something i can attempt.    I don't know that i will actually do anything useful but i would like to give one more thing back to the world before i return to my fiery kingdom below.   

Thanks

Crazy Pete

[Inverted18650]

Dave,

The Amazon store link under this video is not working.

[Electro Detective]

Dave,

The Amazon store link under this video is not working.

https://store.digilentinc.com/openscope-mz-open-source-all-in-one-instrumentation/

Working fine here, although Quote page:

"Digilent uses cookies to ensure you get the best experience on our websites. By using our site you agree to our use of cookies."

Explanations here > https://store.digilentinc.com/privacy-policy/

 

[chris_leyson]

Well it seems Digilent pulled the Opensope MZ because it was so unreliable, not surprising really. It was no help when trying to debug electrical fast transients because I just couldn't get it to trigger not even once.
Red Pitaya would do the job of debugging transient events where you need galvanic isolation.