4WS no go :-(

[nutshell]

The story so far:

Recently taken possession of a burgundy 91 with 4WS and all the good bits.
It has been regularly maintained.

The previous owner just paid mega to have the steering rack overhauled by the local agents in Auckland (Winger Subaru) at the beginning of Feb.

HOWEVER...now the 4WS dash indicator is on almost permanently! :confused:

Winger have been servicing the vehicle since 2000 and (surprise, surprise) say they never even knew that it had 4WS! :eek:

Obviously the steering rack work has upset something.
But what?

I need to be able to intelligently advise the agents where to poke their probes so they can unravel and rectify their handiwork. :D

Any assistance gratefully received.
Thanks,
Ian

[Earthworm]

Have you pulled the 4WS codes yet?

Instructions in the SVXipedia.

[b3lha]

Hi Ian,

Got your PM. I'll reply here in case the answers help anyone else in the future.

When you say the rack was overhauled, presumably they just replaced the seals etc in the existing rack, rather than swapping in a rack from another car? I'm asking this because the front rack on the 4WS car is a different ratio to the 2WS SVX rack. Turning the steering wheel from lock to lock should be about 2.16 turns (two and one sixth).

The steering angle sensor is inside the car. Put your head down in front of the driving seat and look behind the brake pedal where the steering column passes through into the engine compartment. You should see a round metal thing on top of the column and a little black box underneath with a bunch of wires coming out of it.

As far as I know, this is the only interface between the front wheel steering and the rear wheel steering, so in theory replacing the rack shouldn't have affected this. However, the steering angle sensor needs to be syncronised with the front rack.

What I mean is that the centre position on the sensor must match the centre position on the front rack. Maybe the steering wheel was moved one or more turns while the rack was disconnected. Even though the steering wheel now points ahead, the sensor could be telling the control unit that the wheel is two turns left or right of centre. This impossible steering angle would certainly cause an error code.

I'm just speculating at present. As Earthworm said, pulling the codes is the way to start diagnosis. Try that and let us know how you get on.

This diagram may help you:
http://www.geocities.com/jamsvx/trans4ws.html

[b3lha]

One final thing. I notice that you are in Auckland, NZ. There is an SVX guru called Trevor Sheffield and he lives in Auckland too. I'm sure I remember him saying that he had some technical notes on the 4WS system. Unfortunately he doesn't come to this site anymore because somebody was rude to him (long story). I'm sure that somebody here must have his email address or contact details.

[dwolfe71]

Do the american version have 4ws??? I once read that my 94 has passive 4ws. Is their anyway that such a system can be added to the car if it is really just passive for american versions???

[immortal_suby]

Do the american version have 4ws??? once read that my 94 has passive 4ws. Is their anyway that such a system can be added to the car if it is really just passive for american versions???

just passive 4ws on the american models. its a suspension trick.
I haven't heard of anyone adapting the system for active 4ws.

[nutshell]

I'm just speculating at present. As Earthworm said, pulling the codes is the way to start diagnosis. Try that and let us know how you get on.



Thanks guys. I am sorting out the legacy (no pun intended) "park-by-feel" dings at present, and I have spoken with the agents. Once the body is up to speed they have agreed to put it on their diagnostics.

I have identified the two sensors and think it's likely they moved the wheel and reset the steering incorrectly when they did the re-alignment.

lf the 4WS fail-safe comes on before traveling too far the steering wheel is perfectly aligned for the duration of the journey. But if the light stays off for a while, then when the failsafe bites, although the vehicle apparently tracks fine, the steering wheel is cocked to the left about five degrees.

If I knew exactly what I was looking at in the wiring down by the internal fusebox I am sure I could run the codes, but a brief examination of the bird's nest wiring loom reminded me why I am not an auto-electrician. :)

Ian

[b3lha]

Checking the diagnostic codes is easy. Have a quick look of the two documents below.

http://www.subaru-svx.net/photos/files/How-To/33055.pdf

http://www.svx-iw.com/svxiw/article/?article=83

[svxistentialist]

One final thing. I notice that you are in Auckland, NZ. There is an SVX guru called Trevor Sheffield and he lives in Auckland too. I'm sure I remember him saying that he had some technical notes on the 4WS system. Unfortunately he doesn't come to this site anymore because somebody was rude to him (long story). I'm sure that somebody here must have his email address or contact details.

