Paint is the last job on the list and the first thing anybody notices. You can spend two winters getting the carbs right, and nobody will ever know. Chip the cases putting the engine back in the frame and you will look at that chip every single time you walk into the garage for the next ten years.
It's worse on a cafe racer. Strip a CB550 down the way this series keeps recommending — fender off, signals off, airbox gone, seat cut short — and you've removed most of the things that used to hide the motor. What's left is a tank, a pair of side covers if you kept them, and a large black-and-aluminium engine hanging in plain view. The finish is the bike.
Three threads on the SOHC/4 Owners Club forums cover this better than anything else we've found, and this series hasn't touched any of them. One is a short, funny, surprisingly technical argument about putting motorcycle parts in your kitchen oven that has been read seventy-six thousand times. One is a four-year-long forensic investigation into why VHT engine paint chips off some people's cases and not others, which ends with a professional powder coater telling everybody they've been asking for the wrong product. And one is a member teaching himself candy and metalflake from scratch with a Harbor Freight spray gun and craft-store glitter, in public, over two weeks, until he's matching factory Sunrise Orange.
Read together they answer the question every 550 builder gets to eventually: how do I get a finish that survives being ridden, without sending the whole bike out to a shop?
The oven thread: heat is just time you didn't have to wait for
The oven thread opens in December 2009 with a member who has primed a battery box and a tool tray and half remembers reading something about baking parts — forty-five minutes at about 100 degrees, was it? He also mentions, almost in passing, the reason he's asking: last time he painted the tool tray, the paint chipped when he put the bolts back in.
Both halves of that post get answered, and it's the second half that turns out to matter.
On the temperature: 100°F does essentially nothing. The best correction in the thread came from a member in Texas, who pointed out that every painted object in his state spends the summer curing at 100 degrees or more for days on end and still takes two or three months to get properly hard. Heat isn't magic. Heat is a substitute for the calendar. If you want the oven to do real work, you have to go meaningfully above ambient.
The numbers people actually use cluster tightly, and most of them land on their oven's lowest setting:
- Around 180–200°F for 20–35 minutes, after the paint has already dried to the touch — several members specified waiting a full 24 hours before the first bake.
- One member curing a rear rim for a CB500 project baked at 200°F for 30–35 minutes at every stage — primer, colour and clear — watching it and cracking the oven door periodically to make sure nothing bubbled.
- Another heat-cycled Duplicolor's 500-degree ceramic engine enamel four times up to 250°F, letting the parts cool completely between cycles. He then dropped the finished transmission cover two feet onto a concrete garage floor and didn't put a mark on it.
Three cautions came with it, and they're the useful part.
First, bake the right paint. The original poster was planning to use ordinary oil-based Rust-Oleum, and more than one member warned that it could go soft in the oven rather than hard. Engine enamels — the Duplicolor ceramic 500°F, the VHT products — are formulated to be heat cured, and their instructions say so. Ordinary hardware-store enamel isn't, and the advice was to test on scrap before you commit a part you care about.
Second, think about whether you want to bake between coats. One member who paints for a living argued against it unless you're going to scuff the base afterwards: leaving some solvent in the previous coat is what lets the next one bond into it chemically. Fully cure each layer and you've turned a chemical bond into a mechanical one, and now you need tooth.
Which brings us to the third and most important point, and the actual answer to that tool tray. The thread turns almost immediately from curing to prep, because the original chipping problem was never about hardness. Sandblast it. Give it tooth. If you don't own a blast cabinet, sand to around 400 grit — a wire wheel can leave a surface that's clean but too polished for anything to grip. As one member put it, adhesion will be poor no matter how hard your paint gets. That sentence is the whole thread in nine words, and it's about to become the whole next thread as well.
Two practical footnotes. Baking is the only realistic way to dry parts in a cold garage in January, which is when most of us are doing this work. And on the domestic question — several members were emphatic about not using the oven you cook in. The smell is serious, ventilation is not optional, and when someone asked VHT directly about VOCs from home-oven baking, the company's answer was that they don't recommend it. Three years after one member joked about the lump on his head from baking a chain guard, another member noticed he hadn't posted since, and drew the obvious conclusion. A cheap second-hand toaster oven or an old range in the garage solves the entire problem.
