Alfa Romeo Forums banner

Speedometer woes

6K views 11 replies 5 participants last post by  Craig 
#1 ·
I installed a new speedo sensor this spring as I was having problems with the old (original) speedo sensor. But I still have the same probelms. The problems are very intermittent with random frequency. Problems include a bouncing needle (it bounces from between zero and 150!), and reading lower than actual speed by either 10% or 50% - almost exactly (eg, if actual speed is 80 then it might indicate 80, 72, or 40)! It never reads zero though, except when I'm stopped :rolleyes: One strange thing is that even when the speedo is reading at 50% it appears that the odometer is still counting off the right distance.

One thing I noticed when removing the old and installing the new was that the 22mm nut was tight but the speedo sensor was still a little loose in the housing. The new sensor is the same - nut tight, but the sensor is a little loose. Trying to remedy the problem I've used contact cleaner on the plugs in the engine bay and on the instrument panel and the two adjustment pots on the speedo itself.

Any ideas what the causes might be?
 
#3 ·
Do you have a pic of where this solder joint is? Is this a common problem? One on of my other cars (it's german with VDO gauges) there is a common problem of bad solder joints on one of the gauges that leads to all sorts of other problems as its part of the "central info" system. Maybe this could be a reason why my 164's cruise control has never worked...
 
#5 ·
You might want to closely inspect male and female terminals in connector to sensor at tranny. Be sure female terminals are not unlocked or spread apart in connector. I know it is hard to see them with Air flow meter and inlet tube installed but I had a broken female one and it caused havoc with speedo readings with intermittent connection.
 
#6 ·
Hmm I wouldn't just yet

look for a bad solder joint in the instrument backplane. Only because in my limited experience, once you take the doggone inst panel out, once you put it back, squeaks seem to develop.

The principle of operation as i recall (i am away from my manuals so most if from memory, please forgive me if this is incorrect and by all means the experts on the board correct this!) the speedometer gauge is an analog gauge that gets its input originally from a digital pulse train, then a D-A converter changes it to a current (70% sure its a current source).

My car (95LS) reads about 7% fast on the gauge versus true speed. Generally, not always, this is due to reference voltage drift in the D-A circuit. There is usually a pot to adjust the ref voltage in these types of circuits, and I believe there is one on the instrument gauge panel although I havent looked (yet). This drift is a common occurence as components age, big time D-A converters have self -biasing ref voltages in a feedback circuit so the intrinsic error is minimized. I think this is likely the source of your 10% error as it is my 7% - ish error
The bouncing reading of 50 to (90% of 100) is maybe not a bad solder joint. I would suspect something called 'wind up' -- similar to that which occurs in a cable-type speedo. These senders are comprised of a mechanical shaft that turns a 'chopper wheel" (wheel with open and closed areas that allow light to pass or be blocked), a light source on one side, and a light sensor on the other. (an alternative design is a magnetic "hall effect" that subsitutes magnetic flux for light, but the basic principle is the same.

If either the mechanical interface is loose or misaglined, each 'revolution' can end up as an edge (inclomplete transition) and the subsequent revolution can end up as a "double". happens all the time in motion control systems that are misaligned.

I would remove the sender carefully (as noted in the digest and the BB) and look inside with a flashlight and check the gear interface. Follow the service manual instructions and re-install it -- I do not think it should be loose --STEVE IF I AM OFF THE RESERVATION HERE LET ME KNOW---
bob
 
#8 ·
I would remove the sender carefully (as noted in the digest and the BB) and look inside with a flashlight and check the gear interface. Follow the service manual instructions and re-install it -- I do not think it should be loose --STEVE IF I AM OFF THE RESERVATION HERE LET ME KNOW---
bob
Don't ask me you as you are talking way over my head when it comes to electronics end of the 164.
 
#7 ·
I guess I need to explain the caveat (where I said 'based on this') I posted earlier.

Since Frank (I assume it's Frank) stated that the odometer was correct while the speedo was off (at least in one case) this would indicate that the odometer was receiving a correct signal from the sender. And since the speedo and the odometer both get their signal via the same wire/connector/pin it would be reasonable to assume the everything was correct up to the point where the signal splits (somewhere on the circuit board).

Now it may be that the observation about the odometer was incorrect, maybe it was a coincidence and the odometer was running fast and slow but happened to average out when checked, in that case I'm wrong.

As for how the speedo on the 164 works, well it's just a voltmeter, more specifically is a moving coil galvanometer and so are the rest of the gauges. Also it looks like both the speedo and the tacho share the same mechanism.

As the speedo is a voltmeter it is most likely reacting to the voltage provided by the D-A. While contaminants (like dust) may slow the movement down it is unlikely to bounce as it is quite well damped.


Speedo adjustment (caveat, on your head be it)

As Bob hinted, the speedo (and tacho) can be adjusted. There are two pots provided for this.
The pot circled in green is for constant offsets (ie: the reading is off by the same amount at all speeds) and can be accessed without disassembling the instrument cluster (you must take the cover off and the pot can be adjusted through the slot at the top left mounting screw).

The pot circled in red I believe to be for adjusting the linearity (ie: reads fast/slow at medium speeds but is correct or has a constant offset at high/low speeds) But I have not yet confirmed this.
 

Attachments

#9 ·
Craig interesting

Thats good info, I have not looked at the instrument panel for the 24 V but I have one I can take out and look at. What leads you to suspect that those are te 2 pots for bias and linearity? Would make sense to me that there would be -- do you see them connected to an op-amp or other active device in the meter circuit (based on looking at the trsaces of course)!
 
#10 ·
G'Day Bob,

The top pot (bias) is actually in series with the speedo and yes I have used it to make such adjustments.

The other one is a guess but seems reasonable, I have to make some test equipment before I can confirm.

There is a SAF-1092 which is listed as a "Speedometer and Mileage Indicator IC" but I can't find a datasheet for it. Pity as it would probably give the whole circuit.
 
#11 ·
There is a SAF-1092 which is listed as a "Speedometer and Mileage Indicator IC" but I can't find a datasheet for it. Pity as it would probably give the whole circuit.
If you search for the SAF1091 it usually includes the 1092.

I'd upload the one I have but I seem to be having trouble uploading files right now.
 
#12 ·
G'Day Tom,

Thanks, that got me the ITT information sheet covering the 1091/1092, sak215 and tca700y which provides some info but is not what I would call a datasheet. If you've found more you could email it to me via the site feedback link on my site.


It seems Alfa (or Jaeger) is using the 1092 to do the maths for the odometer and using discrete components to drive the stepper motor (for the odometer).

For the Speedo they are using an sak215 (yes, a tacho IC) plus discretes.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top