Home Made ABS?

[fallenarch]

I have been playing around with the idea of making an aftermarket ABS system older bikes like the BUSA. I’m going to tell you all of these ideas because under no circumstances am I going to market something like this (liability, who knows if any of this will actually work). It seems that the ABS systems on most bikes are overly complex because they have pumps and separate the rider from control feel through the brakes. I was thinking to just use actuators to “correct†bad inputs from the rider.

Basically there would be 2 master cylinders (MC) at each wheel. The primary MC is the one you already have on the bike and that your control lever operates. The second MC would be linked opposite of the primary. Rather than being normally “open†and then closing to increase pressure in the hydraulic line like a typical MC, the second MC would normally be “closed†and would be “opened†to reduce hydraulic line pressure.

The secondary MC would be controlled by an electronic actuator, which would be controlled by a microprocessor. When you apply too much braking force to the lever and primary MC, the secondary MC would reduce the pressure, “correcting†your input so the wheel does not lock up. The advantage of this setup is that it could be very light, maybe a couple pounds for each wheel. Second, you would have much of the feel of the normal braking system as you are still linked to the caliper through the hydraulic line (some systems, like the Honda system you are “braking by wireâ€).

Now for the sensors. The microprocessor is pretty cheap – about $100. We will need a very good industrial linear actuator - $200 or so. I propose to create a GPS speedometer for the bike that would be a part of this system. This is also pretty cheap, probably $200 or so. Using the GPS speed indicator gives an absolute speed, independent of wheel slip, gearing changes, etc. It also gets us the ability to include lean angle in our control algorithm. We will use a standard induction type wheel position indicator to monitor wheel speed at both ends of the bike.
The microcontroller would compare wheel speed at each wheel to the bikes GPS speed. If a wheel is spinning a certain amount too slow relative to the GPS speed, the secondary MC is opened up until the wheel speeds are again consistent. It would be easy to allow a certain amount of slip, especially in the rear wheels, and you could allow some configuration by the rider here. The system could also be turned off completely, at which case the brakes would be essentially stock in their function and feel.

If we really wanted to get slick, we could add lean information. For example, if the bike is upright a small amount of wheel slip could be allowed. If the bike is leaned over, the system would act aggressively on any slip. The lean angle would also allow the system to adjust how quickly or aggressively it reacts to wheel slip. For example if you are riding in a straight line and grab a hand full of brake. The system would sense that the bike is straight up and allow some slipping to get maximum braking performance.

Fault tolerance is easy too. If a sensor is not sending or sending erroneous information (negative speed or excessive speed) the system would shut off and close the secondary MC, so the brakes would work as stock.

All of this might sound like techno-mumbo-jumbo but i'm convienced it’s possible with off the shelf stuff. If anyone is interested in playing around with it PM me. I don’t have the hydraulic experience I need to get the secondary MC part done but I think I can handle the electronics and programming.

 

[lankeeyankee]

I've messed with linear actuators and hydraulic flow controls, mac valves, Dayton timers, etc. for a throttle and clutch automated control system for drag racing.
Here is a link for Stamp Controllers
BASIC Stamp Programming Kits
Linear Actuators, This company is pretty good I have dealt with them.
Mini Actuators - Linear Actuators, Linear Motors, Micro motors, Electric Motors

 

[superbeagle]

Makes sense to me.

 

[AJAY]

Kinda makes sense / kinda doesnt.

GPS is way too slow to measure your speed. Refresh rates are about 0.5 to 1 second. You will need to measure much faster than that for control purposes.

As for the "linear actuator," isn't that pretty much how it's done already? Ok, it's done on the hydraulic fluid directly to modulate pressure. I'm not sure if you're proposing a mechanical linkage opposite the MC or not.

Are you talking about some method to eliminate the "pulsing" of conventional ABS? The pulsing is needed because you never really know what the limit of the braking force is, so the ABS takes it to the point of slipping, then backs off, then repeats the cycle. Why don't you know the limits of the braking? Because bikes have different tires, you're on different road surfaces (concrete, asphalt, tar and chip, gravel, dirt, sand, leaves), different road conditions (wet, dry, icy, oily), tire traction changes (temperature, lean angle, G-loading, front-rear wheel loading, side loading)

I think you could take an existing ABS system from a comparably sized motorcycle and retrofit it to a Busa. It would take some tuning to make work, but that would likely be your safest bet for a workable system. It would probably be the cheapest route as well.

