Electro-mechanical Brake Actuator

I'm working on a next-generation solar car and am investigating electro-mechanical braking.
Reasons are numerous but include the ability to integrate ABS for
maximum effectiveness of brakes and to preven tspectacular results during vehicle performance testing at scrutineering. :-)
The approximate actuator force required is 2kN if the numbers on the back of the envelope are right.
Low power consumption and mass are primary constraints; after the obvious safety one!
A geared motor with worm drive uses no power unless it's moving, so that's good.
Coils will require much more current but can act faster.
I'm open to suggestions as to the type of actuator to use.
Force feedback on the pedal would be wasteful of energy. I'm thinking of using accelerometers to measure actual deceleration and match that to the pedal position by varying the braking force; at least until the onset of brake lockup.
A progressive spring on the pedal should give enough "feedback"; though I can imagine using the dive under braking to increase the base return force at the pedal.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
I'm missing something. An average hydraulic brake might apply 800 psi to a slave of about 1 sq in piston surface area. That's human power. Modulating that braking with ABS involves bypassing the slave. This seems like a lower power faster method than your electromechanical mode. So I am missing some part of your requirement. Wonder what it is?
Brian W
On Wed, 04 Feb 2004 14:25:09 +0800, Bernd Felsche

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

[snip]
Sorry; I don't think in psi... the piston diameter will determine the force if hydraulic pressure is used as a guide. A quick check on pressure conversion and typical area puts your clamping force at about 10kN - which is by my reckoning is way too much in this application because of virtually assuring wheel lockup whenever the brake is applied.
The clamping force on the caliper needs to be around 2kN for the desired braking effect.

ABS requires a secondary power source to re-apply pressure when the wheel starts psinning again. That means some sort of pump and pressure accumulator.
That pump and accumulator weigh quite a bit; I doubt I'd be able to get anything under 2kg in mass. That'd be almost 1% of kerb mass and a 1% increase in rolling resistance; all the time. 1% extra energy required.
I'm not even considering the electrical power consumed by the pump because I have a hunch that the electro-mechanical brake will consume about the same amount of power as the ABS pump.
The actuator I'm looking for isn't hydraulic or electro-hydraulic, but electro-mechanical; e.g. a fast stepper motor on a worm gear.
Conti-Teves and Bosch have been working on electro-mechanicalk brakes for a number of years now, but the literature appears scant on detail.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Wed, 04 Feb 2004 21:58:09 +0800, Bernd Felsche
////

///
800 pounds per square inch bearing on one square inch is 800 pounds so 800 lb X 0.455 kg/lb X 9.81 newton/kg = 3600 N about....
Take a look at the usual hand lever operated hydraulic master on a motorbike, feeding a disk calliper/slave cylinder.
If this won't come out light enough, a Bowden cable is the next lighter approach, I suppose....
Brian W
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Sorry; I didn't notice the tiny piston size in your posting. Caliper pistons are usually a bit bigger than 1 square inch. The piston's I've seen have been about 40mm diameter for a motorcycle.

Doesn't provide ABS. Bowden cable is also NOT reliable at the sort of forces required to stop a car unless it's very strong cable and the lever is moved a long distance (as in hand-brake lever).
The reason for ABS is because the tyres for a solar car are very thin to minimise hysteresis (rolling resistance); not much rubber at all so any significant lock-up destroys the tyre immediately.
P.S. Please use Usenet standards when quoting. Putting /// in quoted text is very unconventional.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Thu, 05 Feb 2004 12:30:30 +0800, Bernd Felsche

