Diy Haldex controller........Haldex Controller taken apart with pics

The unit is now further apart exposing the controller PCB within (I had to order a small Torx screwdriver which I didn't have).

The question now is which route to go down with this, in terms of do we (a) just rip out the controller board, and replace it with a simple board that just controls the stepper manually from an external box with buttons on it, or (b) attempt to interface with the original controller PCB?

To do the latter would enable a standard unmodified Haldex unit to be used, but I have zero idea how much work effort would be involved. I assume it's a CAN bus based thing, so building some CAN controller could be done, but I have no idea if anything about the interface or bus messaging is published anywhere at all. Anyone know?

The former idea means the Haldex unit has to be ripped open and the board discarded and replaced with a simple stepper motor controller. Sambo is that what the intention has always been?

 

Exactly that Trev

Rip out Pcb and either install new Pcb in or just extend stepper motor wires out of the Haldex enclosure then new controller can be placed anywhere in the car

What is your thoughts of the built in spring that drives the pin back home each time stepper motor is unpowered?

Remove or overcome its tension if possible?

 

Hi found this thread while searching the internet we have a T5 with 4 motion the haldex coupling has to much grip which is causing transmission wind up which in turn wears the tyres at an incredible rate ie 7000 miles between new tyres, we have removed the fuse for the 4 motion system and the tyres wear normally ie 7000 miles and the tyres look like new.
Has any one found a mod for the coupling i would like to reduce the power to the coupling during normal running without the loss of grip by the coupling when wheel slip is detected.
I have asked about the problem with ** but they say there is nothing wrong that is the way they are just keep putting tyres on it or if you prefer to remove the fuse thats fine by us.

 

OK I've made a bit of progress on this now.

I've used very cheap off-the-shelf PCBs from eBay to do this:

(1) An Arduino Uno board. Actually the one I have is a non-genuine clone PCB, and it cost just five quid off eBay.
(2) An L298N stepper driver PCB, again about five quid from eBay.

Interfacing between L298N board and Uno board as in the following guide: http://tronixlabs.com/news/tutorial-l298n-dual-motor-controller-module-2a-and-arduino/

Then, installing the Arduino IDE and using some of the supplied stepper motor example programs immediately gave some results.

I've done a simple example program that shows how the stepper can drive the rod back and forth. Also, to find a 'home' position, I think the best way is to keep stepping backwards a minimum number of times to ensure the spring is fully uncompressed. Then, it can be determined how many steps forward are available until the spring is fully compressed. If you step forward too many times, the spring suddenly uncompresses due to a ratchet mechanism inside.

In answer to Sam's question:

As you found, if the entire stepper controller's power is switched off then the spring forces the rod back. That's because the stepper needs to have one of its two coils energised to keep it in position.

But, it's OK as long as one stepper coil remains energised. In other words, if the driver is powered up but the stepper isn't being made to step, the rod will stay where it is. During this time about 1A or 2A is required for the motor coil. So the rod will only retract under the spring when you turn the whole system off.


In its present state the Arduino board is just running a continuous loop that initially steps back the motor to contract the spring fully as fast as it can (it steps back about 80 times to ensure fully compressed; it just sort of ratchets when it has fully compressed the spring) and then it slowly steps forward, gradually uncompressing the spring. There is a long pause at some point to prove the rod can be held at an intermediate position for as long as you like as long as power remains to the system.

The next step is to turn this demo into something practically useful in the car. Any suggestions? I could easily add buttons to the Uno board to command it to, say, fully compress or relax the spring, or go to some intermediate position. What's needed? It's over to you guys now to suggest what this controller should actually do, as I haven't a clue TBH

 

I quickly bunged a video of it working on the Facebook group by the way: https://www.facebook.com/groups/clubgti/

 

Great stuff!