Make it your business to find Trevor. He is a real nice guy.

As Phil says, he was upset by some intemperate comments, and is not too enamoured of the [probably very young] gung ho brigade, who know very little, but are arrogant in their ignorance.

Trevor knows his stuff, and is sorely missed on here.

Joe

[floatingkiwi]

Hi NUTSHELL, and welcome to the best kept secret on all sides of the Tasman.. (the car and this site of course!)

My family has two of these beauties down south of you in the King Country, but unfortunately neither of them have 4WS, so I can't help you much! (then why am I posting??):D

Matt

[floatingkiwi]

By the way Ian, (hopefully I'm not trying to teach you to suck eggs here) but your comment about the birds nest at the fuse box leads me to think you're slightly in the wrong place. You've got to pull off the triangular shaped, carpeted panel just to the right of where your right foot lies. It sortof comes off in a backwards-ish direction after pulling towards the transmission tunnel first to pop off the little button that holds it in place. Once you're in there the right connector is fairly apparent. You'll see the two blue wires/probes taped up to some other part of the loom. Use either one... The large connectors are the same colours for us as they are for the yanks If I remember rightly (I'm in Europe right now). Certainly the Orange one is not the right one...

How is the 4WS (when it's working)? My next will be a 93 with everything I think. Currently a 92 with the usual Jap options - no sunroof.

Matt

[floatingkiwi]

AND (sighs..) a final post: Don't forget to register yourself on the SVX map of the world. Us kiwis are sorely represented in the downunder section.
Follow the link..
http://www.frappr.com/svx

Matt

[nutshell]

Thanks Matt. Appreciate the sentiments.
I'll try to track Trevor down.

[spinn360]

I want 4WS!!

[ensteele]

I want 4WS!!

Can't have it. :D

[ensteele]

Make it your business to find Trevor. He is a real nice guy.

As Phil says, he was upset by some intemperate comments, and is not too enamoured of the [probably very young] gung ho brigade, who know very little, but are arrogant in their ignorance.

Trevor knows his stuff, and is sorely missed on here.

Joe

I would agree with that, he is sorely missed. :(

[nutshell]

Got the car back. Bodywork now almost ding free.

The diagnostics on this vehicle are located by the throttle pedal, under the fuse box, not alongside the transmission tunnel.

Ran the Codes: came up 36 -Left/Right Lock exceeded.

As I thought, looks like the steering wheel was rotated when the rack was out being repaired.

Now the question is: how do we make sure it is centered correctly?

[b3lha]

The first thing that comes to mind is to disconnect the column from the rack, rotate it exactly one turn, reconnect and try again. With any luck you could find the correct position after a few turns in each direction. But this is clearly not the right way to do it.

You said that the light is not always on. Presumably it triggers when the steering angle exceeds what the sensor believes is full lock. So why not drive at 40km/h+ to activate the 4WS and then see how far you need to turn the wheel in each direction in order to trigger the light. It might give you an idea of where the sensor believes the lock positions to be. The centre position will be halfway between the two lock positions.

Finally, the correct procedure is undoubtedly in the JDM service manuals and will be known to Japanese dealers and probably also the Japanese SVX clubs and mailing lists. I've had success emailing them in the past, but I don't have any contacts at present. Try emailing some Japanese websites and with luck one of your messages will reach someone who can read English.

http://poyo.biz/svx/
http://groups.yahoo.co.jp/group/svx/
http://www.takenet.or.jp/~subaru/
http://obihirosubaru.co.jp/
http://www.asahikawasubaru.co.jp/
http://www.946subaru.co.jp/
http://www.phoenix-c.or.jp/subaru/

If you find the correct procedure, I'd love to know what it is. I guess that you probably set the wheel straight ahead and then zero the sensor somehow.

