Post a reply to the thread: Can you recommend REST API controllers?
You may choose an icon for your message from this list
Please enter the name by which you would like to log-in and be known on this site.
Please enter a password for your user account. Note that passwords are case-sensitive.
Please enter a valid email address for yourself.
Will turn www.example.com into [URL]http://www.example.com[/URL].
Originally Posted by MaxBurn This is a level of reliability we can not have in commercial controls. Understatement. Have systems I installed in the 90's that are still in service today that the owner still doesn't want to pony up to replace. Stuff designed in the mid-late 80's Originally Posted by MaxBurn This is why controls are a lot dumber, they are paired down to only what they need to be reliable. While there is truth in that, don't discount build quality. OEMs certainly can make a product that lasts the life of the building and they have. That seems to be something cost "optimized" more often in newer gear from what I see. There really is no motive for OEMs to build a 20+yr lifespan into a product were the customers are use to the disposable cell phone thinking culture of today. Design to exceed your warranty by a safe margin and call it good enough. Sells more widgets and the general public is getting more acceptable to this BS. Best, cheap, fast...pick any two.
Security system alerted me I left the garage door up yesterday morning on the way to work. No problem, once at the office I went into the home network and reset the alarm, worked. Told the alarm to lower the garage door, did not work. Tried bypassing the zone in order to simply arm the motion in the garage and leave the door up, didn't work either so looks like commands aren't working for some reason. SSH into the raspberry pi and tell it to reboot. It never came back. Power cycled it last night when I got back home and it still didn't boot. This is about the third time in four years or so that has happened to this. There have been other failures too, one time the NIC stopped working on the pi 2, swapped to a pi 3 I had. About the third SD card I've tried too, currently on lexar 633x high endurance, as was the last time it failed too. This is a level of reliability we can not have in commercial controls. This is why controls are a lot dumber, they are paired down to only what they need to be reliable. Sort of like the space shuttle flight computer was.
webinar: The Life & Times of an API https://www.nginx.com/resources/webi...ife-times-api/ This webinar focuses on the key considerations for companies to take into account when building out their API strategy, including what to look at when evaluating the various API management solutions on the market right now. About the Webinar According to a recent report from Akamai, API calls now represent 83% of web traffic. Join Ed English and Miles Martin as they discuss how the majority of these calls are coming from custom applications in support of digital transformation and cloud-based application deployment. With a myriad of complex licensing models for API management solutions on the market, Ed and Miles discuss how to avoid getting locked into a platform that doesn’t provide the technical or commercial flexibility your API ecosystem may require in the future.
Originally Posted by orion242 Seems pretty poor but I assume its all about exact conditions and what the I/O on that memory looks like. Have seen them puke like this though. Have a few toss away units running wireshark captures 24/7 close to a year now and they just keep working...knock on wood. While they are 24/7 use, the amount of traffic they are capturing is very small, aka little writing to the memory. Power & temp are conditioned, etc, etc... Used them more demanding cases and saw this and other issues much more. Would agree, not 'controller' grade plain and simple. Good day Orion242, <snip> I assume its all about exact conditions and what the I/O on that memory looks like. Agreed. I suspect it all depends upon what you are using the RPi for. In my testing I am writing to the SD/MMC regularly (every 5 or so minutes) in a data logging application test. Consequently, my tests are stressing the design's weak point, but that was the intent. Power & temp are conditioned, etc, etc... This is a must, otherwise problems start to manifest a lot quicker. Quite honestly, I am surprised at a few of the RPi "solutions" that appear to have little thermal design efforts applied to them. I guess the design architects missed the class on Thermodynamics and how temp affects semiconductor/electronic component aging, reliability, and longevity. I do have to say that I am impressed at how well the RPi does work given that it was really only targeted towards consumer use with a focus on low cost. Although my testing does reveal issues, I have seen some much higher end and expensive designs perform similarly and so the RPi is not alone. Cheers, Sam
that's certainly true in regular times, but with what is going on now, everyone is gonna be fighting for scraps soon.
Originally Posted by FreezerGeezer You are an optimist, I’ll give you that. In my experience managers usually have a poor grasp of what someone’s actual usefulness could be. And more than once I’ve worked in companies where to get a raise you had to resign and hope they wanted to keep you, if you wanted to stay. Exactly, that's why average tenure in tech sector in a lot of areas is under a year. People gain skills quickly and get lured away so fast it's darn near impossible to retain. If they aren't actually looking for another job/raise they end up underpaid.