Hmmm...so bowden cable isn't reliable enough for stopping cars except when it's used for stopping cars. Gotcha! :-)
Good luck with the project!.
Brian W
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Parking brakes are only used to stop a car in an emergency; when main brakes have failed.
0.2g stop is all that's required to be delivered by a parking brake and only from lower speeds, IIRC which means an "eternity" in an emergency from high speed. Not that the solar car will be going very fast; but it'll still have to do better than typically 0.38g (average) stops.
We had enough trouble meeting that requirement with the non-designed brakes of an earlier vehicle.
Typical hand-brake movement is 10 to 20 cm, which results in about 1 to 2 cm of cable movement. Foot brakes can work with less "stroke" because the force applied can be greater.
Cable stretch with repeated use is problematic partly because of the limited effective displacement of the cable.
Does anybody remember cable brakes in the early VW Beetle? I've heard nothing good about them; requiring constant adjustment.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Bernd Felsche wrote: <snip>
Check out go-kart brakes
Check out hydraulic brakes on motorbikes.
Do not go near mountain bike disc brakes.
I am surprised to hear you are worried about locking your tyres, I don't remember that as an issue on our solar cars (there again we retired with brake failure once). If you are braking all wheels then use a brake proportioning valve (or some other scheme) to keep the force low in the rear, then you'll get understeer if the dummy locks them.
If you were entering the World Solar Challenge Engineering Excellence Race then it would be worth developing an ABS, maybe. If you are desperate to have ABS then bear in mind that ADAMS comes with an ABS module which you could use to set up your control algorithm.
Of course if you are feeling cunning you could use your drive motor to prevent that wheel from locking up, sort of traction control in reverse.
Cheers
Greg Locock
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Will do.

BTDT.
Some are OK... well; I've only found one that's really a candidate and that's the Magura Gustav M which has a steel disc and is the only one approved for tandem MTB.
As with the motorbike hydraulics, it doesn't make ABS easy to implement. I've download a few SAE papers to absorb over the next couple of days.

We definitely locked up; real good. Finished a bit sideways. If you were there in October, you might be able to figure out which vehicle with the cheeky grin did that! :-)

Rear brake will be separate; Bowden cable acting on drum. Fail-safe in case of front brake failure. Yes; I know that that presents a stability issue if the driver gets too heavy-handed.

ADAMS? Wassat? I'd bear it in mind if I knew what it was. :-)

Yes... and perhaps snap the chain or break the drive-shaft in the process. :-( Well, maybe that's an exaggeration, but braking torque can be far greater than drive torque from a "3kW" motor.
I had thought of it. Great if you have a wheel motor. Our total budget may just buy a wheel motor.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
I've got it! The low-mass, low-energy anti- flat spot solution, that is.
Assuming you've got spokes or wheel teeth or barcode on the rims, and optical or magnetic prox sensors or something to detect them, you use simple hydraulic brakes, and program a PIC to watch the bars and stuff go by, and when it detects an incipient lockup, say as a rapid reduction in edge rate, it just pumps a few milliamperes through the driver's thigh muscle to reduce the applied pressure.
"Here, son, shave your leg and apply these, er, anti- lockup patches.."
Okay, there might be some regulatory hurdles if you want to put it in production, but you're just making the one car, right?
If there's no existing rule that bans it, there will be the second time you try to use it, but it's good for one race...
;-)
-Mike-
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
If this vehicle has a human cargo, or operates in the presence of people then fail safe is mandatory. You will need proportional control of the force, unaffected by wear. If you skid and start facing sideways your accelerometer will get the wrong idea. For ABS, you monitor wheel speed, and look for the very rapid deceleration of a wheel locking up. If you went for electrical actuators, you would have to design a spring applied break with electrical relive, so designed that cutting the electrical supply would allow the spring to apply the break.
Jonathan
Barnes's theorem; for every foolproof device there is a fool greater than the proof.
To reply remove AT

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
{top-posting fixed]

[snip]
I'm allowing for independent brake circuits for power; and some NiCd cells near the actuator to allow for a couple of stops if main power goes.

In the first instance, I'm controlling via the output variable, which is vehicle speed change.
That is what you'll get by equating brake pedal travel to deceleration. There's the proportional bit; unless you're interested in making a system overly complex by trying to compensate for the dynamic effects ahead of the corresponding error signal (deceleration != set-point) being detected.
A fail-safe mode of the controller will have to determine nonsensical data from the accelerometer and fall back to a mode where more pedal pressure means brake harder and less pressure means brake less. Similar to what you have in hydraulic brakes in a conventional motor vehicle of the past century; sans pedal feel.

