2006+ Honda Civic Forum banner

1 - 20 of 66 Posts

·
Registered
Joined
·
42 Posts
Discussion Starter #1
I’ve a weird problem with my vehicle speed sensor. If I haven’t driven the car for a couple of days the speedometer stay’s at 0 km/h while driving. After about 8 kilometers the speedometer suddenly kicks in and after a minute of up and down readings it starts working without problems. These problems got worse the last couple of weeks to a point it doesn’t work completely anymore.

So, I figured the VSS (vehicle speed sensor) was broken.

The car: A 8th gen 2006 Honda Civic 2.2 cdti sport with manual 6 gear transmission.

The I knew the owner before me had some VSS troubles, but I don’t know the details about it. I ordered a new VSS and when I wanted to install it I discovered the original speed sensor had two washers between the gearbox and VSS mounting hole. The VSS looked fine. Since the workshop manual does not mention the use of washers I thought the distance between the magnet in the gearbox and the VSS was too big. So I removed one washer to push the sensor 1mm deeper into the gearbox. Started the car, pushed it into gear… Still no reading. Pulled the VSS out of the gearbox and discovered this:



Totally broken. That’s why the previous owner put washers between it. So, there’s definitely something wrong with the magnet on the shaft inside the gearbox. The gearbox works fine by the way. I’ve replaced the VSS and put it in the gearbox with the two washers between it. Still no reading from the speedometer. Tried waving a magnet over the VSS but still no reading from the speedometer. I figured the waving wasn’t fast enough to show at least 1 km/h.

I’ve discovered the car handles a lot of things with the information of the VSS.
The consequences of the malfunction VSS lead to:



  • A stutter and jerky propulsion when driving slow in 1st and 2nd gear.
  • The power steering stays to much engage when driving fast. And eventually gives an error warning light on the dashboard.
  • Hard to push into 5th and 6th gear because the automatic reverse lock does not encage
  • Music does not go automatically louder when driving faster.
  • Can’t use cruise control
  • The car sometimes goes completely into limb mode and can't accelerate above 2k rpm
To save me from a costly repair of the gearbox I had the idea to replace the VSS with a microcontroller and GPS. I’ve done projects with these things before, so technically I can handle this. But before I try something like this, I want to know what can go wrong if I choose this route.

My main problem I can think of is that the speedometer won’t work in tunnels. But that’s an issue that I’m willing to accept. I’ve already discovered the cruise control just disengage itself when it gets a 0 reading from the VSS. Maybe I can calculate the speed with the microcontroller with the RPM and gear it's in if it doesn't get a good GPS fix eventually.

But are there other things that can lead to a problem when I replace the VSS with a GPS? It’s not reading the pulses directly from the gearbox, so maybe the ABS also malfunctions if the real speed pulses are still fed to the car computer but the gearbox isn’t turning because of a slipping tire? Maybe problems with the VSA?

If I go through with it. What kind of signal do I need to emulate? I haven’t measured voltages myself but from studying elsewhere on the forum and google I’ve discovered this (please confirm this):

  • The VSS gets 0 and 12V and sends a 5V pulse to the car computer.
  • The pulse wire is the middle wire on the VSS.
Some extra questions:

  • Let’s say I want to display 100 km/h how many pulses per second do I need to send to the Car computer (approximately of course).
  • Can I “Steal” 5V somewhere from the car? The microcontroller and GPS I want to use, needs 5V to operate. Otherwise I need to use a step-down converter.
  • Why does my car have a stutter and jerky propulsion when driving slow in 1st and 2nd gear?
 

·
Registered
Joined
·
42 Posts
Discussion Starter #2
Okay, I've used a multi meter to try to figure out what the signal-wire is. According to this review:
http://www.civinfo.com/forum/how/8346-calibrate-speedo.html

with this picture:

The blue/white wire should be the signal wire... but my multi meter say's its +5V. according to my measurements the Green wire is the Ground, Blue +5V and the Yellow the signal (I'm measuring real low voltage's between green and yellow and blue and yellow).

Can anyone enlighten me?
 

·
Registered
Joined
·
42 Posts
Discussion Starter #3
Okay, I'm really lost at the moment about the wires. Can someone tell me what these wire's supposed to be?

  • Green with black stripe: Ground
  • Yellow with blue stripe: 12V ? This wire is "dead". Almost no voltage between the green/black and the blue/white. I've measured all small fuses under the bonnet and dash, all good. Does this wire supposed to have 12V?
  • Blue with white stripe:Signal wire. This wire has 5V on it, so I guess the VSS pull's it to ground? Can I test this wire by shorting it repative to ground to see if the speedometer gives a higher reading than 0?
I can't find any schematics on the VSS circuitry. And I don't want the risk blowing up my ECU.
 