Good Luck

Phil.

[AvPPoW]

I would agree with that, he is sorely missed. :(

Apparently, he's back... :confused: :D

[oab_au]

Be carfully turning the steering wheel when it is disconnected from the rack. The wiring to the controls on the column are wraped around the steering shaft, if you wind it around too far the wiring can tear off the shaft.

Harvey.;)

[nutshell]

I have arranged to take the car into the agent first thing tomorrow am so they can 'sort it out'.
After all, it was working fine until they 'fixed' it.

Hopefully they have all the manuals, in English, and so I trust I can report a positive result.

Btw: with the car stationery; if the wheel is turned all the way to the Right, the light stays out. If it's turned all the way to the Left, it comes on as the wheel reaches full travel. And with the light on, before moving off, the steering is ok - the wheel is centered.

I'll go for a run and see how far it has to move before failing "on the trot'"
Later:
Hmmm - very strange. At normal speeds, on normal roads, the light stayed out, unless I was turning hard left at an intersection. Initially, with the light out, the steering wheel was cocked about five degrees. The light stayed out at speed until I turned left - at an intersection, which requires more turns than usual - and came on.
I stopped, turned the engine off and restarted. The light stayed out, but the wheel was still cocked.
When the light came on again - at an intersection - I stopped and restarted.
This time the wheel stayed straight!:confused:

With the light out the car felt as though the back was working in corners; very flat and more precise than with the light on.

I will be very interested to hear what the agents have to say on the morrow.

Ian

[paddlesnz]

Finally, the correct procedure is undoubtedly in the JDM service manuals and will be known to Japanese dealers and probably also the Japanese SVX clubs and mailing lists. I've had success emailing them in the past, but I don't have any contacts at present. Try emailing some Japanese websites and with luck one of your messages will reach someone who can read English.Phil.

I'm a member of the Japanese SVX club and fluent in Japanese, so if you can't get the problem sorted, PM me and I'll post a question on the Japanese site.

The 4WS service manual is only in Japanese, so I doubt any NZ agents would have a copy.

[nutshell]

Thanks mate

[floatingkiwi]

Got the car back. Bodywork now almost ding free.

The diagnostics on this vehicle are located by the throttle pedal, under the fuse box, not alongside the transmission tunnel.


Yes, what I meant was in the direction of the transmission tunnel - as in pull the panel directly off without moving it forward or back. Anyway, you've obviously got it.

[b3lha]

Be carfully turning the steering wheel when it is disconnected from the rack. The wiring to the controls on the column are wraped around the steering shaft, if you wind it around too far the wiring can tear off the shaft.


Very true Harvey. I'd forgotten about the clockspring wiring.

The problem is that the steering was turned while the rack was disconnected, so the solution is to turn it back the opposite way. Zeroing the 4WS sensor to the current position will not be correct because the clockspring will be off centre too.

Btw: with the car stationery; if the wheel is turned all the way to the Right, the light stays out. If it's turned all the way to the Left, it comes on as the wheel reaches full travel. And with the light on, before moving off, the steering is ok - the wheel is centered.
<snip>
At normal speeds, on normal roads, the light stayed out, unless I was turning hard left at an intersection. Initially, with the light out, the steering wheel was cocked about five degrees. The light stayed out at speed until I turned left - at an intersection, which requires more turns than usual - and came on.
I stopped, turned the engine off and restarted. The light stayed out, but the wheel was still cocked.
When the light came on again - at an intersection - I stopped and restarted.
This time the wheel stayed straight!

So it looks like your wheel was turned to the left while the rack was disconnected - because you can exceed the left limit, but not the right limit. So maybe try disconnecting it and giving it one turn to the right.

The effect of the steering wheel being cocked is probably because the rear wheels do not automatically return to centre when the 4WS controller goes to error state. They just stay where they are. It might not be obvious to look at because the maximum rear steering angle is only a couple of degrees.