Only if there's only one accelerometer. There have to be at least two in order to account for the slope of the road. If there's any camber on the road as well, then you need three. A yaw-rate sensor is desirable.
Acceleration in the original direction of travel cannot be discontinuous. It means working out two main vectors; that of gravity and that of vehicle velocity change. If the initial vehicle direction changes, then without compensation, the system will tend to apply more braking force because the deceleration component in that direction will be reducing.
Also remember that there's a driver on board. It's not an autonomous vehicle.

Yes. You can also look at the deceleration to check that it's not beyond a level where lockup can be expected. And you can use the acceleration figure, integrated over time, to determine vehicle speed during braking to control the maximum amount of slip. Those are all options.

No need for that. Conventional hydraulic brakes in motor vehicles have no such _undesirable_ facility.
Redundancy and reliability is how safety is achieved.

Consider that if a car had failsafe hydraulic brakes that locked up on line failure; on one wheel at speed.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

As you like :-) ( I have sniped a bit )

what movement is required

Very heavy, and how is the rotary output used to produce the force on the break pad ?

A couple of stops ?? if you louse power the car must stop and stay stopped till the problem is fixed.

No it's not... your output variable is wheel torque resistance. this may or may not effect vehicle speed, the relationship between vehicle speed and applied break force is not predictable in the real world.

the relationship includes tyre grip, slip ratio and lots of other terms you would be hard pressed to measure under controlled laboratory conditions. in the real world you have no chance.
There's the proportional bit; unless you're interested

Real smart.... just like a computer game, except " game over " is a bit different.
and fall back to a mode

BIG assumption that the vehicle is travelling in the direction it's pointing

a BIG button to tell the computer to F**k off, lets hope the driver remembers to press it in an emergency..... but then how does it link the break pedal sensor to the actuators ?

Lock up expected... How the hell do you know,,, a little patch of water on the road and your system gets it all wrong.
And you can use the

You had better armour all your signal cables to be as tough as hydraulic lines, make any joint very waterproof, and have redundancy.

You forgot mechanically robust, and in the case of electrics, able to keep water out.
Electronics have curios failure modes, unpredictable, and software is worse.

Look at the air breaks on lorries, have you seen the great streaks on the road a trailer with a failed line leaves.
Hydraulics on cars are both dual system, and required to have a mechanical back up ( hand break )
-- Jonathan
Barnes's theorem; for every foolproof device there is a fool greater than the proof.
To reply remove AT
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Very little; a few millimetres.

Doesn't have to... After a lot of googling; I'm glad I didn't have to invent it all:     http://www.estop.de/en /
The trick is reducing the braking force when required; for ABS or simply to allow the car to move. A few SAE references:
2002-01-2582 eBrake(R) - The Mechatronic Wedge Brake 2003-01-3331 Modelling and Validation of the Mechatronic Wedge Brake 980600 Modelling and Control of an Electromechanical Disk Brake 1999-01-0482 Clamping Force Estimation for a Brake-by-Wire Actuator 2003-01-0325 Wheel Slip Control for Antilock Braking Systems Using     Brake-by-Wire Actuators 2001-01-2478 Electronic Braking System of EV and HEV -     Integration of Regenerative Braking,     Automatic Braking Force Control and ABS
The first two papers are available from the eBrake site.

Hopefully the power won't get loused. :-)
The car must not stop immediately. That's a very important safety issue. The car must still be able to stop when all other power is gone, but when the driver wants it to stop.

See paper 2003-01-0325 which is based on experience with the Conti-Teves system.
Aiming for a fixed slip amount appears to be a good strategy. Determining the amount of slip by comparing vehicle speed (given by accelerometers) against the wheel speed produces a slip figure that is the controlled variable.

Apparently not.

Why? You cannot trust all sensors, all the time. Redundant data, even of lower quality can be used to determine if a primary sensor has failed.

Three axis sensing is envisaged; though Conti-Teves seemed to make do with two.

