Jump to content

Andrew at MEA

Registered
  • Content Count

    84
  • Joined

  • Last visited

  • Days Won

    14

Andrew at MEA last won the day on July 24 2017

Andrew at MEA had the most liked content!

Community Reputation

13 Good

2 Followers

About Andrew at MEA

  • Rank
    Participant

Personal Information

  • Organisation Membership
    EA
    IEEE

Recent Profile Visitors

523 profile views
  1. Andrew at MEA

    Lessons from Myriota IoT Developers

    Tales from the Trenches Part 2 More learnings from the three Myriota sensor-to-satellite systems installed in the MEA Test Garden are told in the latest yarn at the bottom of Page 3 in “What Does it take to be an IoT Engineer?’ in this forum. ‘A Hat Trick in the IoT’ covers issues with over-zealous data feed rate, the vagaries of UARTs and glimpses into topographical issues caused by the Mt Lofty Ranges. Oh, and one more thing: check the sealing around the Myriota SDK antenna connection where it enters the enclosure. Sometimes the O-ring seal is under insufficient pressure to prevent water ingress if the unit is placed outside under rainy conditions. Simply tighten the connector nut inside the enclosure to fix this.
  2. Andrew at MEA

    What does it take to be an IoT engineer?

    A Hat-Trick in the IoT. Myriota have repaired our third recalcitrant sensor-to-satellite modem and sent around one of their engineers to hand-deliver it. Good companies are tested in bad times. This has been a superb experience in dealing with Myriota’s support team, and we have gained confidence in them rather than losing it. Facing down technical issues is a learning ground; one is forced to set aside specious hand-waving and theorizing and to let the problem do the talking. Thus, what started with some manufacturing faults by one of Myriota’s suppliers turned into deeper insights at MEA into how this new technology really works. First up, we’d been trying to account for missing records. We’d assumed these were simply poor radio transmissions that hadn’t gotten through to the satellite, or could not be decoded. More careful analysis showed that they weren’t actually missing, but up to thirty-hours late. How could that be? Turns out we’d over-stretched the pipe, by trying to send four records twice a day for each pass of the single satellite. Dropping back to three records per pass (six log records per day at four-hour intervals) found us moving data reliably every 12 hours. Next up were issues in our own data – inexplicable glitches that suggested we were dropping digits, causing ugly spikes and unseemly drop-outs in our lines graphs. This turned out to be coding issues in using the Myriota low-power UART to obtain serial data from our sensors. Implementing a system of error-checking and re-tries fixed this. Finally, a first glimpse into the mysteries of topographical interference in transmissions to passing satellites… The single Myriota satellite that we see here in Adelaide is in polar orbit, with the Earth rotating under it twice per day. Even though the satellite completes an orbit every ninety minutes, we are only likely to see it in our field of view for between five and eight minutes during each fly-by. And for MEA - sitting up close to the foothills of the Mt. Lofty Ranges - this means that satellite transits to our east may be blocked on some passes. This is mere supposition at this stage, but it's something we picked up listening to Myriota’s engineers discussing our connectivity issues. Older (MEA is now officially 35 years old), wiser and somewhat less brash, we re-install our repaired tank level Myriota transmitter, complete with data re-tries, slower four-hour log interval and a wary eye to the East. The first rains of autumn appear from the west shortly thereafter, and now all three MEA-Myriota test stations work together to show the measured rainfall, the consequent lift in soil moisture in the surface layers, falling temperature and rising relative humidity associated with this cold front and – best of all – a rise in the level of the rainwater tanks as they begin to recharge.
  3. Andrew at MEA