I've got some interesting info in my locker about the theory of how in-phase 4WS works. It's from a Mitsubishi service manual. Their system is hydraulic rather than electric, but the theory still applies to ours.
http://www.subaru-svx.net/photos/user.php?b3lha|33362

[nutshell]

The thot plickens:

Took the car into the agents... and after 4 hours deliberation they concluded the column was 150 degrees out (to the left), and that the Sensor on the column had failed! Of course, nothing to do with the earlier rack work!
Naturally they claim the sensor cannot be removed and repaired and is integral to the column. So they suggest we replace the column with a second hand one, which will (hopefully) have its Sensor working.

Meanwhile the vehicle is still very mobile, with the 4WS temporarily disabled.

hmmm

Ian

Btw: I spoke with Trevor. He is truly a font of knowledge

[b3lha]

The thot plickens:
and after 4 hours deliberation they concluded the column was 150 degrees out (to the left), and that the Sensor on the column had failed! Of course, nothing to do with the earlier rack work!
Naturally they claim the sensor cannot be removed and repaired and is integral to the column. So they suggest we replace the column with a second hand one, which will (hopefully) have its Sensor working.

Meanwhile the vehicle is still very mobile, with the 4WS temporarily disabled.

hmmm

Ian

Btw: I spoke with Trevor. He is truly a font of knowledge

If they turned the column 150 degrees left then the steering wheel would now be at the 7 o'clock position when driving straight ahead. Unless they removed the steering wheel while changing the rack or rotated the outer sleeve of the column where the sensor is attached. Neither of which sound likely.

As they have examined the car, they are in a better position to say what is wrong with it than anyone else, but doesn't their conclusion sound a little strange to you?

If the sensor has failed then maybe it failed because it got turned 150 degrees past it's limit as a direct result of them installing the rack incorrectly? Perhaps you can use that argument if they try to make you pay them for fixing a problem which they themselves probably caused.

I had a look at the sensor on my car. Certainly the part underneath where the wires are connected is attached by bolts and would be easy to remove. The aluminium bit on top looks like it might be part of the column but that doesn't necessarily mean it can't be dismantled. Impossible to tell for sure without removing the column from the vehicle.

Phil.

[nutshell]

I guess "column" may not be the technically correct word to use to describe what is 150 degrees out of whack. Suffice to say something has gone seriously awry. Rest assured I shall be closely examining the replaced components. Aren't automobiles fun! :-)

[nutshell]

Yet more mystery

Took the vehicle in for a long and expensive analysis.

Here's the tech report.

"Check 4WS warning light coming on.
Plugged in select monitor and checked for fault codes. 15 and 21.
Carry out individual systems test, front steering showed incorrect settings.
Remove steering column and rest (sic), recheck. Fault still present.
Replace steering angle Sensor. Remove steering rack and reset.
Reassemble and clear codes. Road test."

After all that you might expect success. Alas - not so.

4WS now works fine, up to about 110 kph (two short freeway runs), then the light came on and the wheel cocked to the left as before. Stop, key off - restart - all ok. Until the next on ramp.

Ran the codes. ;-)
Shows zip, nada, blinks steady at regular intervals.

:confused: Any suggestions? :confused:

The one that comes to mind is to disable the 4ws by pulling the fuse, but that may work or not, and anyway, it defeats the purpose and does not fix the problem.

Rear axle speed sensor? [but no code alert]
I believe there's 4ws parts close by.
Ah, and it now 'clunks' at the rear when it settles and stops (with the 4ws active) - didn't do that before.

[b3lha]

:confused: Any suggestions? :confused:


First thing I would suggest is to drive it some more and see if you can get it to produce a code. Failing that you'll have to get them to hook up the select monitor again.

It seems to me that this "clunk" would be a good place to start searching for the problem.

Do you know the capabilities of the select monitor? Is there some sort of test function which would allow them to activate the 4WS while the car is up on stands without the requirement to drive at 40km/h?

Maybe PaddlesNZ can dig up some info from the JDM service manuals.

Phil.

[nutshell]

Finally. I can report my SVX is now running as it was designed.

