19 replies to this topic

[simonr]

Most of the threads about this kind of thing on the internet seem to result in the poster getting flamed for even thinking about doing something as outrageously dangerous. If you're considering flaming me - at least read the whole post and then, ideally, keep it to yourself!

I'm considering buying a vehicle fitted with ABS but not traction control and it strikes me that the ABS pump / valve block is potentially more useful than just as ABS. Also, ABS without traction control is of limited use off road IMHO.

What I'm proposing is to make a new, manual controller. When you want to go off road, you unplug the ABS ECU from the pump/valve block and plug in your manual controller. I suppose the change-over could be via a multi-pole relay, perhaps when Low Range is engaged - but I figure that unplugging / replugging does not constitute modifying the ABS for use on the road.

Most ABS units seem to consist of a pump plus two solenoid valves per wheel. One allows pressure from the pump to enter the caliper when energised and the other stops it leaving when energised. With neither energised, the master cylinder connects to the caliper. There are some additional non return valves which do the same if any of the components fail to operate.

My question is: Is the pump in an ABS unit actually a pump, or just a modulator - i.e. is it's purpose just to locally increase and decrease the pressure as it rotates which would in turn make the caliper grip then release even if you are standing on the pedal - or is it a true pump and it's the valves which are alternately connecting the caliper to the pump / exhaust to achieve roughly the same thing?

If the vehicle has traction control, it has to be a pump, proper - I'm just not sure for ABS Only?

Assuming it is a pump, is it high enough pressure to stop the wheel on it's own or does it rely on the master cylinder pressure to 'top it up'?

The ABS unit could make a pretty good set of fiddle brakes or, by reading the wheel and steering position sensors as well, an improved traction control?

If one un-plugs the Valve Block, does it cause the ECU to throw errors (The Three Amegos type thing)

So - I'm not proposing 'modifying' any of the braking components, just unplugging and re-plugging one of them when the vehicle is being used off road. This should avoid the pedestrian holocaust that other threads seem to think inevitable.

Si

[Bowie69]

I don't know much about other ABS systems, but both my Audi and my Rangie use a pump (electric/hydraulic in the Rangie, and engine driven on the Audi) to charge a nitrogen sphere which then basiclly acts as the servo, charged to something daft like 2000PSI.

On my Rangie, it doesn not have traction control, HOWEVER... I believe I read somewhere that adding TC to a Classic that already has ABS is just a case of swapping the ECU over, so this answers your question about whether the valve block is able to send high pressure fluid down the line to a caliper without relying on the brake pedal effort.

Thinking about it, for fiddle brakes you should just be able to find the right wire for the left and right rear caliper solenoids and put a MOM switch to send high pressure fluid down the line -damned sight easier than fiddle brake levers and two sets of calipers, brackets etc.....




*EDIT* It got me interested and I googled.... the valve block is slightly different when an RRC has TC..... :


1) PCRV valve moved to below booster unit mounted in inner wing
2) accumulator - has direct feed to ETC valve block. accumulator bleed valve relocated to ETC valve block.

[zoltan]

My question is: Is the pump in an ABS unit actually a pump, or just a modulator -

We've had one apart to use as part of a pressure test rig for brakes and the unit we attacked contained a pump

[elbekko]

I've had the same idea - it can't be hard to come up with a good traction control system, and since I like fiddling with things, I'd make it so it's controllable via computer (with a shiny interface, of course).

Also, you'll obviously kill a busload of nuns!

[LandyManLuke]

For what it's worth, I'm half way through developing/testing this already.

ABS block has been re-plumbed to sit in the same position as fiddles, ie 1 inlet, 2 outlets, due to an interesting issue discovered looking at the pump/block schematics.

Have sourced a connector, and test wiring loom is ready to go.

Just haven't had time to test the principle.

PIC and FETs are on the shopping list to make a controller, probably with a left/right/both joystick, and maybe PWM, if it's required/effective.

Ho hum, I guess you'll get their first Si.

Luke

[simonr]

I doubt it Luke - it's just an idea at the moment!

The initial intention is just a selection of buttons which lock individual wheels AKA Fiddle brakes but longer term, what I fancy is, if you have a steering position sensor, you can calculate for any amount of steer what the relative speeds of each of the wheels should be. You then force the wheels to rotate at those relative speeds.

The system only becomes 'active' when one or more wheels looses traction in which case it brings the speed back to what it's supposed to be. Otherwise the valves sit idle.

What I think this will give you is the equivalent of a diff-lock for cornering as well as driving in a straight line. In a situation like driving on ice where there is almost no traction on any wheel - it should give the best chance of steering the path intended.

Si

[Superpants]

The one I've had to pieces in the past was a wabco defender unit and was definately a pump. The displacement was very small as it is really designed to top up the pressure lost when activating the ABS, not provide the full pressure to run the brake system.

The way I understood it was it could get to full pressure to apply the brakes, but this may take a bit of time to do- certainly wouldn't be instantaneous.

