WatchUSeek Watch Forums banner
1 - 10 of 10 Posts

9,657 Posts
Discussion Starter · #1 ·
Writing this review has been difficult: not so much that I don't have anything nice to say, nor that the company or watches aren't interesting, but much more that there is so much going on here, technically, that it's hard to know where, exactly, to start. This is going to be a long review: there is a lot going on here that I'd like to explain as well as possible!

First, let's think of the watch industry as we know it: you have ETA, the monopoly supply of movements (Ébauche in the parlance, meaning generally a preliminary sketch, but in the watch business it's a complete movement, ready for modification or, by adding case, hands, dial and crown, becoming a watch assembly company, not to be confused with an actual manufacturer, who makes their own movements), maker of such well-known and used movements such as the 2824 and its variants, the 7750 and its variants, the 2893 and its variants, as well as many quartz and other mechanical movements. Let's be honest: these movements are good. They are relatively inexpensive (in volume, of course), perform very well right out of the box, and can be worked on to provide some top-notch watches.
But, alas, they are not perfect. There are limitations what can be done with mass-produced watches, there are limitations because they are designed to be mass-produced (largely design limitations that preclude better mechanisms in the interest of manufacturing simplicity), and there are fundamental limitations based on their actual design.

What limitations am I talking about? Well, IANAW (I am not a watchmaker), but let me give you my take on this. First, the perfect watch movement would leave nothing to chance. Literally nothing: from winding to the final movement of the hands, there would be no aspect of the movement that makes compromises, that allows something to move in a haphazard or random way (and yes, I am talking about things like micro-second jitters and sub-millimeter tolerance sloppiness because it's hard to get the necessary hard-core precision. Second, there are small, but subtle aspects of the watch's movement that have to be paid attention to: torque, for instance, from the mainspring, can put a distortion into the power curve because it is both nonlinear and the power curve itself is an artifact of the mainspring, rather than of design. Other design aspects that are often left uncontrolled because of the difficulty in controlling them include escapement movement (where the escapement moves freely based on the hairspring oscillation, rather than being carefully limited in its movement range); the fact that the escapement's function is not merely to provide timing, but also to provide motive power for the driving of hands; the way the hairspring itself is attached to the balance wheel (where even small kinks in the hairspring mean loss of control, resulting in deviations from the planned norm; the list goes on.

What's the point, however, of getting these admittedly small differences and deviations under control? Well, the first is that these errors are generally cumulative: while some deviations from the design norm might, by accident, cancel each other out, the usual result is that the watch movement will not perform to its ultimate ability. You can see this with your average 2824: the basic movement generally performs ±15s/day, with the TOP ébauche bringing this down to ±10s/day or so, and the COSC performance down to +6/-4s/day. The COSC performance means that this is probably the very best you can ever hope to get out of a 2824 movement because the movement itself is not designed to perform better. We all know the triumphant owners who claim to have a movement that is +1s/day or similar performance: however, this is a statistical anomaly, rather than by design.

Of course, the more attention paid to such a movement by a master watchmaker, the better the results: let a master watchmaker overhaul your 2824 with carte blanche to improve the timing, and you will get a movement that will be better than COSC in six positions, especially if the movement is new. But he will literally spend days working on the movement, fine-tuning how the balance spring works, rebalancing the balance, working on the drive train, all the fun stuff. You'll end up with a movement that performs extremely well, but will never be really perfect: the design of the movement doesn't allow better accuracies.

Now, that is exactly what some termineurs (those who put watches together) do: they know the design limits of what you can expect with a movement from the 2824, 7750 or 2893 family (which is one of the reasons the latter is so well regarded by watchmakers: you can do a lot with that movement!) and they have been given the freedom by their bosses to make those movements stand up and sing. Of course, that costs: it is one of the reasons why you can pay $10'000 for a watch with a 2893 inside, while the "same" basic movement can be found in a watch costing $600. Sure, it's not the only reason or even the primary reason, but it is one of the reasons.