For three days last week technicians at Subaru's Auckland dealership (Winger Subaru) labored to correct the problem caused when they removed, reconditioned and replaced the steering rack in February, without knowing the vehicle had 4ws.

To get the adjustments right, the techs referred the problem to Subaru New Zealand. It seems the problem was that the rack was reassembled wrong. The steering pinion was two teeth off the centre of the rack.

Once the rack and pinion was correctly centered, and all 4 wheels correctly aligned, it was possible to finally reset the steering angle sensor to zero.

The results are darn near miraculous. Not only does the car now track like glue through corners, its gas milage has improved and it actually "feels" as though it is all together again. Those strange clunks and groans seem to have vanished, and it is settling down nicely.

My thanks to you guys for your support.
If it were able, I am sure the car would thank you too.
:)

Now I can get on with the fun of learning to drive it. :D

[floatingkiwi]

Bloody marvellous! How did you convince them to do that? They finally figured that they'd screwed it up in the first place?
Well done, justice feels good doesn't it!

matt

[Electrophil]

I want the police special with the twin turbos! Varoom! Varoom!! :D

[nutshell]

How did you convince them to do that?
matt

Basically they took several hundred bucks off me then worked out they had to fix the problem.

I could argue they should have done the job right from the get-go, but if they had I doubt the previous owner would have sold it to me.

Sort of catch 22.

[Trevor]

Basically they tools several hundred bucks off me then worked out they had to fix the problem.

I could argue they should have done the job right from the get-go, but if they had I doubt the previous owner would have sold it to me.

Sort of catch 22.

All very alarming, as I thought Winger Subaru knew their stuff. After all the fix analysis was quite logical, taking all of what they had done into account.

As there appears to be little data on offer, a short explanation of how it all works may prove of interest.

The four wheel steering system is completely electrical, there being no mechanical connection between the steering and the rear wheels. A potentiometer type transducer is fitted to the bottom of the steering column as the actuating control device.

The rear wheel steering is mechanically actuated by means of an electric motor and a position sensor is incorporated at the output of the associated reduction gear box. The controller is mounted as a unit in the rear luggage space. The motor is switched by means of field effect transistors within the controller. The circuit is continuously alive and does not turn of with the ignition for obvious reasons. As a result, this should be taken into account in the event of a flat battery problem, which is otherwise not accountable.

[b3lha]

Hey Nutshell, that's great news. I was wondering about your problem.

Actually, a few days ago I took my SVX in for it's annual MOT test and the mechanic told me that there was a little bit of play in the steering rack. Maybe it would need reconditioning for next year's test. :eek:

Given your experience, that suggestion scared the living the crap out of me. I'm not letting anybody mess with it unless it becomes a genuine safety issue.

Gotta love the way these 4WS cars corner. :D

Glad you got it sorted.

Phil.

[b3lha]

The circuit is continuously alive and does not turn of with the ignition for obvious reasons. As a result, this should be taken into account in the event of a flat battery problem, which is otherwise not accountable.

Hi Trevor,

Could you explain please? The obvious reasons are not obvious to me at this time of night. Are you saying that turning the steering wheel with the battery disconnected will cause the front and rear wheels to lose syncronisation?

Phil.

[Trevor]

Hi Trevor,

Could you explain please? The obvious reasons are not obvious to me at this time of night. Are you saying that turning the steering wheel with the battery disconnected will cause the front and rear wheels to lose syncronisation?

Phil.

Greetings Phyl,

No I am not saying that. The position sensor and the controller will retain continual syncronisation regardless.

Problems could occur if the ignition switch was turned off when the car was moving. That is if the supply was not arranged as it is.

By the way the controller and motor are from Mitsubishi Electric.

Trevor.

[nutshell]

Just thought I'd update this record. All positive.
The 4WS took me about 18 months to get really comfortable with.
Now it makes me wonder why Subaru dropped it after only 500 vehicles.

[svxistentialist]

Just thought I'd update this record. All positive.
The 4WS took me about 18 months to get really comfortable with.
Now it makes me wonder why Subaru dropped it after only 500 vehicles.

