I'm a mechanical engineer by trade and training, but am not afraid to get my hands dirty urr zapped in regards to electricity. Forgive me if my terminology is off. I know this isn't as cool an endeavor as some other projects, but I want to document it for future reference. Because I'm a mechanical guy, electrical crap is always a project to me.

For the sake of science, I'm creating a device that can be input between the manual transmission VSS sensor.
Why? A myriad of reasons. First and foremost, a lot of people throw different height tires on our cars (DR's, etc); this will allow a user to correct the speedometer in either a + or - manner. This will give the user a more accurate speed reading for both performance and legal reasons.

Deliverable
A programmable circuit and associated documentation of such so that it can be easily reproducible by Joe-Blow off the street. This circuit should be plug and play into the harness, both easily installing and removing, allowing the user to adjust a percentage bias on the signal coming out of the VSS to calibrate the VSS. It is done in this manner rather than through the ECU of the car as altering the ECU is more likely to void a warranty.

Getting started, I gave my mechanic a ring. Looking up some info, here’s the pertinent information:

Knowns
1.) The voltage going into the VSS is off the B+ bus of the car. I’m assuming this is 12VDC. The connection scheme is as follows:
- (VT/L-BU) – Circuit 5294 – Powertrain Main Relay Fused Supply (5)
- (L-GN) – Circuit 2098 – Output Speed (Digital) Signal
- (BK) – Circuit 850 – Ground
2.) The connector going into the VSS is PN: 15373426. Mouser has it currently at $1.45ea HERE. The entire connector kit including the pins is GM PN: 19119347 which can be found on gmpartsdirect HERE for $22.32. It would be easier to order this as to not have to order a billion (exaggerated, but there’s a min order qty) terminals.
3.) I’ll be utilizing an Arduino micro, both for size, as well as having a better timer. You can get one off Mouser HERE for $22.49. These require USB (5VDC) power which we have in our cars.

Info currently under research:
1.) I need to find out what signal is coming out of the VSS. There has been a lot of talk of it being an AC signal. The only reason I could imagine this being done is for signal degradation. Although, with the cable lengths involved in a car, I don’t see how this is an issue. (Someone more knowledgeable please explain). From a manufacturing standpoint, converting the signal to AC would be a headache in regards to design, adding complexity for not much, if any, gain. Hall effect sensors already spit out PWM DC. For such, I’m under the presumption that the signal is in fact a PWM DC signal, which would be stupid easy to scale. If it is indeed AC, I’m pretty sure I’ll have to include some fast diodes and a mofset into the design. If anyone has more info on this, it would be greatly appreciated. Else, I’ll be hooking up a scope onto the sensor.

2.) I need help finding the male connector, the connector that’s on the VSS itself. Usually, suppliers are pretty good about the connectors counter parts in related products but I’m unable to find anything. I skimmed through the data sheets as well but couldn’t find anything. I’ll probably give Mouser a shout this morning to figure out what it is.

If you have any info pertaining to either of these, I’d love to hear it. As of now, my immediate goal is to get all these parts ordered in double qty’s. For this, I really need point 2 under the current research.

That's all for now.

 

Love it! Was going to get you the other info, but looks like you have it sorted.

Unfortunately, with this being a Delphi part instead of TE/AMP (and being EOL doesn't help us, either), it's going to be harder to find the mating connector, assuming there even is/was one.

Here is the molded part, itself, to offer some visual help:



On another odd front, I've seen AC crosstalk in my projects, but couldn't fathom the use of it in the car. It almost certainly has to be PWM, though it is not indicated on the schematic (see below).

Here's what you (most likely) already have in front of you. If not, it at least confirms what you already know regarding the physical parts and connections.



This should give you the rest of what you're looking for, along with your scope readings. I'd offer to help, but it's too hot here to be having any fun in the garage.

 

Called Mouser this morning and got the connector ordered. The male counterpart connector is PN 829-15326813 and manufacturing PN: 15326813 which can be found HERE
EDIT: looking at that, I don't think its the right one. . . oops. We'll see when they come in.

 

That's a good part number on the receiver end - the part, itself, has some oddball tabs molded, but as you see in the female plug, the keys are left wide open.

The part you just mentioned may be a good match - it appears to have the correct landing and pin width. Worst case, you'd have to remove the keys, but it appears both ends of the connector are actually fairly open to facilitate several keyed connections.

Love Mouser for this reason, alone. They'll find something that works, or figure out how to make it work. Lots of fellow EEs there that love a challenge.

 

Connectors came in. . . VSS end isn't a match. On the good side, I have the harness size. I'm just gonna hook up a scope on the power side to make sure its indeed providing 12VDC. If it is, I'm just gonna rig one of these up w/ a power supply set to that to take the rest of the readings. . .

 

Bump to the top.

