1992 Jeep Wrangler with 350TPI/700R4 VSS Options

[misterpat]

Please forgive my noobness. I searched and found some advice, but no solutions.

Here is what I have so far.

Vehicle:

1992 Jeep Wrangler with original 4 banger efi with 5 speed manual

Swap engine:

Blueprint Engines 355 1 piece rear seal. Forged crank/rods/pistons. Dart heads and a tpi friendly cam.

Fuel Injection:

GM TPI from a 1989 IROC. Painless wiring with a 1227730 ecm running speed density.

Plans:
​
Remove all original fuel injection related wiring from Jeep harness and replace with painless and keep as neat as possible. I would like to retain all GM TPI emissions so I can get through inspection. Also keep all original Jeep gauges working along with speedometer. Basically keep it as neat and factory looking as possible.

As far as I know, the stock gauges will work with the GM senders. Oil, water, and Volt that is. Speedometer is a different story.

VSS Problems:

As far as I can tell from searches is the original Jeep VSS is a 8 pulse sender of the "Hall Effect" type. The GM TPI is looking for a 4 pulse VSS AC signal, I think?

Possible Soultions:

1. Use and aftermarket ECM that will accept the 8 pulse "Hall Effect" VSS?

2. Find someone to burn me a custom PROM that will work with the 1227730 ecm to accept the 8 pulse "Hall Effect" VSS?

3. Find a complete aftermarket ECM setup to use the 8 pulse "Hall Effect" VSS?

4. In 1991, Jeep used a mechanical speedometer and a VSS. The sender for this year had the mechanical speedometer and electric vss in one unit. I think I can use the VSS part of the sender to run the original Jeep speedometer. Where the mechanical speedometer would normally hook up, I can use a Painless 4 pulse generator for the 1227730 ecm.



I hope this makes sense. Sorry if I have been rambling on, but I'm just trying to wrap my head around this one issue I have with figuring out the VSS issue.
If I had my way, I would like to have option 2. This way I can use as many oem parts and senders.

Thanks for reading.

Pat

[fastacton]

I believe that Dakota Digital has a VSS signal converter that will work for you.

[misterpat]

That's definitely an option. I was trying to keep it to OEM parts to keep it simple to fix if needed on the road.

I found another option. Maybe someone can tell me if this will work. Install a VSS in the T-Case that will output the required 4000ppm for the GM ECU.

In this post here----> http://ls1tech.com/forums/conversions-hybrids/1233072-5-3-into-jeep-vss-question.html#post12811090

Quote from the thread.


- The Jeep uses a 3 wire hall effect VSS that outputs a square wave @ 8000ppm. Since the Jeep speedo (http://www.autoanything.com/gauges/77A2309A3715310.aspx?kc=AFFCJ&mr:trackingCode=D9E02D26-E47C-E111-9D2A-001517B188A2&mr:referralID=NA)is spliced into the VSS signal wire the speedo must need a square wave @ 8000 ppm.

- The GM ECM (http://www.amazon.com/gp/search?ie=UTF8&camp=1789&creative=9325&index=automotive&keywords=gm%2Becm&linkCode=ur2)has a square wave speedo output ,with software can be adjusted to 8000 ppm output.

Then use the output from the GM ECM to drive the factory Jeep Speedo. Can that be obtained through a custom prom?

[JeepsAndGuns]

I have been going through the same issue as I am planning on swapping the jeep ecm for a gm ecm to run my 4.0 in my 93 YJ.
The vss has been my biggest hurdle. After much research and different ideas, I have come down to this:
Replace the tailhousing on the NP231 with the tailhousing from a NP233 from a S10 or S10 blazer. These tailhousings have the required gm speed sensor and a 40 pulse tone wheel. I am then going to install a 4 out buffer/drac module and then also install a aftermarket speedo.

[misterpat]

Jeepsandguns,

This sounds like the way I'm going to go as I need a slip yoke eliminator anyhow.

JB Conversions makes a kit for the 231/233 with a GM sender.
http://www.jbconversions.com/products/sye/np231c_short_sye.php

My only question now is can a custom prom be programmed to output 8000ppm instead of 4000ppm on the speedometer output?