So, let's talk about Heritage Watch Manufactory. It is a recently founded private watch company in Neuchâtel, Switzerland, backed by German investors, and who engaged Karsten Frässdorff, well known for his movement designs with FDMN (Fabrication De Montres Normandes), the French company dedicated to re-creating high-end French watchmaking. The watch design is by Eric Giroud, a product designer who has done some extremely innovative work with companies like Harry Winston, Swarovski, and who has won the Grand Prix d'Horologie de Genève in 2007, 2009 and 2010, as well being a red dot design winner in 2010 and 2008 Watch of the Year award in Japan. Suffice to say that the company has two significant talents working together.

I'm only going to talk about one watch here: their Tensus model, since as a watch geek I find it to be by far the most interesting. It's not the only watch they make: Centenus, Magnus, Magnus Contemporaine, and the Magnus Classic are the other models, with the Magnus being perhaps the most interesting, using Karsten's precision balance design. But their Tensus model combines no less than five patented innovations that together create a very unique timepiece.

This is what it looks like: sub-dial at 9 is the power reserve indicator, with subseconds at 6, a classic design.

First and foremost: this is a chronometer design, characterized by a relatively large and heavy balance wheel running fairly sedately at 18000 beats per hour, using a constant-force escapement system, a double-barrel mainspring, a mass-balanced balance wheel, a precision system for setting the hairspring without damaging the spring itself, and a very remarkable system that allows the escapement itself to be set very precisely. That's five major technological advancements in the movement, each dedicated to improving the isochronicity of the movement.

So: what makes the watch tick? What are these five innovations, each patented, that makes such a difference?

Let's take a look at the movement:

In talking briefly with Mr. Frässdorff at BaselWorld 2011, the first question I had was how these innovations actually performed in the real world. His answer was, very correctly so, "It depends". When he puts the watch together, he gets accuracy, for every-day wearing, of around 1 sec/day; he expected that when Heritage ramps up production and uses "normal" watchmakers (we are, after all, talking about a Neuchâtel company!) that out-of-the box accuracy will be in the order 2-3 seconds/day. That can be perhaps improved some by having the watchmaker adjust to 6 positions, but getting accuracy down to that level direct out of the box is remarkable. Remember that anything else out there doesn't come with that sort of accuracy built-in to design, but rather is the result of significant reworking and tweaking of movements. Here the design of the watch movement results in the accuracy.

So, let's take a look at what makes this watch movement show such isochronicity.
First: that large and relatively heavy balance wheel moving at a moderate 18000 BPH. Simple mechanic theory says that a moving large mass shows greater inertia than a moving smaller mass, all other things being equal. While 18000 PBH is slow by modern standards, using a large and relatively heavy wheel means that the trigger stone on the balance wheel will show a smoother sinusoidal curve, the result of a more stable oscillation. The mass of the balance wheel is fairly considerable: 126mg/mm³, or around the density of rhodium (which is a tad lighter) or mercury (which is quite heavy for a metal), which means that once the oscillation sequence starts, the stability of the oscillation curve will be helped significantly by the higher inertia of the balance itself.

Now, given that the balance is moving so smoothly, the actual balance of the balance wheel itself becomes a factor. This is where Heritage uses what they call the VIVAX balance assembly, which is a mass-balance system. While it is basically a system to adjust the balance of the oscillating mass, the way that this is done is unique. Fundamentally, when the balance oscillates, it should do so in such a way that it repeats in a constant and consistent way, since this is basic to time-keeping precision. There are two major reasons why this doesn't happen: non-constant influences due to changes in the power curve of the mainspring driving the oscillations, for instance, will change the beat of the balance, while changing the effect of gravity on the balance wheel (by actually wearing the watch!) creates changes in how the balance pivots press against the balance wheel and cap stones. Normally the watchmaker - or more exactly, the balancer, a special position amongst watchmakers - uses regulating screws to control these influences by the balance itself, or for less expensive movements, small amounts of material are physically removed from the underside of the balance wheel to achieve a relatively stable oscillation. However, and this is a big however, the "normal" way of balancing a balance wheel is static: once done, it remains a constant until the balance wheel is re-balanced.