 

[fallenarch]

I'm not going to argue any of your points because I'm not sure one way or another. I do know that slipping can be picked up with sensors. I was told that the pumping was a means of error correction. By going up to the limit of traction and releasing it's less risky than finding the point of slipping and backing off slowly as you would when you do it mannually. I don't think the more sophisticated systems pump anymore.

The GPS units I am referring to stream data continously (115000 baud). That's way faster than needed. Now is it accurate enough to work is another story.

OOPS! Turns out the GPS units update at 10 hz (10 times a second) and communicate at 115000 baud. So this might not be fast enough. However with the wheel speed sensors this might still work as the GPS is not the primary sensor for slippage. Humm - I suppose that's why know one has done it!

 

[AJAY]

You got it Arch. The wheel speed sensors pick up the change in wheel speed. A slip is detected as a very high rate of wheel decceleration. Call it a "microslip" more than a slip. I believe the Bazaaz aftermarket traction control system measures rapid wheel speed acceleration by engine speed, probably coupled with a clutch sensor, so maybe you could try a trick like that, at least for the rear wheel, but I'm not sure that could give you a fast enough control loop for braking, especially throwing in the effects of chain slop and suspension.

GPS has come a long way if it's up to 10 Hz refresh rate. That may have as much to do with processor speed to do the calculations as anything. However, even aircraft, with slower dynamics that an ABS system, do not use GPS for primary stability. GPS is used for guidance (a much slower process), but inertial sensors (gyros and accelerometers) are used for stability. (AJ - a former gyro guy)

 

[Dennis]

So what happens when the secondary mc lets enough fluid pass that the brake lever is fully pulled against the grip? This is why abs systems have a pump.

 

[ruder]

Rather work with Clicks & Ticks. Each wheel has a sensor that counts clicks. The processor counts the ticks. If the clicks per tick change either a braking or spinning condition exists. If the number of clicks have increased for a wheel it is spinning. If the number of clicks have reduced the wheel is braking. If the clicks have stopped the wheel is slipping. The more clicks that are counted per wheel revolution, the more accurate the system will be.

The second important component is an inline electric valve. This valve must be normally open to allow fluid to pass unhindered through the line. The second position is closed, blocking fluid and preventing a rise of pressure in the line. The third position is bleed, to reduce pressure in the line. If the wheel is slipping the valve must switch to bleed to reduce the pressure until the sensor counts clicks on the wheel. The valve must then switch back to open. As long as the slipping condition exists, the valve must switch between bleed and open. The faster it can do this, the more accurate and effective it will be.

The existing master cylinder can be used. The valve bleed line is fed back to the master cylinder.

The downfall with the simple system above is that brake line pressure falls. The rider will find he has to increase lever pressure continually to maintain braking. For a better and safer system you can add a lever position sensor and a pump. Once the valve sets to bleed, the lever sensor reads the position. As long as this position reads over a certain value (>80%) the valve cycles between bleed and closed, and activates the pump every closed position.

 

[AJAY]

Thanks, ruder. I wasn't sure if motorcycle systems use the full modulating valve/pump setup the way cars do. The systems are much smaller than cars, so I think the pump may be just a large modulated piston that can overwhelm the master cylinder. Here's a description of the BMW setup. BMW description of ABS2

General Description of How the ABS Control System Works:

The closed-loop control circuit consists of signal processing by the sensors and automatic control electronics which activate the electro-motor-powered pressure modulator in the event of overbraking by the wheel.

If the ABS control unit detects that wheel locking is about to take place on the basis of impulses being received from the rotating-speed sensors, it activates the motor of the clutch shaft. At the same time, the electro-magnetically controlled friction clutch of the respective brake circuit mounted to the clutch shaft is activated.

Depending on the frictional torque generated by the control unit in this clutch, the plunger piston in the pressure modulator is pulled downwards against a spring resistance by a precisely defined amount.

The resulting enlarged volume above the plunger piston results in the pressure decrease required in the brake circuit. At the same time, the movement of the reciprocating piston is recorded by the displacement sensor and forwarded to the ABS control unit, which obtains the information needed concerning the position of the piston in this manner and thus has a reference as to the relative brake pressure in the brake circuit. This scanning takes place within the closed-loop range every 8 ms and in the open-loop range every 4 ms.

The wheel is optimally decelerated just below the locking limit. This control sequence is repeated every time the wheel is on the verge of locking.

This system is the first system to enable recognition of the relative pressure by virtue of the displacement sensor of the reciprocating piston, with its corresponding position consequently making it possible to operate the motorcycle in accordance with any brake pressure curve in keeping with the respective riding conditions.