The VHT thread: why the same can works for one man and not the next
The second thread is the one to read if you're about to paint your engine, and it opens with a member at the absolute end of his patience. He has painted his cases with VHT engine paint two or three times. He glass-bead blasted them, soaked them in solvent, scrubbed them with soap and water, wiped them with acetone and tack-clothed them. He tried it with VHT's own primer and without. Every single time, it chips — from tightening a bolt, from the cam chain brushing the case while he rotated the engine on the stand. He called VHT, who sent him free cans. It chipped again.
What makes the thread valuable is that the replies split almost exactly down the middle. Half the board reports the same disaster; the other half has used the same product for years across bikes, karts and race cars with no trouble at all. One long-time member painted his 550 cases and head from the same can, prepped identically, and got zero chips on the cases and a couple on the head.
So the thread stops arguing about the brand and starts working the problem. These are the theories that survived four years of testing, roughly in order of how much evidence backed them.
Aluminium is genuinely hard to paint. It oxidises the instant it's exposed, and oxides are about the most chemically stable thing there is. There's very little for paint to bond to chemically, which throws the whole load onto mechanical adhesion — and rattle cans have no activator, so you don't get the toughness of a catalysed system to compensate.
Polishing is counterproductive. The single best piece of forensic work in the thread came from a member who sanded a set of parts up through 400 grit and then polished them before painting. He shot Duplicolor primer, colour and clear on a front hub, a rear hub, and several covers, all at the same time, in the same shop, at the same temperature. He laced the rear wheel without a single chip. On the front wheel, the paint chipped everywhere a spoke touched it. Picking the hub up days later he noticed the detail that mattered: every chip went straight to bare aluminium. The colour hadn't failed — the primer had never stuck to the metal in the first place. And the reason was that he'd taken the tooth off.
Solvent trapped in the metal. One member's suggestion was to blast the part, bake it bare at the recommended temperature to drive off any solvent sitting in the pores, then paint with clean gloves and bake again on schedule. It's an unusual step and nobody ran a controlled test, but it's consistent with the number of failures in the thread that followed a heavy solvent wash.
The wrong stripper. If you use chemical stripper on aluminium, use aircraft-grade or aluminium-specific stripper. The general-purpose stuff leaves a heavy oxidation layer behind, and you've now made the adhesion problem worse than when you started.
Too much paint. Repeated by several members with good results, including a couple of the board's most experienced hands: thin mist coats, just enough to get colour, and stop. Thick paint chips easily; thin paint doesn't. One member goes so far as to use no primer at all under VHT.
Etch primer. The most promising product answer in the thread. A member who used Duplicolor's self-etching primer on an aluminium top triple clamp reported that even brake cleaner wouldn't lift it when he wanted to clean the fork-tube bores. Its one drawback is that it's a greenish-straw colour, so it has to live under something. Another member offered the aviation industry's version — Alodine 1201, a conversion coating that turns aluminium gold and, importantly for a cylinder head, doesn't act as a thermal insulator.
Time. A member who'd been finishing furniture professionally for fifteen years gave the least exciting and most useful advice in the thread: prep twice as much as you think you need to, always spray multiple thin coats, and always protect your colour with a clear topcoat. Another, who spent three weeks cleaning and degreasing an engine before he sprayed it and then left it a full week between primer and colour, has had his Duplicolor cast-iron finish survive both a drop and being hit by a car. His warning is the one most of us need to hear: rattle-can paint keeps hardening for days and weeks after it looks dry. Don't jump in and test the finish.
The thread also produced a real cost-benefit reckoning. VHT's Flame Proof has a published cure schedule of three cycles at 250, 400 and finally 600 degrees. Most home ovens stop at 550, and no reasonable person wants to put a set of crankcases through 600°F anyway. A member who tried to do it properly and gave up partway reported that even his small parts, prepped and painted and cured together, came out inconsistent — one would chip and its neighbour wouldn't. When he emailed VHT about the missing 50 degrees, the answer was that unless the directions are followed completely, they can't promise anything. His verdict, which is hard to argue with: for the time and money involved, the juice wasn't worth the squeeze.
Which is where powder coating enters — and where the thread's single most counterintuitive piece of advice appears, from a member who coats parts for a living.