By fixing what is otherwise a dynamic process - the balance wheel does not move at the same speed during the oscillation cycle, but rather speeds up and slows down, dampened by the hairspring and driven by the power from the mainspring - you make compromises. What Heritage has done, in order to create an automatic dynamic adjustment of the poise of the balance, is to add so-called affixed to the periphery of the balance wheel: these are attached like arms. When the speed of oscillation of the balance wheel increase, basic centrifugal force exerted on the arms of the affixes causes the affixes to open up slightly, introducing the ability to influence how the balance wheel speeds up and slows down by the use of additional mass screws (since when the affixes open, the speed can be influenced by an additional mass attached, which, when done correctly, self-regulates the balance, since a balance wheel moving too quickly will receive a slow-down effect from the additional mass that is moved by the affixes). These mass screws can also be attached closer or further from the fulcrum point of the affixes and provide an additional adjustment point.

The VIVAX balance uses two additional mass screws that can be set for five possible positions: this allows, by design for the mass inertia of the balance itself to be regulated: the result is a self-regulation back to the original, designed oscillation. This means that the Caspari effect (i.e. isochronous deviation) becomes significantly easier to control over the long-term: when the amplitude of oscillation of the balance where changes, the affixes and the additional mass provided by the mass screws (and their position on the affixes!) uses basic physics - centrifugal force - to self-regulate the oscillation.

In other words, the design of the balance wheel means that basic physical forces are used to help reach a higher temporal accuracy that, once set, will always return the balance oscillation back to design specifications automatically.

So, that in and of itself is pretty cool: what about the other innovations?

Well, in looking at the balance oscillation, the whole system is powered by the mainspring, with the balance wheel "merely" serving to drive the escapement and break the energy of the mainspring into tiny, exact spurts of energy that drives the gear train and ultimately the movement of the hands of the movement. So what about that mainspring?
Mainsprings have what is called a power curve: it starts off strong - the spring has the greatest tension - and as the energy stored in the mainspring is slowly released, weakens over time until the spring has lost so much energy that it can no longer drive the balance wheel and the watch stops. Simple, right?

Oy. Not that simple at all.

The force exerted by a fully wound mainspring is not trivial (try disassembling a watch without releasing the mainspring and see how far parts fly…). This force is not completely transferable, but rather torque effects put pressure on one side of the shaft of the main spring: this is not a trivial amount of energy and is a classic wear point, which is why fully jeweled watches will put a stone on each end of the main spring shaft. Given the usual position of this shaft, heavy wear here can be expensive to fix: the bottom main spring pivot is fitted directly into the base plate, for instance, and if the choice of material is poor there will be wear that is almost impossible to fix without using a new base plate. Here the problem is both inaccuracy and poor efficiency: the power curve of the mainspring becomes distorted and the energy contained will be spent working against the hole in the base plate, rather than to move the escapement.

Heritage addresses this by using what they call the Pariter dual-spring mainspring. By incorporating two main springs, inverted by 180°, the torque effects are cancelled out, especially if, as Heritage does, it uses two mainsprings from the same piece of original spring material for exact material behavior: any inner-metal variances due to materials inconsistency will be, since they are 180° out of phase, cancel each other out. The springs are independent of each other, separated, but both act on the same spring core in the barrel shaft of the mainspring barrel: by displacing the point of contact in upper and lower springs relative to this shaft, the torque generated by the mainsprings is not applied to one side of the shaft, but both sides (and, given the fact that the springs are so carefully calibrated, this really means that there will be virtually no force applied).