[JeepsAndGuns]

Got a little bored today as they called off work because of the weather.
I have one of those 91 YJ speed sensors. It is magnetic and has only two wires. It also should be a 8 pulse.
Anyways I hooked up a 7727 (cause I do not have a 7730, but they function the same) to my test harness and tried hooking the jeep vss to the vss inputs. It did not work. Then I remembered a old thread where there was a issue when hooking up a magnetic vss. The cure was to hook up one wire to ground, and the other wire to pin C6 on the 7730 (or pin B21 on the 7727) and then select the bit "magnetic vss installed". I gave this a try and ta-da, it worked! The adx file showed speed when I spun it by hand.
At one time I tried hooking a 7730 to my cherokee with MPFI instead of a modded 7427, the speed sensor didnt read in the adx. After I did the above mod, the speed read correct. I am running a 2 pulse JTR vss though, so a 8 pulse is gonna read much faster.
I do not know however, if the 91 only speed sensor will drive the 92-95 jeep speedo. The 92-95 sensors are hall effect just like you said. The 91 only vss is magnetic. Those are completely different signals. I suppose you could try, but I would be afraid it might hurt the speedo. You might try asking around on some jeep sites and see if a 91 vss can drive the 92-95 speedo.

However, now that I think about it, your idea of trying to drive the jeep speedo with the vss output of the ecm is a interesting idea, if it does in fact output a signal the jeep speedo could use. I wonder if it would work? Only thing I can think of, is they are using a newer (LS) pcm. I am not sure if the older 7730 can do this. If the input could be changed to take a 8 pulse vss instead of 4, would the ecm then output 8 pulse also? Someone much more knowledgeable about these ecm's will have to answer those questions.

[lionelhutz]

If you can switch the bin to use C6 instead of B9&B10 for the VSS then I would expect the hall effect sensor will work too. The issue may be that a 0-5V signal from the VSS isn't enough to drive the C6 input. It'd be a simple matter of using a transistor and 2 resistors to change the signal level to be 0-12V.

In the LS PCM's you can set the pulses per mile of the speedometer output pin. Default is 4000 but it can be set to any number. You'll have to open the bin/mask in Tunerpro you are using and see if you can change that output for the 7730 with the mask you are using.

The other option is that you set it up so the PCM calculates the speed as double what it really is then that output would become 8000 pulses per mile. You'd have to adjust every table with speed to be double what it normally would.

Also, I'm pretty sure the output of the PCM will be 0-12V so this may hurt the speedometer since the output of the Jeep VSS is 0-5V. Just using 2 suitable resistors can fix that problem.

[misterpat]

Ok, please correct me if I'm wrong.

A Hall Effect sensor is of the square wave type.
The Jeep speedo gets its signal directly from the VSS, so I assume a square wave signal will make it work.
The output from the gm ecm is square wave.

Looking at this screenshot of TunerPro, If the "Vehicle Speed Sensor Instrument Panel Pulse Divisor " is set to .5 shouldn't it output double the inputted signal?


http://i274.photobucket.com/albums/jj269/misterpat1975/tuner_zpsv5jjtj0v.jpg


Thanks for all this help with this. I never dug this deep into tuning before. Swapped a TPI once before, but it was a chevy in a Chevy. Piece of cake.

[JeepsAndGuns]

If you can switch the bin to use C6 instead of B9&B10 for the VSS then I would expect the hall effect sensor will work too. The issue may be that a 0-5V signal from the VSS isn't enough to drive the C6 input. It'd be a simple matter of using a transistor and 2 resistors to change the signal level to be 0-12V.

I'm not sure about having to set the bin, but I know I have hooked the jeep 8 pulse two wire (91 YJ only) sensor to pins B9 and B10 and the ecm would not read it. It did not matter if I checked or unchecked the flag for "magnetic vss installed" neither way worked. But hooking it to C6 does work.

So if this vss works, could you not set the value at 0x15 (vehicle speed sensor) to 8000 instead of what it is now (4006.96)? Would this not make the ecm use 8 pulse instead of 4, and therefore all the parameters that are speed dependent still work without changing them?

I do also have one of the 92-95 hall effect vss's and I also have a old XJ cluster I can use to test this with. The same year range XJ's used the same hall effect sensor as the YJ, so the speedo is looking for the same signal. If I have some time this weekend, I will try to wire it up and hook it up and see if it works.