Do not ask for high-temperature powder on your engine cases. He'd stopped offering it: it's finicky, it can flake off in sheets the first time the part meets a pressure washer, and he was redoing enough jobs for free that he pulled it from his catalogue. High-temp powder exists for parts that genuinely see 1200°F. Standard powder coat handles spikes to around 500°F, which is far beyond anything your crankcases will ever see — and if your cases are near 500 degrees, flaking paint is not your problem. If you truly need heat resistance, ceramic coating is the right answer, not high-temp powder. He also noted that Harley-Davidson and S&S now powder coat their engines rather than paint them.
And for the historically curious: one of the board's resident Honda men explained why factory engine paint from the 1970s looks better after fifty years than most of our refinishes look after two. It was lacquer. It isn't available any more, and it isn't coming back.
The practical shortlist that fell out of four years of this, for anyone who wants to skip the argument: bead or soda blast, don't polish, use aircraft stripper if you use stripper at all, consider a self-etching primer or skip primer entirely, spray thin, bake at 200–250°F in cycles with cooldowns, clear coat it, and then leave it alone for a month. Products members reported real long-term success with included Duplicolor's ceramic engine enamel, POR-15's engine enamel (which is also brake-fluid resistant, so it does calipers and master cylinders too), Eastwood's engine paint with a urethane activator through a gun, Rust-Oleum engine enamel, and — from a member two years in with no chips — plain Krylon Dull Aluminum, on the theory that the aluminium content carries the heat well enough for a CB.
One last thing the thread agrees on: accept that the paint under a bolt head is going to mark when you torque it. Everybody's does. That's not a failure, that's assembly.
The orange flake thread: a $40 tank that reads like a $1,000 tank
The third thread is the most fun of the three, and the most useful if you want to paint your own bodywork. It's July 2012, and a member decides to teach himself candy and metalflake from zero. His equipment budget is close to nothing: a Harbor Freight detail gun, a bottle of House of Kolor Tangerine concentrate, some cheap clear, craft glitter from the hobby aisle, and a home-made "flake buster" built from a vitamin bottle, a T-fitting and a rubber grommet.
He also worked out the cheapest test-panel supply chain in the hobby: sixteen-cent white ceramic tiles from the hardware store, a scrap sheet of plywood, the back of his shed, and — for practising candy, where you need to see through the coats — an old piece of window glass. A set of table legs got sacrificed to metallic silver along the way.
The first attempt was, by his own description, an orange disco ball. Adding a pressure regulator to the flake buster changed everything: instead of full glitter coverage, he could lay down a hint. But the real breakthrough was archaeological. He found a scratch on an original tank, looked closely, and realised that factory Sunrise Orange is not a heavy flake job at all. It's a shimmering silver metallic base with very fine flake — on the order of .002 to .004 — under candy. The reason it looks so alive in the sun isn't quantity of flake; it's the candy over the top doing the work.
That reframing is the thread's biggest single takeaway for anyone chasing a period-correct 500 or 550 tank, and it makes the job easier and cheaper than most people assume.
The rest of it is technique, and it came mostly from two experienced painters who joined in. The parts worth writing on the wall of your garage:
Flake goes under the candy, not over it. Candy is a topcoat and it's what makes flake glow. Shoot flake on top and the flakes look like leaves floating on a pool — and at a distance the part just reads as dirty. Coloured flake over a candy of a different family tints the base colour, and the shift gets stronger the brighter the light. Pearls are the exception; they're soft enough in reflection to work over candy.
Craft glitter is a real option with a real limit. The dry-flake method with craft glitter goes back to the 1970s. But the flakes are large and, used sparingly, sit too far apart, which reads as a kid's craft project up close. If you use it, commit — the pros want minimal space between flakes, and whatever space there is should be the same colour as the flake.
Run each coat 90 degrees to the one before. This is the fix for tiger striping, the banded look you get from overlapping passes. It applies to candy, wet flake and metallic alike.
You cannot spot-repair candy. Candy depth comes from the number of coats, so a touch-up is always a darker patch. Get it right or start over.
Distance is about transfer, not inches. The original poster spent days chasing orange peel before working out that he was holding the gun too far back and the paint was drying in the air. Close and square, watching the paint actually make contact with the surface, gave a glass-smooth chain guard. Six to eight inches away on the same part left visible pitting from dry spray.
