Hacker News Reader

Comment by mygnu | Hacker News Reader

I am considering adding a watchdog logic in the microcontroller that holds a mosfet gate, if there is any signs of failure in logic it would pass through the normal light signal.

View on HN

A lot is also dependent on how it's wired into the vehicle, as in the main electrical failure modes like open circuit, short to ground, stuck at +12V, etc. Need to be careful about where the responsibility (and liability) is.

I have a second concern re: posi-tap as a connection method. Personally, I would not use a insulation displacement connector like that in a high-vibration environment, at least not without many many hours of accelerated lifetime testing under high vibration load. It's possible they're fine, but at the end of the day you're driving a couple of knives partway through a wire in a way that's non-inspectable.

For field installs perhaps the low-temp soldering butt splices would work well? That would be my go-to in your shoes. It does require cutting and stripping wires, which has its own issues if done by the layperson, but it does connect and encapsulate the joint in a way that should be very robust to vibration and to environmental contamination/corrosion.

I think the ideal here would be what good trailer brake adapters do in cars that are not factory equipped with a trailer light connector: they have an adapter for each car model that plugs directly in to the stock wiring harness at the taillight. No soldering, no stripping wires, no piercing insulation. However, this complicates the product because you need an adapter for each bike manufacturer and possibly each bike model or even model year.

Another advantage of this is it simplifies installation for the bike owner. Just mount the controller and connect the plugs.

You have to run the business case (at the end of the day all the work you do has to support this).

What's the NRE/tooling and marginal cost to produce a harness variant, and how many units with that variant will you sell? Will people buy the device without the harness adapter?

OE connectors are hard to get, they're typically not some stock Molex or Deutsch dealy-o. I've worked on projects where we 3D printed whole connector blocks to try to do get mating to OE ECUs. It kinda sucks, and you need a whole lot of volume to support that kind of engineering.

Agree, it has to be practical. The better trailer light adapters do it but probably not for every obscure older car/truck model. And really the whole product idea is something that only applies to older low-end bikes. Newer/higher end bikes increasingly have this feature from the factory.

Genie soldering is not preferred for high vibration environments. Crimping is considered more reliable. Proper strain relief and anti corrosion treatment is key.

It's not preferred for inline connections, but context is important. In a field installation, you can do up a solder-type butt splice with a lighter if you really need to, whereas crimping requires specialized tools.

True enough, but I would take it one farther and say that a properly executed hand splice is arguably preferable to solder if vibration is a factor. Also, even less tools needed, though not everyone carries a little tube of conductive grease with their electrical tape. (But they should, it makes a decent splice immune to corrosion and a solid, “permanent” repair. )

I have had many, many soldered splices fail on generators, marine engines, other long running equipment.

They just break where the copper meets the solder joint. Copper generally has poor fatigue characteristics, depending on the alloy.

I thought electrical wire was generally ~100% Cu.

Some of this also comes down to harness design and how it's put back together after the splice. Ideally, the exposure of the splice itself to flexing is very low. I reckon if it's strain relieved/mechanically fastened well and encapsulated, it doesn't really matter what style of splice you use.

Yes, strain relief is key. Even more so for soldered connections where vibration is significant. But really, no joint is immune to problems without some kind of strain relief to prevent stress concentrations at the rigidity transitions.

I agree that soldering probably won’t produce any issues if the wire is not subjected to significant flexure or vibration, or if it properly strain relieved and secured to a bulkhead so there is no movement at the limit of solder wicking. It’s just that I’m usually too lazy to do that perfectly, so I tend towards more forgiving solutions in vehicles and equipment. But sometimes, solder is the one true way, and I always feel better about soldered joints even if I know I’m statistically wrong.

As for wire being 100 percent copper? I’m not sure, but I have definitely noticed different rigidity of copper fibers. Might be impurities, annealing, or work hardening from the drawing process… but there is definitely variation. Some wire can be really susceptible to fatigue cracking, while others seem relatively immune. Not sure of the why.

Also, don’t even get me started on copper plated aluminum fiber wires.

Consider adding logic to test if the bulb is open. That would be a killer feature for sales.

With BLE integration and an app showing you great deals on bulbs.

And also ads for unrelated products down at the bottom of the page in the app.

ads are amateur hour - gotta make it a subscription and really milk those users.

Also offer a lifetime subscription, then the API disappears after 6 months

it's all LEDs now, i would think, is "open" still a thing?

Reply on news.ycombinator.com