    Lessons from Myriota IoT Developers

    Lessons from Myriota I’ve written three blog posts on how MEA started up our adventures in moving IoT data over Myriota’s nano-satellite link to our Green Brain web application. Those stories can be found at the end of my forum posts entitled “What does it take to be an IoT Engineer?” on this Engineers Australia IoT site. I’d like to hear how other engineers are travelling with their new Myriota SDKs (Software Development Kits), so jump onto this thread and contribute. Let me kick-off by mentioning that I’ve got three sites set up moving soil moisture, climate and tank level data – about 350 messages so far in the past five weeks. Proof-positive of five weeks of soil moisture tension data linked from MEA's test garden to Myriota's tiny satellites The folks at Myriota have been wonderfully helpful as we walked into all the usual bear-traps, sharing “Ah, blast!” moments on either side of the fence. Here’s two: MEA has been jamming too many messages into the pipe, at the rate of 8 per day. This constipation results in lost messages. We’re dropping back to six messages per day. Our tank level monitoring site has not transmitted a single message (see that last story “The IoT via Satellite”). Turns out this was due to multiple instances of dry solder joints on that particular SDK, particularly under battery pads and around the UHF aerial-connections corner of the module (pads 19 to 29). Lesson: Keep an eye out for hardware faults if you are not getting connectivity. The Myriota folk are taking this up with the SDK-kit manufacturer. OK, open to Myriota IoT stories from engineers working with this gear...
  4. Andrew at MEA

    What does it take to be an IoT engineer?

    The IoT via Satellite In less than a week after receiving two more Myriota satellite modems (in SDK form) we’ve built two more operational IoT devices, tripling the rate of data flow and improving our chances of understanding this brand-new technology as we race to create working products for new markets. Modem #2 is a mini-weather station, capable of accurate measurements of rainfall, air temperature, relative humidity and their derivatives – dew point and frost point temperatures and vapour pressure deficit. Data compression allows meaningful data to be pushed up via the very slow and narrow pipe that the Myriota technology offers, essentially 14 bytes per three hours. Modem #3 is doing something a bit more exciting; we’re measuring tank water level using a downward-looking ultrasonic depth sensor mounted in the tank roof, which is the perfect mounting spot for upward-looking satellite transmissions. This new technology makes use of our latest ADC technology and one of a series of generic sensor interfaces we’re developing to plug-and-play with sensor front-ends and different comms back-ends in MEA’s small IoT world. Calibration checks use far less sophisticated technology; a tape measure with me up a ladder hoping my glasses won’t take the deep dive to the tank bottom. This is the end of playtime; now we have to get serious about building real IoT products. For that we need a ‘reference design’ from Myriota to help us with specifics like antenna types, special radio-frequency PCB layout issues and so forth. It is commonplace now for manufacturers to use their own engineers to help the applications teams in the customers’ world to get to market as quickly and painlessly as possible. All this focus on the hardware and data transmission technologies hides one of the fundamental truths of the IoT; without a web-based data platform to connect data to customers all this ground-based techno-effort is for nought. MEA’s Green Brain is the final repository for the data that’s left my test garden, gone straight up 800 kilometres to the low-earth orbit satellite whizzing by at many kilometres per second twice per day, to be carried in polar orbit for another forty minutes until the satellite passes over Norway where it downlinks the raw radio data for web-based processing using Myriota’s special algorithms. Finally, the sorted and scaled data is hopped across to Green Brain from whence I can track things – just metres away from where it all began. We corral all these three demo sites into a shared link to Green Brain for Myriota so that their engineers can see what we see from their own offices, and share that link in turn with others so that they too may play and be convinced. In the meantime, we’re running a careful watching brief on the demonstration systems. Almost immediately we’re reminded of how easily one can be lulled into complacency; only one of the two new units is getting data through. We recheck battery sensor and aerial connections, press the RESET button on the Myriota modem board, check that we haven’t made a typo in the Device ID, try to get radio transmissions to the PC dongle in the dev kit, reflash the code, and wait through various twice-daily transmission windows. Everything checks out. No fault found at our end. Finally – by standing outside in the MEA car park – we realize that we’re not getting a GPS fix from this module. Without that, nothing starts up. We are brutally reminded that our understanding of this new data transmission technology has all the depth of a car park puddle.
  5. Andrew at MEA

    What does it take to be an IoT engineer?

