Piano World Home Page Forums Our Most Popular Forums Piano Tuner-Technicians Forum Polishing metal capstans in the field

Joe, I just completed a capstan polish in the field using a variable speed hand drill, a small buff wheel, and the Brasso that Johnkie mentioned.

The variable speed allows the polishing without spraying the brasso all over the action, keyboard and everywhere else.

So Dan ........ how did they end up ....... like 88 mirrors with 88 Dans ?

Best wishes - John

Thanks, Dan.

I'll try this first at the bench. to get the hang of it.

Thanks,
-Joe

Pretty much the case Johnkie, although I would have to state that we could all do without another 87 Dan’s…one is more than enough to tolerate I believe…actually, seeing as there is the original version already that would be minus 88...my math is no good...no wonder I can't do this job....
Cheers,

Joe,
squeeze some out of the container and leave it on the bench for a while so some of the water comes out and it gets a bit thick. That way it doesn't spray too much off the wheel...

All:

My piano is 42 years old. And even if it had a 50 year warranty, there'd be no hope in running back to the manufacturer either, as this was the only South African piano manufacturer and closed shop more than 20 years ago... To think that Ibach was their flagship brand - I wonder what the cheaper products looked like.

I did, however notice this: the heel felts sitting on top of the marred capstans are not really worse for wear compared to the others. The indentations and the amount of felt dust are more or less the same. I seem to remember someone here (Del?) mentioning that there's very little lateral movement between the capstan and the felt/cloth in the first place, i.e. very little friction.

Joe:

I don't mind spending time on my instruments. To the contrary, I enjoy it and it's therapeutic.

John:

The bridge pin placement is just as bad... Almost unbelievable that Ibach actually granted this factory a licence to manufacture them locally. But it sounds nice nevertheless, and the Renner action drives nicely too.

To remove burrs try a NOGA NG1700. I have used this tool for years! The best!
All the best!

Hi Mark,

There's nothing wrong with filling the gouges as a learning experience. In fact, this could even be an appropriate repair for certain museum-quality pianos where you want to keep as much original as possible.

But, I do think it's worth a quick look at the Schaff catalog to see how much the replacement parts wholesale for. In fact, since I think you mentioned that other capstans also have burrs around the perimeter, perhaps a new set of capstans might be a good investment to maintain consistency from note to note.

Whatever you decide... Best Wishes

-Joe

changing capstans one may have the luck to get the same thread and thickness, which is not always the case - a few keys could be doweled and drilled again, but it is not as easy as it looks to get the same position and square bore.

Indeed the friction is not that high on the surface, (if the action is at the correct height and the keys too) but the burrs on the edge if the contact the cloth are adding much. ALl depends of the shape of the whippen heell then.

It is easy to put a little felt marker or carbon paper ink on the top of a few capstans to locate where the friction is, if you cannot look at the movement from the side..

Finally we get to the reality of the situation. Yes, those capstans pictured above look terrible and the company installing them should be—but probably wasn’t—embarrassed. But the bottom line is that it hardly matters in the real world.

Unless the key-to-action interface is really screwed up there is virtually no sliding motion between the capstan and the capstan block (wippen heel). If there is any detectable sliding motion it is an indication that the geometry is badly off and needs to be corrected. And when this is the case you will find the capstans already nicely polished.

ddf

Well Dell, only on adequate height instrument with also adequate setup. On small pianos with no capstan heel ther is friction.

On the tall ones there can be much also (front to back motion of the capstan)

Capstan shape differs, in the book "the key and whippen" those motions have been deeply analyzed and indeed in a good setup the sliding motion can be reduced much. in any case it is never something large, due to the concordance of the profils.

But on one side you have a 12 cm lever, which is bringing in motion a lever approx half this size. The profiles may match, but one is moving faster than the other.

WHat state very finely Mr Pfeiifer is also that, due to the foces and the fact that one side of the profiles is cushioned, whatever the original shape or design is, when in motion, the soft part of the 2 profile have always a concave shape which is taking the shape of the hard one.

He sketched the paths on the conjugate levers contact point, and the sliding friction/motion is well present..
His conclusions are exactly at the opposite, he demonstrates that no motion between key and whippen take place without a slide path (which can be large or extremely small)

Now is it so important as we think ? certainly no, but we are having the habit when trying to undertsand the action, to consider parts as if they where non elastic and suffer no deformation, and it is not the case.

WHat the design points are providing are ways to minimalise the friction, to conjugate the levers variation during the stroke so they help acceleration or adsorb better the friction moments during the stroke.

As a recommandation he state that if basics are respected one may just have a rounded part on one side, and an adequate plane on the other, and that should suffice. He find more problems with less than optimally placed line of centers, or (on verticals) jack center deported a lot from the direction of forces, hence giving more effort to the whippen center.