·
Registered
Joined
·
42 Posts
Discussion Starter #4
Found an interesting post on another forum over here:
Help Please , Civic 8th Gen Weird odometer Problem - 212774

The first thing is to check the wiring harness at the VSS and see if all the continuities are okay.

Remove the 3 pin connector at VSS and check for each pin one by one.

Black wire (Ground) should have a 0 resistance to chassis (If not, rectify open wire between ground)
Black/Yellow SHould read 12V when ignition switch is in on( If not, rectify open wire or blown fuse between FuseBox)
BLU/White should read 5V or more when Ignition switch is on( if not, rectify short in wire from VSS to ECU)
But this is not about the European Civic. And the color's are a bit different. Still no one over here who can share some information? Or can someone tell me where the fuse is for the vss Yellow with blue stripe 12V wire?
 

·
Registered
Joined
·
42 Posts
Discussion Starter #5 (Edited)
Okay... I've short-cut the blue 5V wire repeatedly to ground, and the speedometer displayed different speeds. So if I replace it with a microcontroller it needs to pull down the 5V reference signal.

Still no one who can tell me where the yellow wire with blue line goes to? What fuse and does it suppose to be have 12V on it (instead of 0 like it has now)? :confused:
Maybe the previous trick with the washers is still working fine if I fix the fuse somewhere to get 12V back on this wire. But I really like some insight knowledge about this before I risk letting the precious smoke escape the ECU .

This thread is becoming a bit of a monologue ;)
 

·
Premium Member
Joined
·
114 Posts
I think you're facing quite a unique challenge and no one can offer any prior experience. I suggest you download the service manual and check for wiring diagrams.
 

·
Registered
Joined
·
42 Posts
Discussion Starter #9 (Edited)
Hi guy's, at least thanks for the moral support! :D

Well... if I can fix it with a microcontroller and a GPS module that will cost me about 30 / 40 euro's hardware and some hobby time. And a nice extra is that I can have accurate speed readings. The speedometer was about +6km off.

If I need a garage to fix the whole gearbox just for some stupid magnet misalignment it's probably going to cost me into the thousand euro's. And if I'm going that far, it probably wise to overhaul the whole gearbox and pressure plates as well. Don't want to know what that's going to cost.

But, maybe to washers on the sensor to pull the sensor a couple of mm out of the gearbox is enough (it worked before this way) and the only thing I need to do is put 12V on the sensor. But without confirmation I'm a bit afraid to put 12V on a sensor that sends a signal to a 5V ECU. I have the garage manual, but can't find any schematic about the VSS or fuse too the VSS :(

I've managed to program a microcontroller and control the speedometer. The frequency of the pulses are way higher than I expected.
9 km/h (Real speed) / 10 km/h (Display speed) is 191 hz
194 km/h (Real speed) / 200 km/h (Display speed) is 4125 hz

... and, have you ever seen this on your display?

...priceless :D
 

·
Super Moderator
Joined
·
7,124 Posts
You either have your display and real speeds set wrong or you typed it out wrong.

When you are doing 96km/h the speedo will show 100km/h

You can use 21.2Hz per 1km/h as a base but there will be some error.
 

·
Premium Member
Joined
·
1,929 Posts
I did 999mph once! My converter was loose I was so gutted was really hoping to time travel :(
 

·
Premium Member
Joined
·
2,202 Posts
I've subscribed to this thread. I love how the OP refuses to fix the box and is trying to reinvent the wheel.

Sent from my GT-I9300 using Tapatalk
 

·
aka Paul
Joined
·
4,524 Posts
This is an interesting way of addressing the problem, but I'd advise against it.

It will probably invalidate your insurance, but more importantly it will be very confusing for your ECU due to the slow reaction time of the 'speed sensor'. I would say it would be potentially unsafe.

If you're sure it's the box, not the sensor, get the box swapped/repaired.
 

·
Super Moderator
Joined
·
7,124 Posts
This is an interesting way of addressing the problem, but I'd advise against it.

It will probably invalidate your insurance, but more importantly it will be very confusing for your ECU due to the slow reaction time of the 'speed sensor'. I would say it would be potentially unsafe.

If you're sure it's the box, not the sensor, get the box swapped/repaired.
VSA may also have issues since data from the VSS is conflicting with data from the ABS and WSS data, and if you are starting to lose it this is very very very very bad.

Go into the city and GPS signal will be bad or may even drop out.
Go into a tunnel and signal will drop out.
 