    The IoT on a Shoestring. With MEA’s 35th birthday now only weeks away, we find ourselves cobbling together demonstration systems at a furious rate, calling up snippets of technology from our IP grab-bag to make ‘Minimal Viable Products’, or MVPs. The problem facing MEA is that we don’t know these new markets that the Myriota ultra-low-cost satellite IoT technology has opened up before us – field measurements beyond the bounds of the cellular IoT network. Rather than spend our limited marketing resources wandering expensively into the Outback to survey the views of potential agricultural customers, we are assembling working prototypes around the modems we have been able to obtain from Myriota and hooking various applications across to our Green Brain web platform. We’re after rapid feedback from a select few. These MVPs are only 80% operational, but our target audience – Early Adopters – prefer this so that they can be the first to play with this new technology. They are even prepared to pay a premium for that privilege, and to put up with MEA’s engineers hovering anxiously nearby as we tweak out obvious bugs. By the time Myriota can get into solid volume production of their technology, MEA will also be ready with the application technologies to take full advantage of it. This is only possible because of the ‘home-town advantage’; both Myriota and MEA are Adelaide-based companies, and I’ve been with them along their journey since they first came out of the University system some years ago with a wad of patents, seed funding and the drive to disrupt the current IoT paradigm, where satellite links have traditionally been prohibitively expensive. Marketing folk call this ‘networking’; I just call it ‘staying alert’. And so – in a matter of weeks – we’ve hacked together satellite-connected demonstration models for five typical applications in both irrigated and broad-acre agricultural applications: 1. ultrasonic and hydrostatic tank water level monitoring 2. sub-soil moisture tension measurements in grain-growing areas 3. standard soil moisture content profiling in irrigated agriculture 4. frost and micro-climate recording systems 5. rainfall monitoring MEA’s two IoT teams proceed at a swift pace into this new arena, working on a shoestring budget but drawing upon a deep well of experience and proven technologies that took us years to develop. So here they are – young and old: - The younger team are our Green Brain developers: Matthias (Green Brain back-end), Brent (Green Brain design front-end) and Dominic (business development and customer insights) The product development team builds all the ground-based stuff not in the Cloud: Andrew (electronic engineering), Terry (production engineering and phone apps), Jack (mechanical engineering and document control) and Raf (firmware engineering)
  6. Andrew at MEA

    What does it take to be an IoT engineer?

    ‘Straight-Up’ with the IoT. A new year has fired up at MEA and the troops have rested. I wish I could have rested, but with MEA closed over the Christmas-New Year period, I inevitably spent that down-time reading up on all those technologies I’d sidelined in my head amidst the daily blur. This year I got up to speed on the latest ultrasonic technologies for level sensing, just in case we set about marrying Myriota’s remote-area satellite IoT technology with tank level monitoring and sensing cattle movements. And it’s been hot ‘Down Under’, so the Christmas break also found me conducting a large-scale study in the MEA Test Garden on the impact of current-collapse in our IoT solar panels at temperatures above 40°C. To complicate things, the summer sun climbs high into the sky during the summer solstice, with shorter charging days for north-facing solar panels exacerbating solar-battery charging difficulties caused by radiant heat load. January and the return to work saw us hosing down last-minute alarms in our new Bluetooth-gypsum block IoT technology, destined for shipment to some remote tropical island off Australia’s north-east coast where the Green Turtles are breeding. Things that shouldn’t have gone wrong did go wrong. The ADCs and direct memory access functions in our MSP430 embedded controllers didn’t work as either the data sheet or Errata notes suggested they ought. Occasional data gaps were appearing in the logged record because measurements didn’t complete. We fixed this with a re-write using basic code and much cursing. Then the outdoor air temperature hit 46.9°C (116.4F) in the MEA Test Garden in Adelaide – with over 52°C inside the enclosures – and suddenly we had data spikes. That last-minute glitch was also hosed down and we entered production. With all that done, we got back to debugging our Myriota satellite IoT test system, which had stopped transmitting following a rain event. That too got sorted, teaching us yet again that one can’t be careless with enclosures even when hacking stuff together to test first principles. February rolled around and we dug deep into our back pockets for intellectual property – hardware and firmware – to lash the Myriota modem to our newly-developed smart gypsum block interface. Sometimes it’s good not to be a start-up, as there is no need to develop every darn thing from scratch before you can get down to business. Test beds evolve quickly. Six weeks into the New Year in the MEA Test Garden and we’re monitoring the soil moisture tension profile in the orchard and beaming data straight-up to those whizzing satellites that pass overhead twice-a-day. We’ve already learnt that we can send eight 20-byte packets by queuing them and letting the Myriota modem tackle the business of attempting multiple packet transmissions at each pass. This will stop-gap us against otherwise poor sampling rates until more satellites are launched. Once again we find ourselves under time pressure. Does this satellite IoT technology really work as expected? Can we reliably design product around it? Where are the pitfalls? Can we compress the data sufficiently to match the very small payload? How do we connect data across from the Myriota server to MEA's Green Brain in the cloud? Only squads of data – and time to collect it – can help us here, and we need all sorts of weather events to shake loose nasty gremlins that will surely bight us if they get through to production and our customer base. In the meantime, my Christmas sabbatical has fired my imagination and I’m spending the wee morning hours in my home lab testing new sensor ideas. There just never seems to be enough time in this IoT race!
  7. Andrew at MEA

