The facts behind Honda’s secret gearbox

Dailies -

By

Honda-gearbox-1

“Obviously HRC is not using anything illegal, but it’s true that we have something new on the transmission that currently allows for faster gear changes,” says HRC vice president Shuhei Nakamoto. “We are not talking about a DCT, but I can’t tell you how it works…because I don’t know!”

HRC has twice now admitted to the existence of a “secret” (their word) gearbox on its factory MotoGP bikes and speculation in the paddock is its this which is giving them such a significant lap time advantage in pre-season testing. Here’s what we know about Honda’s secret.

Update: David at MotoMatters had the genius idea of comparing sound files of the Hondas shifting next to recordings of the Yamaha and Ducati. Makes for a great comparison.

The MotoGP rule book expressly forbids the use of more than six gears, dual clutches, continuously variable and automatic transmissions. The rules do, however, allow, “manual transmissions with gearshifts assisted by quick-shifter systems.” If Honda is playing by the rules, that means no DCT, CVT or full-auto, but it does leave the door wide open for other solutions.

What it looks like Honda has done is find a way to eliminate or at least reduce the break in acceleration which occurs when you shift between gears. That would go a long way to explaining why their riders have been so dominant in pre-season testing. Take two equal bikes and the one that can spend more time accelerating is going to have an advantage. But where might that innovation come from? Two gearbox companies, Xtrac and Zeroshift offer two remarkably similar transmission systems which deliver DCT-like seamless shifting.

Xtrac’s MotoGP gearbox.

Connecting a few dates makes me believe that Honda’s most likely using the Xtrac Instantaneous Gearchange System. Honda has an existing tie to the company, which builds the transmissions for its IRL cars. Additionally, Xtrac issued a press release expounding on the benefits of “two years racing…with professional teams” and also advertises a “MotoGP Transmission” on its website.

Regardless of whether its Xtrac or Zeroshift, the solution works in the same way. Both systems hinge on a ratchet system between each gear which allows two gears to be selected at the same time, but with only one engaged. When the transmission is shifted, the engagement of the next gear causes the disengagement of the previous one. Rather than a revolutionary change in transmission technology, this is more evolutionary, bringing the benefits of DCT without breaking the rules.

The Zeroshift selectors are visible in red and blue.

Current gearboxes require the shifter fork to deselect by means of moving a slider, then selecting the next gear by moving a slider into contact with it. The slider doesn’t have to travel very far, but it does fully disengage and this causes a gap in acceleration while shifts take place. With both Xtrac and Zeroshift, a ratchet assembly replaces the slider, selecting two gears at once. The trick is that they don’t move as a whole during shifts, simply extending its ratched to the next gear and, when it engages, the previous gear freewheels because it was only ever engaged by that ratchet.

While these transmissions do allow truly instant shifts, the bike itself won’t be capable of just slamming from one gear to the next at full power. A crankshaft with 230bhp behind it simply cannot decelerate from 19,000rpm to 15,000rpm instantaneously. If that was happening, some other part of the system would be giving — the clutch would slip, a transmission ratchet would fail or the tire would break traction. Any of that would, at best, decrease performance and, at worst, cause a crash.

Instead, in order to take advantage of these new gearboxes, an electronics package would have to be included which could control throttle opening, spark, fuel and clutch actuation. Zeroshift confirms this, saying, “the energy spike in the driveline…is managed by dissipation through a combination of engine and transmission management, clutch control and system compliance.” The Honda RC212V already boasts traction control, launch control and a highly-sophisticated slipper clutch. Adding a program like the above, capable of feathering the clutch on upshifts and including a revised engine map and TC settings is all that would be necessary.

So how much advantage is Honda gaining? Assuming 30 shifts per lap and a .03 second advantage per shift, the advantage of .9 to 1 second claimed by Zeroshift seems reasonable and its right in line with Casey Stoner’s advantage over the fastest non-Honda rider at the Sepang tests:

1    Casey Stoner              Repsol Honda Team             1:59.665
2    Dani Pedrosa             Repsol Honda Team             1:59.803
3    Marco Simoncelli      San Carlo Honda Gresini    2:00.163
4    Andrea Dovizioso      Repsol Honda Team            2:00.541
5    Ben Spies                     Yamaha Factory Racing      2:00.678
6    Colin Edwards            Monster Yamaha Tech 3     2:00.966
7    Jorge Lorenzo             Yamaha Factory Racing      2:01.003
8    Alvaro Bautista           Rizla Suzuki Team               2:01.194
9    Hiroshi Aoyama          San Carlo Honda Gresini    2:01.328
10    Hector Barbera         Mapfre Aspar Team              2:01.346

However, both Honda and Yamaha are downplaying the laptime advantage just the gearbox brings.