Ian, as you know the active AWS model was released only in Japan. The top of the range models were the CXD variants, which came "fully loaded" with practically every available option as standard, including AWS.

Feedback from Japan suggests that poor sales of this top model resulted in the CXD variant getting dropped from the lineup early on. Looking in from the outside it seems that most Japanese buyers are wary of complexity, at least of over-complexity. They already have a regime that forces them back to the dealer after a few years to replace seatbelts and other expensive components, or the car does not pass their Shaken test. It looks like this test is fashioned to make it expensive to retain an older car, thereby making it cheaper or more economically advantageous to buy a new car. An imposed inflationary measure to boost the economy.

Combined with teething problems in some of the early release AWS cars, I think this may have put the seal of doom on the more advanced hi-tech SVXes as far as customer sales were concerned. I could be wrong in this, but that's how it looks.

Enjoy your car,

Joe :D

[nutshell]

Thank Joe

I knew there had to be a logical answer, but for the life of me the work (not to mentuion cost) that must have been applied to develop the AWS...and then to drop it.

Amazing!

Trust your beast is running fine.

[svxistentialist]

:D

Beasts, Ian. Plural.

Yeah, all going well here, thanks Ian. I probably have two cars I should sell, but I like them, and the current prices on offer are insulting. It does mean I have a little work to do starting and running them periodically to keep them fettled, but that's a pleasure as much as a chore.

Graham is using the AWS one in England, and the steering on that one gave problems. I'm never sure if it is working right or not. I suspect it may be related to changing the clock from reading kilometres to miles, possibly compounded by adding an aftermarket alarm. It stops and goes and does everything it should, so I expect it's OK. Soon we will be giving it a fresh engine from another car, so it is due a big amount of mechanical work.

Possibly when we get that work done I will ship it home to Ireland, but that would have to be after selling the good '95 car I have here.

Joe :)

[PaulDexler]

Just thought you might be intersted in a somewhat pertinent side note. While SVXs with all-wheel steering were never offered in the U.S., Hondas and Mitsubishis were. I have had the experience of driving all three variants.

First was a Honda Civic, which had mechanical all wheel steering. When the steering wheel was turned only a few degrees, as with normal highway driving, the rear wheels worked as they do on an AWS SVX, that is, they turned parallel to the fronts, but only a few degrees. However, the steering box was set up so that when the wheel was turned beyond a certain amount, the rears would steer opposite to the fronts, enormously reducing the turning circle. It was a delight to drive, the on-road handling was excellent, and parking was a dream, it could make a u-turn in a parking lot aisle. The system used a shaft from the front steering rack to a steering box at the rear axle. All mechanical, and relatively simple.

Honda next went to an electric system which was offered on the Prelude. I believe this is the same system that was offered on the Mitsubishi Galant the same year. (I think 1989, 1990, perhaps.) This system did not produce the opposite lock for parking, but did have the parallel steering similar to what was offered on the SVX. On the Mitsubishi, it worked like a champ, creating that feeling of extra stability. It also did on the Honda, except for a strange defect on the one car I drove, which Honda refused to explain to me when queried about it. I had taken the car on a 130 mile highway trip, with no problems. On the return journey, I was changing lanes on the freeway when a loud "clunk" came from the rear, and the car seemed to be travelling diagonally down the road, similarly to an old body-on-frame car which has been in an accident and the frame was improperly straightened. As I slowed, the steering came back to normal. There was a switch to turn the AWS off, and I did so. The car seemed perfectly normal, so after a few miles I turned it back on again. And again, shortly, came that clunk, and the car was travelling diagonally. I turned the AWS off, and drove home in a less responsive, but more trustworthy, Prelude. A couple of days later, when I returned the car to Honda, I told them about the problem, and asked to know what they found. I never heard, and when I queried them a week or so later, all they would tell me was, "there was a problem." I knew that.

I drove the Galant model with no problems, but both systems were discontinued the following model year.
Paul