Do clutch rear derailleurs add drivetrain friction?

Clutched rear derailleurs have long been a fixture in the mountain bike world, given how the one-way pulley cage pivots dramatically enhance chain security and reduce noise. The road world has largely ignored the technology on two counts, though: weight and friction. The added mass relative to a non-clutched rear derailleur is indisputable, but does a clutch really add drivetrain friction, too? We enlisted the help of Jason Smith of CeramicSpeed to settle that debate once and for all — and the answer is a bit more complicated than originally expected.

What’s a clutch, and why would I want one, anyway?

Most mountain bikers won’t need any explanation on this subject, but for everyone else, here’s a quick primer on clutched rear derailleurs. Standard rear derailleurs have a spring-loaded pulley cage that is allowed to swing freely both forwards and backwards. But in a clutched rear derailleur, there’s a small device in the pulley cage pivot that lets it pull the chain rearward as usual, but doesn’t allow the cage to freely swing forward.

Under normal riding conditions, this means the chain can’t bounce around excessively on bumpy terrain, which not only reduces noise but also helps keep the chain securely engaged on the chainrings and cassette sprockets. However, a sufficient amount of force — from a shift lever or suspension movement, for example — can overcome that clutch and allow the cage to swing forward when necessary so the rest of the bike still functions as usual.

Currently, Shimano and SRAM both offer clutched rear derailleurs across nearly the entirety of their mountain bike drivetrains, and both companies have a few clutched road rear derailleurs as well. Only Shimano offers clutched road rear derailleurs for two-chainring transmissions, though — the Ultegra RD-RX800 for mechanical shifters, and the the Ultegra RD-RX805 for Di2 electronic systems — while SRAM’s clutched models are limited to single-ring drivetrains. Neither SRAM nor Campagnolo have a clutched rear derailleur for two-chainring systems, although the former is rumored to be adding one to its upcoming Red eTap flagship groupset for 2019.

Not surprisingly, clutch-equipped rear derailleurs add some weight relative to non-clutched rear derailleurs, to the tune of around 40g. That part is easy for anyone to verify at home, but there’s also a commonly held belief that clutched pulley cages also add drivetrain friction.

And here is where the folks at CeramicSpeed come in.

Friction foundations

There are several main sources of drivetrain friction: the bearings inside the bottom bracket, rear hub, and derailleur pulleys; the side plates of the chain as they slide and rotate against each other; and the chain as they rub on the cassette sprockets and chainrings.

Of those, though, the chain is the biggest contributor, and that friction can be further separated into four distinct sections: the upper span between the cassette and chainring; the lower span between the chainring and lowermost rear derailleur pulley; the short span between the lower and upper rear derailleur pulleys; and the bit between the upper rear derailleur pulley and the cassette.

The lowermost most that we’re interested in here. The friction there is dependent on how hard the pulley cage spring is tensioning that section of chain, and it’s that span where a clutch would most influence the drivetrain friction. The friction of the upper segment is highly dependent on pedaling input, with higher power inputs generating more friction, and while the tension on the spans of chain running through the rear derailleur is also somewhat influenced by the cage spring force (and possibly the clutch), they don’t vary as much, and they’re too short to make much of a difference.

As with the upper span, more tension yields more friction. But whether a clutch actually adds more tension — or if it just keeps the cage from swinging forward — is the unanswered question here.

The test

Jason Smith is the chief technology officer at CeramicSpeed, but he’s also the founder of independent bicycle drivetrain friction test lab Friction Facts. CeramicSpeed acquired Friction Facts in 2014 — and Smith’s expertise along with it — but, conveniently for us, the Danish firm still provides him with leeway to pursue tests that may not be directly relevant to CeramicSpeed’s business.

In this case, I contacted Smith to help answer the question of whether a clutched rear derailleur really does add friction to a bicycle drivetrain, and he was willing and able to assist.

Based on Smith’s previous (and ongoing) testing, chain-related drivetrain friction is critically tied to chain tension in the four different spans of chain: the upper one between the cassette and chainring; the lower one between the chainring and lower rear derailleur pulley; the one between the two rear derailleur pulleys; and the one between the upper rear derailleur pulley and the cassette.