Originally Posted by fixithvac Those who can not adapt may lose jobs, and those who can adapt will get a raise. You are an optimist, I’ll give you that. In my experience managers usually have a poor grasp of what someone’s actual usefulness could be. And more than once I’ve worked in companies where to get a raise you had to resign and hope they wanted to keep you, if you wanted to stay. Originally Posted by fixithvac It can be frustrating, but what is your alternative? 20 psi in the main line and you are good? No. The alternative is a utopia where something significantly better than any existing protocol comes along & finds universal acceptance. Where remote access software & security software all plays nicely. And installers do a good job - including RTFM! 😉 In the meantime I suppose we will all trundle on as usual. 🙂
And good luck pushing some roll your own raspi solution <insert device of the week here> as a serious option. There is a snowballs chance I would talk to any serious IT department and try and pitch that idea.
Originally Posted by s2sam In my testing I found that SD and MMC corruption happens about every 4-6 months in 24/7 use. Seems pretty poor but I assume its all about exact conditions and what the I/O on that memory looks like. Have seen them puke like this though. Have a few toss away units running wireshark captures 24/7 close to a year now and they just keep working...knock on wood. While they are 24/7 use, the amount of traffic they are capturing is very small, aka little writing to the memory. Power & temp are conditioned, etc, etc... Used them more demanding cases and saw this and other issues much more. Would agree, not 'controller' grade plain and simple.
Originally Posted by MaxBurn Interesting, I hadn't run across Dingo before. Looks like they might have something that could be a gateway for this stuff. https://www.raspberrypi.org/forums/v...c.php?t=161166 http://dingo-iot.io/iot-bacnet/ Me neither. Not to be confused with digo.
Originally Posted by numbawunfela Another awesome contribution Sam, Thanks for this. Good day numbawunfela, You are most welcome! As I mentioned, my testing results may vary depending upon how well the hardware and software/firmware is designed, but also what the RPi application is doing...effectively YMMV (your mileage may vary). Sadly, these type of stats may be the new norm given some of the newer processor and memory silicon technologies currently used. The latest silicon tech is very sensitive to a lot of things that could be ignored in the older tech stuff. For example, one of my first elevator controllers (based on an old Motorola HC11) that I designed and deployed in 1996 ran 24/7 for 17 years with 0 failures or even needing a reboot and was still running without issue when it was decommissioned. Sadly, I will not have my own results of the new tech for a while, as I am just starting to deploy the latest and greatest now. Cheers, Sam
Originally Posted by s2sam In my testing I found that SD and MMC corruption happens about every 4-6 months in 24/7 use. Another awesome contribution Sam, Thanks for this.
Originally Posted by jschulze Yes, there is a version of the Dingo that uses a Raspberry Pi, but look closely, it's actually the Raspberry Pi Compute Module, which is aimed at industrial applications and uses eMMC storage instead of an SD card. This is mounted onto Dingo's Computer Board (DCB), which I would assume adds necessary protections and considerations for operation in its intended commercial or industrial application. This is quite different from using a standard Raspberry Pi in commercial or industrial applications, as many users here have already stated is a bad idea. Good day All, Just a few comments... The Compute Module (CM) is basically a stripped down version of the "usual" RPi, but in a more embedded format that uses MMC. I have been testing the RPi inhouse for several years now (down and dirty prototypes, etc) and I find it quite good overall, but certainly not a platform I would use for high reliability applications. In my testing I found that SD and MMC corruption happens about every 4-6 months in 24/7 use. The effect of the corruption can vary from a complete system crash... which necessitates a full rebuild of the file system, to corrupted files on the filesystem. However, adding more design efforts (better hardware, better fail-safe software/firmware, etc) may overcome these issues. Even so the processor is not an industrial rated processor and so would limit where I could use RPi. Secondly, the long term availability of the CM is shown to be 2026 which is too short for my customers. My view of this is that if I have to design around the limitations of the RPi with external boards, etc then I might as well design the complete assembly and be in full control and availability of the hardware. As for the Dingo... There is not a lot of info on the Dingo's hardware and so it is not possible to know if their products address the RPi's weak points. Looking at the hardware pics on their website reveals little to no over voltage protection, etc. However, it is possible that these protection circuits are elsewhere (back of board, external board, etc), but for now I would assume no protections are present. Interestingly, when I attempt to download files on their site, a web 404 error results... so there are broken links, lots of them. This