AssholeDesign

First things first: it is replaceable. I outlined how to replace it.

Second, the battery doesn’t define the lifespan of the equipment. The upgrade cycle it’s designed around defines the lifespan of the equipment.

But let’s say it didn’t. Let’s figure out how to make an rtc battery field replaceable:

It will need an access panel and a battery holder. We’ll need to limit out selection of battery chemistry to those that don’t corrode around the contacts. We gotta standardize the location and type of battery so that means standardizing the type of rtc circuit and since the whole point of choosing a replaceable battery is to lengthen the life of the equipment, we need to pick a rtc circuit with a long design life too and integrate that choice into all our future designs. I say that because most of the time nowadays there’s so much stuff on any given board that you gotta go out of your way to get a pcb that doesn’t already have some kind of rtc integrated into some kind of chip you picked for a different reason somewhere.

Now let’s put our battery door on. Wait a minute, this things 1u. The front panels all the way out because it’s taken up by 48 Ethernet jacks, six leds, two usbs, two fcs, two more ethernets, a serial port, a button and somehow a corporate logo around the edge. Okay let’s turn it around: crud, all taken up by two huge bus connectors, two power supplies and their vents, two power inlets, the quick release tabs for the power supplies and bus connectors, the mandated visible ul/health and safety sticker and another fan grille.

Guess we’ll have to put it on the other sides. No big deal, just isolate the place on the pcb that the battery holder sits on so its standards compliant, add a few plasma cuts and a dimple to the machining steps for that panel and add a little plastic door (captive with a screw or hinge so it can’t be lost) and we’re done!

Except we just made it worse.

Imagine going to replace that battery. It’s field replaceable, right? So what’s that process? Oh it’s real straightforward, you just bring down or reroute the systems reliant on the unit, unplug 48 Ethernet jacks, two usbs, two fcs, two more ethernets, a serial port, two bus connectors from the back, and two power inlets from the back, pull it out of the rack, unscrew the battery door, replace the battery, put it back in the rack, plug in 48 Ethernet jacks, two usbs, two fcs, two more ethernets, a serial port, two bus connectors into the back, and two power inlets into the back, turn it on, securely set the rtc (this is incredibly important!) and reroute systems back to it or bring them back up.

And if nothing goes wrong, if we didn’t damage a connector or receptacle, drop something or catch an esd event then we just doubled the lifespan of our device!

Except we didn’t. The battery had a ten year service life in this environment, and we replaced it at nine to be safe. Cisco though, stopped releasing security patches too, so now our router has ten new years of being unavoidably owned by every sk the world over.

Well, if we just went through all that work, probably the firmware policy would be updated and the support window lengthened to match, right? So now they’re pushing security patches for 20 years! Certainly that cost won’t be borne by the customer and we’ll be rewarded for our good design work!

On a gigabit router. For the rack mount environment.

And we also have a standard rtc and battery type to integrate into all future designs, training material and equipment.

I didn’t just waste thirty minutes of a day off writing that up to make fun of you, but to illustrate how the conditions under which specifically telecommunications equipment operates, is designed and managed dictate what decisions the designer makes about stuff like rtc backup batteries, which are a security feature btw.

If we had say, a planned economy, we could expect devices like these to be designed with a longer lifespan and an extant recycling process in mind. Under our current system that’s just not possible.