Of the four, the only one that could potentially be affected by the pulley cage clutch is the one between the lower rear derailleur pulley and the chainring. The upper one varies with rider pedaling input, while the other two remain constant regardless of pedaling force or pulley cage spring tension.

Smith started with his standard “full-load” drivetrain friction setup, and first recorded baseline measurements with a number of different rear derailleur models. Shimano clutch-equipped rear derailleurs were tested with the clutch engaged and disengaged (since Shimano derailleurs incorporate a handy toggle switch), while I custom-modified SRAM clutched rear derailleurs for Smith so he could get an apples-to-apples comparison there as well.

To look more specifically at the effects of the clutch itself, Smith then positioned two different eccentric pulleys along the lower span of the drivetrain, in between the lower rear derailleur pulley and the bottom of the chainring, so as to simulate the effects of riding on bumpy terrain. All of that data was recorded as well, with the same derailleur models.

The results

So what did all of that data tell us?

Surprise, surprise: clutches don’t add any drivetrain friction — at least in the lab, under ideal conditions (more on that in a bit). And in fact, a clutch-equipped rear derailleur can actually produce less friction than a non-clutched rear derailleur, given the proper setup.

According to Smith, one of his earliest questions was whether derailleur clutches were truly one-way, or if they added friction in both directions. This was key to the findings, since he had already established that drivetrain friction on that lower span of the chain was primarily related to how much it was tensioned. If the only thing a clutch did was keep the pulley cage from rotating forward — and didn’t add any spring tension pulling it rearward — then it stands to reason that a clutch wouldn’t add any drag.

Lo and behold, that’s what Smith found. But ultimately, the answer to the question we originally asked wasn’t quite so simple, and Smith discovered several pieces of interesting information along the way.

According to Smith, a clutched rear derailleur only doesn’t add any drivetrain friction if it’s designed properly: i.e. only if it’s built with enough clutch force so that the pulley cage never (or at least very rarely) swings forward under typical riding conditions.

“If the cage of a clutched rear derailleur is allowed to rotate, then the clutch rear derailleur indeed will create increased friction,” he said. “Friction is created by the non-elastic torque to rotate the clutch itself, and by the chain being allowed to oscillate to a greater degree. When the mass of the chain gets moving in the vertical direction, it takes more horizontal force to bring the moving mass back to equilibrium. If the chain is not allowed to start a vertical oscillation because of heavy clutch force, the overall horizontal force (chain tension) is less.

“In this test, the off-center wheel physically push the chain up and down, and therefore always cause the clutch to engage, and show higher friction numbers,” Smith continued. “However, in real-world applications, the chain will only bounce up and down if the potential bouncing forces of the chain are high enough to overcome the clutch force. If the clutch is super strong, it won’t allow the chain to bounce. The clutch won’t engage, and therefore the additional friction caused by clutch engagement doesn’t happen.”

Riders on SRAM clutched rear derailleurs won’t have to worry much about this since the clutch force is pre-set (and non-adjustable) at the factory. The clutch force on Shimano’s rear derailleurs can be tuned, but Smith recommends maintaining the stock setting — or, if anything, increasing it, provided it doesn’t adversely affect shifting or suspension performance. Decreasing the clutch force not only reduces its effectiveness in terms of chain control, but also adds drivetrain friction since the clutch is more likely to be overcome when riding on bumpy terrain.

Smith says that having a properly adjusted clutch on your rear derailleur also has a potential, and unexpected, side benefit. Since the clutch inherently keeps the chain from flopping about, there’s no longer a need for lots of pulley cage spring tension. And since it’s that pulley cage tension that generates drivetrain friction, reducing the spring tension can increase drivetrain efficiency.

“The [pulley cage] spring itself, when at equilibrium, puts tension in the chain, and this, in general, creates friction,” Smith explained. “But for a spring to do what a clutch does, it would have to be a very heavy spring, which adds a ton of ‘all the time’ friction.”

SRAM rear derailleurs don’t allow you to make any adjustments in that area, but Shimano derailleurs do, so by repositioning the cage spring to the lower-tension setting, you can not only get better chain control, but reduced drivetrain friction as well.

“The perfect clutch rear derailleur is one with very light back spring tension,” Smith concluded, “and very heavy forward clutch force.”

