A partial list, compiled over several years and at some personal expense, of the things the carburetor turned out not to be.
The carburetor is innocent. We want to say that up front. The carburetor is a precise and elegant piece of engineering that has been blamed for more problems it did not cause than any other component on a motorcycle, with the possible exception of the battery, which we will get to. The carburetor gets cleaned, gets rebuilt, gets replaced, gets synchronized, gets blamed in listings and in forum threads and in conversations in driveways, and in a meaningful fraction of these cases it was not the problem to begin with and cleaning it did not fix anything and now you have a clean carburetor and the same problem you started with and a workbench that smells like carb cleaner.
This column exists because we have been there. More than once. The third time it happened — clean carb, bike still wouldn't pull cleanly above four thousand, two weekends gone — the senior editor sat down and wrote a list. Not an angry list. A methodical one. Everything that produces symptoms that feel like carburetor problems and is not the carburetor. It got long. This column is that list, organized and explained, offered here in the hope that it saves someone a Sunday.
We are not saying don't clean the carb. Clean the carb. But clean it last, or at least not first. Check these things first.
Valve clearances. This is the one that looks most like a carb problem and least like a valve problem. A tight exhaust valve will hold the valve open slightly past when it should close, letting hot combustion gases back past the seat, which erodes it, which makes it tighter, which makes everything worse on a timeline you don't notice until it's well advanced. What you notice instead is that the bike runs rough at idle, stumbles off the bottom, and clears up at higher rpm where the valve timing matters less. You clean the carb. Nothing changes. You clean the carb again. Nothing changes. You check the valve clearances, probably six months later than you should have, and one of the exhausts is tight enough that you can't get a feeler gauge under the follower at all. This is not a carb problem. It never was.
Ignition timing. Timing that has drifted — and on a points-equipped bike, it drifts, continuously, as the rubbing block on the points wears down and the gap closes — produces symptoms across the entire rev range that feel exactly like a fueling issue. Too retarded and the bike is lazy off idle, runs hot, and has no enthusiasm for anything. Too advanced and it pings under load, pulls hard but briefly, and makes a sound under hard acceleration that you do not want to hear regularly. Either one will send you to the carb first. Neither one will be fixed by the carb.
Plug condition. Not just whether the plug fires — a fouled plug fires badly and intermittently and at different points in the compression stroke than it should, which produces a misfire that varies with rpm and load and feels, from the seat, like a lean bog or a dirty pilot circuit. Pull the plugs. Look at them. The color and deposit pattern tell you things. A plug that's been in there since the previous owner is not a data point, it's a liability. New plugs are cheap. They are the correct first step and almost nobody takes them first.
Plug caps and HT leads. The resistance in an old plug cap climbs with age. At low rpm, the coil has enough dwell time to overcome a high-resistance cap and fire the plug adequately. At higher rpm, where dwell time is shorter, the resistance wins and the spark weakens or misfires entirely. The result is a bike that idles fine and falls apart above five thousand. This is mistaken for a main jet issue with reliable frequency. A new set of plug caps costs almost nothing and takes ten minutes. It is also one of those fixes that feels too simple to be the answer right up until it is the answer.
The petcock. A petcock that is partially blocked — varnish in the screen, deteriorated diaphragm in a vacuum-operated unit — restricts fuel flow in a way that only shows up under demand. At idle and low throttle, the float bowl keeps up. Push it hard, ask it for full throttle for more than a few seconds, and the bowl empties faster than the petcock can refill it. The bike leans out, stumbles, may cut out entirely, and recovers when you back off and give it a moment. This is a textbook petcock symptom. It is routinely treated as a main jet symptom. The jets are fine. The jets were always fine.
The tank. Rust in the tank finds the petcock screen reliably, in pieces, at the worst moment. A tank that looks clean at the filler may have a floor that doesn't, and disturbing it — a long ride, a stretch of rough road — floats debris directly into the fuel supply. The symptom is intermittent, maddening, and completely unresponsive to carb cleaning because the carb is not where the problem lives. Drain the tank. Put a light in it. Look at the bottom, not just the top.