    More Australian IoT satellite launches

    Hi Geoff Yep - exciting times, indeed. We've had ours since last year and have been transmitting data from the MEA roof via satellite about four times per day since then. I can hardly believe it's possible! And so many interesting applications opening up for IoT (truly-) anywhere... The Myriota applications team have been quick and helpful sorting out bugs and queries. All the best with your developments. Cheers for now Andrew
  8. Andrew at MEA

    What does it take to be an IoT engineer?

    Another Christmas in IoT Land Weariness pervades all as another working year winds down here at MEA. With the pressure to release new IoT products and the need to shore up the ravages of Mother Nature among the old ones, it’s been a tough year in Product Development. We learnt this year to prototype and manufacture in China, speeding up PCB manufacture while driving costs down in a competitive arena. In 2018 we designed and built a Bluetooth logger for SDI-12 digital sensors, then immediately set to work on a specialized version capable for making AC resistance measurements in gypsum blocks (a simple 78-year-old soil moisture tension sensor technology that we’ve sold by the tens of thousands into Australian agriculture). We finished that gypsum block logger only this morning (just hours short of our close-of-business); we already have orders for a whole raft of them in the New Year. Test jigs and the business of building production lines was a mixed bag in 2018; we’re going to have to start that process earlier in the product development cycle and allocate a greater portion of resources to it. Next year we’ll do better. Our Green Brain web app for farmers passed a billion records in 2018 and boasted over 99.8% up-time and a 90% customer retention rate. Google Analytics tells us that most irrigators with our gear are hitting up their Green Brain account on a daily basis. Adelaide-based Myriota have progressed their satellite IoT technology to the point where they released their software development kit (SDK) in November 2018. Within weeks of purchasing ours we’d established a link via satellite from the rooftop at MEA via the Internet back to MEA. Next year we’ll turn that ground-breaking IoT technology into practical agricultural applications. In my own small corner of MEA’s engineering world I’ve completed my three-decade effort to build a crop-water stress (CWS) sensor for use in horticultural crops. These self-contained solar-powered instruments make use of all that new 2018-technology for data logging with Bluetooth data pickup. In the wee morning hours, I’ve also solved the sensor insertion issues needed to bury the tiny probe into the xylem tissue of vines and trees. The prototype CWS sensor has run flawlessly for two solid months in a grapevine in the MEA Test Garden. The production run of the first 25 operational units ran smoothly. As part of that CWS instrument design we created a sensor shelter to house the vapour-pressure deficit sensor. This specially-shaped nylon housing costs under $60, was designed by MEA’s mechanical engineer and manufactured in Brisbane in less than a week on a HP Jet Fusion 3D-printer. Commercially, 2018 was a hard year for us at MEA as we transitioned to second and even third generation products – this effort absorbed huge portions of our profitability but kept us in the marketplace. There’s no standing still in IoT land. However, help was at hand from the Australian Federal Government – in November 2018 MEA learnt that we had been awarded an ‘Accelerating Commercialisation Grant’ to build a new generation of on-farm IoT products. That will be the job in 2019-2020. With the official MEA year at an end, I’ll be heading home shortly to drink a beer on the back porch and look out over the MEA Test Garden where I spend my engineering down-time growing things. This garden is the secret sauce behind my working life in irrigated agriculture, just as my remote hillsides at Pine Hut Knob were the secret sauce behind the field measurements in wind and solar energy that fueled MEA’s early decades. It’s inside this huge garden that I will begin the shakedown testing of the crop water stress sensors over the Christmas-New Year period and into 2019. Just like the CWS sensor, my tree crops and vines represent decades of patient cultivation - avocados, figs, almonds, oranges, lemons, grapefruit and mandarins, bananas, mulberries, plums, peaches, nectarines, wine and table grapes, apples and pears, kiwi fruits and passionfruit – all the trial crops surround me, and my home lab stands in the middle. There’s plenty of scope in my own back yard for a quick and broad test of the CWS technology across a whole gamut of horticultural crops. So where will I be over the holiday season? Hovering anxiously over all these new sensors in my garden, nurturing and mothering and brooding away, willing all this magic to sing in unison for the greater good. And keeping a close eye on my irrigation through the heat of an Australian summer…
