Internal gear hub comparison test(Sram S7 vs Shimano Nexus)
 

Now that I have both systems, I feel it is fair to submit a comparison.

My Nexus 8 (SG-8R20) is the stock hub on my Trek L200, with a roller brake mounted on it and shifted by the combined brake/grip shifter unit (SB-8S20).

I've also ridden a Shimano Alfine bike at my LBS shifted by a bar-end shifter/travel agent combo.

My SRAM S7 is the drum brake unit, run with a triple in the front and tensioned by a short cage Suntour Superbe Pro derailleur.

For comparison, I've ridden Sturmey AW hubs for years.

1. Efficiency

The Sturmey AW is more efficient than all of the above in all gears. The most inefficient gear on the AW (Gear 3) is more efficient than Direct Drive on the Shimano Nexus (Gear 5).

The SRAM S7 is nearly as efficient as the Sturmey AW in gears 3-5, with efficiency in gears 1 and 7 noticeably reduced. With my particular configuration, with the triple in the front, I have a wide band of efficient gears, with emergency bailout gears for climbing and descent gears when necessary.

The main problem with the Nexus 8 is that its least efficient gear is Gear 4 (based on subjective tests, around 80% efficient). The Nexus 8 is a derivative of the Nexus 4, which had 4 gears with 1st gear as direct drive. On top of the relatively efficient Nexus 4, shimano added another epiciclic gear. Thus, gears 1-4 have another 5-7% loss on top of the top gear losses. Gear efficiency, from lowest to highest seems to be: 4, 8, 3, 7, 2, 6, 1, 5. The main advantage to this design is that the primary climing gears are more efficient than those on the Sram S7. The main disadvantage is that some of the most used gears (3,4) are rather inefficient. Most of the time, I even prefer to avoid gear 4 because it is so draggy.

For comparision, some of these losses have been mitigated with the Alfine hub. Needle bearings, shared with Nexus Premium, reduced losses in all gears. The roller clutch in the Alfine further improves efficiency. Nevertheless, efficiency is still lower than Sram S7 in gears 3-4, and slightly lower overall.

2. Cable connection

All 3 speed designs have much simpler cable/shifter systems, so I'll leave them out here.

The Shimano Nexus cable is easier to setup on a new bike. Cut cable to length, install nut, adjust. The installation of the cam nut into the shift mechanism inboard of the chainstay is a bit of a pain without tools.

The Sram S7 Shifter/Clickbox combo comes pre-cabled and set up with an initial length of 1.65m. Once set up, the Clickbox and its quick release is a joy to use compared to the Nexus cable nut. However, the clickbox sits exposed outside the chainstay, making it vulnerable to impact. This is slightly mitigated by the included bumpguard for the clickbox. Cutting the shift cable to length is much more of a pain with the Clickbox than with the Nexus cable.

3. Shift under load

No question here. Nexus is significantly better than the others at shifting under load, with Alfine further improved with its roller clutch. The only thing to watch for is the 4-5 shift, where it shifts 2 sets of epicycles simultaneously (soft pedal).

Sram S7, while more efficient in gear, requires intentional soft pedaling while shifting. Pretty similar to Sturmey AW.

4. Weight

Of course the 3 speeds are lighter.

Between the Sram S7 Drum and the Nexus w/Roller Brake, the S7 is about 400g lighter.

5. Braking performance

The roller brake on the Nexus is stronger than the drum on the S7. However, the roller brake is more affected by weather. During winter rides, the brake mechanism had a tendency to freeze in place.

6. Shift gaps

The S7 has a very consistent shift gap, with smaller shift gaps at 1-2 and 6-7. The Nexus generally has good shift gaps, except the annoying 5-6 shift, with its 28% shift gap.

7. Chaincase compatability

The inboard shifting mechanism on the Nexus lends itself better to chaincase use. From the research I did briefly, none of the Hesling or De Woerd chaincases typical on dutch bike are compatible with the Clickbox on the Sram S7. The Hesling Sabre on my Trek L200 could be improvised to work by removing the right sprocket cover.

8. Noise

The Nexus is dead-quiet, except for some very slight ticking in gear 3. Freewheeling is also incredibly quiet, and completely silent in some gears.

The Sram S7 is even louder than the Sturmey AW. So much amazing mechanical ticking noise, but it may be annoying to some.

9. Shifter

The Nexus gripshift with integrated brake lever is an improvement over the regular gripshift. The Sram S7 shifter has a lighter shift, with a more definite click. Matter of preference, but I prefer the S7 shifter.