As you state you find the capstans polished (and the cloth in dust) and you feel it much under the finger, if there is too much motion.

AN easy experiment is to move those wooden capstans that are used in tall vertical actions, and change the leverage that way. At the same time a friction and an acceleration change is well perceived, as the line of centers is passed sooner or later.

Rules and ideal situations are mostly theories, but I agree you are right pointing that.

Now polished capstans AND nice cloth are felt under the finger.

I like to add a layer under the cloth if I dont have budget to change it, then the whippen height is back, this only provide often a very nice change in touch (raising the back of the key is in the same kind of idea)

I recently finished a project in which I analyzed the key and action geometry for several vertical pianos ranging in size from 114 cm to 131 cm. (I did several grands as well, but this thread is mostly about verticals so I’ll stick with those.)

Before the changes I made to the key balance points, capstan locations, etc., in the worst case the action had a slide path between the capstan and the capstan block on the wippen of approximately 0.1 mm. After the changes were made (mostly to the key geometry) the worst case has now a slide path of approximately 0.04 mm. In neither case—old or new geometry—is the slide path sufficient that sliding friction is going to be an issue no matter what the surface of the capstan looks like.

ddf

I agree that this is not an issue but I am surprised by the numbers, so I checked my documents

slide paths as you said are low 0.3 to 0.4 mm in a vertical, 0.13 to 0.20 in a grand. (that is, when respecting the rules)

Anyway as said Mr Pfeiffer, the soft side have a cup shape, when playing normally.

SO I believe that the better transmission is signaled by the lessening of the theoretical slide path, is not it ?

I have not been able to find any vertical piano configuration with a slide path that great.

We’re talking about the intersection of two arcs—actually, the very slight overlap of two arcs—when the action is at rest. In the actions I’ve been working with the model with the shortest key is 114 cm piano and this is the model with the longest slide path. And this is only 0.04 mm. I can’t imagine what the geometry must be like to have a slide path of 10 times that. (These numbers, incidentally, are coming directly from my CAD drawings of the actions.)

Del

I had the data from the Pfeiffer's book, I will scan you the page(s) if you wish. The demonstrations and figures are taking a whole book, where line of forces, frictions, heigh of travel, weight and paths are analysed. (it is an English version you may possibly find it at Bochinsky in Germany)

The conclusions are however similar to yours in the fact that those are so little in any case that they can be neglected (not seen as a problem) particularly the difference between sharps and white keys is considered not important.
Bad configurations are also noticed - different profile shapes and orientations.
You find method to calculate the height of pivot points in regard of the paths without CAO

I would trust your CAD program nevertheless

And I admit I could not go far in the computations despite it is standard geometry and trigonometry and the formulas are explained step by step. I mostly jumped to the conclusions, but enjoyed the demonstrations and sketches.


Pfeiffer did a very complete job on the subject

I will try to scan the most interesting computations, sketches and explanations if you like. let me know.

BTW Del could you compute me the torque in Nm provided by a 1.00 mm wire tension 800 N (or 75 Kg something approx..) on a 7.00 mm diameter pin (in theoretical perfect conditions)

Or provide me the formula , which is simple I suppose.

Thank you A friend told me alittle less than 3Nm

I have the English version of both of Pfeiffer’s books. It’s been a few years since I read through them, though.




Yes, he did, but if his formulas indicate a slide path as high as you indicated I’d question his methodology. (I don’t remember just what that methodology is; as I say it’s been years—decades—since I’ve really studied his books.) In a CAD drawing it is fairly easy to simply draw the arcs and measure. Which is what I’ve done. I’ve taken the actual CAD cross-section drawings of the actions, sketched in the arcs of travel and measured the overlap (out to two decimal points—would anyone like 13? that is the accuracy of the system).

ddf

I don't think I've ever had a formula for this. It's not something I've ever investigated.

ddf

The path he compute is the one of the contact point between the profiles, its path is S shaped if I understand well . I would be very curious to see the CAD graph if not too secret.

Too bad, I believe it is a simple formula, a point on the perimeter of a circle 4.00 mm for half the diameter, 75 kg or 800 N force, what is the torque ?.

I suppose it is a simple translation taking in account the lenght of the lever, and providing the correct measure unit for torque (Nm)..

I find T = F*dist 800*0.004 = 3.2 Nm , similar to what my friend told me. Due to the lenght of the pin and the fixture, the torque is applied to more than one direction, (and not as if the pin was a circle or even a cylinder fixed in its center, we will try to obtain a decent computation of the deformation of the pin (he know where to ask )