  9. It's all happening Down-Under; three Aussie companies have launched four satellites for IoT connectivity in the past few weeks. https://www.theage.com.au/national/how-a-tiny-box-attached-to-a-cow-s-ear-is-making-waves-in-space-20181205-p50kas.html
  10. Last week, Adelaide startup Fleet Space Technologies launched its Proxima I and II CubeSats from New Zealand, marking the company’s first commercial satellite launch and laying the foundation for free global connectivity for the industrial Internet of Things (IoT). https://www.createdigital.org.au/adelaide-startups-cubesat-launch-is-an-australian-first/?utm_source=ExactTarget&utm_medium=email&utm_campaign=EDM-20181120 This story appeared in createdigital - an Engineers Australia eNewsletter
  11. Andrew at MEA

    What does it take to be an IoT engineer?

    Producing the IoT I’m a self-confessed ‘data-sheet junky’. This is an inexpensive and innocuous habit, allowing me to indulge in circuit daydreams at odd moments while the digital whiz-kids are struggling with coding and software and Internet issues. Many hundreds of data sheets sit quietly on my phone and are backed up in the Cloud, weighing nothing, costing nothing, crammed with applications information and a free education in how to build things with electronics. These data sheets keep my dream of being an electronics engineer alive while I do my day job in the IoT business. In the past four months – since official product launch of our latest on-farm IoT product – MEA has been running field trials with live customers, testing phone apps on every conceivable Android and iOS device and generally getting beat-up as per usual. While all that has been going on a whole new engineering project has been underway: construction of the production line. No IoT product development has been completed until you’ve built the programming and test jigs and are churning out stock to load the shelves. Boxes and boxes of built-up circuit boards arrive from remote PCB factories and pile up on the incoming stock shelves. Our operations manager hovers nervously on the fringes of the engineering team, waiting for the gun to go off. Designing the product is only half the game; the second half is running every device through electrical tests, charging batteries and load-testing them, loading firmware into various on-board microcontrollers, testing Bluetooth and other comms channels, reading sensors, milking out those all-important identifying MAC addresses, logging them and printing labels then storing test results to our server. Then there’s documentation, staff training and meshing-in all those other manufactured components such as enclosures, packaging and ancillary items. We’re building low-cost volume IoT products, so this whole production process has to be automated and run by any kid old enough to serve hamburgers. Profit margins are slim as we compete in this over-hyped and over-heated world where the final arbiter of success is a signed purchase order, positive cash flow and a light service load. Our product development team is a small enclosed world demanding high skill levels and intense concentration. Upstairs is a very different reality; our marketing folk are trying to get the knots out of the customer experience, simplifying the creation of Green Brain accounts and the downloading of the Retrieve app that runs our latest Bluetooth loggers and connects sensors and loggers through phones to our Green Brain database in the Cloud. Endless talking, phone calls, stress, tweaks to Green Brain operations – just the usual daily round while management tries to keep the MEA ship on an even keel and the books balanced. So that’s the scene around me. Except my mind is off wandering, scouting ahead to the next generation of products that will one day create an internal furor of their own. I’ve come up with a cunning new circuit for solving the Yamartino formula in the measurement of wind direction for newer faster MEA weather stations. I need a PWM-to-DAC converter and I just know I’ve got a data sheet for that somewhere…
  12. Andrew at MEA

