Speedo fails at 75mph - driving me crazy!

[greg123]

Hi guys,

My turn to ask for help rather than answer for a change http://briskoda.net/images/smilies/B_thumb.gif

I have now resolved the other issues apart from these on this 2002 1.8T auto (01M) Octavia (Golf 4 platform):

1) The car is faultless, totally, other than this: The speedo is faultless, UNTIL I hit about 75mph. Then the needle starts to waiver. Sometimes 85mph is needed to 'break' it, but once it's broken it drops to 70-80mph tops, even if you are doing 120mph which is very disturbing as essentially you can speed drastically without realising. The speed signal is rock solid UNTIL the needle starts to waiver, I know this because cruise control is flawless till I see the needle start to drop down and fluctuate, at which point the car starts to lurch around as the cruise trys to keep the speed steady with what is obviously a constantly changing speed signal. This 'waivering' once it's started won't quit till I drop down to about 50mph. If I speed up it will start waivering a lot more easily - leave it parked up a while and it will take a highish speed of say 80mph to cause it to 'fail' again. If I drop down to 40mph the needle is rock solid and cruise is smooth again, but forget using cruise again once the problem has occured on the freeway the speed signal is all over the place!

There are NO fault codes related, no can bus fault, no instrument fault, no gearbox fault etc etc. The only fault code I have currently and never managed to cure is intermittant no comms to radio & radio rear speakers not connected (someone disconnected the tweeters I think). I tried it with the radio out - same problem. The radio & cd changer are flawless & the radio volume speed adjustment works but seems to be overly loud as you speed up. It's always the same though and never suddenly drops the volume up or down, it's always gradual transition from very loud=freeway to quiet at the lights.

I have put a brand new and superceeded speed sensor in. I thought it had fixed it - I hit 90, then bang the needle dropped to 70 and waived around all over the place. After fixing the other faults on the car (coolant sensor, can bus error rear door) I took the battery out and stripped down the main earth and the 4 other smaller earth wires. None were bad, though not great, all connections to the wires looked fine. I put them back better than new and well greased, put it back together and went for a spin. Thought I had cracked it, hit 95mph rock solid needle. Then on the way back bang - needle waiving. Currently not as bad as it has been, but there is still an issue there.

The car is a dream to drive and this is the only operational fault, I have put so much work in during the last year building it up this last fault is drivig me crazy.

Has anyone any suggestions? I have VCDS so I can carry out any test or log. Output test of the instruments is perfect, full needle deflection. Displaying the speed on the climatronic made the speedo needle slow to respond and the speed incorrect/too slow and just no sense so that didn't help.

UP TILL the problem occurs, the speedo is perfect, crisp, solid, instant response.

I'm actually thinking of taking out the instruments, coding in another one, seeing if that will do the same thing, then swapping & coding them back (I have ********* and I have skc's and a spare vehile of a friend I can borrow from - it's not mid line and is diesel not petrol but is VDO and hopefully the speedo should work just swapping them over).

Any suggestions? Anyone got a schematic for the speed signal?

Greg.

[Crasher]

Have you observed the speed readings in Measuring Blocks as seen by the dash, box and the engine ECU to see if they waiver?

The speed signal comes from the hall sender on pin 2 (white/blue) and goes to a 14 pin connector in what is described as “engine compartment left) where it changes from a white/blue wire to a green. Further inside the car, it splits off at a junction where other devices get their speed signal from such as the box ECU. The manual says,

The speed signal from the speedometer sender G22 is prepared by the speedometer and distributed to the following systems,

Cruise control system
Radio
Engine management
Selector lever lock
Automatic box
Air con

I would rig up a test lead and scope the box sender signal at speed to see if it breaks down.

[greg123]

Thanks buddy!

Okay I can now confirm that with a different instrument cluster and without the radio fitted the problem remains. For anyone interested I can also confirm that a diesel non-mfa VDO immo iii cluster works fine in my petrol immo iii vdo equiped car that came with MFA (half height info display)! I can also confirm that a diesel manual cluster gave correct rpm/mph when in a petrol auto body which is pretty cool.