HRC’s Nakamoto points out, “Aoyama on the satellite bike doesn’t have this new transmission and he is still very strong so it means the bike as a whole is simply getting stronger.”

Yamaha engineer Masao Furusawa told MotoMatters, “[Honda’s gearbox] will only give a small advantage on the track. Going on to suggest that it’s actually the RC212V’s low-end torque that’s making the most difference.

And Honda’s advantage has decreased at this weeks’ Losail tests:

1    Casey Stoner              Repsol Honda Team             1:55.681
2    Dani Pedrosa             Repsol Honda Team             1:55.745
3    Ben Spies                    Yamaha Factory Racing       1:56.294
4    Marco Simoncelli      San Carlo Honda Gresini    1:56.433
5    Andrea Dovizioso      Repsol Honda Team            1:56.439
6    Randy De Puniet       Pramac Racing Team           1:56.445
7    Jorge Lorenzo            Yamaha Factory Racing       1:56.707
8    Colin Edwards           Monster Yamaha Tech 3      1:56.716
9    Nicky Hayden            Ducati Marlboro Team         1:56.726
10    Hiroshi Aoyama       San Carlo Honda Gresini    1:56.740

A less obvious advantage from the secret new gearbox is less obvious than outright shift speed, but could still help. With such a system as we described above and its accompanying electronics, the instability caused by shifts while leaned over would no longer be an issue. Subsequently, riders would be able to carry more speed through and take a more ideal line through corners requiring an upshift.

So how long can we expect Honda’s advantage to last? Well, Yamaha is already working on a similar gearbox solution.

“”Sure we are working on a system back in Japan, but it is not ready yet,” Furusawa told MotoMatters, which goes on to discuss the problem in load transfer discussed above, saying that’s the problem Yamaha is currently working on. Furusawa continues, “I am very interested to know their solution, but I have not been able to work it out yet.”

Nakamoto knows that’s where the real threat lies, “[Yamaha is] achieving very fast lap times and seems very consistent, Jorge is able to keep his same low lap time on long runs, both here and in Sepang, and Spies seems to be stronger than last season – so I expect them to be very competitive.”

Sources: MotoMatters, Speed, Xtrac, Zeroshift, HRC

  • jwinter

    The only thing that sucks about Honda being so strong is the danger of the races becoming really processional again this year.
    Free practice in Qatar wasn’t as decisive as testing, aside from Stoner that is.

  • Myles

    Why outlaw autos and DCT? Shouldn’t top-tier racing be about advancement?

    • http://twitter.com/hagus Luke

      Same as in F1, it’s about keeping runaway costs down and the playing field relatively even. Take the limits off and you will have teams with carcinogenic beryllium brake rotors in a heartbeat. They already struggle with grid numbers.

      The failure modes also get more dramatic the more esoteric you get. Witness what happens when the ground effect for old F1 cars “stopped working”. Weee!!

      • Myles

        But beryllium brake rotors don’t tie into real world vehicles – while non-standard transmissions do.

        One of the great things about racing is the trickle-down effect we all get to enjoy. ABS, traction control, crumble zones, and more efficient engines all come directly from racing. It would be sweet if we could get an awesome transmission also.

        I understand the argument for rules in racing, but I don’t understand the specific rule on transmissions.

        • Ducky

          I think this might actually be a cheaper method of integrating completely “clutchless upshifting” in modern bikes, with electronic controls. Watch for it in a decade or two.

  • http://twitter.com/hagus Luke

    This article is hilarious.

    “Assuming 30 shifts per lap and a .03 second advantage per shift”

    What part of the lower intestine was the 30ms per shift plucked from? It’s just unadulterated numerology designed to match up with the lap time deltas.

    Do you think that a 30ms shift speed improvement actually knocks 30ms off your lap time? Guess what, while the engine is momentarily cut, the bike doesn’t stop moving entirely. You’d need to compute the acceleration loss *during* that 30ms. We’re talking a poofteenth of a millisecond savings here.