·
Registered
Joined
·
42 Posts
Discussion Starter #17 (Edited)
You either have your display and real speeds set wrong or you typed it out wrong.

When you are doing 96km/h the speedo will show 100km/h

You can use 21.2Hz per 1km/h as a base but there will be some error.
You're right, I've typed it wrong. (corrected the post)

This is an interesting way of addressing the problem, but I'd advise against it.

It will probably invalidate your insurance, but more importantly it will be very confusing for your ECU due to the slow reaction time of the 'speed sensor'. I would say it would be potentially unsafe.

If you're sure it's the box, not the sensor, get the box swapped/repaired.
I'm going to use a gps with 10hz update rate. So I don't think the update speed will be a problem.

VSA may also have issues since data from the VSS is conflicting with data from the ABS and WSS data, and if you are starting to lose it this is very very very very bad.

Go into the city and GPS signal will be bad or may even drop out.
Go into a tunnel and signal will drop out.
These where also my concerns (see the openings-post). But there are other people over here that altered the VSS signal without problems:
http://www.civinfo.com/forum/how/8346-calibrate-speedo.html

At the moment there are two things wrong with the VSS:

  • There's a unknown magnet misalignment inside the gearbox. But the previous owner had this fixed by putting some washers between the VSS and the gearbox. This probably worked fine until problem 2 (which I discovered while working on this project):
  • There's no voltage on the yellow with blue stripe wire. I think there supposed to be 12V on this wire, but I don't know where this wire comes from or where it is fused. It's a easy fix to create my own fused 12V wire and attach it to the VSS... But I want to be absolutely sure the VSS needs 12V to operate. I don't want to fry the ECU.
Still it's an interesting project. If I get the VSS fixed I still want to use the GPS inside the dashboard console to change the pulse signal to the display, if there's no gps-fix, sent the original (or corrected) pulse to the display otherwise the real GPS speed. I'ts so stupid constantly have to retract 6km from the speed reading to know the real speed.
 

·
Clutch pedal guru
Joined
·
5,153 Posts
There's obviously a problem with the sensor wheel on the gearbox. Just fix that and all your problems will be fixed.

Instead of trying to fix the symptom, fix the root cause instead.
 

·
Bye bye Honda...
Joined
·
5,892 Posts
Interesting!

You could overcome the problem of GPS signal loss in tunnels and 'urban canyons' by programming your microprocessor interface to simply maintain the most recent output frequency when the GPS receiver reports loss of satellite reception.

Better still, could you instead use the signal from one of the wheel speed sensors? Or even better still, take an average of all the outputs from the wheel speed sensors... tap into the signal lines, run these through frequency-to-voltage converters, derive an average of the 4 output voltages, then run this voltage into an appropriately-scaled voltage-to-frequency converter. Make this last stage adjustable so you can calibrate the system using a GPSr.
 

·
Registered
Joined
·
42 Posts
Discussion Starter #20 (Edited)
There's obviously a problem with the sensor wheel on the gearbox. Just fix that and all your problems will be fixed.

Instead of trying to fix the symptom, fix the root cause instead.
I understand, but as I stated earlier, removing the complete gearbox just for fixing a stupid magnet inside it is a costly operation. Besides that, still symptom 2 exist:
There's no voltage on the yellow with blue stripe wire. I think there supposed to be 12V on this wire, but I don't know where this wire comes from or where it is fused. It's a easy fix to create my own fused 12V wire and attach it to the VSS... But I want to be absolutely sure the VSS needs 12V to operate. I don't want to fry the ECU.
So maybe it isn't even necessary to fix the magnet inside the gearbox.

Interesting!

You could overcome the problem of GPS signal loss in tunnels and 'urban canyons' by programming your microprocessor interface to simply maintain the most recent output frequency when the GPS receiver reports loss of satellite reception.

Better still, could you instead use the signal from one of the wheel speed sensors? Or even better still, take an average of all the outputs from the wheel speed sensors... tap into the signal lines, run these through frequency-to-voltage converters, derive an average of the 4 output voltages, then run this voltage into an appropriately-scaled voltage-to-frequency converter. Make this last stage adjustable so you can calibrate the system using a GPSr.
The wheel sensors... that's an interesting idea... probably only need the average of the two front wheels... and in theory, is there a difference between a signal from the output shaft (=vss) and the average speed from the two front wheel sensors (if we ignore the difference in pulse speed)? Their both mechanically linked to the same shaft. And if we average out the two front wheels speeds we exclude the differential "error".

Though, The frequency to voltage and back converters is a bit complicated. I can easily calculate average of the pulse width from the wheel sensors directly with a microcontroller.
 
1 - 20 of 66 Posts
Top