Anyhow, back to my problem. A different instrument cluster and the problem remains, so...

I have read the speed in all the other blocks however they don't make much sense, the digital values changed too slow and when I'm viewing speed in any of the measuring blocks, the value gets significantly lower than the speedo the faster I go and starts off km/h slower to begin with. Also the speedo responds VERY slowly if I'm viewing speed data in VCDS, eg when I'm stopped the speed is still coming down from 20mph the speedo is so slow to catch up! Exit the measuring block and the speedo is 'lively' and instant again. Same performance with a new instrument cluster. The viewed speed in any block eg in the climatronic, is always lower than displayed on the speedo and significantly so at motorway speeds, maybe 75% of the actual speed. Despite the slow response, the speedo continues to show roughly the correct road speed as far as I can make out, below70mph anyhow, when viewing the speed data through VCDS at the same time.

I have not got anything that can scope the wire. I may need to come you to buy some time/hire something that can test that if draw a blank with anyhing else.

Where to go next? I'm wondering about cutting into the loom in the connector to the speed sensor, cutting the signal wire and running a new one into the cluster having cut the wire from the cluster to the loom. However it's a 3 wire connector and I don't know if it's some sort of twisted pair that by running a fresh lead I'll have so much interfearance it won't work.

Also I just thought that if it's supposed to be a solid connection from the sensor to the instruments, I could run a long lead and measure the resistance, if the resistance is anything other than a short circuit see if I can find the bad connection and wiggle/fix it till I get a 0 resistance.

Comments appreciated,

Greg.


Have you observed the speed readings in Measuring Blocks as seen by the dash, box and the engine ECU to see if they waiver?

The speed signal comes from the hall sender on pin 2 (white/blue) and goes to a 14 pin connector in what is described as “engine compartment left) where it changes from a white/blue wire to a green. Further inside the car, it splits off at a junction where other devices get their speed signal from such as the box ECU. The manual says,

The speed signal from the speedometer sender G22 is prepared by the speedometer and distributed to the following systems,

Cruise control system
Radio
Engine management
Selector lever lock
Automatic box
Air con

I would rig up a test lead and scope the box sender signal at speed to see if it breaks down.

[Crasher]

That is odd that the speedo response is poor when in Measuring Blocks 08, I will have to try that on mine later.

[greg123]

It's very odd. I'm odd. The world is odd. I think I need a :beerchug:

Actually it seem to be in any measuring block, eg if I'm viewing vehicle speed on the screen in VCDS and drive the car, the speedo needle acts like it has taken a kilo of vallium.

I'm looking into some tests/replacing the wiring to the sensor, if I'm having no joy with that I'll have to get a scope/2nd opinion from you as by that point I'll need to rule out anything I might have overlooked via a 2nd pair of eyes.

Greg.
PS good quote, been to any of the Churchill lectures at the Churchill Museaum last year? The talk from his secretary was very funny.


That is odd that the speedo response is poor when in Measuring Blocks 08, I will have to try that on mine later.

[Crasher]

I went around the Cabinet War Rooms in the summer and the attached Churchill museum, very interesting.

[greg123]

Okay here's the update for anyone watching. Tonight I had some time and did the following:

1) I can confirm that the speed signal wire going into the instruments is a DIFFERENT wire to the one going out of the instruments to the radio GALA signal. So basically, the instruments receive a raw signal, process it, push out the processed signal to the other modules on a different wire.

2) I can confirm I'm looking at (and replaced) the right gearbox sensor. I pulled the plug and drove - no speedo movement at all.

3) I can confirm the plug is a 3 pin and has 14v across the other terminals when the engine is running, so I don't think there is a supply problem.

4) I can confirm that the signal wire from the gearbox speed sensor does indeed go from pin 2 (white/blue) and goes to green wire coming into the multi plug behind the tacho. It had a resistance of 1.2 ohms and I would have liked it to be a bit lower, but at that level even a slightly dirty probe on the multimeter gives a similar reading. It certainly wasn't open circuit.

