[johnlear]

Another brake related thread has caused me to become quite curious about the ‘self energizing’ effect, so I’ve been giving it some thought...

The diagrams below are representative of GTA pads, i.e. the pad length is 138mm and the thickness is 16mm (including backing plate). This is just an example, the fundamental principles apply to all brake pads.

The following is only my personal understanding based on my own analysis, FWIW, starting with some diagrams:

Apologies for diagram quality, something makes them a bit fuzzy when exported from the native program...

All disc brakes generate a relatively weak but not insignificant self energizing effect which acts to increase the force with which the pads are pressed against the disc faces. This effect is inherent and ‘automatic’, and could also be described as being an ‘auto-servo’ effect because to whatever degree it reduces the pedal effort required to achieve a given braking effect. The effect is geometrically generated at the pads and creates force that is manifested unevenly over the pad area, more strongly at and near the pad leading edge and less strongly farther from the leading edge, i.e. the effect is stronger at / near the leading edge but very much weaker at / near the trailing edge.

The pad force created by this effect is additional to the piston force. The effect is broadly similar to the self energizing effect generated by ‘leading shoes’ in drum brakes. The effect in leading shoe drum brakes is far stronger than it is with any disc brake (the geometry which creates the effect is a lot more extreme with drum brakes), being the reason why drum brakes typically do not require artificial force amplification, whereas disc brakes typically do (other than in extremely lightweight cars).

With disc brakes the effect is created by friction induced drag (between disc and pad) being resisted at a point which is physically offset from the source of drag (i.e. offset from the disc face). This creates an effective lever arm which redirects some of the ‘drag force’ more or less laterally (this effective ‘lever arm’ is represented by dimension ‘C’ in the diagrams)..

The self energizing effect may be relatively weak with disc brakes (compared to drum brakes), but still strong enough that the pads are not evenly pressed against the disc faces, even when ‘staggered’ size multiple pistons are used (i.e. smaller pistons near the pad leading edge and larger pistons near the trailing edge). This causes uneven temperatures over the pad area, and in turn this causes uneven wear, manifested as longitudinal taper wear (i.e. greater nearer to the leading edge, and lesser nearer to the trailing edge). This tapered wear in turn creates free play and sponginess within the caliper, manifest to the driver as free play and sponginess at the brake pedal.

One end of each very rigid pad backing plate bears against the caliper body (represented in the diagrams as points ‘A’), being where rotational force created by the disc / pad friction is resisted (i.e. where the pad is rotationally constrained in the caliper). Points ‘A’ are in effect points of articulation which act similarly to a hinge, allowing a pivoting action, or would do if the pad were not wedged between the pistons and disc face.

The disc face is laterally offset from the point of resistance (A). Friction occurs at the pad / disc interface creating force that is laterally offset from point A, and so creates a force vector (blue arrow) which acts tangentially to the disc face, ‘pulling’ the leading edge of the pad inward against the disc. This is the self energizing effect, caused by friction and drag acting via an effective lever arm causing the pad leading edge to ‘push’ harder against the disc. This is somewhat simplified, because the self energizing effect acts on the entire pad face, but while it will be relatively strong near the leading edge, it will be very much weaker toward the trailing edge.

Note the angle of the blue arrows relative to the disc faces, representing force vectors. The more parallel this vector is with the face of the rotating disc the weaker the self energizing affect will be (Figure 4), and so conversely the more angled to the disc face the stronger the self energizing effect (as in Figure 3). This is entirely a function of the dimension ‘C’ (backing plate to disc face offset), and the pad length, i.e. the distance between the pad leading edge (point B) and the where the end of the backing plate abuts and ‘pivots’ on the caliper body (point C).

Note that hypothetically, if the force vector (blue arrow) were to be truly parallel with the disc face then the self energizing effect would be zero, but this is not possible, unless all of the pad material were to wear away completely.