By ignoring the accelerometer and ABS circuits. Pedal position is simply used as a command for brake pad position.

In that respect, it functions like all other ABS.

Independent circuits, yes. Brakes on each wheel for sure. One mechanically actuated by the driver.
Hydraulic lines aren't very tough... Routing signal and power lines so that they aren't exposed to the worst that is dished out is part of sensible design.

The solar car doesn't drive in the rain. There's HIGH VOLTAGE everywhere.

Ahem... last time dwon the centre, we had a brake pad fall out over a stretch of 60 kilometres somewhere North of Alice Springs and the driver had no warning until brakes were required.
Software and electronics are as bad as you make them. If you don't apply Microsoft quality metrics (if it sells, it's good enough) and instead go for paranoid programming only essential functionality in well-defined building block that are tested independently, then software can be boringly reliable.
Just like mechanical systems.

And massive blow-outs leaving bits of rubber on the highway over several hundred metres... not something you can afford to do in a 3-wheeled solar car running on tyres with less rubber than a condom.

Not dual system; independent circuits operating on different wheels. You lose braking ability when one circuit fails.
P.S. Can you please learn to spell "brake"?
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
"Bernd Felsche"

And then to re apply it when the wheel has regained speed

stopped
Things do tend to get loused [ up ] , and with brake systems hope is not enough :-).

O.K., but should the system allow the driver to re start the vehicle without repairing the fault, and how about warning the driver, If he is relying on the ABS, and it's not operational.......

This works well, but it's fouled up if you are tracking speed with accelerometers and you spin, ( assume a 360 spin, your record of forward speed is well out ).
For your purposes where a 360 spin would probably shred all your tyres it's maybe not that important.

??? are you agreeing with me, that in the real world you have no chance, or disagreeing with me that there are to many uncontrolled factors to predict the force.

Where are you getting the redundant data from ? what is measured ? A yaw sensor would go a long way to resolving my reservations.

pad position.... will this be related to break force in a consistent way ?

Most ABS systems only monitor wheel speed, and look for the wheel deceleration as the primary trigger for reducing the braking effort.

It's surprising how often common sense is in short supply.
I have seen a hydraulic line on my car rub through to it's braid, if it had been an un-armoured cable it would have gone through to the cores... O.K. so the rubbing should have been picked up earlier, but that is the way life is.

Is the driver in a rubber suit in case of an unexpected shower ?

Yike !! :-(

This assumes you program for all conditions...

Split circuit and mechanical back up....

*Another* post it note on the monitor....:-( Dyslexia rules K.O. -- Jonathan
Barnes's theorem; for every foolproof device there is a fool greater than the proof.
To reply remove AT
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

No trick in re-applying. It's the same as applying.

Lack of ABS isn't a safety issue unless the driver doesn't know about it. If the system can figure out that ABS functionality isn't "sane", then it can equally warn the driver of that. That's not rocket science.

Don't spin. :-) _Why_ would you spin?

No; it's not important.

I'm not agreeing with you. And I've given the reason why. Conti-Teves have evidently undertaken significant real-world testing that's published in the cited literature.
If you have some publication to the contrary, then do let me know about it.

Didn't I initially mention a yaw sensor? It was only after reading the literature that I put that idea on the shelf.

The way that standard, "dumb" hydraulic systems work is by the stroke of the piston in the master cylinder resulting in proportional stroke of the slave-piston/piston in the brake.
That is the _simplest_ model for "fail-safe". Attempting to sense or estimate braking force introduces complexity.

The trend is to use total vehicle dynamics for ABS. Especially as this provides "free" stability controls.

Tis a rare quantity indeed.

At 45 degrees C? No way.

The "world" as seen by individual components is constrained by the range of data that's presented at their interface. Out-of-range data, either at the input or at a calculated output, immediately flag an error condition.
The error condition comes up as a system fault and the total system goes into fail-safe mode where the driver is required to exercise more precise control.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Not true. ABS reduces the effectiveness of brakes to allow an unskilled driver to steer while braking.