Don't follow the tank around a corner. This one cost him two tanks to runs. Paint the side square and parallel, stop at the front of the tank, physically move yourself, and then paint the front. The moment you swing the gun round the curve, your distance changes and you get a run. Block the tank up off the bench so you can get the underside too, and have a route — bottom, sides, front, top — and stick to it.
Metallic base and candy flow like thin milk. Keep the gun moving. Half a second stationary is a run.
Intercoat clear has a ceiling. Around four coats before it starts running. Let the flake and candy coats flash off between passes.
Put a marble in a siphon gun. It keeps the flake agitated and stops it clumping and binding the gun — a trick from a painter who'd been shooting flake since the drag racing days. A 1.4 tip will pass 400-micron flake; the flake makers themselves recommend 1.3 to 1.4 for the small "baby" sizes.
Mixing candy by eye. Find a shiny crescent wrench that roughly matches your metallic base, put a couple of drops of tinted clear on it, and look. Add concentrate until you like it. It will never go darker than that drop. Concentrate rather than pre-mixed candy is what gives you that control — a quarter-ounce of Tangerine was enough to do an entire tank.
Prep and post are where the money is. Wet sand your primer to 600 before candy — it needs a genuinely smooth surface. Keep a separate gun for primer. Use quality finishing filler for small dings rather than generic body filler. And then cut and buff: a 1000-grit wet sand and a machine buff and glaze after the final coat is what makes a cheap paint job read as an expensive one. One member on the thread had already proved it from the other direction — after paying $400 for a bad paint job on a tank, he did it himself with about $40 of Duplicolor "Metal Specks" over a light base and a couple of coats of 2K clear, and the only criticism anyone could make was that he hadn't cut it afterwards.
The line that sums up the whole thread came from one of the pros: there are people getting show-bike finishes out of rattle cans and yacht enamel, because it isn't really about the paint. It's about the surface you're spraying onto and how smooth you get it afterwards. The best painters he knew sanded between nearly every coat.
And wear a respirator. More than one member watched the instructional videos and winced at the bare skin and the missing mask. Modern 2K clears contain isocyanates, and they are not something to be casual about in a home garage.
The through-line
Put the three threads side by side and the same finding turns up in all of them, from completely different directions.
Almost nothing in these threads is actually a paint failure. The tool tray that chipped in 2009 chipped because a wire wheel left it too smooth. The engine cases that chipped in 2011 chipped because the primer never bonded to polished aluminium. The tanks that came out orange-peeled in 2012 did so because the gun was too far away, and the ones that ran did so because the painter swung the gun around a curve. In every case the can did what it said, and the surface underneath it didn't.
The second finding is that time is an ingredient, and it's the one nobody wants to buy. Rattle-can paint keeps hardening for weeks. Baking is not a shortcut past that, it's a way of buying some of it back — which is why the members with the best results bake in cycles with cooldowns rather than blasting one long cure. The three weeks somebody spent degreasing a motor and the extra week between primer and colour are the reason his paint survived being hit by a car.
The third is a matter of scope. If you want the engine on your cafe racer to look right in five years rather than five months, the forum's honest answer is standard powder coat — not high-temp powder, and not a rattle can. If you want the tank to look right, the forum's honest answer is that you can absolutely do it yourself in a garage for the price of a decent dinner, and that the difference between your result and a professional's is mostly the hours you're willing to spend sanding between coats.
It's the same shape as every other lesson in this series. The unglamorous half of the job is the half that decides the outcome.
Credit where it's due
None of this is ours. It belongs to the members of the SOHC/4 Owners Club community, who spent years baking parts in ovens they weren't supposed to use, ruining hubs, chipping cases, arguing with paint manufacturers, and posting the results in public so nobody else has to repeat the experiment. Those threads are worth reading in full, and the forum is worth supporting — there is no substitute for an archive that deep.
And when the paint has cured and you're ready to bolt the bike back together, that's where we come in. The Thirteen Motorcycles CB550 catalog is built for exactly this stage of the project — parts made for this bike and made to work together, so the finish you spent all winter on gets to sit on something worth looking at.
0 comments