Given the way the springs are set up, there is no additional power available, but rather the power that is available will have a very smooth power curve, much smoother than it would have with a single mainspring. Given, however, the design of the system - two springs fitted over each other - there is one drawback: the mainspring is slightly thicker than two "normal" mainsprings stacked above each other (an alternative would be, like Favre-Leubre did back in the day, to put two barrels in parallel, but this doesn't address the mainspring barrel torque problem), making the watch thicker than it would be otherwise. In this case - no pun intended - the choice makes sense: the removal of a long-term maintenance problem and a power curve that is extremely smooth.

So, let's review what we have so far: a self-regulating balance wheel and a mainspring that reduces wear and tear while have a very smooth power curve. What's not to like? J
The hairspring is next: this gets a tad technical. There are basically two ways of attaching the hairspring to the balance so that it can smoothly and reliably "breath". One is to simply glue or fix the hairspring with a pin to attach it (relying on this to adjust the position of the hairspring to meet design specifications); the other is to attach only the end of the spring, relying instead on using a hairspring that is carefully designed and manufactured and hence can be attached "freely", i.e. a free-sprung hairspring. While there are good arguments for using the latter (since it reduces mechanical complexity and makes design work much simpler), using a free-sprung hairspring means that you can't adjust the watch by adjusting the position of the hairspring (fairly simple and usually done in-place), but rather have to adjust the balance wheel instead, which, as the watch ages and needs to be re-regulated, is considerably more work to achieve basically the same result.

The problem with using a "normally" attached hairspring is that by attaching it the hairspring itself is deformed and damaged, which, even if the damage is very small, still negatively affects accuracy. If done poorly the hairspring will have to be replaced (or will fail to perform to specifications) and the greater the damage, the more difficult it will be to have the watch be even a fair performer, regardless of the original quality of the movement. Heritage uses what they call the Tenere system, which, unlike classic solutions of gluing or using a mechanical pin, uses a mechanism that resembles that used in a free-sprung balance, clamping the hairspring between two surfaces and leaving it completely undamaged. In contrast, though, to a free-sprung balance, the length of the hairspring can be adjusted when the watch is regulated: the regulation mechanism itself is also unique. A different mechanism allows the shifting of the effective length of the hairspring and its geometry (determined by the fulcrum points of attachment), allowing the regulation mechanism to adjust for deviations due to positioning, as well as to finely adjust the symmetry of the hairspring itself.

As mentioned, being able to fine-tune the escapement is also critical: this is where the power transmission takes place between the mainspring and the hairspring, with a constant supply of power needed. Mainsprings, due to their nature as mechanical storage devices of energy, have a power curve, starting out strong and weakening over time. One might think that the alternative is then to use as little of the power per beat as possible, but that reduces amplitude of the hairspring beat and hence accuracy. Put simply, the greater the oscillation amplitude of the watch balance, the more accurate it becomes, but this requires more use of energy, not less, as you need more power to regenerate the oscillation cycle. Automatic watches tend to do well here, as constant wear generally means staying at the top of the power curve, only winding down at night when the wearer sleeps, but once you take the watch off, the power curve comes back into play as the watch winds down.

Rather than try to fight this, Heritage took a different tack: the balance isn't driven by the mainspring directly, but rather indirectly, using a small spring between two anchor wheels. When the pulse is sent to the balance, the mainspring re-tensions this intermediate system, but since this is, essentially, a single, almost binary act (either the mainspring has the power to re-tension the spring between the anchor wheels or it doesn't), the power curve becomes significantly less important. Heritage calls this the Sequax escapement. The anchor wheel that is powered by the intermediate spring, driving the balance, actually doesn't have any gear teeth, removing torque concerns that are transmitted through the transmission train. This means that as long as the mainspring is wound every day, regularly, this system results in the balance oscillation being perfectly constant, providing the foundation for watch accuracy. While getting into the design further would require quite a bit more than I'm willing to write right now, suffice to say that this escapement has two anchor wheels as well as three anchors with six pallets: the two anchor wheels are needed to both isolate and move the impulse from the mainspring side to the balance wheel side, and three anchors (and, of course, six pallets) to provide the anchor wheels with something to work with.