Haldex starts engaging as soon as the control pin starts moving against the spring due to oil pressure building up in the clutch pack

So i guess if you power up unit and lets say haldex is disengaged as default then we need some buttons to send some drive to rear wheels as and when needed

So couple of buttons to control pin is all that is needed i guess

If you can have some sort of indication where the pin is that would be helpful but if its too dificult then its not a biggie

 

The Haldex unit has now been handed back to Sambo at Curborough for it to go back onto his car, so we are on the lookout for another Haldex unit to develop this controller against.

Control of the stepper from the Arduino Uno and the L298N board has been very trivial to do. What's needed next is a control interface and finished software to make it all work.

I looked on eBay and found that a PCB with LCD and buttons can be had for under five quid. This simply plugs onto the Arduino Uno.

http://www.ebay.co.uk/itm/Arduino-L...887?pt=LH_DefaultDomain_3&hash=item43d2390477



My plan therefore is this:

- Arduino Uno board has the above LCD / keypad board plugged onto it, and is put into a plastic box which sits on the dashboard.

- The L298N stepper driver is mounted into another small box which sits on the rear differential next to the Haldex valve unit.

- Fused high current ground and +12V supply wired to the stepper driver box.

- Stepper driver box is wired to the stepper inside the Haldex valve unit. The original circuitry inside the Haldex valve unit is discarded/unused.

- Cable then routed from the stepper driver box on the diff, through the car, to the dash unit. This carries stepper control signals (sufficiently buffered from Arduino board I hope, due to long cable run etc.) and power supply for dash control box.

- Simple bit of software to control stepper/valve position based on key presses, with LCD showing valve position. There could be functions to recall memorised positions.


The beauty of this is that if the project is completed and all published on here with the source code, I think this could be put together by anyone with a bit of electrical experience using 15 worth of parts from eBay.

 

Anything alse happen with this or did it die with the thread??

 

Thanks for bringing this back up

I've ordered a few parts just need some more info please trev

 

I've been playing around with Arduinos quite a bit recently, so may be able to help if required.

 

I found this as I hadn't looked into this for some time, maybe helpful for some to understand it's workings

https://youtu.be/F-1sfpeTFPc

 

any help would be helpful lol

 

Writing the code is going to be my downfall

 

As I'm running emerald ecu with switchable maps could i intergrate into the controller, so I could have map 3 with full power available and switching control to the haldex via L298N etc

Or am I getting to excited as I'm laying on the sofa with a rotten cold!!

 

No problem. I've used Arduinos to power a CAN dash from a VW using a CAN shield, and my next project will be using one as an EGT monitor for (hopefully) 4 EGT sensors, so this shouldn't be too difficult.

I've had a look at what's there, and, as Trev says, controlling the stepper motor is arbitrary. The tricky bit will be working out how many steps equate to 0% rear bias, 50% rear bias, 95% rear bias, and 100% rear bias. I think Trev has done some work on this already?
If we can work that bit out, we can make a start on the code.

It's what Trev and I do for a living.

You probably could (depending on what the Emerald can output), but that wouldn't be the aim of this project, it needs to be ECU independent.

 

Regarding rear bias there wouldn't be 100% to the rear would there, or was that just figure of speech so 50/50 split as the original haldex doesn't allow full lock up? Is that correct?



Thanks much appreciated

Regarding emerald it was only a question as I was doing nothing last night bored!!!

What sort of enclosure are you using? I was thinking it could mount in the stereo position in a mk2 but obviously there are going to be other vehicles it's used in

 

Sorry, I wrote that out, then lost the whole post, so wasn't as clear as I could have been. The 0 to 100% refers to the movement of the pin. 100% pin movement = 50/50 drive split. I'm pretty sure that's possible, but I may be mistaken.

I wouldn't use an enclosure as it's a race car. It would be zip tied somewhere out the way! lol

 

Thought that's what you meant

Fair enough that will do me!

 

I have a couple of diffs does anyone want me to strip the stepper motor out and see if I can count the revolutions from 0-100%, would that help?

 

If you want to send me one, I can use it to start testing. I've ordered a motor shield already, and will check with Trev if he has any info on the number of steps required. If not, I know someone who has done something similair, so should be able to get some info from him.

There's no rush, as I won't be able to look at it until after Autosport.