could be a simple oversight, but it does not speak well for a company to have so many active broken links on their website. If one attempts to purchase the hardware you are directed to an ordering page that states that you must order through their International Distributors where only one is listed and is based in the UK (https://www.innovelec.co.uk/). Searching this Distributor site does reveal Dingo products with datasheets, but pricing is via a quotation only. Another concern of mine is that Dingo does not list their company location, address, e-mail address, or phone number ...only an message submission page and links to twitter, facebook, or linkedin. I would ask how one accesses support? To me twitter, facebook, etc is not suitable for tech support. With that said if anyone is using Dingo's products, then by all means please post up your experiences. Cheers, Sam
Interesting, I hadn't run across Dingo before. Looks like they might have something that could be a gateway for this stuff. https://www.raspberrypi.org/forums/v...c.php?t=161166 http://dingo-iot.io/iot-bacnet/
Originally Posted by fixithvac The Dingo runs on Pi too ) Yes, there is a version of the Dingo that uses a Raspberry Pi, but look closely, it's actually the Raspberry Pi Compute Module, which is aimed at industrial applications and uses eMMC storage instead of an SD card. This is mounted onto Dingo's Computer Board (DCB), which I would assume adds necessary protections and considerations for operation in its intended commercial or industrial application. This is quite different from using a standard Raspberry Pi in commercial or industrial applications, as many users here have already stated is a bad idea.
Originally Posted by FreezerGeezer Ha! Good philosophy. I keep reminding myself of the mantra that ‘Uncle’ Brendan used when I was a journeyman “If the job was easy, they wouldn’t need us!”😁 It doesn’t quite cure the frustration of spending an hour or two to do what aught to take 15min. But hey ho. 😉😇 So anyway, still waiting for proof that IT is a good thing for our industry, people? It can be frustrating, but what is your alternative? 20 psi in the main line and you are good?
Originally Posted by numbawunfela I am a firm believer that if things get more complicated, then I suddenly become the only one who can fix them and then I get a raise. So I tend to be very zen about the whole thing. I hear ya, I agree much of the time, I am just not grumpy. Make it hard, weird whatever. It only makes me more indispensable. Those who can not adapt may lose jobs, and those who can adapt will get a raise.
Originally Posted by FreezerGeezer Devil’s advocate time: Can anyone explain to me EXACTLY how IT-centric equipment / software is GOOD for our industry, given that I lose several hours of productivity a week to fighting with VPN’s, open / work profiles on my company phone, SSO which TOTALLY isn’t (I signed on about 7 times today alone. Usually when already SSO’d to something). Networking engineers paid literally double my wage who swear blind everything is ok. So we find multiple kinks in their network cable. Bypass that junk & suddenly a JACE that used to fall over ever month or more, has been rock solid for a year. IT security departments that tell me our own version of Workbench is not an approved software & it has to come off my laptop unless I can justify having it. And so on. And so on. Ad infinitum. In other words, I find in the IT world the left hand almost never knows what the right is doing. Or often even what it’s supposed to be doing. Most of us work in HVAC. That’s comfort stuff. Sampling room temp every 15min is entirely adequate. Why do we NEED to have gigabit comm’s? Come on. Prove your argument. There is always this Security - Usability equation. You need to find your middle ground that works for you. Regardless of what you do you need good relations with the IT guys and patience.
Originally Posted by theintegrator We get a lot of requests from peers and clients as to why we aren't using things like the Raspberry Pi. We've got a lot of reasons, pick whichever fits your need....warranty and factory support, ease to install (decent sized screw terminals...no soldering), designed and tested for the ambient conditions of the equipment they control, we don't have a guy sitting on a bench in his basement making them, IT security requirements (everyone has different requirements), and nothing is ever $30. Even if something was $30, no one would sell it for that. You have assembly costs, programming costs, overhead, design, warranty costs, etc. Others have mentioned it. The VAV controller doesn't cost $200 to build, it is probably a tenth of that or less. But every one is the same and covered by the manufacturer. They test them, get them UL/BTL/other certified. None of us sell a $200 controller for $200 either. Because we own or work for businesses that need to make profit to continue growing and hiring people that sit on HVAC-Talk all day. 20 years ago Linux was getting the attention of the public. They had the same arguments as you - what is the business model in offering the software for free? What about support? Today the internet and the cloud runs on Linux. To get pro support you need to pay. You can get community support for free.
I agree ) This will change the business model too and a lot of people will not be happy.
Forum Rules