    It might materialize to some measurable delta over race distance, but the per lap saving would be annihilated if you forgot to shave your stubble before the race. For Simoncelli you’d never see any savings at all!

    • http://hellforleathermagazine.com Wes Siler

      That number comes from zeroshift. Do you have some other problem with the article? Small things like this are what adds up to a competitive gp bike these days, hence the cost.

      • http://twitter.com/hagus Luke

        Ok, so the figure comes from Zeroshift. You should call that out because this article is trying to build a case for their being a link between this technology and Honda’s superior lap times. The figure is just hanging in the air with no supporting evidence right now (hence my incredulity).

        I can’t find the figure on Zeroshift’s website. I did find mentions that the shifting takes 50ms (in an article from 2008).

        How the 30ms per shift “advantage” arrived at? The true advantage will not be the shifting time, but the speed increase from not killing the engine.

        How much engine killing are we talking about here? Let’s look at the default configuration of a commercial DynoJet quick shifter, such as I have fitted to my bike:

        1st gear – 60ms 2nd gear – 53ms 3rd gear – 53ms 4th gear – 48ms 5th gear – 48ms 6th gear – 48ms

        (source: http://www.powercommander.com/downloads_general/install/4-112installguide.pdf)

        This is the factory setting, so you can assume they will be relatively conservative. The average ignition cut time is 51 ms. The MotoGP version will be a lot better.

        Let’s assume MotoGP bikes are currently using a squid-level DynoJet quick shifter which takes 50ms per shift on average. They adopt the Zeroshift system and it reduces shift time to 0ms. Let’s assume there is absolutely no acceleration break at all.

        On average, per shift, what will be the momentum loss of the bike during 50ms? Regrettably I was doing, er, other things during physics class and I can’t compute this myself. But I think what we’re talking about here is more like a 50ms saving *per lap*, if you’re *lucky*.

        Someone can punch me in the face with Math if they are better at it than me, but I presently believe the numbers in this article do not build a case of any sort for the shifter being responsible for Honda’s pre-season lap time dominance.

        • http://hellforleathermagazine.com Wes Siler

          Did you miss this part of the article?

          “However, both Honda and Yamaha are downplaying the laptime advantage just the gearbox brings.

          HRC’s Nakamoto points out, “Aoyama on the satellite bike doesn’t have this new transmission and he is still very strong so it means the bike as a whole is simply getting stronger.”

          Yamaha engineer Masao Furusawa told MotoMatters, “[Honda’s gearbox] will only give a small advantage on the track. Going on to suggest that it’s actually the RC212V’s low-end torque that’s making the most difference.”

          • Myles

            But he’s saying quicker shifts don’t provide equally quicker lap time numbers – and he’s right. The article kind of makes it seem like .03s shorter shift or whatever makes you .03s faster.

            Go run a 1/4 mile in any vehicle.

            Run it again and clutch in for a full second anywhere in the second half of the run.

            You won’t run a full second slower. It’s not a direct correlation because you’re still moving and covering ground, even if you aren’t accelerating.

            Does a better transmission make for faster lap times? Hell yeah, we’ve all played GranTurismo – but the correlation isn’t as direct as the article makes it out to be. But HFL is still awesome because Honda’s getting time on the front page and HRC is making awesome bikes and all is well in the world.

            • http://hellforleathermagazine.com Wes Siler

              Ah, right, I see. .03 isn’t the shift time, it’s the aggregate advantage to overall lap time accrued. Allegedly anyways.

              • http://twitter.com/hagus Luke

                Exactly. The manufacturers are downplaying it, but the article (Zeroshift?) alleges a full second per lap advantage. The math there doesn’t seem to add up, and I can’t see how it ever could.

                If it was read wrong and the advantage was 0.03 per lap, that wouldn’t even be a single second over an entire race distance, to put that in perspective. The advantage of the system would have to be elsewhere, such as “feeling” better to the rider or being more tolerant of mistakes or wearing better.

                In terms of why Honda is fast: never rule out rider talent. Casey is fast on the Honda. Guess what? He was goddamned fast on the Ducati, day in day out … when he stayed on the thing. If the switch sees him stay at the same speed but finish more races, he will walk away with it this year.

                The greatest rider we’ve ever seen, with the greatest mechanics in the paddock, who can supply the very best engineering feedback, hasn’t been able to get the Ducati around the track anywhere near the same times as Stoner. Looking at the time sheets, *that* is the only big surprise here, not the fact Casey and Dani are really quick as always.