It can be seen in diagrams 3 and 4 that the vector angles decrease as the pads wear thinner, so newer / thicker pads must have a stronger self energizing effect than worn / thinner pads. All else being equal, this might help explain why new full thickness pads seem to have significantly more ‘bite’ and generally a noticeably stronger braking action compared to substantially worn pads...

The braking effect is generated by total hydraulic clamping force added to the self energizing force (or vice versa...), so if self energizing force is stronger then the braking effect (actual retardation) will be the same for a lesser pedal effort, assuming all else to be equal. So, for a given pedal effort the braking effect will be stronger with new pads, because they will generate a stronger ‘auto servo’ effect than pads which have worn significantly thinner...

Looking at the diagrams, it appears obvious that pads of shorter length (distance between leading and trailing edges) must also generate a more angled force vector, and so tend to self energize more than pads of greater length (therefore be more prone to worse taper wear). Pad chamfers in effect shorten pad length (as seen in diagrams 3 and 4), so will increase the self energizing effect, but decreasingly so as the pads wear thinner, causing the chamfers to gradually become smaller, until they disappear altogether, and the pad reaches its' maximum effective length.

A stronger self energizing effect may well reduce pedal effort, so in at least this aspect might be considered desirable. However I don’t think it is necessarily an unalloyed ‘good thing’ in all respects. The effect is ‘automatic’ and so not under the drivers’ direct control, and therefore cannot be directly modulated (only the hydraulic clamping force can be). So while it will tend to make the brakes feel more powerful and decrease the pedal effort, the downsides may be increased taper wear and for the brakes to possibly be somewhat more ‘grabby’ in their action.

This is a fault shared by most drum brakes, which typically have a very strong self energizing effect created by the shoe leading edges and a much stronger self energizing geometry than exists with disc brakes. ‘Grabbiness’ is a real issue for drum brakes, one of the reasons why disc brakes are superior.

With disc brakes the degree of grabbiness created by the self energizing effect is probably no big deal most of the time (far less than for drum brakes), but what if the car were braking at the extreme limit (either emergency or racing), and a brake locks up, especially on a slippery surface? Perhaps it would not have if the self energizing effect had been a bit weaker...?

My gut feeling is that it is probably better to have less rather than more self energizing effect, so while pad chamfers may be helpful (supposedly..) to reduce pad squeal around town, they may be significantly counter-productive for pad longevity, for pedal travel and feel (after some early taper wear), and perhaps for predictability in more extreme use...

Regards,
John.

 

[johnlear]

I thought it might be prudent to find / provide some independant opinion to support my argument above. The following text I found here:

Brake Shoe and Pad Fundamentals

Brake Shoe and Pad Fundamentals 28.4.1. Brake Shoe Self Energization The drum type brake consists of two internal semicircular shoes, which ...

theengineersjunction.blogspot.com

"Disc Brake Pad Alignment
..........The thickness of the pad provides a small offset between the pad/disc interface and the pad's back plate reaction abutment within the caliper (Fig. 28.12C). This produces a couple, which presses the pad harder against the disc at its leaning edge compared to the trailing edge. Consequently this effect causes a very small self-energizing servo action, due to which the wear rate at the leading edge is relatively higher than that at the trailing edge."

I think that pretty much concurs. I'm not suggesting the 'self servo' action is great enough with disc brakes to have a particularly significant direct affect on braking performance (though it must have some affect, possibly enough to feel at the pedal when new pads are fitted), but is strong enough to significantly affect the pattern of pad wear and so have knock on affects due to that...

Regards,
John.

 

[alfaitalia]

I think I just about have my head around that....will read it again.....be interesting to hear @EBC Brakes views on this.

But I guess that since when I push the middle pedal the car stops that's really all I need to fully understand!!

 

[johnlear]

I've corrected incorrect diagram numbering and references to diagram numbers in the text.

Regards,
John.

 

[johnlear]

be interesting to hear @EBC Brakes views on this.

I think you will find that almost universally it will be said that a self energizing effect only exists with drum brakes. I don't think this is correct, but it isn't all that far from being so. I suspect that with disc brakes the effect is small enough that generally it is either ignored or not recognised to exist, because disc brakes are not significantly self energising, while drum brakes typically are very significantly self energising ...