So, I decided to run a new cable from the speed sensor signal to the dash. I ran a temp cable and temporarily scotch locked it in at the gearbox end. I cut the green wire into the cluster and fed the new wire into the green wire with a connector, so that other than my temporary connections the path was now straight through to the instruments with no breaks for wiring harness connectors.

I tested it with the multimeter and it was okay so I went for a test drive.

1) the speedo was very slow and laggy. Just like it was previously while viewing the speed in VCDS (if you were not viewing speed data in VCDS the speedo would respond quickly to changes in speed and follow the road speed accurately), only now it was laggy all the time even with VCDS not logged in. What I mean by this is come to a rapid stop and while stopped the needle is still floating down from 20mph. Similarly accelerate up to 40mph and sit there at 40, waiting for the needle to catch up from 30 for a second or two before the needle gets there. Once the needle caught up, it said the correct (slightly optimistic) figure.

2) At higher motorways speeds the problem came back just like before, needle waving around up and down round 75mph, regardless of how fast I was going. Again like before took a few miles to come on, once the problem came on it got progressively worse if I kept the speed up.

3) I viewed in HVAC the speed reading and it was 5kmh lower than the speedo at low (20mph) speeds and around 20kmh lower than the speedo reading at say 70mph. The faster you go, the more the speed readings in a VCDS measuring block would be lower than actual speed. This was also an existing problem.

So, where to go from here. New wire, new sensor, new cluster, same problem.

It did occur to me that if the signal is very small as I expect it is, if the problem is partly shorting to ground or similar in the old wire I should cut the old wire at the gearbox end, like it is at the instrument end. Currently only my new wire is connected from the speed sensor to the instruments, the old wire goes from the gearbox speed sensor to the loom behind the instruments and is cut off there doing nothing. So I guess maybe it's possible that the problem is some interfearance/shorting on a very low level in the 'old' wire and by not disconnecting it at the gearbox end I haven't really helped the problem other than adding an extra 3m of wiring into the sysem on what is a very low signal circuit. I didn't want to cut it initially as it's hard to get to and if this didn't help I was going to put it back stock.

Maybe I should just cut the old wire out and properly solder/heat shrink the new one in when I find somewhere to feed it from the engine compartment up into the dash and likewise solder/heatshrink the connection to the cluster multi-plug.

If that doesn't work, I'm really struggling. I guess I could test the ground/b+ on the plug to the speed sensor with a small load such as a 5w bulb and check voltage isn't breaking down with a little load, other than that..... Normally if you probe a plug and it has 14v you'd say the supply is okay. But normally you wouldn't have a crazy speedo that only breaks over 70mph even when all the bits are new...

Greg.

[dpoonwassie]

Hi Greg,

I'm having the same problem with a a3. I had a friend look at ti and we're going to change out the spped sensor on the gearbox, Audi are asking £46 +vat, which sounds steep to me, what do you think?

I don't have the experience you guys have with the setup, but I do have almost exactly the same symptoms. Once I get a fix I'll let you know what I find.:confused:

[Crasher]

Is the A3 an auto?

[dpoonwassie]

Yeah, it's an s reg 1.8T sport auto a well known 3rd party parts supplier has quoted £35 delivered, I think I'll go for that. will let you know if it sorts the problem.

[greg123]

Sorry for the late reply.

I'll pm you re the speed sensor.

My issue seems to be the drive cog so far. Hopefully your issue is not the same. Crasher gave me another idea to test which seemed to confirm it, I have been through everything else now (which is why I have spare new speed sensors).

I'm working on a resolution that does not require a box strip/removal, if I am successful I will post back.

Greg.

[dpoonwassie]

Got the part today, will fit at the weekend and let you know what happens....

[greg123]

Did the new speed sensor fix yours? Looks like mine was the drive cog, but the box has other issues, just as I figured out a fix... never rains but pours :aargh4:


Got the part today, will fit at the weekend and let you know what happens....