                • http://hellforleathermagazine.com Wes Siler

                  I agree that talent plays a huge part. But it’s unprecedented to see the ENTIRE field of Honda riders faster than EVERYONE else, including the world champ. There’s got to be an explanation for that advantage and so far it’s looking like this is it.

                • jwinter

                  It’s not a big surprise. Rossi is still hurt and has barely had any riding time on the bike. There are tons of other factors that make straight comparisons less valid. Not to take away from Stoner, he’s really talented. I just think he needs a bike that can get away if he wants to win consistently. I do love watching him ride Philip Island though.

    • jwinter

      All the comments I’ve seen out of Qatar say that the difference is noticeable just from watching. The smooth shifts seem to be making a real difference. Honda are desperate to get an 800cc title before it all goes back to 1000cc also so they are going to throw everything they have at that goal.

    • Sean Smith

      30ms is just a number that Zeroshift put on it. The only way to really know for sure would be to have one of the riders do back to back sessions on one bike that has the new tranny, and one that does not. Even then, it would be hard to put a number on it because the speed advantage is more than just faster shifts.

      Imagine, if you will, that you know you will need to up-shift just after the apex of a corner. you’ll be leaned over roughly 50 degrees, and traction is precariously balanced between the front and rear tires. Grabbing a gear will temporarily halt all acceleration, and toss a bunch of weight onto the front tire. You can’t take an ideal line, have the bike leaned over as far as you should, and as a result, you’ll be going slower in order to make that shift without losing front traction.

      With a system that doesn’t require a break in acceleration, you CAN take that ideal line, and you don’t have to worry about low-siding nearly as much. It’s not magic, and your acceleration won’t be completely linear, but the break that wreaked havoc on your front tire will be almost completely gone.

      When you start to consider those advantages, you realize that time saved isn;t so black and white. The whole package changes, your lines change, and even the way you ride the motorcycle changes.

      • http://twitter.com/hagus Luke

        At the risk of feeling like a troll because I can’t let it go :) …

        The first question one needs to ask – especially a strong, independent publication like HFL – is how that 30ms number was arrived at. I’m simply not buying it, for all the reasons outlined in my other posts. There is a decent chance some basic miscommunication has occurred, but there is not one second a lap in pure shift speed. There just isn’t. Someone prove me wrong, using The Science™ and The Math™, not a press release from a gearbox vendor.

        Improvements in “feel” are worth discussing. But you need to be more skeptical here as well. Bear in mind that all MotoGP bikes already have quick shifters that almost certainly shift faster than 50ms currently. We are not starting from slouchville.

        Even before that, shifting at the apex is not going to be really common in MotoGP. They make mountains of power and the circumstances where you will be fully cranked over and need to snatch another gear will be the exception, not the rule.

        Even then, the hardest thing about shifting fully cranked over is not the engine cutout, but left hand corners. Most riders will plan a left hander knowing that tapping the gear lever isn’t an option until the bike starts to stand up a bit. A fancy gearbox isn’t going to solve that particular concern.

        So what about all the other situations where riders would otherwise choose not to shift because the <50ms engine cutout time produces an adverse situation. It would be great if someone with some skills could calculate the deceleration produced by cutting acceleration for 50ms, accounting for the momentum of a minimum 150 kg motorcycle. Will that be "tossing a bunch of weight" onto the front end? Will that "wreak havoc" on the front tyre? I just don't think <50ms is long enough to produce that effect.

        Furthermore one has to assume that current MotoGP bikes are relatively "dumb" in this respect and the problem – if it exists – of unwanted weight transfer mid turn due to gear shifts hasn't been smoothed over by some other form of electronics.

        As for "low-siding nearly as much" – good lord. I'm calling shenanigans on that one. Find me a single instance of any rider of professional level who blames a mid turn gearshift, using a quick shifter, for their low side. I certainly can't recall a single instance. I think there was a time either Barros or Biaggi *upshifted* mid turn and binned a podium as a result, but my Google-fu fails me. I do remember that ended their race dramatically though!

        I buy that there could be improvements gained through this system, but all proposed magic performance improvements are guilty until proven innocent, using hard science. As fun as the speculation might be …

        • Sean Smith

          lol, I’m not trying to claim that they’re absolutely one second faster because of some transmission voodoo and lowered shift times. That 30ms number is almost certainly the product of fuzzy math, but it’s all that was out there to quote.