Overall general recommendation:

Get Nexus 8/Alfine for Hilly terrain, where gears 3-4 will be used less and where its efficient climbing gears (1-2) are very useful.
Get Sram S7 for flatter terrain, where climbs are more rare, where you can take advantage of its more efficient mid-band.

Hubs I'd still like to try:

Nuvinci CVT
Sram P5 5 speed
Sram I-9 9 speed
Sturmey XRD5 5 speed

Sturmey XRD8 8 speed (new model) if mounted on folding bike.

Fantastic comparison - I love hearing about things like this from people who really have both.

I own/use bikes with the Sachs/SRAM 7 speed hub, Nexus 7 as well as the Sachs PentaSport 5 speed and have commuted in the past on bikes with S-A 3 speeds AW and TCW as well as 5 speed S-5 and numerous models of single speeds. All except the AW and S5 have coaster brakes. I could only feel more drag on the Sachs PentaSport 5. All the coaster brakes worked quite well for my purposes except the TCW.

How did you determine the differences in efficiency between gears and models?

For an efficiency comparison between Sram S7 and Sturmey AW, you can reference this article:

http://www.ihpva.org/pubs/HP52.pdf

The efficiency of gears 5-8 on the Nexus 8 should correspond roughly to the chart for the Nexus 4. Gears 1-4 on the Nexus 4 should be derivable from the efficiency of gears 5-8 multiplied by an efficiency for the reduction gear, which I estimated as 0.95-0.98.

Let me rephrase my question, could you actually feel the theoretical differences in efficiency when cycling? Do you think in a blind test you could tell the difference between one type or the other? Especially a theoretical difference in efficiency of a couple of %.

I would guess that over time in a long distance race such a theoretical efficiency advantage might make a difference, but not in commuting unless the difference was significant and that would be obviously noticeable.

You absolutely CAN feel that gear 4 on the nexus 8 is significantly less efficient than gear 5.

Gear 5 has approximately 90-92% efficiency. Gear 4 is around 80%. You certainly feel it, and it is quite easy to test with both gears being next to each other.

Likewise, you CAN feel that gear 1 and gear 7 on the Sram S7 have a reduction in efficiency compared to the midband.

One of my LBS's has a Swobo with a Sram I-Motion 9. I'll give it a spin when I have the time. Keep tuned for an update.

It's indeed true that you feel the 4-5 change in the Shimano 8R25/Alfine.

I would really love to try the new S-A 8 speed.

How did you test for efficiency? How did you come to the conclusion of 0.95-0.98 efficiency of the reduction gear? Which Nexus 8 are you comparing? How did you come to the conclusion that the Nexus 8's efficiency is a product of the Nexus 4 and the reduction gear? I didn't see anything in that article about the SRAM S7.

For those that say they can feel the difference in efficiency of just a couple of % points, I am amazed. I can't tell a difference of how much energy I am using and how fast I am going in one gear. What I mean is that if I am pedaling along and going 15.2 mph then I go to 15.5 mph I don't really notice a difference in the power I am using. Even if I did, I would think the road might have something to do with that. I know I wouldn't be able to do it over more than one gear. You people must be very sensitive to power output, speed and road elevation.

I assumed the efficiency of the reduction gear based on typical efficiencies in industry for reduction gears, which can approach 98% with a helical gear and proper lubrication.

As for why it is similar to the Nexus 4 speed, it is my analysis of how Shimano came up with the design for the 8 speed.

I am comparing to the regular Nexus 8, so the Premium and Alfine should be better. Premium and Alfine should have similar efficiency, but Alfine shifts better under load.

The SRAM S7 in the article is the Sachs 7 speed.

While it is difficult to notice a difference of 1-2%, jumps of 5-10% are certainly noticable, like gear 4-5 on Nexus8/Premium/Alfine.

I am confused about what it is that you can actually feel when riding; the difference in gear ratio when you shift up or down the gears (which is no surprise, otherwise why shift?), or the theoretical difference in efficiency between one hub model (or one gear) and another?

You can feel the some grinding and dragging present in 4th gear, while 5th gear is perfectly smooth, as the direct drive gear should be.

So, the SRAM S7 is identical to the Sachs 7 speed? They haven't made any changes at all since 2001? I did not know that. I don't keep up with the development of hub gears that closely, I just assumed that they would have tried to make some improvements.

Why should the direct drive gear be any more efficient than the other gears?

