Tacho Not Functioning Correctly

[Trozzle]

My tacho used to intermittently play up immediately after turning the key to IGN. With my ECU (ViPEC v44) it's set up to sweep from 0rpm up to 7000rpm and back to 0rpm when you turn the key on before starting the engine. When playing up, it will go immediately to 7000rpm, then drop and point directly downward. Turning the ignition off then back on again will see it flick back up to 7000rpm, but fall down again. The first few times it did this, turning ignition off then back on again a few times would sort it out. A while later when it did it again, it didn't fix itself until the engine was brought up to roughly 3500 - 4000rpm at which point it would flick to the correct rpm immediately, then proceed to function correctly. Unfortunately now, it won't play ball at all. When I rev it out slowly, it will begin to climb ANTI CLOCKWISE from around 4000rpm, up until reaching 7000rpm (going backwards though...) at which point it will remain there until the revs drop - even upon hitting rev limiter.

It doesn't seem to be affecting anything adversely, so my car still runs fine. I haven't got a laptop to plug into the ECU and watch what it says about the RPM while the engine is running though, but given it runs at all the fault has to be somewhere between the ECU output and the tacho itself.

Anyone got any suggestions? I'm thinking the tacho itself is farked, which sucks but isn't the end of the world. Just wanna see if anyone has had a similar issue - particularly with this ECU.

video:

[PM-R33]

Only thing I can think of is this, however it is kind of hard to explain.

If you have ever had your dash cluster out and played around with the needle on the tacho, there is a point of resistance where the needle stops. This is at zero RPM. You can freely move the needle by hand but when you let it go, it will automatically fall back to zero. Maybe spring loaded or something. If you wind it to far forward there is another point of resistance. If you go passed this point of resistance it ends up changing the initial point of resistance.

Only reason I know this is because I have fitted white faced dials which means you have to take the needles off and put them back on again.

Could it possibly be that this spring loaded function is now broken and therefore the needle just falls to the bottom due to gravity and now doesn't want to move?

Test the tacho by changing the tacho sweep on the Vipec to different points: 2000, 4000, 6000 and see if it still sweeps to those points.

Quite strange never the less.

[Trozzle]

Only thing I can think of is this, however it is kind of hard to explain.

If you have ever had your dash cluster out and played around with the needle on the tacho, there is a point of resistance where the needle stops. This is at zero RPM. You can freely move the needle by hand but when you let it go, it will automatically fall back to zero. Maybe spring loaded or something. If you wind it to far forward there is another point of resistance. If you go passed this point of resistance it ends up changing the initial point of resistance.

Only reason I know this is because I have fitted white faced dials which means you have to take the needles off and put them back on again.

Could it possibly be that this spring loaded function is now broken and therefore the needle just falls to the bottom due to gravity and now doesn't want to move?

Test the tacho by changing the tacho sweep on the Vipec to different points: 2000, 4000, 6000 and see if it still sweeps to those points.

Quite strange never the less.

While that sounds like a good idea, I've never had the cluster out. Never had a problem in the 1.5 years I owned the car prior to the ECU upgrade, but then didn't have the issue after the upgrade for a while....Also, the way it moves so erratically when IGN is turned on suggests it knows something is wrong...also the way it falls down doesn't move like it would be gravity imo

Thanks for the input though, will consider it.

[Trozzle]

Also phil, did you watch the embedded video? Now that I look at it again myself, it is actually dropping slightly "further" than dead vertical.