From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.racing,rec.bicycles.tech,rec.sport.triathlon
Subject: Re: Mavic Mektronic soon to be delivered ??
Date: 21 Apr 1999 16:02:47 GMT

Alex Gian writes:

>>> As usual, USCF was ahead of them -- the USCF board adopted a rule
>>> change on February 7 that made the Mektronic units legal for U.S.
>>> racing.

>> Well maybe we'll get some field testing of this derailleur but I'm
>> afraid that those who spend the money won't come back with an
>> unbiased report when they have a few miles on them.  We can always
>> hope.  The basic flaw with the original ZAP does not seem to have
>> changed, only the wireless feature is truly new.  Their life
>> expectancy for the transmitter seems unusually long considering
>> that it has a speedometer in it that runs continuously.

> I am quite optimistic, I had a chance of trying the Mek in Milano
> last year during the Bike Show and I must tell that after attempting
> to abuse them in all possible ways to see if they would fail (for
> what was possible on a stationary trainer) they performed without
> any problem.  Interestingly enough almost everybody was trying to do
> do the same (Italians are skeptical and like to challenge new
> things) and I never saw any repair or adjustement performed to the
> derailleur while I was there.

I don't know what you tested for but the two main problems are lack of
sufficient chain take-up for a 40-52CW and 13-24 cluster, and
fragility of the mechanism at high chain speeds.  These are related.

From an earlier evaluation:
---------------------------------------------------------------------
Electronic Shifting

A reader asks whether the Mavic Mektronic is any better than the
earlier Mavic Zap electronic shifting.

New styling didn't fix the basic problems of this device, although it
has an elegant speedometer and controls.  The same basic problems
remain in the derailleur mechanism that shifts by means of a ratchet
pushrod that moves in and out with each idler wheel rotation.  The
faster the chain moves the faster it pumps.  A shift occurs during 1/2
revolution but primarily in 1/4 revolution considering the motion
profile of sinusoidal motion.  The stroke takes place in about 35
milliseconds when pedaling a 52t chainwheel at 100rpm.  This heavily
loads the small electrically activated pawls, one for up and one for
down, that engage the up or down ratchet of the pushrod.  The opposing
ratchets of the pushrod have teeth space exactly one gear apart with
little overshoot.

Besides the ratchet problem, the upper idler must lie on axis with the
derailleur pivot, a feature that reduces chain slack take-up.  Today
derailleurs have the pivot offset from and between the two idler
wheels, and use a slant parallelogram (low friction) movement.  The
Mektronic uses a sliding post (like early Simplex derailleurs) that
resists motion when chain tension loads it with torque.  Moving it is
similar to pulling a socket wrench off a nut while tightening it.  A
rubber boot covers the mechanism that must run in an oil bath.

Drawing power to shift from the chain is both the novelty and the
fault of this design.  The novelty is that only control power is drawn
from a battery while power for shifting comes from the chain and only
while shifting.  The fault is that to make this possible the function
of the derailleur is compromised.  Because it can support only a short
tensioning arm due its sliding post, it cannot take up large chain
differences typical of large to small chainwheel shifts.  Most
seriously, pushrod velocity is too great to be reliable at speed.

Jobst Brandt <jbrandt@hpl.hp.com>

From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.tech
Subject: Re: Mavic Mektronic soon to be delivered ??
Date: 22 Apr 1999 20:54:14 GMT

Stacey J(who?) writes:

> Would you find the design more sound were Mavic to change the shift
> actuation from 1/2 rotation to 1 full rotation?  Granted, this
> wouldn't affect the crashworthiness at all, but it seems that it
> would address the biggest design flaw that you have pointed out.  If
> not, is there anything that you can think of that would make Mek
> worthwhile?

First, that is not possible any more than pedaling only once for two
crank rotations, and second, that is not the only problem.  That this
is a sliding post derailleur makes it like the early Simplex ones that
needed two cables because they would not return reliably by spring
force.  That is why Campagnolo took over nearly all derailleurs with
the parallelogram linkage.  Returning to the plunger brings back the
old problem, although encasing it in a rubber bellows insures better
lubrication.  The shift mechanism demands enclosure anyway.

I don't see how the shift speed increase with chain speed can be
overcome, and it is this mechanism around which the whole invention
lies.  Making this ratchet and pawls reliable may be possible but I
don't think the solution has been found yet.  We'll wait and see.

Jobst Brandt      <jbrandt@hpl.hp.com>

From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.tech
Subject: Re: Mavic Mektronic soon to be delivered ??
Date: 22 Apr 1999 21:29:58 GMT

Seth Moore writes:

> My understanding is that Campy already has a patent (or has applied
> for one) on their version of an electric shifter. Don't know if it's
> wireless transmission though.

I can't imagine what they would patent.  The clever feature of the
Mavic systewm is that it extracts its power from the chain and does
not rely on a motor(s) to move the derailleur.  THat is a novel
solution to something that others could not solve.  Unfortunately it
constrains the design to the two drawbacks that I have outlines and
that are its feet of clay.

If you are willing to carry A-size batteries, this can be done by
motors that also weigh something.  I don;t see that shifting by hand
is worth the complexity, weight and minimal advantage if any.  This
whole issue may soon go the way of Breathe-rite nostril spreaders or
Softride saddle mounts.