Boy, if this could work, it would save me the time, headache, and money of having to swap tailhousings on my transfer case and buy/install a aftermarket speedo.

[lionelhutz]

'7730 TBI applications used C6 with a buffered speed sensor while '7730 TPI applications used B9 & B10 with an AC speed sensor.

It appears the Jeep sensor produces a square wave signal with a 5V amplitude.

I believe the GM TBI buffered speed sensor produces a square wave signal with a 12V amplitude.

So, if you want any chance of the Jeep sensor working you need to be able to switch the TPI bin use the C6 input and not the B9 & B10 inputs. Then, hook the Jeep sensor to the C6 input. It appears that switching that "magnetic speed sensor" bit will do this.

Still, you may have to deal with converting the Jeep 5V signal to a 12V signal for it to reliably work. A transistor and 2 resistors would do this for you.

I highly doubt the 4000ppm output signal is just a pass through of the C6 input signal. Remember, this PCM will also produce the 4000 pulse output when using the 40 pulse AC sensor. So, I expect that the PCM will calculate the MPH (or some higher resolution internal reference representing MPH) and then use MPH to calculate the 4000ppm output.

[JeepsAndGuns]

OK, I got some time and played around with the test harness.
I hooked up my 7727 and hooked up the 92 and up jeep vss. I hooked it to pin B21 (C6 on the 7730). I hooked it to a 5v reference and a sensor ground. I then hooked the XJ speedo to power and hooked the signal wire to pin B8 (not sure what that is on the 7730) Thats suposed to be the 4k out.
I loaded a bin to my emulator and gave it a go. With magnetic vss installed checked, it reads nothing. With it unchecked it reads! However, the jeep speedo does not move. I noticed I could spin the vss by hand and could get the speedo in the data stream dash to max out. I figured it was because it was a 8 pulse and not a 4 pulse. I tried changing the value in the bin from 4006.96 to 8000. It did not make a difference. I unhooked the signal wire from the ecm and hooked it to the jeep speedo. It worked normally. I could not get it to go past 15-20 mph by hand. Hooked it back to the ecm and I can max it at 130-something mph. I tried adjusting the parameter for vss all the way to its max and it did not make a bit of difference.
Also, every bin I looked at, the parameter for instrument panel pulse divisor was 0.00. I tried changing it to 1.00, still nothing on the jeep speedo. I tried 0.50, nothing. I went all the way up to 50.00 and still nothing. I pulled out my multimeter and tested the output wire from the ecm on both ac and dc voltage. I get nothing what so ever from that wire.

So my findings so far is that somthing does not jive. Its strange that it will read the jeep vss, but changing the pulse number does nothing. Also nothing is coming out of the ecm on the speed output.

[lionelhutz]

Try feeding both the speedometer and the PCM with the VSS.

[1project2many]

Looking at this screenshot of TunerPro, If the "Vehicle Speed Sensor Instrument Panel Pulse Divisor " is set to .5 shouldn't it output double the inputted signal?

That's a creative solution. Unfortunately the ecm is not created to do that math. The circuitry inside the ecm is very specific about what is allowed. I'll do my best to get this out correctly on the first try but it's been years since I've had to deal with it.

When using the "optical VSS" square wave input, the signal into the ecm is routed directly to VSS signal in the ecm. This signal was originally designed to work with a 2 pulse per revolution sensor on the speedometer cable. The cable is calibrated to spin at 1000 revolutions per mile so factory parts originally produced a 2000 pulses/mile signal to the ecm. It is not routed to the ecm/pcm 2k or 4kppm output pins.

The "magnetic VSS" signal is an a/c signal. The ecm circuitry creates a pulse each time this signal crosses 0V. As a reluctor tooth approaches then passes the sensor the signal goes from 0V to +V to 0V to -V and returns to 0V. This cycle creates two pulses in the ecm so the magnetic VSS input has a built in divide by two circuit to keep 1 reluctor wheel tooth equal to 1 pulse. The circuit is then routed to the ecm's internal VSS signal and also, separately, to a divide by circuit for conditioning before being routed to two 2kppm and one 4kppm ecm output. The magnetic VSS input does not create a pulse if the signal does not cross 0V so a voltage shift is probably required to use a square wave signal at that input.