Now, there is just one further detail: controlling the escapement movement. Normally you use stop pins (there are other mechanisms, but they all share one aspect: they are static). By having static pins, this means two things: first that if you try to adjust the movement range provided by the pins, this always leads to irreversible deforming of the pins, or, second, that you have to use a complicated escapement design that doesn't lead to precise results, as striking points are then tilted laterally, rather than being presented to the stop surfaces of the escapement at the right angle. There may also be vertical play in the escape anchor, which then means that there are differing degrees of opening for the anchor: again, these small deviations result in worsened accuracy.

What Heritage has done here is a mechanism, called the Sectator, which precisely and separately constrains both the input side and the output side (remember, there are two anchor wheels in the escapement and three anchors, hence you have both an input and an output side!). A lever arm is used to constrain the movements of the escapement very precisely via a screw held against the lever arm with a spring. This screw can be adjusted by tiny, tiny holes drilled sideways in the screw head: these can be accessed with a special tool from both the back of the movement and the dial side. This allows the watchmaker to adjust the escapement with an unusual degree of precision, with the ability to change this at any time, reversing if need be, to prevent any deformation of the mechanism (we are, after all, talking about very, very small size parts!) when adjusting the escapement constraints.


There's a lot going on here: what Frässdorff and Heritage are doing is nothing less than systematically reducing, one step at a time, all the ugly uncertainties that watchmakers face from a design perspective, by controlling each of the variances. This means that the watch movement, in this case (no pun intended) the Heritage Cal. 880, is designed to run extremely precisely out of the box, as it were, without a watchmaker having to spend inordinate amounts of time fixing design problems that leave inherent uncertainties in the watch movement's operation.

I wish that I could provide you with a breakdown of all aspects of exactly how this works with diagrams and photos: that would make this even more unwieldy than it already is.

But here is the next best thing: a HD video of the Tensus in action.


So, I hope that this has been informative and any errors are mine entirely.

Heritage can be found here.

Oh, and a final word: the price. This is not your average watch and it is clearly aimed at the luxury watch crowd. That said, what you are getting for your money isn't just precious metals and fine workmanship, but also some fairly amazing advanced horological work that you simply can't get anywhere else.

7,751 Posts
Superb review, thanks for sharing your ideas/opinions! The way Mr. Frässdorff has gone about in a systematic manner to minimize all the possible sources of timing variability is pretty amazing. I've been eying these watches ever since Baselworld 2011, hope to own one someday ...

2,153 Posts
Sensational review. You have done a remarkable job at highlighting the technical brilliance of this piece. You have certainly piqued my interest. Does any member own one of these pieces ?


234 Posts
unreal review, nice job and thanks

3,433 Posts
Sorry for bumping this but this is just utter jibberish. Please correct me if I'm wrong, but this is the ultimate gimmick watch.

I have not considered every "innovation" but some are just too obvious to ignore.

Line art Circle Diagram

I'm not sure what problem they're trying to fix here, but they've merely replaced it with another. Instead of a force in the tangent of the arbor at one point, they've replaced it with two opposite (unbalanced!) points. Sure, if they want mainspring wobble, be my guest. The real "solution" to this "problem" would of course be to have three layers of springs, the one in the middle being as thick as the two others combined. This way the force would be balanced and thus not make the barrel wobbly (which is not even a problem if properly taken care of). In four seconds I already came up with a better idea for a patent (a garbage patent nonetheless).

Gimmick #2:
Clutch part Circle Auto part

This is quite astounding, one of the most stupid ideas put into practice. What is the point of this? It acts as a governor, but to what effect? If consistency is the aim, a solid balance wheel is by definition the best. This will just make sure that these flanges themselves start to oscillate, probably at different rates. Thus making the balance wheel unbalanced, unpredictable, jittery and the time keeping will undoubtedly suffer as a consequence.

Can't find HWM's website or information anywhere, have they gone bust? That wouldn't surprise me given their level of incompetence.
1 - 10 of 10 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.