Carpokes

Back here with another gauge cluster question, to see if the collective mind on here has any ideas or solves to a bouncy oil pressure needle at idle. 1988 944 S, for context.

I’ve tried replacing the sender (twice, the first one leaked a bit) with an OE VDO unit. After replacing, the needle calmed down for a drive or two, but then went back to being bouncy, once the engine is all warmed up. I’ve cleaned what I understood to be the gauge cluster grounds both at the cluster and under the dash, to no avail.

From what I understand, it may be some sort of mechanical dampener that goes bad over time in the cluster itself, but the fact that the gauge occasionally DOES chill out leads me to believe that’s not the issue.

When cold, and when driving the gauge stays solid around 4 -5 bar, but once warmed up and sitting at idle, it bounces rapidly between 2.5-3 bar, ish.

Curious if anyone has solved this, and if so, what did you do?

Could be a sticky oil pressure relief valve.

Also could be worn engine bearings. An out of round bearing will let the pressure drop when the oil feed is on the loose part of the bearing and will allow the pressure to rise on the tight part.

Check the pressure with a mechanical gauge.

Mine has always done this too. Two different motors and many different senders. You've motivated me to do a little investigation... We had a conversation a while back here on CP about how Porsche wildly over-manages the gauge reading on the modern cars. Coolant temps on my 992 are displayed as 194c pretty much unless/until the car starts to over heat. They are kind of the 'there, there, nothing to worry about' of automotive gauges. But it makes me think -- a bit of smoothing might be quite easy and provide for more meaningful readings... Stay tuned...

A funny story, not specifically Porsche related follows:

My daily drivers tend to be Audis (specifically for the quattro drive, since my Porsches are not suited for snow driving). In the 80s, Audis all had oil pressure gauges, but, in the 90s they phased those out and installed idiot lights instead. The reason that I heard was that there were too many warranty claims for low oil pressure when the engines were warm. Using idiot lights solved that problem.

gb951 wrote: Sun Oct 12, 2025 10:22 am A funny story, not specifically Porsche related follows:

My daily drivers tend to be Audis (specifically for the quattro drive, since my Porsches are not suited for snow driving). In the 80s, Audis all had oil pressure gauges, but, in the 90s they phased those out and installed idiot lights instead. The reason that I heard was that there were too many warranty claims for low oil pressure when the engines were warm. Using idiot lights solved that problem.

It seems Porsche solved the same problem with a slightly different solution -- they give you the gauges, but they just program them to show comforting numbers unless/until very close to the idiot light range...

I've started digging in on this a bit, and believe I have a fairly easy fix to offer that eliminates the bouncing needle (regardless of root cause). I'll post that in the next day or two when I have time to write it up and more fully test it. For now, I thought I'd share some basic info about the gauge since I couldn't find it anywhere on the internet and had to sort it out to get to the bottom of the needle jitters.

The gauge has three inputs: 12 volts, ground, and the sender pressure signal (post G, blue/white wire). The pressure signal is just a variable resistor to the case/block; the higher the pressure, the higher the resistance. For bench-testing the gauge itself, here are the pin-outs...

The pin numbers are to edge connector C behind the cluster (the one on the oil pressure side of the cluster). You can test by applying 12v and ground to their respective terminals, and connect the sensor input to ground via a potentiometer or resistors to make sure the gauge reads what it should (per the chart below) at the various resistance levels.


Here's a similar VDO gauge diagram. The top connection is just a light for the gauge, since this diagram is for a stand-alone gauge. Other than that, this is the same as the 944 oil pressure gauge. Note the warning light is not really part of the gauge at all. The WK post on the sender simply shorts to ground when the pressure is too low, completing the circuit to the warning light bulb in the cluster.