IIRC correctly you will only get a fault on the warning lights if you unplug the ECU for the time it is unplugged. The fault will clear when plugged back in.

There was some good PDF info on the wabco website about how all of this worked and how to wire it up.

The same basic block is used in a number of applications- there are a number of different ECU part numbers for different vehicles, and I too remember being told that traction control is implemented in some.

[MECCANO]

Same here, looked into this.

You probably already mentioned it but the solenoid/valves are all fail safe, so you just end up with standard braking in an un powered position.

Wabco ( or how ever you spell it) have a whole load of manuals on their systems that give you flow diagrams, wiring schmatics, pin numbers etc... if you haven't already located them. It only tends to be 4wd's that have 8 solenoids. i think most front wheel drives have 6 with the rears paired.

I think the pump maybe there to also help retract/ release the brake pressure from the caliper quicker when the system registers a skid, but could be wrong.

There is also a sprung loaded resivior damper to smooth out the shock impulses from the rapidly alternating valves in there some where.

[q-rover]

The ABS unit could make a pretty good set of fiddle brakes or

But isn't that what TC is really?

[Steve_M]

It may be worth asking on Outerlimits4x4, there was a defender on there built for Tuff Truck called 'lockless' that had a home made traction control system.

[GBMUD]

if you have a steering position sensor, you can calculate for any amount of steer what the relative speeds of each of the wheels should be. You then force the wheels to rotate at those relative speeds.

DSC (other names may apply) works like this using steering input as well as wheel speed (and probably giros, accelerometers etc.) to register and correct understeer/oversteer as well as just doing traction control stuff. I expect this is preaching to the choir...

Chris

[A Twig]

The initial intention is just a selection of buttons which lock individual wheels AKA Fiddle brakes but longer term, what I fancy is, if you have a steering position sensor, you can calculate for any amount of steer what the relative speeds of each of the wheels should be. You then force the wheels to rotate at those relative speeds.

The system only becomes 'active' when one or more wheels looses traction in which case it brings the speed back to what it's supposed to be. Otherwise the valves sit idle.

That sounds like a fun project, X-Eng dont have a job vacancy for an electronics engineer by any chance do they?

[Ed Poore]

Not trying to put anyone off (I'd have a go if I had the opportunity) but does anyone honestly thing they can better Land Rover's own traction control systems?

Haven't tried a Defender with TC / TerrainResponse but did get to drive a Disco4 last year and it's truly incredible what the traction control on that thing can do.

[Anderzander]

The Defender's traction control is very crude compared to the Disco set up.

[RobotMan]

Not trying to put anyone off (I'd have a go if I had the opportunity) but does anyone honestly thing they can better Land Rover's own traction control systems?

Haven't tried a Defender with TC / TerrainResponse but did get to drive a Disco4 last year and it's truly incredible what the traction control on that thing can do.

I agree the newest systems are astonishing. I had a go in the new Rangie earlier this year and was well impressed. The reality is that a lot of us can't afford the newer ones certainly not as rough and tumble off roaders. If a kit could be put together for similar costs to that of a locker or such aftermarket toys then it becomes very interesting to bolt to say a challenge motor or such.

[TheRecklessEngineer]

Not trying to put anyone off (I'd have a go if I had the opportunity) but does anyone honestly thing they can better Land Rover's own traction control systems?

Haven't tried a Defender with TC / TerrainResponse but did get to drive a Disco4 last year and it's truly incredible what the traction control on that thing can do.

Yes - mostly due to the constraints of cost and development time required for OEM systems. Look at Megasquirt as an example.

[simonr]

Not trying to put anyone off (I'd have a go if I had the opportunity) but does anyone honestly thing they can better Land Rover's own traction control systems?

Err....Yes!
I had a long conversation with a Software Engineer who worked on Traction Control Systems (I can't remember who for though). The thing you have to realise (which to be honest I didn't) is that you have many other considerations in addition to the traction it delivers. The most important are:
Economy - if your brakes are constantly partially coming on to assist traction, you are loosing energy.
Brake Pad Life - If you are prepared to change your pads every couple of months, you can do a much better job.
The latter of these obviously impacts on safety as no pads means no stopping!
Noise - ABS & TC make a racket!
Reliability - the more the TC is operating, the shorter time it will last.


If you are building a purpose specific vehicle where neither of these is a consideration - then yes, of course you can make something which performs better!

>That sounds like a fun project, X-Eng dont have a job vacancy for an electronics engineer by any chance do they?
Much as I might like, I doubt I could afford you!

Si

[inaine]

so use it purely to keep the rear or front or both axles rotating at the same rate. so bulit in difflocks and fiddles, all on a neat control panel!

[A Twig]

>That sounds like a fun project, X-Eng dont have a job vacancy for an electronics engineer by any chance do they?
Much as I might like, I doubt I could afford you!

Oooh, I come very cheap

Well, if you need someone for some contract work or whatever give me a shout!

[calle-fas]

I really like the idea of "steering assist". Really useful on sideslopes when lockers will do more harm than good.