Online learning and online tutorials have made some nice progress, but I don't think ~WE here~ are necessarily keeping up. Quite. Might be a good suggestion for DigiKey/DesignNews to consider -- if they can "Up-tech" their interactive classes (make the "interactive" a bit more flexible, say) their attendance and satisfaction will increase.
Thank you for your comment about making these webinars more interactive. In Friday's chat and the last notes page for Friday, I am listing links to webinars on this subject which are more technical.
@LevitonDave, interesting you mention the "embedded" bit. I am currently unemployed, and have been surprised (and aggravated) by the extent to which anyone SEEING "embedded devices" in my resume seems to automatically assume that I'm some C+ guru...
Hmm. Anybody know what the text "IRON" next to a username means?? Given that the most common alternative seems to be "BLOGGER", "IRON" is obviously a category we most-all fit within... Only thing that comes to mind though is "What you get when you remove Y(ou) from IRONY"... LOL
The issue of interaction is driven by the facilities and interactivity available in the working environment. I have to go "around" the corporate network to be able to hear the audio so higher bandwidth/higer interactive options may not be available to everyone.
But it is an interesting suggestion. Rich are you listening?
@flared0ne - I agree there is some "real world" learning that can't be avoided. However, providing useful alternate approaches to organizing the "big picture" can help the newbee's and novices to take their game to a more productive level.
Personally, I'm not a test engineer. I do embedded software and as a result there are some differences and similarities. I have, however, been involved with sufficient test equipment to understand the issues you are describing.
I'm leaning toward checking out some of the group-shared workspace packages available out there; don't know if they're readily/freely available as browser add-ins, and they tend to have a bit of head-end expense (if you pick one which is based upon using a whiteboard, for example). But I've been seeing commercials on public TV for some of these packages with sketch-pads and graphics etc.
Online learning and online tutorials have made some nice progress, but I don't think ~WE here~ are necessarily keeping up. Quite. Might be a good suggestion for DigiKey/DesignNews to consider -- if they can "Up-tech" their interactive classes (make the "interactive" a bit more flexible, say) their attendance and satisfaction will increase.
@LevitonDave, anytime someone does a presentation about "Software can do it all!", the issue of "IF you have experience and know what CAN happen" comes up. If you DON'T have experience, all these slides about "This kind of module needs THIS additional functionality, because later you will regret not having it" aren't going to get much traction. "Tough love", "no pain, no gain", etc. Can't recognize the speedbumps unless you bite your tongue going over one at least once.
Skipping back to "formulating the question" -- one major area where "experience counts" is when you are using generic devices to interface with a DUT -- devices which don't ALLOW you to vary signal transition levels, or measure I/O currents, which don't allow you to vary rise/fall timing on an input data/clock/trigger/etc... "roll-your-own" test setups will fail to detect a large percentage of some fairly common failure modes; and you won't know that happened until you get a spike in DOA or infant-mortality stats. SOME test gear allows you to treat a test-signal source-point as a signal generator, with tight control over all specs. YOU have to know/decide when that expense is justified, considering it can damage your credibility (and job security) to have your tested devices falling over in the field.
flared0ne - Your approach has been learned through experience. I was hoping that the topic of "developing sofware for your test application" would bring up some of these issues as part of the presentation to help others avoid some of these "learning curve" issues.
I worked with Hewlett-Packard/Agilent extend some of their "generic testing environment" back in the '80s. Some of these concepts were supported/enforced as part of the environment and made reuse/re-sequencing straight forward. I can't believe that this hasn't been used by others in the past 20+ years.
@LevitonDave -- you've hit on an issue which can bite you. Because in some cases a failure mode doesn't show up until after you have departed from that "common stable state" -- I've encountered device failures which were "prevented" by a board level reset, when ordinary useage didn't ALLOW for a board-level reset at the necessary point (which would have lost other necessary state info anyway).
In the past, I've often been amused/bemused when, yet again, it becomes clear that "properly formatting a question tends to almost automatically evoke the answer". In other words, when I have properly laid out a full description of ALL the functional elements of a device I'm tasked with testing, that full description can become the checklist for the test suite. And I mean a FULL description -- input current specs, input voltage specs, thermal range restrictions, output current specs, noise level budget, etc -- ANYthing which SHOULD trigger a "failed" response needs to have a condition check.
Rick - The approach to software modularity is a combination of the tools, supported modules and sequences to achieve a consistent, reordable testing environment.
Labview is an application development environment that supports both good and bad organization of software testing. This is true whether using purpose built instruemntation or software defined generic tools. LabView ADE is intependent of the software defined approach vs. purpose built instruments.
Software organization will greatly influence the ability to use/reorder/modify/extend tests. This will greatly drive the way the software testing modules/code are created.
Do you have any experience or suggestions on how to structure the software testing within the ADE?
For example, does it make sense to approach tests as a sequence that contains a "setup test", "execute the test", "tear down" to have the DUT in a common stable state between tests.
