You’re likely familiar with the concept of “sunsetting,” where a connectivity standard or application is scheduled to be phased out, such that users who depend on it are often simply “out of luck.” It’s frustrating, as it can render an established working system that is doing its job properly either partially or totally useless. The industry generally rationalizes sunsetting as an inevitable consequence of the progress and new standards not only superseding old ones but making them obsolete.
Sunsetting can leave unintended or unknowing victims, but it goes far beyond just loss of connectivity, and I am speaking from recent experience. My 2019 ICE Subaru Outback wouldn’t start despite its fairly new battery; it was totally dead as if the battery was missing. I jumped the battery and recharged it by running the car for about 30 minutes, but it was dead again the next morning. I assumed it was either a defective charging system or a low- or medium-resistance short circuit somewhere.
(As an added punch to the gut, with the battery dead, there was no way to electronically unlock the doors or get to the internal hood release, so it seemed it would have to be towed. Fortunately, the electronic key fob has a tiny “secret” metal key that can be used in its old-fashioned, back-up mechanical door lock just for such situations.)
I jump-started it again and drove directly to the dealer, who verified the battery and charging system were good. Then the service technician pulled a technical rabbit out of his hat—apparently, this problem was no surprise to the service team.
The vampire (drain) did it—but not the usual way
The reason for the battery being drained is subtle but totally avoidable. It was an aggravated case of parasitic battery drain (often called “vampire drain” or “standby power”; I prefer the former) where the many small functions in the car still drain a few milliamps each as their keep-alive current. The aggregate vampire power drawn by the many functions in the car, even when the car is purportedly “off,” can kill the battery.
Subaru used 3G connectivity to link the car to their basic Starlink Safety and Security emergency system, a free feature even if you don’t pay for its many add-on subscription functions (I don’t). However, 3G cellular service is being phased out or “sunsetted” in industry parlance. Despite this sunsetting, the car’s 3G transponder, formally called a Telematics Data Communication Module (TDCM or DCM), just kept trying, thus killing the battery.
The dealer was apologetic and replaced the 3G unit at no cost with a 4G-compatible unit that they conveniently had in stock. I suspect they were prepared for this occurrence all along and were hoping to keep it quiet. There have been some class-action suits and settlements on this issue, but the filing deadline had passed, so I was out of luck on that.
An open-market replacement DCM unit is available for around $500. While the dealer pays less, it’s still not cheap, and swapping them is complicated and time-consuming. It takes at least an hour for physical access, setup, software initialization, and check-out—if you know what you are doing. There are many caveats in the 12-page instruction DCM section for removal and replacement of the module (Figure 1) as well as in the companion 14-page guide for the alternative Data Communication Module (DCM) Bypass Box (Figure 2), which details some tricky wire-harness “fixing.”
Figure 1 The offending unit is behind the console (dashboard) and takes some time to remove and then replace. Source: Subaru via NHTSA
Figure 2 There are also some cable and connector issues of which the service technician must be aware and use care. Source: Subaru via NHTSA
While automakers impose strict limits on the associated standby drain current for each function, it still adds up and can kill the battery of a car parked and unused for anywhere from a few days to a month. The period depends on the magnitude of the drain and the battery’s condition. I strongly suspect that the 3G link transponder uses far more power than any of the other functions, so it’s a more worrisome vampire.
Sunsetting + vampire drain = trouble
What’s the problem here? Although 3G was being sunsetted, that was not the real problem; discontinuing a standard is inevitable at some point. Further, there could also be many other reasons for not being able to connect, even if 3G was still available, such as being parked in a concrete garage. After all, both short- and long-term link problems should be expected.
No, the problem is a short-sighted design that allowed a secondary, non-core function over which you have little or no control (here, the viability of the link) to become a priority and single-handedly drain power and deplete the battery. Keep in mind that the car is perfectly safe to use without this connectivity feature being available.
There’s no message to the car’s owner that something is wrong; it just keeps chugging away, attempting to fulfill its mission, regardless of the fact that it depletes the car’s battery. It has a mission objective and nothing will stop it from trying to complete it, somewhat like the relentless title character in the classic 1984 film The Terminator.
A properly vetted design would include a path that says if connectivity is lost for any reason, keep trying for a while and then go to a much lower checking rate, and perhaps eventually stop.
This embedded design problem is not just an issue for cars. What if the 3G or other link was part of a hard-to-reach, long-term data-collection system that was periodically reporting, but also had internal memory to store the data? Or perhaps it was part of a closed-loop measurement and control that could function autonomously, regardless of reporting functionality?
Continuously trying to connect despite the cost in power is a case of the connectivity tail not only wagging the core-function dog but also beating it to death. It is not a case of an application going bad due to forced “upgrades” leading to incompatibilities (you probably have your own list of such stories). Instead, it’s a design oversight of allowing a secondary, non-core function to take over the power budget (in some cases, also the CPU), thus disabling all the functionality.
Have you ever been involved with a design where a non-critical function was inadvertently allowed to demand and get excessive system resources? Have you ever been involved with a debug challenge or product-design review where this unpleasant fact had initially been overlooked, but was caught in time?
Whatever happens, I will keep checking to see how long 4G is available in my area. The various industry “experts” say 10 to 15 years, but these experts are often wrong! Will 4G connectivity sunset before my car does? Abd if it does, will the car’s module keep trying to connect and, once again, kill the battery? That remains to be seen!
Bill Schweber is an EE who has written three textbooks, hundreds of technical articles, opinion columns, and product features.
Related Content
- Powering the autonomous car
- The beneficial ripple effect of new automotive power devices
- Vampire Drain: The Bane of Today’s Cars
- Being Honest About Your Power Source
- Battery Self-Discharge is Real — But One Case Has Me Mystified
References
- Wexler Boley & Elgersma LLP, Subaru 3G Battery Drain Lawsuit
- JND Legal Administration. Subaru Battery Settlement
- Legacy GT, Subaru Battery Drain due to shutdown of 3G net
- Subaru/NHTSA, Telematics DCM Replacement
- Subaru/NHTSA, Data Communication Module (DCM) Bypass Box
- Subaru Outback.org, That annoying battery drain problem on the Outback and other models
- Subaru Outback.org, Dead Battery: Telematics Data Communications Module (DCM) Warranty Extension 8yr/100k miles
- EV Guide, What Is Vampire Drain?
The post Did connectivity sunsetting kill your embedded-system battery? appeared first on EDN.