There are two constants that are related to the speed signal and they are frequently confused. This is because they are for two different purposes. There is a "road speed sensor constant" which I've also seen labeled as "vehicle speed pulses per mile "or VSS Pulses." This value is used within the ecm calibration to calculate vehicle speed based on input pulses when the magnetic VSS option is selected.

The second constant is the instrument cluster or IP pulse divisor. This constant is used for the divide by circuit which modifies the incoming magnetic VSS signal to produce a square wave out out for speedometer, cruise, or other external components. Although there are 8 bits in this value, only three are used to select the proper divide by ratio. Selecting a wrong value often results in no speedometer signal. Divide by ratios are as follows:

A B C
0 0 0 Divide by 1 (ie for a 4000 ppm VSS)
0 0 1 Divide by 9 (36000 ppm VSS)
0 1 0 Divide by 7 (28000 ppm VSS)
0 1 1 Divide by 11 (44000 ppm VSS)
1 0 0 Divide by 6 (24000 ppm VSS)
1 0 1 Divide by 10 (40000 ppm VSS)
1 1 0 Divide by 8 (32000 ppm VSS)
1 1 1 Divisor disabled, no output

These bits responsible for controlling the divisor are the highest three bits in the byte. So the values to enter into your tuning software are as follows:

0 0 0 Divide by 1 (ie for a 4000 ppm VSS) (0 through 31 or $0 through $1F)
0 0 1 Divide by 9 (36000 ppm VSS) (32 through 63 or $20 through $3F)
0 1 0 Divide by 7 (28000 ppm VSS) (64 through 95 or $40 through $5F)
0 1 1 Divide by 11 (44000 ppm VSS) (96 through 127 or $60 through $7F)
1 0 0 Divide by 6 (24000 ppm VSS) (128 through 159 or $80 through $9F)
1 0 1 Divide by 10 (40000 ppm VSS) (160 through 191 or $A0 through $BF)
1 1 0 Divide by 8 (32000 ppm VSS) (192 through 223 or $C0 through $DF)
1 1 1 Divisor disabled, no output (224 through 255 or $E0 through $FF)

More to follow...

[maxpowerta]

It is my understanding that when you use a square wave sensor into pin C6 of the 7730 you do not get the speedo and cruise outputs at B11 and C1 those only work if you are using a 4000ppm mag sensor into B9 B10 (or whatever the equivalents are on the 7727) There is a nice diagram that illustrates this in the wiring diagram section of the EBL site.

[1project2many]

It is my understanding that when you use a square wave sensor into pin C6 of the 7730 you do not get the speedo and cruise outputs at B11 and C1 those only work if you are using a 4000ppm mag sensor into B9 B10

Correct for the 7730 ecm. Since this info applies to several different computer numbers, I've edited my previous post to clearly say the optical VSS input is not connected to the 2k and 4k outputs.



Anyways I hooked up a 7727 (cause I do not have a 7730, but they function the same) to my test harness and tried hooking the jeep vss to the vss inputs. It did not work. Then I remembered a old thread where there was a issue when hooking up a magnetic vss. The cure was to hook up one wire to ground, and the other wire to pin C6 on the 7730 (or pin B21 on the 7727) and then select the bit "magnetic vss installed". I gave this a try and ta-da, it worked! The adx file showed speed when I spun it by hand.

By wiring to the optical input you basically forced the ecm to read only the positive side of the magnetic sensor's a/c pulse. This might work, but it could also lead to a problem because the mag type speed signal voltage increases as the vehicle speed increases. The ecm may have trouble if the pulse voltage gets too large.



Also, every bin I looked at, the parameter for instrument panel pulse divisor was 0.00. I tried changing it to 1.00, still nothing on the jeep speedo. I tried 0.50, nothing. I went all the way up to 50.00 and still nothing. I pulled out my multimeter and tested the output wire from the ecm on both ac and dc voltage. I get nothing what so ever from that wire.

The XDF data definition and conversion formula for this value should not allow floating point numbers. There should also be a note about what that value means. I believe we could edit the formula so it will only allow meaningful changes but I don't know if TP can be setup to only use specific values. I'll explain this more if needed.