The caveat

Yet if Smith’s testing definitively concludes that clutches don’t add drivetrain friction, why do so many riders report otherwise (myself included)? As it turns out, they’re both right.

Smith’s testing confirms that clutches do, indeed, operate in a truly one-way fashion, meaning that the force they generate only acts when the pulley cage is trying to swing forward. However, while the clutch does only work in one direction as intended, the way it works isn’t always purely binary. In other words, it’s not simply a matter of the clutch being engaged or released; some derailleurs have a certain amount of grey area in between, and it’s here where additional drivetrain friction can be created.

To put it in more specific terms, Smith measured each rear derailleur’s stock clutch force to be around 9lb, meaning that anything above that figure would cause the clutch to release, and allow the cage to swing forward. But it’s also possible to load the clutch at a level below that release threshold.

This is exactly what happens when you shift from a smaller rear cassette sprocket to a larger one, and why, for example, the drivetrain friction in a 50x17T combination feels different when you shift into that gear from the 15T sprocket versus when you shift into it from the 19T. When you shift from a larger sprocket to a smaller one, the pulley cage swings rearward to take up the additional chain slack, and the clutch essentially resets.

However, when you shift from a smaller sprocket to a larger one, the cage is forced to swing forward to accommodate the larger diameter, with the force required to overcome the clutch threshold (about 40N, or 9lb), provided by your finger (or derailleur motor). Once the cage swings forward and settles into its new position, the clutch resets, and the process repeats as needed.

Well, that’s how it works in theory, anyway.

SRAM’s Type 3 clutch design does actually work this way, and Smith’s testing showed no increase in drivetrain friction regardless of gear ratio or shifting sequence. As previously mentioned, the clutch truly works in a one-way fashion, as intended, but it also operates in a binary on-off fashion as well.

Shimano’s Shadow Plus mechanism, however, has more of a grey area.

Whereas the SRAM clutch doesn’t move at all when loaded at sub-threshold levels, the Shimano clutch has enough slop in the mechanism that smaller shifts to larger-diameter sprockets simply load up the derailleur structure with extra tension (and friction) instead of releasing the clutch. Our test cassette had two-tooth jumps between most of the gears, and each single downshift revealed drivetrain friction increases of about 3W. Once we hit a total downshift of around six teeth, though, the Shimano cage finally swung forward enough to overcome the factory-set clutch threshold, and then the system reverted to baseline levels.

All is not lost with the Shimano system, though.

Interestingly, Smith noticed that that built-up tension in the Shimano clutch would release over time, returning back to non-clutched friction levels after about a minute when the clutch was set at factory levels. That rate of change was the same when the clutch tension was increased, too, meaning that tighter clutches would still release that tension over time, but would just take longer to do so.

There’s a footnote here, though: Smith’s test setup has some out-of-round runout in both the chainring and cassette mounts, so there’s always a small amount of cage movement when the system is in motion. As a result, it’s unclear whether that slow return to baseline friction levels on the Shimano clutch comes about simply due to chain tension, or from the small cage movements induced by the test apparatus.

That said, one could argue that those small tester-induced cage movements might also simulate riding on real-world road surfaces, so it’s hard to conclude that those movements wouldn’t happen while riding, anyway. But even so, there are some lingering questions here that remain unanswered.

So are clutches good? It depends

The mountain bike community has wholeheartedly embraced clutched rear derailleurs, and for good reason: they dramatically improve chain security, and the associated downsides aren’t significant enough for most riders to trade that security for a slight gain in drivetrain efficiency and suspension sensitivity.

But on the road, the waters are a little muddier. Riders who frequent rougher terrain and value chain security and/or reduced chain slap more than a few watts of drivetrain friction will likely still find that clutched rear derailleurs work better for their riding situations. However, the argument for clutches isn’t nearly as strong for traditional road riders that more regularly find themselves on smoother surfaces. For them, the added friction will likely compound more noticeably over time, and there are fewer concerns with chain security there, anyway.

Nevertheless, clutches will invariably continue to improve, and SRAM’s new Red eTap clutch may change the landscape further still. But as things stand now, a little chain slap every now and then won’t be a deal breaker.