I recently purchased a new HVAC system for my house. My circa 1985 heater was well past its prime having been slated for scrap since 2005. As a stats and features guy, I was very impressed with the use of DC variable speed fans, modulating gas valves and relatively massive compressor coils that create efficiency of 95-97%. The Pièce de résistance was the Wi-Fi touch screen thermostat. This awesome piece of technology was solving at least a half dozen problems I didn’t know I had.
The problem is it doesn’t work right. It’s not the interface, which is beautiful. It’s not the organization of the screens and functions, they are very intuitive. It’s the underlying code and hardware that suck. The product hit all the sexy marketing points needed for an innovative product, but failed to work reliably. As a customer, I feel like I just purchased the polished turd.
Far too Often Companies Plan Big and Execute Small
You may not know this but RTA helps a lot of companies create the technology behind interfaces and scheduling solutions. I can tell you that many of these projects are slated to miss their goal from the start.
It’s because established device manufactures have catalogs full of products that do similar things but operate very differently. Take a generic HVAC manufacturer. Every few years they release a new thermostat. Over time they end up with thermostats that need to support multiple physical layers, signals and protocols. These companies continue to support the legacy features because they don’t want to upset an old customer or their sales people. Plus they figure the retrofit market can add a bit of volume to a new launch.
In the end each new thermostat is bogged down with the features and functions from 20-30 year old technology. The development effort takes four times longer than it should because they have to implement the legacy technology along with the new features. To add insult to injury, sales decides that the new, vastly advanced, touch screen thermostat still needs to cost under $200 or they can’t compete with other high end thermostats.
To meet those costs they pick processors and components that have less horse power. Then they end up spending months trying to shoe horn the expected functionality into a platform that isn’t powered to handle it.
Attempting to Develop the Solution
The engineer assigned to this project is actually salivating. Finally he gets to create a product from scratch, and because of the basic components and low budget, all of the code and subsystems will have to be created from scratch. He will be the hero and supreme architect of the entire project!
Instead of using established software tools to create the solution, the engineer will write all the code to avoid costs. Any up front savings from not buying a software solution(s) is lost 3-6 times over when you factor in the development hours, on-going support, and maintenance.
As a final straw, the single engineer who was assigned this project, big enough for a team of 6 engineers, is fired or leaves and all of the collective knowledge behind that solution is lost.
Through thrift and misplaced ideas, you end up with a house of cards that ensures you will go through the exact process the following year.
This leads me to 4 Simple Rules to Avoid Nickel and Diming Innovation in Technology Products:
Of all the support requests we get there is only one that still makes us cringe. It happens every time we hear, “The serial port on the RTA gateway stopped working.” We cringe because 99.9% of the time the failure was not caused by a manufacturing defect but by the environment in which our gateway was used.
You can imagine how hard that message is to explain to a customer with an application that is down. “The hardware we provided is perfectly fine, even if we sent you 100 different replacements, they would all fail in your environment and application.” That is the truth, but it’s not a fun message to deliver. If the hardware is not the issue what is the cause of these failures? Serial ground loops and transient currents.
Serial Ground Loops and Transient Current
Serial ground loops and transient current issues are rarely witnessed, and one of the most misunderstood points of failure in communication. Even though ours is an industry of seasoned engineers, theses antiquated serial issues are not common knowledge. Why is that you ask? In part, because of how weird these serial faults can appear. You could literally have 100 identical machines in a factory and only 1 could have a grounding issue. It appears as though, for no apparent reason, serial communications fail, or worse, the entire device fails. These failures are caused by excess current flowing where it is not supposed to flow. The Million dollar question is always where does this excess current come from?
The frustrating fact is that there are a number of potential sources and because the cause is often intermittent, the sources are sometimes very hard to pin point. The first step is to understand the problem.
Why Faults Happen:
When two serial devices are at different ground potentials, current flows from the high potential to the low potential. If the ground at one part of the circuit is at a higher potential than the ground at another part of the circuit, unwanted current flows from the serial device with the higher potential to the serial device with the lower potential. Since serial devices seldom provide the capability to manage this additional current, it’s a matter of time before the ICs literally burn up, causing a device failure. Depending on the severity of the differential, the device may work normally for hours, days, or even weeks before the failure occurs.
Another possible problem source is transients on the serial data lines. Transients are short bursts of energy which are either wirelessly induced on a serial line, or enter the system from an external source. Lightning is often a cause for transients, but power surges courtesy of your local power company, induced transients from large motors, and improper wiring can all add transients to your serial data lines. These signal distortions can be minor and destroy only the data signals, or major enough where they destroy the physical serial receivers.
What Can You do To Stop Faults?
There are a number of mechanisms used to minimize the risks of ground faults and transients. Two of the best are Isolating the Power Supply and optically isolating the serial communication lines.
An Isolated Power Supply transformer couples power from the input to the output. Electrical noise in the ground plane from lighting, magnetic field issues, and nearby equipment is prevented from entering the communication system. Though this works well to help avoid transients and some ground loops, it does not prevent noise induced wirelessly.
Another approach is to optically isolate the serial communication lines at each device. In this method, a line powered optical isolator is connected to the serial communication lines at each device. As both the serial transmit and receive lines are connected through optical connections, no additional current can flow across the communication lines. The disadvantage of this approach is that optical isolation is required at each device. It a costly pill to swallow as a preventative step, but an absolute must if you are experiencing issues.
This information might not save your from losing a device on the line but at least you will know how to prevent it from happing in the future. The trick is to buy products with good warranties or isolate your power supplies and serial communication lines.
If you have any questions, send us an email at DrewandScott@rtaautomation.com.