I hooked up my 7727 and hooked up the 92 and up jeep vss. I hooked it to pin B21 (C6 on the 7730). I hooked it to a 5v reference and a sensor ground. I then hooked the XJ speedo to power and hooked the signal wire to pin B8 (not sure what that is on the 7730)
J&G, I'd Tee the speed signal line off the sensor so it goes to both the speedometer and the optical VSS input. That signal may drive both devices. If it works then the next issue is the speedometer pulses per mile value. With the optical VSS bit checked the code normally ignores the VSS pulse count that you tried changing and does it's own internal check to determine the number of pulses per mile. The check is very simple: The code looks at the speed sensor pulse count and determines if it's one or two pulses per mile. If it's neither of those, it calculates the speed as if the sensor is three pulses per mile. There are several possible solutions to allow an optical speed sensor with other pulse counts to work in non-stock applications but the easiest answer for this situation might be to change the three pulses per mile value to 8 pulses per mile. If you can use TP to manually edit your calibration in the hex editor, you could change $3F93 and $3F94 from 0566 to 0E67 and see how that works. If it's good we could change the XDF to allow modifications to this location by other users for pulse counts greater than two. If you'd like me to make a change to a .bin for testing, lmk.

[JeepsAndGuns]

Holy cow info overload....

I had about come to the assumption that the vss outputs from the ecm only worked with the magnetic vss inputs but wasnt sure, you however have confirmed it.
So what exactly is different about the signal needed for those inputs Vs what the magnetic jeep vss is making? The 91 only jeep vss is only two wires, so I would assume its magnetic, however hooking it to the magnetic vss inputs, the ecm will not read anything. So something must be different.

Teeing the signal from the 92 and up jeep vss would work, as that how its done in factory jeep systems, it tees off to the speedo and ecm. However that wont make the ecm happy as I can max the speed in the ecm by hand.


If you can use TP to manually edit your calibration in the hex editor, you could change $3F93 and $3F94 from 0566 to 0E67 and see how that works. If it's good we could change the XDF to allow modifications to this location by other users for pulse counts greater than two. If you'd like me to make a change to a .bin for testing, lmk.

I'm up for giving it a try. I am very uber noob with hex. I really dont understand any of it. All I can understand is how to find addresses and can change them if I have the proper value.
On the bin I was using to test with (AXXD, 91 Fcar 5.0 manual), it has the value of B6 at $3F93, and 80 at $3F94. What should I change them to?
Just for the heck of it, I looked at AUJP (90 Fcar 5.7 auto) and AXCN (91 vette 5.7 manual) the were both 05 at $3F93 and 66 at $3F94. Any reason for the different values? I actually tried all those bins in the tests yesterday, just to rule out the chance it could be a issue with only 1 bin and not others.

[1project2many]

Holy cow info overload....
Yeah... I know. I figured if I put this into one thread it would save people including me from having to look all over the 'net for resources the next time this comes up.


The 91 only jeep vss is only two wires, so I would assume its magnetic, however hooking it to the magnetic vss inputs, the ecm will not read anything. So something must be different.
I agree. But I don't know what it is. It'd take some research to come up with an answer. I have a fairly expensive aftermarket sensor that was guaranteed to work with the GM ecm but it would only read correctly to 30 mph or so. I never did figure out what was going on, although I happened to find an article the other day that talked about needing a bias voltage for some GM VSS lines so I might try and follow that up.


On the bin I was using to test with (AXXD, 91 Fcar 5.0 manual), it has the value of B6 at $3F93, and 80 at $3F94.
Neat find! AXXD has 11 additional bytes of IAC control code located at $B1D6 (shows up at $31D6 in the hex editor). This moves the location for the speed sensor ppm value 11 bytes higher in the calibration, to $3FA7 and $3FA8.

[JeepsAndGuns]

Neat find! AXXD has 11 additional bytes of IAC control code located at $B1D6 (shows up at $31D6 in the hex editor). This moves the location for the speed sensor ppm value 11 bytes higher in the calibration, to $3FA7 and $3FA8.

Interesting. So if it is in different locations in two different bins of the same mask, then I am guessing that adding a parameter to change this value is not going to work and it will have to be done manually.
Not knowing hex that good, what is the hex values I need to put in to equil the numbers you posted above?


