The curse of the intermittent fault

c: | f: Technology / Design

When your boiler’s broke. And your water’s cold. Who you gonna call? Ghostbusters? A heating engineer? Or someone with a soldering iron and dogged determination?

I think I’m incompatible with today’s throwaway society because I dislike chucking stuff. Maybe I’m just old school, but I like to be sure that something is beyond repair before I consider getting rid of it. This is in stark contrast to most people in the retail environment, such as Apple Genius bar employees or, in this case, gas engineers who would rather I buy new.

There’s no blame implied here. They’re trying to make a living just like me. But if something – especially electrical in nature – exhibits faults, and its disposal and replacement comes at a huge cost to the environment, I remind myself of the WEEE man at the Eden Project and the vast amount of stuff that each person throws away in their lifetime.

The relentless drive to own the latest gizmos and hence line the pockets of corporations and service providers, has pushed us where we are. Insane marketing and advertising budgets impress upon us the need to progress or be left behind. Dubious legislation with its heart in the right place makes “old tech” obsolete by virtue of its supposed environmental impact.

But it rarely take into account the true cost of replacing something – the raw materials, mining, manufacture, packaging, energy consumption, disposal/recycling, transport, and storage that all go towards ensuring you can buy something on demand. And that scale drives manufacturing costs down: good for consumers in one regard, but it usually cuts corners in long-term quality because the products no longer need to last more than a few years.

Gas appliances use electricity too

Like a lot of people I have a condensing boiler, which switches on when we run the hot tap, or when we choose to start the central heating. Except it’s been on the blink for ages. It’d fire, the gas would ignite, flames leap, things get hot. Then after some time – five seconds, ten minutes, half an hour, it goes out and won’t come back on. Maybe later it’d come on and be fine, perhaps work flawlessly for days, then it’d go back to being flaky.

Of course, I’ve had someone look at it. The first guy thought I was a fantasist because it worked flawlessly when he came round. Naturally, for an intermittent fault. The diagnosis of the next guy was that it was knackered and I needed a new one – something he was only too keen to sell me for £1200. I remained unconvinced so we lived with it, despite the number of times it shut off mid-shower and I ended up racing to remove the traces of soap in the last twenty seconds before the water turned to icy needles.

In the end I sought a third opinion. The boiler had never really been serviced so there was every chance that it just had years of crud gumming up the workings. The engineer stripped the front off, cleaned it out, and put it all back together. And though it still didn’t work, he was far more pragmatic. Instead of trying to flog me a new one straight away, he analysed the failure sequence, ruled out sensor failure and concluded that it was most likely the circuit board. With the best will in the world, the best he could offer was to replace the PCB for a few hundred pounds.

Heating engineers aren’t generally electricians. The nature of their job requires them to deal with mechanical components like flues, flanges, pressure switches, water, and plumbing. Show them a circuit board and it’s voodoo to them: they’re not electrical engineers.

But I am.

Of sorts.

At least, I spent almost ten years in manufacturing quality and know that the primary source of failure in anything electrical that’s subjected to repeated on/off cycles and heat is… the dry joint.

Don’t bogart that joint, my friend

Nothing to do with the Fraternity of Man song of the same name, this is a brief lesson in manufacturing. Put simply, some clever spark arranges components on a circuit board that does something. These components are poked through holes in the board or, in newer circuits, glued on top of tiny silver pads. The board then passes through a wave solder process.

Imagine a conveyor belt upon which the circuit board travels, as it winds its way through the factory. It goes through machines that place hundreds of components, past banks of operators who add more delicate parts, and when it’s all assembled, one conveyor belt has a small gap in it. A bubbling tank of silvery liquid is beneath, and a thin stripe of this liquid is forced into the conveyor gap. As the circuit board passes over this knife as it’s called, the silver – solder – sticks to the components and thus creates the required electrical contact. It’s far faster than soldering each component by hand.

But herein lies the problem. There are the following variables to juggle to get it perfect:

  • The height of the knife. Too high and too much solder sticks or bubbles through the holes in the board, causing short circuits. Too low and not enough sticks, making the joints weak.
  • The temperature of the liquid. Too hot and it blisters the board or melts the glue that holds them, causing them to drift a fraction and miss the silver pads. Too low and it risks not adhering to the joints evenly.
  • The speed of the conveyor. Too fast and not enough coats the joints. Too slow and you risk scorching the board or putting too much solder on.
  • The variability of the previous processes – manual or automatic – that place the components. If the legs aren’t fully through the holes, there’s not enough component to stick the solder to, and the joint becomes susceptible to mechanical failure over time.

All these variables are tough to balance. Even if you get it just right, the machine that creates the solder knife goes out of alignment from time to time and causes quality issues. These might not be apparent immediately, but they can rear their heads in later life. In someone’s boiler, for instance.

eBay not to the rescue

There are a tonne of places online that’ll sell replacement circuit boards for a fraction of the retail cost. They recondition them and sell them on. You send yours in, they test it, and either repair it immediately or send you a replacement and put yours back in stock when it’s fixed for someone else. Some places even send you one out for a deposit then offer a refund when you swap and send the old one back.

Could I find such a place? No. “Out of stock, sir” or “Which circuit board do you want? There are two.”

How was I supposed to know? The manual’s useless. No part numbers given anywhere. The online places I tried didn’t know which part I needed. I was blind. And stuck with a boiler that was possibly repairable and I didn’t want to throw away, but nobody could help.

Well, okay, one person could help and it wouldn’t cost a penny: me.

So, backed into a corner I didn’t want to be in, I took the thing apart. Slid the tray out housing the boards and looked at the top. The first thing I saw: brown discoloration on the top of the smaller (ignition) board around a couple of components. That’s always a bad sign. It smacks of sparking. Why do things spark? Because the joints are crappy. Why are they crappy? Dollars to doughnuts, it’s a faulty wave solder process.

What harm would it cause to take a look, right? I took some photographs just in case, unclipped the myriad cables and unscrewed the PCB, then wrestled the board out of its seat and turned it over. With a bright LED torch I inspected every single blob of solder, starting around the part of the board that was scorched. What did I find?

Dry joints.

Lots of dry joints.

They’re given away by a telltale little ring around the solder blob, partway up the slope where the silver has cracked slightly and the joint is therefore not sound. I lost count of the buggers but there must have been twenty at least; some more pronounced than others. Bad connection = sparking = intermittent fault. Probably.

So I fired up my soldering iron, reflowed every one I could find with some fresh solder and put the board back in its holder.

The boiler’s worked ever since, without a single issue.

Moral of the story? If you’re ever faced with intermittent faults like this, it often pays to get an opinion from someone whose expertise lie outside the comfort sphere of the expert who deals with the daily grind of the mechanical. Because after software, electrical circuits are the greatest bugbear of the modern appliance.

Leave your mark

(required)

(required, never made visible)

(optional, linked with rel="nofollow")

(required)