Based on the rotating-speed information coming from both wheels, this system is also capable of detecting lifting of the rear brake when the brakes are fully applied and preventing th is by correspondingly regulating the pressure in the front wheel brake circuit.

Owing to the fact that this system employs a plunger piston, the brake pressure can be decreased to 0 bar, thus ensuring dependable control behaviour on difficult riding surfaces, e.g. black ice, gravel etc.

 

[Dorin G Clisu]

I have been playing around with the idea of making an aftermarket ABS system older bikes like the BUSA. I’m going to tell you all of these ideas because under no circumstances am I going to market something like this (liability, who knows if any of this will actually work). It seems that the ABS systems on most bikes are overly complex because they have pumps and separate the rider from control feel through the brakes. I was thinking to just use actuators to “correct†bad inputs from the rider.

Basically there would be 2 master cylinders (MC) at each wheel. The primary MC is the one you already have on the bike and that your control lever operates. The second MC would be linked opposite of the primary. Rather than being normally “open†and then closing to increase pressure in the hydraulic line like a typical MC, the second MC would normally be “closed†and would be “opened†to reduce hydraulic line pressure.

The secondary MC would be controlled by an electronic actuator, which would be controlled by a microprocessor. When you apply too much braking force to the lever and primary MC, the secondary MC would reduce the pressure, “correcting†your input so the wheel does not lock up. The advantage of this setup is that it could be very light, maybe a couple pounds for each wheel. Second, you would have much of the feel of the normal braking system as you are still linked to the caliper through the hydraulic line (some systems, like the Honda system you are “braking by wireâ€).

Now for the sensors. The microprocessor is pretty cheap – about $100. We will need a very good industrial linear actuator - $200 or so. I propose to create a GPS speedometer for the bike that would be a part of this system. This is also pretty cheap, probably $200 or so. Using the GPS speed indicator gives an absolute speed, independent of wheel slip, gearing changes, etc. It also gets us the ability to include lean angle in our control algorithm. We will use a standard induction type wheel position indicator to monitor wheel speed at both ends of the bike.
The microcontroller would compare wheel speed at each wheel to the bikes GPS speed. If a wheel is spinning a certain amount too slow relative to the GPS speed, the secondary MC is opened up until the wheel speeds are again consistent. It would be easy to allow a certain amount of slip, especially in the rear wheels, and you could allow some configuration by the rider here. The system could also be turned off completely, at which case the brakes would be essentially stock in their function and feel.

If we really wanted to get slick, we could add lean information. For example, if the bike is upright a small amount of wheel slip could be allowed. If the bike is leaned over, the system would act aggressively on any slip. The lean angle would also allow the system to adjust how quickly or aggressively it reacts to wheel slip. For example if you are riding in a straight line and grab a hand full of brake. The system would sense that the bike is straight up and allow some slipping to get maximum braking performance.

Fault tolerance is easy too. If a sensor is not sending or sending erroneous information (negative speed or excessive speed) the system would shut off and close the secondary MC, so the brakes would work as stock.

All of this might sound like techno-mumbo-jumbo but i'm convienced it’s possible with off the shelf stuff. If anyone is interested in playing around with it PM me. I don’t have the hydraulic experience I need to get the secondary MC part done but I think I can handle the electronics and programming.

Your cylinder idea seems fine. I know my reply is digging up an old thread but I think it might be relevant for other people at least.

This is what I did on a bicycle for under 50$, more as a technical exercise / for fun rather than safety.

Details here: https://zenontech.wordpress.com/2015/07/23/anti-lock-brake-system-for-bikes/

 

[icemansid]

Wow thats pretty cool. Very good read on an old thread as well. I do like the idea of having a secondary MC that would relieve pressure and re-apply pressure as needed. seems that it could be easy to configure and small enough to be hidden and not too bulky. the ABS on our late model mustangs is absurd. the space wasted by the entire system unreal, not to mention the weight..

 

[Dorin G Clisu]

In fact, I think it would be less troublesome to mount the ABS actuator at the hand lever instead of the caliper. It could be just a hydraulic cylinder pushing back the hand lever directly or an electric servo, each with their pros and cons.

The reason I mounted the actuator at the caliper and not the hand lever is that the friction along the brake mechanical cable would have introduced too much hysteresis to get any useful response time. But on a hydraulic system which I suppose all "real" bikes use, the friction would be much smaller and should not be a problem.