Another question, would the vss from the transfer case of a mid 90's truck work on the magnetic input? (ACDELCO Part # 15547452) Like I mentioned in my first post, I have a tailhousing off a S10 blazer NP233 that will bolt onto my NP231 and use that vss with a 40 tooth reluctor wheel.

[1project2many]

The hex values are 0E67. So you'd enter 0E at $3FA7 and 67 at $3FA8.

The mid 90's sensor will work just fine. I believe that same sensor was used with the 7427 in some applications.

[misterpat]

Info overload is right! I'm still reading all this, but ran out of things to contribute. Worst case senario would be I have to run an aftermarket speedometer. I really do appreciate this. :rockon:

[JeepsAndGuns]

The hex values are 0E67. So you'd enter 0E at $3FA7 and 67 at $3FA8.

I'm not gonna have the time tonight, but I will try to make some time in the next couple days to give this a try and report back.



The mid 90's sensor will work just fine. I believe that same sensor was used with the 7427 in some applications.

Hoping not to go too far off topic, but yes, thats exactly the sensor I am talking about. It was mounted on the transfer case tailhousing and read off the 40 tooth reluctor wheel. It was used in the 7427 year range trucks, as well as many others (S10's, blazers, etc..) Rock auto says it was used from 88-99 in fullsize trucks, and 89-04 in S10's.
Ok, so if this sensor could be used with the magnetic inputs on the ecm, could the speed sensor value be changed to accept the 40 pulse reluctor wheel?
I am still trying to understand the info in post #13. If the above could be done (use the 40 pulse vss) could the divisor be set to output a 8 pulse signal instead of 4?
I'm curious if the ecm output would drive the stock jeep speedo. IF by chance it did, then you could (maybe) fudge the input pulses slightly to account for different size tires/gears, thus in turn maybe making it to where you could correct the jeep speedo. Right now, with my tire size and gears, the closest speedo drive gear for my vss puts my speedo about 3-4 mph too fast. Being able to adjust it to be correct would be cool.
If only I could figure out a way to spin the reluctor wheel and hold that sensor so I could test it.....

[1project2many]

Cordless drill to spin the wheel?
The GM sensor and reluctor wheel should work fine. Speed sensor pulses value should correct sensor reading. Speed signal out depends on divide by ratios and they won't account for 2-3mph off. Try a DRAC?

[lionelhutz]

The second constant is the instrument cluster or IP pulse divisor. This constant is used for the divide by circuit which modifies the incoming magnetic VSS signal to produce a square wave out out for speedometer, cruise, or other external components. Although there are 8 bits in this value, only three are used to select the proper divide by ratio. Selecting a wrong value often results in no speedometer signal. Divide by ratios are as follows:

A B C
0 0 0 Divide by 1 (ie for a 4000 ppm VSS)
0 0 1 Divide by 9 (36000 ppm VSS)
0 1 0 Divide by 7 (28000 ppm VSS)
0 1 1 Divide by 11 (44000 ppm VSS)
1 0 0 Divide by 6 (24000 ppm VSS)
1 0 1 Divide by 10 (40000 ppm VSS)
1 1 0 Divide by 8 (32000 ppm VSS)
1 1 1 Divisor disabled, no output


This explanation doesn't make sense. Just to give examples from 2 of my vehicles, a 28" tall tire with 3.73 rear gear ratio produces 110557 pulses per mile and a 29" tall with 3.36 rear gear produces 96230 pulses per mile. There's no way to create either 2000ppm or 4000ppm signals from these values by only using these divider ratios. So, the PCM simply can't be taking the sensor input and only dividing by these constants.

[1project2many]

The output signal is separate from the internal VSS signal.

GM is known to use external buffers. I'm sure they've run into the same limitations. The DRAC in 4X4's and the yellow buffers in TBI F? cars are some examples. What are your vehicles?

[1project2many]

Info overload is right! I'm still reading all this, but ran out of things to contribute. Worst case senario would be I have to run an aftermarket speedometer. I really do appreciate this. :rockon:

No problem. I think this will work although we might have some trial and error work to do first.