Jobst Brandt      <jbrandt@hpl.hp.com>

From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.racing,rec.bicycles.tech,rec.sport.triathlon
Subject: Re: Mavic Mektronic soon to be delivered ??
Date: 22 Apr 1999 21:47:14 GMT

Andrew Albright writes:

>> exorbitant price to fix it. So I an going to wait until other
>> people have tried it first!!! I would like to try the Mektronic
>> under harsh conditions, such as a riding for a whole week during
>> stormy conditions. Try going through several dozen puddles and see
>> how the Mektronic holds up.

> IF it is like the old ZAP, then you don't need to worry about water
> problems.  I rode in rain and ice and never had a problem.  think
> about it, why would it be so difficult to make something like this
> waterproof?  How often does a Cyclometer go bad in the rain? etc...

There is a difference.  Bicycle speedometers have no rotating shafts,
the ZAP/Mek does and it leaks.  This is one of the dirtiest places
on the bicycle for water and grit spray.

> p.s. I doubt they wouldn't honor a warranty past the 1yr.  I had the
> Zap for about a year, sent it back once.  Got tired of it and gave
> it to a friend about 2-3 yrs later.  He called up Mavic with a
> couple of questions on how to fix it and when he couldn't they said
> send it in and we'll give you full wholesale credit for it.  I think
> he got a pair of Cosmics or something.

I can imagine they didn't want to fix it nor return it to the field.

Jobst Brandt      <jbrandt@hpl.hp.com>

From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.tech
Subject: Re: Mektronic/Zap
Date: 3 Feb 2000 16:51:53 GMT

Bob Judi writes:

> ...  I agree the neatest thing about the design (this is the same RD
> design as in the Mektronic) is the fact that it uses the rotation of
> the upper idler wheel to move the mechanism in/out to shift.  This
> precludes the use of a big heavy battery to move the thing
> electrically.

UCI rules prohibit supplying power to perform any physical function
on the bicycle.  The extraction of shift power from rider input is
an essential constraint on the design.  It is not the engineer's
choice whether to use parasitic power to shift.

> It works quite well, and I have found it to be very reliable.  The
> 'post' JB mentions, and the detents, etc. work very well - any ME
> would love the guts of this thing.  It's beautiful.  Looking at the
> details of how the bearings are arranged along with the load points
> will reveal the problems that JB points out are really not problems
> at all.

When I inspected the ZAP at the bicycle show this problem could be
demonstrated by putting a load on the derailleur to simulate a slightly
higher chain tension.

> Chain tension is on the light side, but not so bad.  By the way - I
> did find the new Mektronic rear derailleur has a bit more tension
> here.  Also, the pins engaging the notches don't create any problems
> either.  The post is hard anodized, and the interface between the
> pins & notches is designed quite well.  My unit showed NO wear after
> 8600 miles.

I didn't mention "wear" but rather pawl failure.  It's a go or no-go
situation.  Shifting at high chain speed puts excessive force on the
shift pawls that are electronically thrust into the ratchet pushrod.
At high chain speed this becomes an impact load on the pawls that can
break off the mechanism or at least the sharp tip.

> But, my biggest problem, one that JB really missed totally, and the
> reason I don't use it on my serious TT bike, is the drag the rear
> derailleur puts into the drivetrain.  The cams which induce shifting
> are constantly moving, whether shifting or not.  And, there is a
> finite amount of drag on the upper wheel at all times.  You can feel
> it with your fingers while turning the wheel by hand.

That is the way I tested it and found that it moved freely.  Of course
I couldn't assess what happens when the idler turns at over 500 rpm,
the speed at which it turns when riding fast.

> The drag is not only friction of the bearing, etc., but is
> compounded by the fact that Mavic uses a u-joint to transfer the
> rotation of the wheel to the plane on which the derailleur moves in
> & out.  Yes - take one apart yourself.  I was disappointed in this,
> but I could understand why Mavic did it.

The universal joint is necessary because the derailleur cage must move
in a slant that approximates the envelop of the sprocket cluster, and
that is not perpendicular to the idler.  As I mentioned, chain tension
prevents the sliding post from moving freely because chain tension is
taken up as torque in this shaft.  The unit demands low chain tension
and a short derailleur cage.

> Again, any ME (especially with an automotive background) will
> recognize the baggage that comes with a u-joint.  We are talking
> about a small amount of inefficiency here, but we're only starting
> with about 1/2 HP produced by the rider.  The cams which are
> constantly moving are moving back & forth parallel to the motion of
> the derailleur body.  So, the u-joint is needed as the wheel spins
> in a different plane.  Anyway - that's the way this thing works.  It
> is a great design, but given the drag it adds to the drivetrain
> (however small it is), those 30 seconds you'll lose at the end of
> your 40K TT won't be so great.

I think a measure of drag would be important to assess ho bad this
effect is.  I'm sure that hidden in some internal report, Mavic knows
exactly how much power this absorbs.  In my estimation, the mechanical
weakness of the shift mechanism are more damaging.  The sliding post
and shift rate proportional to chain speed kill the device, not to
mention the need to be hermetically sealed and lubricated.

Jobst Brandt      <jbrandt@hpl.hp.com>