The fuel cap vent. A blocked vent creates a vacuum in the tank as fuel is consumed. The fuel stops flowing. The bike runs fine for twenty minutes and then dies quietly, as if someone turned a key. It restarts immediately after you loosen the cap. It runs fine for another twenty minutes. This one has a way of taking a long time to diagnose because the symptom is so clean — it doesn't stumble, doesn't run rough, just stops — that it doesn't feel like a fuel starvation problem. It feels like an electrical problem. It is neither. It is a blocked cap vent, which takes thirty seconds to clear, once you know that's what you're looking for.
Valve clearances: Tight exhausts in particular. If you don't know when they were last done, they're due.
Plug condition: Pull them, read them, replace them if there's any doubt. There should almost always be some doubt.
Plug caps: Measure the resistance. More than 5k ohms on a standard cap is worth questioning. More than 10k is worth replacing.
Points gap and ignition timing: If it has points. If you haven't checked recently. If you bought it used and the previous owner's maintenance history is "he said it ran great."
Petcock flow: Remove the fuel line from the carb inlet, put it in a container, open the petcock. The flow should be immediate and generous. If it dribbles, start there.
Tank condition: Light and mirror. Look at the bottom. This takes five minutes and has saved many hours.
The grounds: Always the grounds. Every time. We will say it again at the end of this column and we will keep saying it until it saves someone a weekend.
The grounds deserve their own section, their own column, possibly their own publication. We have given them a section here because this is where we are and the subject cannot wait.
A motorcycle's electrical system is a loop. Current goes out from the battery through the components and returns through the ground path to complete the circuit. On a new bike with clean connections this works as intended. On a forty-year-old bike with original ground straps and frame paint that has crept under the terminal over the course of four decades, the ground path has resistance it isn't supposed to have, and that resistance introduces problems across every system simultaneously in ways that are genuinely difficult to diagnose because nothing fails cleanly. The ignition runs weak. The charging system underperforms. The lights are dim. The bike starts hard when it's warm. The idle hunts. These symptoms point in six different directions and none of them point at the ground strap behind the battery that you could improve with a piece of wire and a wire brush in about fifteen minutes.
The procedure is simple enough that it feels too simple: find every ground connection on the bike, remove it, clean the contact surfaces down to bare metal, reassemble with fresh hardware if the original is corroded, and do it again in a few years. That is the entire procedure. It does not require diagnosis. It does not require a multimeter, though a multimeter helps confirm what you've done. It requires wire brushes and time and the willingness to do a thing that produces no visible change and delivers its benefits quietly, over the following months, in the form of problems that don't happen.
We have seen a grounds refresh fix a hard-starting problem that had been chased through two carb rebuilds, a coil replacement, and a new battery. We have seen it fix an intermittent misfire at high rpm that an experienced mechanic had diagnosed as a cracked carburetor slide. We have seen it fix a charging problem that two separate people had blamed on the regulator-rectifier, one of whom had already replaced it. The grounds were the problem all three times. The grounds are the problem more often than anything except the thing you haven't checked yet.
"The grounds were the problem all three times. The grounds are the problem more often than anything except the thing you haven't checked yet."
— the editors, on the subject they will not let go
After all of this — after the valves, the timing, the plugs, the caps, the petcock, the tank, the grounds, all of it checked and addressed and confirmed — if the bike still isn't right, then you look at the carburetor. And at that point you will look at it differently than you would have at the start. You will know what it isn't. You will have ruled out the things that feel like carb problems and aren't. What remains, by elimination, is more likely to actually live in the carb, which means you are more likely to find it and more likely to fix it. The carb clean you do at the end of a proper diagnosis is a different procedure than the carb clean you do first out of habit. It has a higher success rate. It almost always finds something, because by the time you're actually in the carb for a real reason, there usually is something there.
This column was written, as we said, after the third time. There has not been a fourth time. Not because we got lucky, but because the list exists now and we use it, in order, before we touch the carb. It takes longer up front. It takes less time overall. It produces fewer Sundays standing in a garage holding a clean carburetor and a problem that hasn't moved.
The carb is innocent. Start somewhere else.