    What does it take to be an IoT engineer?

    Constipation in the IoT In his best-selling book “The HP Way: How Bill Hewlett and I Built Our Company”, David Packard provides insights into managing and motivating people and inspiration for would-be entrepreneurs. [The following two paragraphs are an extract from the cover notes…] “From a one-car-garage company to a multibillion-dollar industry, the rise of Hewlett-Packard is an extraordinary tale of vision, innovation and hard work. Conceived in 1939, Hewlett-Packard earned success not only as a result of its engineering know-how and cutting-edge product ideas, but also because of the unique management style it developed – a way of doing things called 'the HP way'. Decades before today's creative management trends, Hewlett–Packard invented such strategies as 'walk–around management', 'flextime', and 'quality cycles'. Always sensitive to the needs of its customers and responsive to employee input, Hewlett-Packard earned massive steady growth that far outshone its competitors' vacillating fortunes, even with radically different products from those responsible for its initial boom.” Yet the pearl for me in reading The HP-Way was some advice given to Dave Packard in the company’s early formative years by an older bank manager, that more companies die of constipation than starvation. The great difficulty of being a player in the Internet of Things arena is the sheer diversity of skill-sets required to implement a useful solution to a customer’s problem. Further, one runs the risk of vacillating to a stand-still, because there are simply too many possible choices to be made between competing sensors, networks, software packages and so on. The particular area causing constipation in MEA’s product development cycle at the moment is in choosing a network operator. When MEA launched its Plexus sensor networks in 2013, choices were few and an on-farm ZigBee network feeding a 3G/4G on-site backhaul made plenty of sense if we wished to provide a one-shop solution. Similarly, there was a dearth of folk who understood cloud-based data-bases and web applications, so we created Green Brain ourselves. We weren’t even sure if mobile phones were going to be sufficiently ubiquitous to allow farmers to view their data from anywhere at any time. (Indeed, we were still creating our own Magpie PC-based data logging software at launch date) And this was only five years ago! Since then, all sorts of folk have made massive investments in parallel network developments designed specifically for the IoT; Sigfox, Taggle, Lo-Ra Wan, Myriota and Fleet Systems just to name a handful. So, from being a provider of our own network solutions, MEA is now spoiled-for-choice with low-cost carriers all anxious for us to feed our data transmissions through their gateways, locking us in to finding and billing customers while the network providers bill us. This new model – of using alternate carriers to Telstra – has plenty going for it in terms of manufacturing and transmission costs, but comes with an uncertainty in coverage. We wind up wanting to build solutions for all of them. (And indeed, they are anxious for us to do so). Constipation indeed! While I was trying to get all this sorted, Green Brain passed the one billion record mark. Perhaps the lesson to be drawn from all this is that one must simply continue to adapt one’s business model to the current technological status quo, and to be fleet-of-foot about it. The Hewlett-Packard model shows that this can be done, while listening to customers and employees at the same time.
  13. Andrew at MEA

    What does it take to be an IoT engineer?