Hmm, this might be an idea -- I'm loading Chrome to be able to LISTEN to the audio. Chrome doesn't allow doing the same kind of "shortcut" link, so I've been clicking on the link in the reminder email and opening an IE window. I think tomorrow I'm going to KEEP the IE window open AND "pause" the non-functional audio player (so it isn't sitting there trying futilely to synch up) and use THAT window to type into Chat. Let the Chrome window sit in the background more-or-less feeding me audio. Diagnostic -- split the problem child into multiple pieces, see which individual piece keeps the problem (or not).
Mr E: Thanks for the comment. A complex microcontroller application, like a car starter, obviously has multiple inputs and outputs. To really exercise the on board coding, it would be prudent to run various test scenarios, and be able to produce reports for each one. I am unsure how the testing instruments would interface to the MCU package, but I can see where one would need to be inventive. For example, if the MCU package had to drive a glow plug directly, then one would want an appropriate resistive load hooked to that output. Deriving an input from this for the testing environment might take some cobbled electronics,like a hall effect current sensor, but would be necessary for real world testing. Assuming Labview could talk to the MCU via I2C or CAN, I think most of the logic response could be captured. However, does LabView provide for individual digital I/O, analog inputs, etc? Thanks again.
Bye for today. Look here for answers to your questions that have not yet been answered. Tune in tomorrow at 2:00 p.m. EDT for the last presentation of this webinar. Download the slide deck first. After the presentation, Email Rich Nass at Rich.Nass@UBM.com to get a copy of the slide deck that includes the notes. On the last page of the notes, and also in tomorrow's chat, I will post my sources, blogs, contacts, Emails, and a request to all of you about how we can make this webinar better in the future.
small software test
Would a set-up like Lab View be an economical choice for testing
moderately complex microcontroller devices. Seems that only a software
system would be able to test all operational scenarios, which would be
essential to uncover bugs in the device design. Thanks
There will be a discussion of this tomorrow. If that doesn't answer the question, look at my list of references and contacts tomorrow, and then if you can't find it, contact the vendor.
@flaredOne - missing posts seems to be a sporadic thing, has happened to be off and on. probably just need to copy text before posting so that if it is lost, just paste and post again
To see my notes, request the version of the slide deck from Rich Nass (by Email) and view the slides under the Notes View. Rich's Email is rich.nass@ubm.com
@luizcosta: Agreed. It would be nice to be able to "reply" to a specific post rather than trying to copy enough context to allow the poster to recognize that you are responding to them.
@luizcosta: When I was working in an environment that dealt with state estimation and Kalman filtering, we would usually have a "reality" model that would either synthesize or replay recorded real-world events in real-time (or occasionally scaled real-time) and compare "reality" (as presented to the state estimator) with the estimated state output. By playing many scenarios, you began to see specific scenarios that were handled incorrectly and could then home in on the problem areas to debug the estimator. Is that the sort of thing you are interested in doing?
Hmm. Accidentally clicked a participant's username, went off somewhere else, and came back -- NOW I get to watch through the entire chat session reloading... auugh!
I've actually had better results doing "cut-and-pAste" than typing directly -- it's just that I usually don't expect a SMALLER message to get lost, so I don't ALWAYS "select all" and "ctrl-C" to snapshot stuff. And then several minutes of thought just vaporizes...
To see my notes, request the version of the slide deck from Rich Nass (by Email) and view the slides under the Notes View. Rich's Email is rich.nass@ubm.com
@flaredOne - missing posts seems to be a sporadic thing, has happened to be off and on. probably just need to copy text before posting so that if it is lost, just paste and post again
Rich posted his email address waay back at the beginning of the Chat "stack" for slides-plus-notes requests. Me, I'm going to hit him up tomorrow for the entire week's worth (either five at once, or five individual msgs, whatever...)
@rfindley: I figured that Agilent and other instrumentation manufacturers had to have products like VEE, but no, I have never used it or even heard of it until you mentioned it, yet the Agilent website shows that it is currently at version 9.3. Go figure!
Hey Rick, I am in a notes view of the presentation and I am not seeing any notes or the table you spoke of. Is the slide Deck posted an older version? Would it be possible to post a more up to date one if it is a version issue?
Odd thing about the audio drops, while he was experiencing difficulties, my speaker jack popped out and I was waiting for him to return. Wierd timing for a failure at my end.
@RICK: As I asked the other day, how do you test if a Kalman Filter or Other similar algorithms are providing the correct state estimate or the correct system's identification?
No one here has used Agilent's VEE? Having used a lot of Agilent test equipment, I got started on VEE, and personally find it more intuitive than LabView.
Resolution: analog value associated with one-bit change in value...
"Drift" (time, temperature, etc) will bite you. Temperature-controlled chambers can minimize thermal effects on test gear. "Test bench stops working if the AC goes off" will frustrate things.
Another thing to take into consideration in noisier environments is, if you HAVE a choice, select devices which use a CURRENT loop (variations in CURRENT are hugely insensitive to most external noise sources).
I had a customer once that had an ethernet wire running from one part of the building over a connecting bridge and into the other building and couldnt understand why they had network communication errors. Theyt not only had stapled through some of the cable but also went passed the ethernet distance limitation. They should have been happy it worked at all.