The scrutineer would be an unskilled driver who is also unfamiliar with the car, right?

If you assume disc brakes.

Discussion so far clearly indicates the vehicle has relatively little mass. I'd take a hard look at a severely cut-down version of electric trailer drum brakes, with a mechanical backup pushing the same shoes. They are naturally self- energizing; a coil fixed to the primary brake shoe attracts and drags on an iron disc attached to the inside of the drum face. The resulting force on the coil housing pushes the shoe into the drum, and the reaction force from that pushes the secondary shoe into the drum. When they're new and clean, they work okay. As they wear, the magnets get covered with iron filings and dust. How long does it have to work well?
Interrupt the electrical power to modulate the brake. Caution; magnetic remanence (I think that's the right word) will keep the brake on for some time after the power is cut, limiting the ABS cycle frequency.
Given the ABS algorithm and sensor requirements that have only been sketched out here, I'd be inclined to adapt an existing control computer, rather than homebrewing one.

It also stays in position when you remove the power, which may not be so good.

Just use a _really_ strong return spring on the pedal, hook it to a rheostat, and forget the ABS.
-Mike-
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@aol.comnospam (Mike Halloran) writes:

In this case, ABS is deployed to avoid wearing through the thin rubber. I thought that was obvious from the solar car application, but maybe not to those who have not thought in that paradigm.

Irrelevant. Scrutineers *never* drive. A scrutineer would never risk damaging a team's vehicle because they could easily trash a million bucks worth of hardware.
One of the team drivers operates the vehicle during the performance testing. The driver who's quite experienced and best at braking, usually.

Yes; hopefully less than 250kg, including driver.

8 to 9 hours between maintenance. Usually little application of brakes.
I'm just a little worried about the binary nature of the trailer brakes. :-)
And a lot worried about the current drawn!
Then there's the mass of the brakes to consider. I don't think our little team is going to have the resources to bake our own carbon brake drums, so steel is the next best thing. I may well be able to shrink an aluminium casing onto a thin "braking ring" of steel (or steel into aluminium) for heat dissipation and stiffness.
Corrosion would have to be watched; but the life-cycle of the vehicle is short and it'd operate in "dry" conditions most of the time.

Existing ones rely not only on a large number of sensors that aren't exactly light, but are usually designed for 4-wheel applications. The better ones use yaw sensors and accelerometers as well.
And to cap it off, the conventional ABS is designed to operate valves and a pump that maintains pressure in an accumulator.

Good for a parking brake. :-)
A backup battery, equivalent to the ABS's accumulator holds reserve power for several full brake operations. The car would never use the reserve for motive power.

Which would mean that I'd have to risk stopping to change a tyre if a wheel locks up during braking; or use tyres that have higher rolling resistance.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

If all you want the system for is to prevent flat spots...
1, You do not need full ABS, just a system to relive break pressure when a wheel rotation stops, ( but not when they all stop :-) ).
2, An inductive sensor can pick up teeth on the edge of a disk, this can detect a wheel stop, a few extra components including a sounder a big red LED and a P3 battery would provide a lock up alarm.
An alarm system would only need to weigh a couple of hundred grams, and would have no safety problems.
Could the driver react quickly enough to prevent a flat spot with this aid?
-- Jonathan
Barnes's theorem; for every foolproof device there is a fool greater than the proof.
To reply remove AT
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Well, that's the biggest risk. I observe that in a panic that very few will actually try to steer around an obstacle to exploit the strength of ABS; rather the exploit its weakness which is the greater than ideal stopping distance.

You mean vent the fluid? I can't think of anything more frightening than the pedal suddendly going to the floor under emergency braking.
Well, nothing more frightening pertaining to braking:-)

A stop is already too late. You have to reduce braking when the wheel decelerates too quickly.

You mean pump the pedal like crazy if it goes all way to the "floor"? Hmmm... do I have enough people who think ill of me?
Yes; the method could work but it won't be popular.
--
/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
\ / ASCII ribbon campaign | I'm a .signature virus!
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.