Certainly drum brakes have a very strong self energizing effect in comparison (at least those with leading shoes, trailing shoes generate zero self energizing), while with disc brakes the effect is relatively quite feeble, not strong enough to make a significant difference to braking performance (why I suspect it would tend to be ignored), but strong enough to affect the manner in which the pads wear.

Regards,
John.

 

[johnlear]

Trying to roughly quantify the self energising effect to vague ball park accuracy:

GTA calipers (as above, just an e.g.) use 38mm and 42mm diameter pistons.
Area of 38mm pistons = 1134 square millimetres.
Area of 42mm pistons = 1385 square millimetres.
The difference is 251 square millimetres.

So, the smaller piston has an area that is approximately 5% less than the larger piston, so the smaller piston will exert about 5% less hydraulic clamping force against the pad than will the larger piston.

If we make the assumption that the different sizes of piston were carefully selected to result in the sum pad force acting against the disc being as near equally distributed over the entire pad area as is reasonably possible, then we might also dare to assume that the self energizing effect may account for roughly 5% of the 'missing' pad force acting nearer the pad leading edge...

I think it needs to be kept in mind that the use of staggered piston sizes in multi piston calipers can, at best, only approximately equalise the distribution of sum force pressing the pad against the disc, and so can only lessen tapered wear, not eliminate it.

The self energizing effect may begin with new pads at X strength, but will lessen as the pads become significantly worn. Therefore it doesn't seem possible that the differences in piston size can fully compensate for the existence of the self energizing effect for the lifetime of the pads, even if they may transiently do so if the pads are transiently at a specific thickness.

If the different piston sizes could fully compensate for self energizing effect, then the pads would wear parallel with the backing plate (at least along the pad length, radial taper wear is a different issue). I've never seen a well worn pad that didn't resemble a wedge, though some worse than others..

It seems to me that the only thing which can plausibly account for pads not 'naturally' wearing parallel to the backing plate (negating the need for different sized pistons) is the existence of an additional force acting to unevenly load the pads, and a self energising force is the only candidate that I can see...

Regards,
John.

 

[johnlear]

But I guess that since when I push the middle pedal the car stops that's really all I need to fully understand!!

Yes, the brakes may well slow the car quite well, but, is the pedal travel erratic and / or excessive? Is it spongy? If yes to any of these questions, then what are the factors that may be causing these problems? These things are resistant to correction unless the actual cause is diagnosed and addressed (and some of the causes are inherent in aspects of the stock caliper design). Just throwing new pads at it won't fix these problems, though the pedal may improve somewhat, for a while, until taper wear sets in again...

New or refurbished standard parts can be thrown at such problems with no real affect, done so often. But, the more that is understood of the why of how the brakes actually work, and why they may not work as well as we'd like, in the nitty gritty detail, the better we can make them, even if we have to deviate from standard specifications (standard specifications are not unlikely to be the root of the problem...).

I completely rebuilt my calipers both front and rear, new internals, new pads, new pins, new bushes, even new hoses. The calipers were as good as new yet still the brake pedal was erratically 'long' and spongy, it felt like intermittently there was air in the system, but there wasn't. The problem was the caliper pins, despite being new and operating in unworn pin bores (could still see the machining marks...).

But now, these problems have been largely eliminated due to modifications I've made, while still being fundamentally stock brakes. The pins have been modified to accept rubber O-rings which eliminate pin looseness, and other minor changes have been made. Pedal travel is now minimal, and sponginess hugely reduced. The pedal is 'right there', firm and responsive after about 2cm of travel. The brake action is WAY nicer than when in completely stock form, because the brakes are well maintained and have had an inherent design issue with the rear calipers corrected (I consider it a design 'fault', the standard slide pins are way too loose in their bores, which adds very substantially to erratically long and variable pedal travel).