It isn't that he's reading from one sheet of paper, he's (generally usefully) jumping from one page to another, following the train of concept/context HE is trying to create for us.
And yeah, having been doing similar things for quite awhile, I'm a -little- disappointed -- mostly because in a perfect world, what HE is doing would be SIMPLER to take advantage of by reading through a sequenced hyperlink document, with extracts and highlighted content. I can read text SERIOUSLY faster than I can LISTEN to text. I.e, I finished the last Harry Potter book in about three hours, while the Books-on-Tape version runs what, eight hours??
Sure, he's reading from various white papers. What I AM seeing is that he IS extracting pertinent tidbits from a wide variety of online sources -- and again, having someone with significant expertise and experience AGGREGATING from multiple sources -- VALUABLE (ahem, mostly, though, valuable if we are given access TO those multiple sources so we CAN in fact expand on the lecture by seeing what he did NOT have time to include).
Basically find some of the other lectures were much stronger. That said the other lectures had some of the same uncoordinated interuptions that these appear to have. In fact I would say they had more of them.
So he's spending a certain amount of time describing the "TEST ENVIRONMENTS" available, then drilling down a bit to discuss some of the stuff that every individual module YOU create should incorporate (again, from experience of having stumbled over "man, I wish I had included THAT from the start").
Any time you have a unique device to be tested, you HAVE to create your own test sequences to fit into "test modules" in whatever testing environment you are supporting.
The good news is that the audience is very varied in its technical depth. The bad news is that it's hard to teach a class at multiple levels at the same time. But I think Rick is doing a good job.
With any luck, if people have ever gone into this subject area and managed to do it at all CORRECTLY, they only need to keep up-to-date.
Several years of experience, scattered across multiple years, multiple projects, multiple companies.
Consider myself intermediate (have rolled my own, have used one or two obscure commercial products). Looking for the high-level concepts about what kinds of adjunct modules SHOULD be included (re the "learning from experience", leveraging institutional knowledge).
Rick, I would say there's some chance you have a slightly loose connection on some cable connector.
@alaskaman66: LabView would let you program your test, but you would still need a complement of hardware to perform the actual interface to the MCU. So for example, suppose the MCU supports an I2C bus, you would need to physically connect something that could assess that the electrical characteristics were correct, and something else to assess whether the communication was within spec, and likely something to deliberately generate various signaling errors and assess the response. LabView could be the "glue" to coordinate all of that test hardware, but of course, so could a number of other tools. I think it would come down to whether you were already committed to LabView or not. Of course LabView would allow a lot of flexibility in terms of exploring levels and probing while running.
Rich, almost everyone who has worked in electronics as a technician in this state for twenty years has worked for the same employer at one point, and usually know each other.
In the previous sessions some asked each time about Python. Python is a text-based language who's motto is "Batteries Included" meaning that everything you need for a program is in the language. Although this doesn't include libraries to talk to hardware they are readily available, including VISA and GPIB.
small software test
Would a set-up like Lab View be an economical choice for testing
moderately complex microcontroller devices. Seems that only a software
system would be able to test all operational scenarios, which would be
essential to uncover bugs in the device design. Thanks
The presentation will begin in 20 minutes. I am logged into this chat, but don't expect me to look at it while I am presenting the webinar. Rich Nass will monitor it, however, so fire away with questions and any audio or other problems. Don't forget to download today's slide deck, as well as tomorrow's. I will be back on this chat after the webinar ends, around 2:30 p.m. EDT.
Using almost 200 light-emitting diodes in the front and back of the new 2014 CTS, Cadillac designers are showing how LEDs can change the character of a vehicle.
We looked at a number of sources to determine this year's greenest cars, from KBB to automotive trade magazines to environmental organizations. These 14 cars emerged as being great at either stretching fuel or reducing carbon footprint.
From Dell / Intel® New Paradigms in Design Work Scott Hamilton, vertical market strategist for Dell Precision workstations, 5/2/2013 3
Early in my career, I worked as a draftsman and remember the days of drawing on vellum with numbered pencils and Mylar with plastic lead. This was a fun experience in the sense that I ...
I've been using workstations for more than 10 years and love finding ways to get more performance from my system. With demanding professional applications that require more power each ...
A lasting memory from my first job as an engineer in an auto assembly plant is standing on hard concrete at six in the morning, vending-machine coffee clutched in hand, listening to ...
A quick look into the merger of two powerhouse 3D printing OEMs and the new leader in rapid prototyping solutions, Stratasys. The industrial revolution is now led by 3D printing and engineers are given the opportunity to fully maximize their design capabilities, reduce their time-to-market and functionally test prototypes cheaper, faster and easier. Bruce Bradshaw, Director of Marketing in North America, will explore the large product offering and variety of materials that will help CAD designers articulate their product design with actual, physical prototypes. This broadcast will dive deep into technical information including application specific stories from real world customers and their experiences with 3D printing. 3D Printing is
To save this item to your list of favorite Design News content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Tweet This
1 