The manual sets out the reference specs as:

1 Bar = 29.6 ohms
2 Bar = 65.3 ohms
3 Bar = 98.9 ohms
4 Bar = 133.6 ohms
5 Bar = 184 ohms

The gauge works by balancing the pull of multiple electromagnetic coils inside the gauge. To visualize the basic idea, you can think of it like this: when powered up, one coil -- let's call it the base coil -- is always powered and creating a magnetic field that pulls the needle up. The sender supplies ground to another magnetic coil -- let's call it the control coil -- that pulls the needle down. This means the control coil gets more and more current, and the needle goes lower and lower, as the pressure and resistance go down. It's more complicated than that, but that's the gist, and explains why a disconnected or dead sensor pegs the needle high -- since there is nothing to counter-act the base coil pulling the needle up. Similarly, if the sender or wire shorts to ground, it supplies max current to the control coil, forcing the needle all the way down. (A clever side-effect of this arrangements is that the gauge is not affected by voltage fluctuations.)

The design is intended to damp the needle so it doesn't over-swing and jitter, instead keeping it locked in a smooth, magnetic tug-of-war. So why does it jitter? I suspect several things can contribute, including a jittery sender, poor connections, pressure pulsing at low pressures, poor grounds, etc. So all the normal advice is worth trying -- cleaning grounds, cleaning the football connector contacts behind the gauge, checking the sender itself, etc.

But if all that fails, I think in many cases the gauge itself is just old. The slightest amount of friction on the needle's axle can contribute to bouncing. And after 40 years, it's a near certainty the axle won't be as smooth and turn as freely as it did when new. So when the needle sticks (ever so imperceptibly), it doesn't move until the force builds up enough to break it free. At that point, the needle breaks free and over-shoots its target, causing the bounce. Kind of like a sticky crank-open window -- to open it one inch, you have to push hard on the crank until it suddenly pops out and bounces back and forth until coming to a rest. That's my theory anyway, and I'm 'sticking' to it The solution I've come up with works regardless of the root cause, however, so even if I'm off base with that theory, my approach will eliminate the bounce. Stay tuned for that...

Tom, blown away by your analysis yet again

i’d read the same theory somewhere on the internet that theres a “damping system” thats gone bad but never actually to the point that i’ve understood what’s going on, so THANK YOU! This makes complete sense

so curious what you solve will be, and also curious in the difficulty of replacing the gauge itself, if scavenging one from a known good cluster.

since my car is an S, the full clusters are tough to find, and expensive at that with the unique redline, but i’d imagine the gauge motor itself is shared throughout the late 944

As I work out a few details, here's a teaser as I test solutions

Holy moly!!

Very, very excited already in your solves - absolutely incredible Tom!

Well, after playing with lots of options, I think the simplest solution is the best. I checked the sender's signal on the scope and saw that it was quite messy, with RFI spike and lots of little short-pulses. What I did was connect an electrolytic capacitor between the sender's signal and ground. Because the sender connects to ground via a variable resistance, the capacitor essentially creates a low-pass filter to filter out the micro-pulses and RFI. The frequency of the filter admittedly changes with the sender's resistance, but within the relevant range it doesn't seem to be a practical issue. Because those little electrical pulses are happening many times a second, they don't really offer much useful information, but they definitely make the gauge bounce like crazy.

I've been playing around with the best way to install the capacitor(s), and am leaning toward making up a little board that sandwiches in behind the football connector on the cluster. Doing this keeps the capacitor in the cabin where it will no doubt last longer than under the hood. I can also add a little jumper this way to give it a low, medium, and high level of damping. It would be cheap and easy get PCBs made -- though I'd probably just stick with my own little prototype board unless others were interested?


Selecting the right size capacitor is part of the challenge, and probably depends a bit on how much your gauge jiggles. That's why I like the PCB above, so that you can play with different damping levels. If the cap is too little, the gauge will still jiggle some, but if you just put a huge capacitor in there, the gauge will seem sluggish. Seems like somewhere in the 680uF-1500uF range is about right (low ESR, automotive 25v or better, something like this one).

If you know what value you need, a cheap and dirty way is just to solder thin eye rings onto the cap and install them behind the football. I give you the "Carpokes Capacitor"!




But if you don't want to pull the cluster, you can also install a cap down at the sender. It's a pretty harsh environment for a catalytic capacitor, but you can make up the little jumper wire with cap and have it installed in very little time -- so maybe as a temporary solution until you are ready to pull the cluster.