Is the only difference between the Nexus 4 and 8 the addition of a reduction gear? I think a proper test would need to be done to be positive about your assumptions. Even if your assumptions are correct, you really need to do a proper test to confirm this. You saying "feel some grinding and dragging", is not a percentage of efficiency. I'm sorry, but I just don't believe that you can feel a percentage of efficiency, especially between gears. I can feel that I am using more energy when I change to a higher gear inch gear while keeping the same cadence, but there is no way I can tell you how much more power I am using. Even if I could, I would then have to be able to "feel" how much faster I am going, add in the air resistance and do the math in my head.

Direct drive means that the sun pinion is driving the hub directly, with no epicyclic gears. The only reason that Shimano might have lower efficiency in direct drive is the very thick grease specified for the Nexus internals.

Nexus 8 does not have the same gear ratios as the discontinued Shimano 4, but the mechanism is very similar.

As for the Sram S7, outside of some cosmetic changes on shifter and hub, it still has the same internals as before Sram purchased Sachs. It will be interesting to see how the innovation in the I-Motion 9 has paid off.

I don't have the resources to do a proper test at the moment. Should someone be interested in performing the test, an electric motor driving the hubs attached to a dyno, as specified in the article above, could be used to perform the test.

I did not know that. I have no experience with the workings of an internal geared hub, never had the need to take one apart. I assumed there was three planetary gears that controlled the gearing, with one being a 1:1. Thanks for the info.

Interesting post. Thank you for the analysis. One possible weakness in your argument: I would expect that Shimano would make improvements when designing the 8 even if it is based on 4-speed. Or am I missing something? How many years apart were they designed? (Maybe the improvements are the differences between the Nexus 8 and the Nexus 8 Premium/Alfine.)

The improvements are indeed the difference between Nexus 8 and Premium/Alfine. Premium adds roller bearings for reduced losses and improved seals. Alfine adds a roller clutch for better shifting under load.

It looks like part of the inefficiency I'm noting in the hub is that the hub is very new. Gear 4/8 efficiency is supposed to improve after 400-500km of riding.

Expect an update to this thread once the hubs are broken in.

Thanks for the detailed comparisons. Much appreciated.

I just ordered a new set of wheels built around the nexus red-band 8. I tried mightily to get two local shops to get the Alfine, but they claimed that their national suppliers said it was 'still not available', which I thought was only said early after the Alfine hub got introduced (they are wrong, right? why the confusion still?). A 3rd shop had rentals possesing the alfine, but the shop attendant could not tell me if they could get me the hub by itself. So I had to settle for the red-band 8, and the additional details listed here make me wish I had pushed harder for the Alfine. Oh well.

My primary commute bike has a Rohloff (Koga Miyata Expression), but the bike getting the new wheels currently has a nexus 3, so the red-band 8 will be a much-needed improvement (Electra Amsterdam).

harriscyclery.com has Alfine hubs available now, I just bought one from them two weeks ago. They aren't local to me, but I still count them as an LBS, philosophically. ;)

I haven't compared it to the Red Band, though I hear most of the internals are the same. Having a disc brake mount was my biggest reason for choosing the Alfine, however.

I've been riding on it about a week, and so far, am very pleased with it. I'm used to riding fixed gear, so having any gears at all is cool, so the wide jumps don't bother me much. I'm used to spinning out a bit, so having to wait a little longer before shifting isn't a problem.

Recommendation: Set up your drivetrain so that you are at a number of gear inches you will use a lot while in 5th gear. That way you'll spend as much time as possible in the direct drive gear, which is the most efficient.

-Sam

If you're still looking for Alfine, they are available here: http://www.hiawathacyclery.com/cart/...3eb3e4106cdd95

This is the bike I test rode for Alfine, also from Hiawatha cyclery:

http://bikehugger.com/2008/04/alfine..._omg_i_wan.htm

Expect some photos to be added to the thread shortly.

Reading the service manuals for both hubs, some of the efficiency loss between S7 and Nexus 8 can be attributed to the following:

The planetary gears on the S7 are oil-lubricated. (Bearings still grease lubricated)

The planetary gears on the Nexus are grease-lubricated. I might break down my Nexus and attempt the oil bath suggested on the Hubstripping page and give it another comparison. I'm thinking that Phil's Tenacious Oil might be a decent substitute for Nexus grease.

All I got to say is that I can't tell a lick of difference in efficiency between any of my gears on my Nexus 7. They say gear 4 is direct but I can't confirm or deny. They all work fine.