I used the good VSS connector PN provided above and the associated pin inserts (mouser PN 829-153223266) to have a pig tail. I then used some paper clips (high tech, I know) to connect the other end to the existing connector on the cars harness. I clipped the pig tail of my improvised device to the paper clips. I first confirmed that the voltage is indeed 12V. I then connected to the (ground and signal) clips w/ my scope (Tektronix MSO4104B). Looking at the pic, you can see a clear PWM signal that's operating at 12V. That sucks since the arduino can only do 0-5V PWM so I'll have to bread board something up real quick to boost the voltage.





Going through the gears@idle, I got the following pulse widths:
1st Gear - 6mph - 2.6ms-2.9ms/+count (kinda all over the place)
2nd Gear - 9/10mph - +count/1.8ms-1.9ms
3rd Gear - 13mph - +count/1.4ms
4th Gear - 15mph - +count/1.2ms

Since these were only the positive portion (+count), you have to double it to get the full time domain/count eg @15mph count/1.2ms is actually 2.4ms in duration. (portion the signal is at 0)

Converting to get more info;
1count/2.4ms * 1000ms/s * 3600s/hr = 1,500,000counts/hr / 15 mile/hr = 100,000 counts/mile.

More math,

Using excel, for the 4 measured values, you get an average of 102,0081 counts/mile. Judging by the speedometers precision, I'm confident to assume 100,000 counts/mile.

Next up is writing some code. Also, I started putting together something that can easily be installed inline. This will involve creating a water tight housing to install the arduino. I have access to a 3D printer at work so I don't know if I will 3D print, or machine it at the shop; the latter insuring better weathering. . .

Some of the models already done:

 

Absolutely fantastic work - there's so much R&D and testing time (and often times a high cost) involved on these types of projects, but it's certainly appreciated.

There are a few Arduino waterproof enclosures, but I've found them all to be rather large (and rather expensive). I had thought about moving over from 'analog+opamp' to Arduino for my other exhaust project, so I had a fair bit of time spent in looking for something watertight. The machining option with a gasket would be the best route, ultimately.

 

This is all fascinating information But EFILIVE and HP Tuners already have software that can correct a speedo when changing tire sizes--Moreover on an auto trans they can both correct the "shift points" associated with a tire size change---
Also although a speedo correction on any Corvette is NOT required with a rear axle change a shift point correction on an auto IS---and again HP or EFILIVE and correct
this

 

While this is true, his intention is to do this without tuning as to keep the warranty in-tact.

 

Correct. And WAAAAY cheaper. InTune (which I have one sitting on my desk in front of me) was ~$350. . .

End goal for this unit is to have any bloke be able to make it/install it. . .

So you have a $20 arduino, $10 worth of connector parts, may be $5 worth of misc parts (transistors/bread board for the voltage boost). . . So I'm at less than $40. . .

And say you have DR's on and going a little faster than you should so johnny law gets ya. . . And you don't have access to your DR's when you go to get your speedo calibrated for a ticket. . . You can use this to account for the error induced by the DR's. . .

This is completely hypothetical however. ..

I just noticed something, you can't type out (less than sign) (dollar sign) (40)

 

Mail order tune $150

 

Umm, I'd say I suck at math, but I'm actually pretty good at it. . . so please explain how $50>$150?

Also, I'm gaining new info on our cars. . .

And also, please lemme know how the warranty works out on those mail order tunes; short of a diablosport, you'll wipe your warranty if anything happens. . .

 

Anyway, drew up the circuit I need to run;

Because the arduino only drives 0-5V PWM, I need to get it back to 0-12. Having a 12V source is great so I'm utilizing transistor. Unfortunately, this will switch the logic so I have to compensate (reverse) the PWM signal in the arduino. I also have to run some resistors as to not short/blow out the ECM in case there's a ground for the signal down stream.

Fairchild offers free samples (yay!) so I don't have to pay $15 shipping on $0.50 worth of parts. I spec'd out 3 of there transistors which should be here tom. (2N7002, PN2222A, and other I forget which)

Here's the circuit as of now:

 

Keep in mind for future orders that Mouser still has a first-class mail option that can cost a few dollars and arrive in 2 days, and they have no minimum. One of the few left that I use for small part orders. Newark stopped being cost effective even with the coupons due to shipping hikes.

Oh, and TI still has their samples program running, just a little harder to find.

 

It has been a while since I have worked with signals like this but it seems to me the designer of the receiving circuit in the ECM would have been just interested in the number of positive going edges plus some minimal pulse width after the edge to ensure they weren't picking up an electrical noise spike.

They could have added a complementary circuit that inverted the signal and thus used the negative going edges of the original signal with an appropriate pulse width to compare the actual signal with its complement to reduce the chance of error.

It seems like you need to do some something similar to reduce the chance of an electrical spike on the VSS signal causing an incorrect signal at the output of your circuit.

Bill

 

They didn't VSS is a 3 wire; they'd need a 4th wire for an additional feedback.

Built the circuit, completely preliminary setup. Gonna write the program for the arduino tonight and hopefully test it tom. . .

(Checking resistance on the high side to make sure I don't insta-fry everything. . .