Why do problems come in pairs?

Article By : Michael Dunn

Computers are infamous for manifesting unrelated problems in groups of two, but now my hardware is doing it too.

So, why do problems often come in pairs. Maybe it’s just confirmation bias. But here’s a pair of them I just fixed.

A few weeks ago, my home heating system started showing symptoms. A pair in fact, and while I was busy theorizing over connections between them, they got worse.

Symptom 1: The thermostat would kick in, then out, then in, then out, then… with a period of maybe two seconds. “Weak batteries” I theorized, and replaced them. Nope. Poor contacts between the mating halves? My standard fix had (d)evolved into unmating and remating a few times in an effort to clean the contacts, and I soon brought out the big guns: my prized Cramolin Blue & Red contact cleaner/preservative. Nope. I had a spare thermostat, so plugged its front panel into the wall-mounted base. Nope. Finally, I just hung the new thermostat, whole, from the wires. Yup. Not sure I can troubleshoot the dying thermostat to component level without a schematic, but it’s there should I ever feel the need to try. Okay, saying I fixed this problem might be pushing my luck.

Symptom 2: Being a hot water heating system, the burner kicks in and out during operation. But lying in bed at night, I can still hear the soft sound of the water as it’s pumped through the radiators. Then, I started noticing its absence. The burner shut down, but…no woosh of water through the pipes. At first, I imagined a thermal cutout somewhere. Perhaps the pump had overheated. The system generally started working again, so I held to that theory for a while. Then one morning, I awoke to quite a chill. The problem needed to be taken more seriously.

I discovered, in time-honoured troubleshooting tradition, that banging on one of the furnace control boxes would generally make a stalled system run again. Okay. I had my culprit. Or…did I?

Now I started theorizing that the thermostat was perhaps not “driving” the furnace hard enough. Connections, right? I figured it was time to really figure out the heating system design. I understood the basics, but because of the three heating zones, there is some extra hardware in the control path. Long story short, the thermostats control the water valves for each zone. As a valve opens, it also actuates a microswitch, which in turn signals the furnace controller (an aquastat). Since the aquastat only has two inputs, two of the zones must be OR’d. No problem, right? Just parallel the microswitches.

But, those two zones also share one pump, separate from the other pump for the third zone. To drive that shared pump, the OR’d switches wire into auxiliary box #1, which contains a relay, transformer, and lame half-wave power supply to pull in said relay. That relay switches the AC line going to the shared pump. It also drives another relay, in auxiliary box #2, which connects to input #2 of the aquastat. EZPZ to grok once the multitudinous bunches of wire everywhere are unzip-tied and traced. Auxiliary box #1 was the one susceptible to knocks, so I resolved to pull it off and take a closer look.

On the bench, wary of the exposed AC line everywhere, I measured transformer and relay voltages. Everything looked fine. I even desoldered that damn rectifier to check it. All looked good. A weak, intermittent relay was becoming my prime suspect. I connected a DMM in beepy mode, thinking a few sharp taps would reveal the relay’s failings, and plugged the AC back in. Nothing. The circuit was dead. Progress? I began probing voltages again, and as soon as I touched one of the transformer secondary’s pins, the relay clicked in! It took a few more tries for me to finally concede that the transformer was the problem child. But how?

Note to self: Next time, perform a better visual inspection. Pull up the magnifier.

A close look at the secondary pins revealed a grainy look to the solder joints. EDN Labs’  fancy new USB microscope told the whole story:

All four secondary joints are wrecked. As the transformer uses two screws to fasten it to the PCB, I must surmise that it was soldered BEFORE those screws were tightened. Oops.

Two resoldered joints and one new thermostat later, all is warmth again. I love troubleshooting. Sometimes. When it works.

Here are Auxiliary boxes #1 & #2:


Related articles:


Michael Dunn is Editor in Chief at EDN with several decades of electronic design experience in various areas.

Follow Michael Dunn on and EDN on Instagram

Leave a comment