    The IoT and the Sensor Black-Hole MEA’s IoT product development road-map is pretty well laid out for the next two years: new CAT-M1, Myriota satellite and Sigfox network cores will connect third-party sensors to MEA’s Green Brain web-app. This in addition to our existing Plexus long-range ZigBee on-farm networks and the new Green Brain loggers with their Bluetooth to 3G/4G hop through farmers’ mobile phones. And of course, there’s always our GDots at the really bottom-end of our product spectrum for farmers replacing IoT technology with human gray matter. This all makes perfect sense from a company perspective, but I find myself off yawning in a corner, quite rightly leaving all the young bucks to work their digital magic. The IoT scenery – so mysterious just a few years ago – has become mundane. Lots of excitement still as the brawls develop in the race to create the networks that will carry all the data from billions of sensors but – yawn – the electronics design challenge has disappeared. Just connect a few wires from sensors to the IoT module and step back while the firmware/software boys step forward to do the real product development. I nailed all the solar-battery stuff years ago, so even that’s not around to interest me anymore. In the marketplace, more and more folk are coming out of the woodwork with a piece of the IoT jigsaw puzzle in their hand, looking for folk with matching pieces. Data analysis folk are looking for databases to work on. Software companies with databases are looking for system integrators to feed them data. System integrators are looking for network modules and network operators to carry their data to the software folk. Sensor manufacturers are looking for system integrators to snap up their offerings and solve the real-world installation issues. Everyone’s looking for customers, agents, distributors or investors/buyers. Yawn. With the IoT dust settling, I’m increasingly conscious of a black-hole appearing in my peripheral vision. Where will all the new sensors come from to interface the stubbornly analog world to this burgeoning digital world? Four decades of studying and using sensors has shown me that this is an area growing at a snail’s pace by comparison to the stellar developments in the digital sphere. New stuff appears only rarely, and then its rather more an adaption of older ideas using newer processes, or is far too expensive to match the low-cost devices in the digital arena. Sadly, analogue electronics is barely taught any longer at under-graduate level and its practitioners – such as myself – are ageing and increasingly sidelined in a digital world. Sure, there are sensors available for most of the regular tasks that can feed the IoT networks, but there are plenty of instances of poor-fits between what we have and what is needed in new applications. Sensor technology then is fertile ground for feeding the ravenous appetites predicted for upstream IoT technologies. Just a few more IoT products to knock out over the coming two years and I’ll be off to the home lab to think about weird and wonderful sensors to solve all those old intractable measurement problems. We analogue engineers are well-suited for just that sort of work. And the technology needed to build and test working circuits is all available for a song on eBay. I’ll happily leave the digital wars for a new generation to fight, while I focus on doing what I love – electronic sensor design. Analogue home labs – like mine – are well-suited to tackle future sensor droughts in the burgeoning IoT era
  14. Andrew at MEA

    What does it take to be an IoT engineer?

