The Movement Room

Rep power reserve went from 40 hours to 68. What actually changed inside the Dandong movement?

By swissreptime · current as of September 2026

Somewhere around 2020 the spec line in every Datejust listing quietly changed. Power reserve stopped reading 40-ish hours and started reading 70, sometimes 72, and nobody explained why.

The number is mostly real. Owners who have actually timed it get 67 and 68 hours out of a fully wound VS3235, and one 2.5-year-old example that has never been serviced still returns 62. But two things about that number are worth knowing before you pay for it. The gain came almost entirely out of one component, and it is not the component the marketing credits. And on a watch you wear to work and take off at night, you will never once see it.

What Rolex did first, because the clone is copying homework

Calibre 3135 and its 3136 variant ran roughly 48 hours, and did so for about thirty years. The 3235 arrived in late 2015 in the Pearlmaster 39 and reached the Datejust 41 reference 126334 in 2016, carrying about 70 hours. Rolex redrew around 90 percent of the parts and filed fourteen patents doing it. The 3285 in the GMT-Master II sits at the same 70; the Daytona’s 4131 at 72.

Two of those changes produced the extra 22 hours, and they are completely different kinds of engineering.

The first is the barrel, and it is almost crude in its simplicity. Rolex kept the barrel’s external diameter the same and halved the thickness of its walls. Thinner walls, same outside, means more room inside, which means a longer mainspring, which means more stored turns. That is it. No new physics — just machining a thinner part and accepting the consequences. The consequence, as any watchmaker who has opened one will tell you, is that the barrel became disposable: the walls are too thin to survive a mainspring change, so the barrel is replaced whole at service rather than opened and re-sprung, the way a 3135 barrel is.

The second is the Chronergy escapement, and it is the opposite kind of problem. The escape wheel is skeletonised, which drops its rotational inertia. The pallet stones are around half as thick, with the contact surfaces on the escape wheel doubled to compensate. The geometry is offset to increase the lever effect. And the parts are formed in nickel-phosphorus by LIGA rather than cut from steel, which also makes them paramagnetic. Rolex claims this lifts escapement efficiency by 15 percent. A watchmaker who services both generations describes that specific figure as partially unverified, which is a fair place to leave it.

Hold those two apart, because the clone treats them completely differently.

The clone copied the barrel and skipped the escapement

Here is the finding that reframes the whole spec line: no rep-side 3235 has a Chronergy escape wheel. Not the Dandong, not the Jinghe, not the Shanghai. A teardown comparing the Dandong and Jinghe 3235 side by side found both run a pallet fork with noticeably larger stones than the genuine part — and the reason is precisely that neither has the skeletonised escape wheel behind it. The fork is sized for a conventional wheel because a conventional wheel is what is in there. It is also why dropping a genuine escape wheel into a clone is not a parts swap: you would need the genuine pallet fork with it.

So the clone factories took the half of Rolex’s work that is a machining problem and left the half that is a materials-process problem. That is not laziness, it is economics. A thinner barrel wall is a lathe setting. A LIGA-formed nickel-phosphorus escape wheel is a production line you do not have.

And the striking part is how little it cost them. A fully wound VS3235 in a VSF Submariner, timed by its owner from a full hand wind on a Wednesday night to a dead stop on Saturday evening, ran 68 hours and 8 minutes. A second owner ran the same test on a 126610LN, wound about 40 turns, and got 67.3 hours against an advertised 72. The clone lands two to four hours short of a genuine 3235 while missing the entire efficiency half of the design.

Which tells you something about where the 22 hours really came from. If the escapement were doing the heavy lifting, a clone without it would fall far further behind than three hours. The reserve came out of the barrel.

What the number tells you about which movement you actually have

A movement’s reserve is a fingerprint of its architecture. Which means that if you wind your watch fully, leave it alone and time it to a stop, you learn what is inside the case without opening it.

An owner with a Datejust badged VR3235 ran exactly that test and measured around 42 hours. He wondered whether something was wrong. Nothing was wrong. 42 is a 3135’s number, and the VR3235 is a 3135-derived platform reshaped to fit a 3235-era case — it carries the older mainplate and gear train and the older split calendar work, under a 3235 name. It was never going to hit 70, because it does not have the barrel that produces 70. The reserve reported exactly what the calibre is.

That makes power reserve the cheapest movement-identification tool you own. Under 45 hours on something sold as a 3235 means you are holding 3135 architecture. Over 60 means the barrel is the new one. No caseback opener, no timegrapher, no QC photo argument.

