How to Debounce a Geo Metro VSS
 

I got my MPGuino installed and working ok but I noticed that it would jitter a lot at certain speeds. I took the speedometer apart to see how it worked and they are really pretty simple. The housing has a magnet inside that spins and makes a reed switch contact. The yellow/green wire is the vss output wire and the other side of the reed switch goes to ground. On my car there was no pullup on the vss side of the reed switch but the cruise control module was pulling it up enough to make it barely work. The automatic door lock thing has been removed from the car and that is probably what normally pulls up the yellow/green VSS wire to 12V. On a scope the signal was pretty noisy.

So my solution involved a couple of 1k resistors. I ran one from switched 12V to the vss wire. And another from the VSS wire to ground. A bit more than a normal debounce but I wanted a clean signal and without the automatic door locking thing I didn't have a clean 12V on the VSS wire. I recommend to do all Metros the same way just to make sure it has a clean signal.

On my site I finally posted a complete list of wiring diagrams for 89-01 Metros. On mine I went from a black/white wire on the same connector as the vss on the back of the gauge cluster for 12V and black/blue for ground. Optionally on the back of the cluster there are clearly marked + and - places so the resistors could be soldered to the back of the cluster as well.

That pretty much cleaned up the signal on the scope and it should work on all Metros as long as you verify the wiring. I spun the tires up to 65 and it read nice and smooth with no jitters in the display at all.

Nice work, Coyote.

I'm surprised to read there was an auto door lock feature in the Metro. I've never seen one with power locks.

It would lock the doors at 15mph but it did not unlock them. The 93 models was the first year for it and I am pretty sure it was also the last. It was really a stupid system. I can't think of any reason any one would want their doors to lock automatically so I killed that feature and removed the solenoids a few days after I started driving the car.

I wonder if the debounce resistors could be added right at the MPGuino instead of taking the cluster apart to add them to the back of it? That would make it a whole lot easier for people to add if needed.

re: add to guino, probably. Was thinking how to use a pin for configurable pullup or pulldown. Do you know what firmware version your guino came with?

I am using the latest firmware, I copied it off the project page when I reloaded the AVR.

The pin could be configured to pullup 5V but that is on the wrong side of the zener and current limiting resistor so it would maybe not have that much of an effect on reducing the jitter coming off the 10-20 year old reed switches. Adding a pullup resistor from the 12V line in to the VSS and one from ground to VSS on the 7805 would probably clean up any Metro as good as doing it on the back of the cluster.

Most cars probably use a better setup than a Metro to generate the VSS pulse so I am not sure if adding it to the standard board would be a good idea. It might confuse some ECUs possibly. I guess the first step would be a standard 1K debounce resistor from VSS to ground. And if that doesn't clear up the signal then adding the 12V to VSS resistor. But so far it looks like the Metros are really the only ones having issues and adding both resistors won't hurt a Metro so it seems ok to try.

I followed the VSS wire in the diagram and it goes to the door interlock, swift automatic seat belts, cruise control, and the ECU. My car doesn't have any of those things anymore so that is why my wire was not getting the full 12V. The audiovox cc that is in it now was slightly putting some voltage out on the line and that was what was letting my MPGuino read anything but I suspect that the 12V pullup on that VSS wire actually comes from the ECU and not the door lock module like I originally thought so anyone running the stock ECU will probably only need the VSS to ground debounce resistor. But it will take a car that is a bit less modified to test out what is actually required.

Forgive my noobiness, but where is the easiest/best place for this resistor, and just a basic 1k resistor?

And I'll throw another question in here: dcb, did you debounce your Metro's VSS?

Not on the 98, pretty sure it is a reed switch. When I went to the "average the last three readings" approach it got a lot smoother, and when I added the speed updates in .75 it got smoother yet.

Ah. OK, thanks. I'm embarrassed to say I haven't yet re-installed my Guino (I didn't bring it to the east coast for the winter) or updated the firmware. It's on the to-do list.

debouncing?
 

Is debouncing defined as: erratic readings for instant MPG? My 95 Geo has great readings for the Tank Mileage and the Trip Mileage, but the instant mileage feature is so up and down that it is really of little use to me.

Is this thread still active?

If it is not steady with a steady throttle it probably could use the debounce resistors to get it to read smoother. You can connect the resistors on the wires before it plugs in to the mpguino. Or really anywhere that it is convenient to do.

Ok, I keep reading this over and over... "So my solution involved a couple of 1k resistors. I ran one from switched 12V to the vss wire. And another from the VSS wire to ground.", ...but am not clear on it.

I wish I could draw you a diagram, but I am not that computer smart either. Do I connect both of these resistors in parallel directly onto the VSS wire? It makes me a bit nervous to add voltage to a circuit. I am trying my best to understand what this will do. Thanks for your patience with me.

Has anyone besides me thought that the bouncing up and down could be caused by the normal function of closed loop operation. I look at my mpguino when I am running down the road at a steady throttle setting (I don't have a cruise control) and the instant mpg's will read 36, 40, 38, 41, 35, 39,.... It kind of looks like what you would expect see with the ECU responding to the O2 sensor. I don't mind the bounce as it shows me I am running closed loop. Or not?