    A ‘BeeTLE’ in the IoT Yippee! – no more USB cables to connect into our IoT devices to set them up! Our new measurement and data logging core – code-named the ‘BeeTLE’ during product development – uses a Blue Tooth Low Energy (BTLE) radio link instead. This is a non-trivial step forward in field measurements, because connectors are the weak-points in field enclosures. At 2.4GHz, enclosure walls are transparent to communications while remaining opaque to rain, sun, dust and wildlife. Similarly, the need for robust portable field computers has finally disappeared; mobile phones – already Bluetooth and 3G/4G enabled ­ – and custom apps pick up that load. For the past five months MEA’s product development focus has been on the development of the simplest possible and lowest-cost IoT device that could read our range of on-farm sensors. This BeeTLE core would be the fundamental logging platform from which future products would step off as we add LPWAN, CAT-M1, NB-IoT and satellite backhaul links. For the moment though, BeeTLEs need only human legs + smart phones + nimble fingers + modest intellect to go out there to fetch data. Data logging fills in the measurement gaps between visits, and the irrigator trades off his travel time against the convenience and higher capital cost of always-on IoT telemetry systems. There are no network Hubs or Gateways needed; the user has those in his pocket in the shape of his mobile phone. Telstra and their ilk provide the longer back-haul of data and send him the bill for those services. Green Brain – MEA’s existing database and farmer web app in the Cloud – provides all the long-term data storage. Once new data is uploaded from a BeeTLE through the mobile phone to Green Brain, graphical data all comes back down over the Internet to the same smart phone for data display. The devil is – of course – in the detail. What happens when the measurement site is out of 3G/4G coverage? (Answer: Unloading takes place normally between BeeTLE and phone, with data stored in the phone’s memory. Upload and data display happen automatically once the farmer returns within range of either Telstra or his home/office Wi-Fi system) Can multiple users – the farmer and his workers – unload the same BeeTLE? (Answer: Yes. Green Brain tracks and automatically synchronises phones and incoming data) And so on and so forth. Good software engineering ensures that the whole process is seamless and ultra-simple for the user, despite lots of clever stuff being needed behind the scenes to cope with all the edge cases. But what’s the really cool thing about Blue Tooth connectivity? The Beetle – now called a ‘GBLogger’ because its powered by Green Brain – needs no On-Off switch! If connectors are a pain, access to power switches elevates that enclosure vulnerability to a whole new level. But no switch is needed; the entire product runs on less than 25µA of current, allowing the GBLogger’s batteries to be loaded and sealed in at the factory with years of shelf life ahead. And Australian aircraft now allow ‘gate-to-gate’ usage of such low-power BT devices, so GBLoggers can be shipped by air while continuing to advertise their readiness to start logging at a moment’s notice. Figure 1 This simple MEA circuit board is at the heart of a battery-powered, Bluetooth-enabled smart-sensor and Green Brain-enabled data logger for the simplest possible MEA on-farm IoT platform
  15. Andrew at MEA

    What does it take to be an IoT engineer?

    The IoT and the Iron Triangle The Christmas break was over and – in mid-January of this year – we regrouped here at MEA to figure out what our next product offering would be. Across the table from the engineering team sat our management and marketing folk who – for some reason – were looking us rather sternly in the eye. “Now” – said our MD – “we don’t set product development time-lines to meet market-launch opportunities but we need new product to go on show at the Irrigation Australia International Conference-Exhibition in Sydney between the 13th-15th June 2018” And so, it all began again… As engineering director, my job is to look blithely unconcerned at these times and to prevent panic breaking out among the younger engineers. Five months, to go from a clean sheet of paper to working prototypes that will utilize new technology across all the diversity of skill sets that form MEA’s IoT product development team – electronics, PCB layout, firmware phone app and web app developers, mechanical engineering and industrial design, manufacturing, packaging and all the usual interplay with our marketing, sales, operations and service people. Not to mention our suppliers… As engineering director, I don’t get to do anything specific in the way of design work anymore, but I’m expected to be across every facet of this IoT ‘sprint’ to an immovable goal post. So, I make use of the ‘Iron Triangle of Product Development’. This rule-of-thumb states that you have to continually make decisions between Quality, Cost and Time-to-Market. But you only get to choose two out of three, because they are all in their way mutually exclusive. What the rule doesn’t say is that different segments of the project get different mixes of choice. Sometime I’ll choose Quality and Speed at some additional production Cost. At other times I need to sacrifice Quality to Time-to-Market and lower Cost. And so on and so forth, literally on a daily basis. So, on this past Tuesday (June 12th, 2018) our marketing team flew into Sydney from Adelaide and Queensland with the expected new product good to launch. True, we only finished firmware a week ago, the phone app half a week ago, and a brand-new version of Green Brain just two days before. And we had to pull a few swifties by 3D-printing new parts of the enclosures rather than the more time-consuming business of creating injection-moulding tools. We took the risky step of prototyping PCB and early-production run PCB assemblies in China for the first time, because local manufacturers are not well set up for rapid and inexpensive prototyping. And we kind of skimmed over the extended field trials, relying on 35 years of data logging and field experience to direct our in-house testing. So, we made it, thanks to a deep well of company experience and rapid prototyping techniques. No doubt next week we’ll be getting the stern eye once again, and we’ll limp once more to the starting blocks.
×