Why your watch never sees 68 hours

This is where the spec line and daily wear part company, and it catches almost everyone.

The same owner who measured 68 hours from a full hand wind reported that in normal use, relying only on wrist motion, he gets around 30 hours at best. Another VS3235 owner wearing his about eight hours a day, every day, sits at 35 to 40 and never reaches anywhere near the advertised figure. Both assumed something was wrong. Neither has a problem.

The mechanism is simple once you see it. Crown winding drives the barrel almost directly. Rotor winding goes through the reduction gearing of the automatic works, so a given amount of wrist motion puts far less into the mainspring than the same effort at the crown — and a desk day barely moves the rotor at all. Meanwhile the mainspring spends at a constant rate whether you are moving or not. Wear a watch eight hours and rest it sixteen and it settles at part wind, permanently, which is why it dies on Saturday afternoon after you took it off on Friday night.

The clone carries a small extra handicap. The genuine 3235’s ball-bearing rotor race is fully packed; the VS3235’s runs seven balls — less efficient winding, on a movement already fighting the reduction gearing.

So the 70-hour number on the listing is a full-crown-wind laboratory figure. Honest, measurable, and a state your watch is almost never in.

The part that actually matters: the tail of the reserve is the worst of it

A mainspring does not deliver constant torque. It delivers most near full wind and least near the end, and balance amplitude follows torque down. Rate follows amplitude. So the last stretch of any reserve is the least accurate stretch — on every mechanical watch, genuine included.

The gap is that Rolex spent money specifically on flattening that curve. The genuine hairspring is a Parachrom with a true Breguet overcoil, and an overcoil breathes concentrically, which keeps the balance’s rate more consistent as torque falls. Index curb pins work against isochronism; a free-sprung balance has none. The rep side has neither advantage in full. The JH3235’s hairspring is flat rather than overcoiled. And while VSF did start shipping free-sprung VS3235 builds during 2026 — the titanium Yacht-Master 226627 in June, the Datejust 41 126334 in July — that version has no regulating weights on the balance rim at all, so it is free-sprung in architecture rather than a Microstella copy; the JH3235 out of Jinghe does carry the weights, and earlier regulated stock is still circulating under the identical name.

Put it together and the practical shape is this: a clone 3235’s rate-versus-wind curve is steeper than a genuine one’s. The watch is at its best in the first day or so off a full wind and drifts further as it runs down. The Monday-morning reading off a Friday-night wind is the worst number that watch produces all week — and it is the number most owners judge their movement by.

Power reserve is also a free service gauge

One more use, and it is the one nobody thinks of. Measure the reserve when the watch is new, write it down, and measure it again once a year.

The 2.5-year-old unserviced VS3235 mentioned earlier returns 62 hours where fresh examples return 67 to 68. That is what thickening lubricant and settling tolerances look like long before the watch starts visibly running badly — a reserve that has dropped materially is telling you the movement is due while a service is still all it needs.

Know the difference between a short reserve and a broken one, though. A clone 3135 that dies in under ten hours is not underperforming, it is faulty, and the thing to check is amplitude: a movement with collapsed amplitude runs to a stop early no matter how much you wind it. Regulating the rate on a watch like that only hides it. A DD3285 owner who chased a −14 seconds a day problem with a regulator adjustment was told, correctly, that a rate moving that far on its own needs a service, not a nudge.

Bottom line

Wind it fully by crown once a week. Thirty seconds, and it moves your watch from living permanently at half wind to spending most of its life in the strong half of the mainspring’s range, which is where it keeps its best time.

Time the reserve once, properly, when the watch is new — full wind, dial up, untouched, note the stop. One afternoon buys you three things: proof of which architecture is actually in your case, a baseline for spotting when a service is due, and a realistic sense of what “72 hours” means for you, which is roughly half that in ordinary wear.

And read a reserve figure on a listing as what it is: a measurement of the barrel, not a measurement of the movement. The clone factories bought most of Rolex’s 22 extra hours with a thinner barrel wall and skipped the escapement entirely. It worked. It just does not mean the rest of the movement came along with it.

Current as of September 2026. The barrel architecture is settled; the escapement is the thing to watch — if a factory ever ships a skeletonised Chronergy-pattern escape wheel, this page changes.

Read next: VS3235 / VS3285 explained · every movement base by factory · which gen swaps are worth it · the full index