The mpg bouncing up and down on the instant mpg is normal. If the speed is bouncing up and down at a steady state speed then it is not reading correctly and needs debounced on a Metro. A lot of other cars won't be able to be debounced because they are not mechanical switches generating the VSS signal.

Something like my Astro has an electronic module generating the VSS and there is nothing that can be done to clean up the signal from it if it was noisy. As far as I can tell the Metro is about the only car that uses a mechanical switch to generate the VSS signal.

So to do it simply get a resistor, it doesn't have to be exactly 1k it can be 1k to 10k and still work so if you have one laying around use it. Connect one side of the resistor to the VSS wire going to the mpguino and the other side to the ground wire going to the mpguino. That will most likely be all that is needed to get the signal clean. If it is still bouncing around some then you can add a second resistor one side connected to a switched 12V source like the radio or heater blower fuse the other to the VSS wire going to the mpguino. You can't use the power wire going to the mpguino since it is 'always on' 12V.

That is how mine acts to moorecomp. Coyotye, can you explain in electronic terms how adding that resistor cleans up the signal?

I have a debounce question. With the latest version of the MPGuino, what is the preferred method to debounce a geo metro? If the double resistor method is not working well, should we consider using a capacitor instead of a resistor between ground and VSS? (see below)

What follows is what I've tried with my '96 geo metro 1.0L. For reference, I just got my MPGuino, so it is version 1.7 or something. I tapped into INJ underneath the airfilter, like someone has posted pictures of, for their '98 Metro. For VSS I went behind the glove box.

Without any debounce resistors, instant speed was bouncing around pretty bad while driving 30 MPH and up.

I tried adding a 1k resistor from ground to VSS (no additional resistors) and it wasn't getting any VSS reading at all.
I tried switching to a 10k resistor from ground to VSS and it worked again, with somewhat less bouncing.
Right now I'm doing 5.6k between VSS and Ground, and 5.6k between VSS and switched 12V. It still bounces some while driving 40MPH+.

The MPGuino is a great little device! It would be great to have a more definitive way to debounce the Metro VSS for it. More resistance? Less resistance? Keep the Ground and 12V resistance equal? etc.

Has anyone tried a Metro VSS debounce circuit with a capacitor?
It is explained at ikalogic.com/debouncing.php
If I understand the schematic right, for this design you would keep a resister between 12V and VSS, but instead of a resistor between VSS and ground, you'd use a 100nf ceramic capacitor. Any reason to not try this? Could it fry the VSS unit? :eek:

:turtle:

There is no VSS unit on a Metro. It is just a mechanical switch that is activated from the back of the speedometer. The VSS on a metro is only used for that idiotic door locking thing on the 93/94 and going to the computer to let it know how fast the car is going. I guess it could also be used for cruise control if someone is lucky enough to find a car with cruise. It is just switching 12v on and off on the wire. So everything connected to that wire can handle 0-12v.

So you can't really hurt it unless you just burn up the switch by grounding it out or something, even that would probably just blow a fuse. Just as an educated guess I would probably just go with a slightly higher value resistor on the ground side than the 12v side. If that didn't clean it up I would use a capacitor on the ground side with the 12v to vss resistor still connected. As these cars get older the reed switch is just going to get worse. So ultimately a new switch might need to be found and put in the cluster if it can't get cleaned up any better.

I have my cluster out right now doing some random work on my car so if I get time I will see if I can figure out where to order a new part from and see how big of a job it would be to change. From what I can remember it was a pretty simple design so it should be easy to put a new reed switch in the cluster. As these cars start passing 250k miles and 20 years that switch is probably not going to make it a lot longer than that and will need replacing to keep the mpguino working.

You can also experiment with the "VSS Delay ms" configuration item. It is the number of microseconds that it sits quiet before looking for another vss edge. Increasing it will reduce the top speed however, try one unit at a time, default is 2.

Thank you Coyote X for the encouragement to try the capacitor, and dcb for the VSS Delay idea. I will try upping the VSS Delay on the mpguino if I need to. The capacitor seems to work!

A Successful Debouncing of 1996 Geo Metro 1.0L w/ 140,000 miles:

• 100nF (same as .1 μF in radioshack units) ceramic disc capacitor connected from VSS to ground.
• 56k resistor connected from switched 12V to VSS

Now the speed is stable at 40+ MPH. Before, using only resistors, I was still getting bouncing of the speed in 40+ MPH range.

Would a polarized cylinder style capacitor work to debounce the VSS? I have a handfull I can choose from. What voltage & uF would be recommended? I'm tired of this noisy wire! Not only is it affecting my MPGuino, but now that I have a cruise control installed, it's affecting that as well.

I just had to debounce another Metro VSS, this one is probably the noisiest thing I have seen. I used pentium0's method of the capacitor to ground and a 47k resistor to switched 12V. I didn't have a 56 but anything in that range will be fine.

Now it is rock steady all the way up to 70mph. So anyone needing to clean up their Metro VSS signal just go ahead and do this method it is the best way to get a clean signal.