I know that periodically sanding the pads to 're-parallel' the faces with the backing plates is very beneficial (i.e. correcting taper wear). I know that the front caliper bushes need to be periodically cleaned, and the pins de-rusted, polished and regreased. And now, from the thought I've recently been giving to the self energising effect, I think there is good reason, given the choice, to avoid pads with substantial chamfers in favour of pads that are 'full faced', because my analysis suggests that this will reduce taper wear.

It's the little details that add up to brakes with a consistently excellent pedal action, that are a pleasure to use. To make my brakes truly exceptional, I think all that remains to be done is to fit braided hoses, because the only significant remaining fault is that the pedal travel is too long only when braking VERY hard...

Regards,
John.

 

[johnlear]

Trying to refine my analysis of this, just because I find it intrinsically interesting...

Resolving force vectors in the X and Y axes involves intimidating mathematics beyond my comfort zone, so to help I’ve used an online vector calculator found here:

Vector Calculator

Again, to provide a real world example (but other pads will be similar in principle) I’ve used the listed and diagrammatic dimensions of a new GTA brake pad (132mm long, 16mm thick). I don’t have a specimen GTA pad to measure so I’m not sure of the actual backing plate thickness, but 6.5mm seems common for relatively larger pads, so I’ve assumed the backing plate to be 6.5mm thick.

The corner / edge of the backing plate face closest to the disc (i.e. “inner edge”) is the ‘pivot’ point defining one end of the force vector, not the centre plane of the plate as I had previously assumed (incorrectly, not that it makes a substantive difference...).

With this particular pad the angle of the self energizing force vector is 5.2° tangential to the plane of the pad / disc faces (ascertained using CorelDraw). This is a product of the lateral offset of the backing plate from the disc, and the distance from the pad leading edge to the farthest end of the backing plate, where it abuts the caliper.

----------

It should be noted that while the self energising effect depends on the pistons to press the pads against the disc, the strength of the effect is not directly dependant on the magnitude of piston force, but rather is directly dependant on the magnitude of the friction / drag at the pad / disc interface. The piston force alone does not dictate the force created by pad friction / drag, which is as much a product of pad material, temperature and etc. It is quite possible to have a weak ‘drag’ at the pad / disc interface, despite a very high piston force, for instance if the pads are too cold or overheated...

The lateral component of the vector force is generated geometrically by the vector angle (5.2° in this example), so the magnitude of the self energized force is always proportional in ratio to the magnitude of the force generated by friction / drag at the pad / disc interface. This ratio doesn’t change unless the vector angle changes, which it does as the pads wear (the vector angle, and thus the self energizing effect, decreasing as the pad becomes thinner).

The force created by pad friction / drag can range from very slight with gentle brake applications, to substantially strong with heavy brake applications, but the ratio of the strength of friction / drag induced force relative to the strength of the self energising force does not alter with changes in pad friction / drag force.

----------

So, according to the above vector calculator:

If an arbitrary value of 100 ’units’ of magnitude is assigned to the force generated by pad friction / drag, then with a vector angle of 5.2° (as per new GTA pad) the portion of force generated by pad drag which acts purely in parallel with the pad / disc face (Y axis) is less than 100 ‘units’, becoming 99.58 ‘units’. The other 0.42 ‘units’ is redirected to act laterally, which tries to push the caliper farther away from disc face (due to the angle of the force vector).

And also according to the vector calculator, another 9.06 ‘units’ of force arises, acting laterally (X axis) to press the pad leading edge against the disc, this force being separate to piston force. This is additional force generated by the 5.2° vector angle, being the lateral pad force created by the self energizing effect...

This force is not directly proportional to the clamping force created by the pistons, rather it is directly proportional to the force created by pad friction / drag. For X force generated by pad friction / drag, the self generating force will always be Y, but the piston force (component) required to create a given level of pad friction / drag is variable depending on pad material, temperature, pad and disc face condition etc.

We can’t easily define what %age of the total lateral force is created by the pistons, and what %age is generated by the self energizing effect, because we do not know (well, I don’t...) the magnitude of the force created by pad friction / drag, which is hugely variable anyway. In other words, the amount of piston force required to generate a given magnitude of pad friction / drag is indeterminate, so there is no 'ratio' that can be used in calculations.