          As for the front end weight transfer, that’s a real thing. The reason you never hear people blaming crashes on mid corner shifts is because they all know what would happen if you tried to pull a stunt like that at the limit. No world-class rider would make that mistake.

          Hard science you say? Good luck. There’s just too many factors to calculate that. Add to that the fact that there’s no way to get the numbers for many of those factors (being unique to the moment they exist in) and ‘hard science’ that breaks down which improvements came from where becomes impossible.

          All we can do is look at the lap times, and speculate based on what the teams and riders are saying. The Hondas were already incredibly fast, but the absolute domination we’re seeing now makes it look like something is up. Honda has said that it has this new piece of tech, and their lap times are dramatically lower than anyone riding another bike. Do they have more horsepower? Probably. Is their chassis more stable and refined? Well, the riders only complaints are for a little more rear end grip. Could their times be from these two things alone? Maybe. But, faster shifting and the improved stability that comes along with it should not be discounted. Any improvement made will contribute to a lower lap time.

    • Ducky

      Not sure what your gripe is. These numbers are theoretical, they come directly from the supplier’s website. Even if you do not achieve a full second, the transmission still cuts shift times, and when free practice has riders in 2nd to 8th place separated by a mere .2 seconds, you need every advantage you can get.

      The smoother power output also helps in other areas. They might be able to shift in mid corner better, they can shift on corner exit smoother, or used in concert with the slipper clutch provide much more stable braking (something that Honda really needs help with right now).

  • TreMoto_Eddie

    http://tinyurl.com/4slhwmg <– Shift-time limited acceleration final drive ratios in FSAE

    These guys find that actual time at ~zero acceleration is close to 0.250s with a pneumatic clutchless upshift system and goes on to explain the development of an automatic upshift system dropping that time to 0.100s. The important distinction here, in my mind, is "shift time" vs. "time at ~0 acceleration". While I have little doubt that factory Hon/Yam can devise a system quicker than that, we are certainly leaving (some) time on the table.

    0.030s @ 200kph covers only about a bike length. So for that bike length the Yamaha is drifting and the Honda is accelerating. Put a different way, the bikes are equal all lap but the Yamaha rider pulls the clutch in and drifts for 1 second on the main straight. How much time is lost? I doubt more than a tenth or two.

    Sounds like Honda has a cool shifting trick (anyone have an audio clip of the new bike under full power upshifts) but it doesn't account for the even the majority of their advantage.

  • KLR_Pilot

    I like pudding.

    • Mr.Paynter

      Me too.
      Only chocolate though.

  • http://twitter.com/hagus Luke

    The final word. I pinged a certain mechanic for a certain rider in a certain MotoGP team.

    Me: “we’re having a huge debate about Honda’s alleged fancy gearbox. How long does a typical upshift take on a MotoGP these days?”

    Him: “a nano. Its not the reason they are fast.”

    Q.E.D.? :)

    Edit: more from the horse’s mouth … “all teams have tested assisted shifting in some form. Lap time never changes? But may help a poofteenth to keep TC smoother”

    • http://hellforleathermagazine.com Wes Siler

      Ha, I defer to Alex Briggs’ greater expertise then.

    • Myles

      Haven’t Honda riders been complaining about problems with stability under breaking for a season or two?

      Maybe the specific way the bikes were set up really made the potential downshifting advantages bigger than they would be with other teams? Maybe in a vacum these transmissions aren’t all that great, but they really helped the Honda bikes. Maybe it’s not the beter acceleration, but the better deceleration.

      Is there lap analysis data available from testing? Where are the Hondas faster? Are they braking later, holding a higher cornering speed, getting on the throttle earlier on corner entry, simply making it up on the straights? Are they taking unique lines? Are they shifting more/less/at different times? With some simple lap data we should be able to see exactly where their advantage lies.

      HFL – it would be cool to do an “anatomy of a motoGP lap” article for at least one rider per team (even the satellite teams). It would be interesting to see the differences for al the bikes, and would give all the fans more ammo to argue about (“team xyz is obviously skirting the rules somehow – look at the power advantage!” “team abc couldn’t build an engine to save their life” “Ducati sucks!” shit like this).

      • http://twitter.com/hagus Luke

        We could probably get Briggsy to engage in a general discussion of that – but top tier teams are going to be cagey with specific data.