However it is posible to calculate the relative magnitude of the self energizing effect as a %age of the force created by pad friction / drag. When this ‘drag’ force is assigned an arbitrary value of 100 ‘units’, then in our example calculation gives a self energising force of 9.06 ‘units’ which is near enough to being 9% of the force strength generated by pad / disc friction /drag.

This could very easily be a quite substantial level of self generated force, given the potentially very large forces generated by pad friction...

------------

Lets compare the same pad when worn down to 3mm pad thickness:

The offset of the ‘inner edge’ of the backing plate to the disc face decreases from 9.5mm to 3mm. The chamfer disappears completely, and so the effective pad length increases from 105mm to about 132mm (length of backing plate, or just a bit less). This causes the angle of the force vector to decrease to 1.65° (compared to full thickness pad force vector being 5.2°).

Crunching these numbers in the calculator, with ‘drag’ force magnitude at the pad / disc interface as before being arbitrarily 100 ‘units’, but vector angle being 1.65°:

The portion of the ‘drag’ force which now acts in alignment with the pad / disc plane (Y axis) increases to 99.96 ‘units’ (new pad 99.58), and the lateral component (X axis) acting to push the caliper sideways becomes 0.04 ‘units’ (new pad 0.4).

More importantly, the additional lateral force generated by the self energizing effect acting the pad leading edge (causing the pad to ‘self apply’ in some degree) decreases from 9.06 'units' to only 2.88 ‘units’, i.e. the lateral force generated by the self energizing effect decreases in strength from about 9% equivalance to the pad friction / drag force strength, to less than 3% equilvalence. So, when these pads wear down to 3mm thickness (about 1/3rd original thickness) the strength of the self energizing effect also reduces to about 1/3rd of what it was when the pads were new.

-----------

The force generated at the pad can be HUGE, so if the self energising effect is around 9% of that HUGENESS (give or take), then it can be a far from a trivial force, even if it is far less than the self energizing effect of a typical drum brake. From this, I would suggest that the very widely held belief that it is only drum brakes which generate a significant self energizing effect, and that disc brakes generate no self energising effect whatsoever, is very far from being true...

Again I’ll suggest that this could well be at least one reason why new pads can create a significantly stronger braking effect than well worn pads, i.e. because new pads have a significantly stronger self energising effect than well worn pads, due to geometric effects (i.e. strong enough to be noticeable). And, why pads wear with a longitudinal taper pattern, and not parallel.

I still think that it is probably better to have minimal self energising effect, to minimise taper wear and maximize pad area, and just maybe perhaps reduce possible ‘grabbiness’, even if this might only at most be a very minor issue for disc brakes (if it is at all...?)..

Regards,
John.

 

[Jetronic77]

I thought it might be prudent to find / provide some independant opinion to support my argument above. The following text I found here:

Brake Shoe and Pad Fundamentals

Brake Shoe and Pad Fundamentals 28.4.1. Brake Shoe Self Energization The drum type brake consists of two internal semicircular shoes, which ...

theengineersjunction.blogspot.com

"Disc Brake Pad Alignment
..........The thickness of the pad provides a small offset between the pad/disc interface and the pad's back plate reaction abutment within the caliper (Fig. 28.12C). This produces a couple, which presses the pad harder against the disc at its leaning edge compared to the trailing edge. Consequently this effect causes a very small self-energizing servo action, due to which the wear rate at the leading edge is relatively higher than that at the trailing edge."

I think that pretty much concurs. I'm not suggesting the 'self servo' action is great enough with disc brakes to have a particularly significant direct affect on braking performance (though it must have some affect, possibly enough to feel at the pedal when new pads are fitted), but is strong enough to significantly affect the pattern of pad wear and so have knock on affects due to that...

Regards,
John.

The Giulietta brembos (the revised ones) have offset pistons to reduce the uneven wear associated with this effect. Also, the pads are relatively thin compared to their length.