[JeepsAndGuns]

Ok, did some more testing when I got home from work.
I hooked everything back up with the 92+ vss and changed the values you said to change. This made no difference. I tried two different bins (AXXD with the values at $3FA7 and $3FA8, and AUJP with the values at $3F93 and $3F94) It still acted the same as it did before.

I was curious about the truck magnetic vss we spoke about above. I had no way to spin the wheel and hold the sensor at the same time, until I remembered the NP205 I just modded to take the exact same reluctor wheel and sensor. So I drug it out (dang that thing is heavy) and hook it up to the magnetic vss inputs and spun the output shaft by hand. Ta-da! It worked. Read way too fast, but worked. I changed the vss value in the bin from 4006.96 to 40000 and it read more normally. Spinning it by hand put the speed at around 4 mph. Sweet, so it looks like this sensor and wheel can be used, and the input pulses probably can be changed to what is needed. So then I got curious, I hooked the jeep speedo up to power and hooked it up to the vss output of the ecm. Hey! It works! However it was reading the wrong speed. When the datastream dash said around 4 mph, the jeep speedo was reading around 20 or more. Now I dont know if either of these speeds would have been correct for how fast I was spinning the output shaft, but it was mainly just to see if it would work, which it does.
I'm thinking, if the right number of output pulses could be sent out of the ecm, it might just work with the factory jeep speedo.

[lionelhutz]

The output signal is separate from the internal VSS signal.

GM is known to use external buffers. I'm sure they've run into the same limitations. The DRAC in 4X4's and the yellow buffers in TBI F? cars are some examples. What are your vehicles?

It doesn't matter what the vehicles are. They are examples of the pulses per mile when using a 40 tooth reluctor wheel and sensor. You posted that the PCM takes the output from the 40 tooth reluctor wheel and divides by one of those constants to produce the 4000ppm signal and the math doesn't work.

There are no DRAC or yellow buffers in the sensor -> PCM -> 4K output circuit.

[1project2many]

Ok, did some more testing when I got home from work.
I hooked everything back up with the 92+ vss and changed the values you said to change. This made no difference. I tried two different bins (AXXD with the values at $3FA7 and $3FA8, and AUJP with the values at $3F93 and $3F94) It still acted the same as it did before.

Ok, just to get this straight...
You had the 92+ speed sensor connected to B21 and the "magnetic VSS installed" box was unchecked, you changed the values in the bin, and the vehicle speed reading in the TP dash maxed out?



It doesn't matter what the vehicles are. They are examples of the pulses per mile when using a 40 tooth reluctor wheel and sensor.
Why the attitude? When a good question comes up it only makes sense to get as much info as possible and start fact checking.


You posted that the PCM takes the output from the 40 tooth reluctor wheel and divides by one of those constants to produce the 4000ppm signal and the math doesn't work.
Not exactly. I did say the 40 pulse wheel should work but when I said "should" I'm not absolute. I was thinking of it in terms of voltage and signal shape... whether or not the signal would trigger a speed reading in the pcm. I don't happen to know of a confirmed application with the 40 pulse ring and wasn't really thinking of it in terms of count. I might even have assumed that your comment in post #10 was right...

Remember, this PCM will also produce the 4000 pulse output when using the 40 pulse AC sensor.
As I sit here thinking about it, I don't actually remember a 40 pulse per rev wheel used with the 7730/7727 in rear drive. As I remember, they first appeared in trucks then moved to cars as the LT1 was installed. Front drivers use a reluctor wheel but they produce a lower frequency signal because they're connected to the axle.

I'm confident the information regarding the handling of the speed signal and the divider is correct. And I'm confident that the internal VSS signal and the 2k / 4k output are separate. If you have an application I'll do my best to follow up and maybe we can figure out how GM made a 40 pulse input work.

[lionelhutz]

It doesn't matter what vehicle or PCM since all RWD vehicles with 40 tooth wheels on the trans output produce the same signal where pulses per mile = tire revolutions per mile x rear gear ratio x 40. So, if you use a 40 tooth wheel and GM sensor connected to terminals B9 & B10 then those are examples of the signal the PCM is starting with.

The schematics for a '7730 TPI F-body shows this speed sensor configuration. The schematics also show the 4K output to the speedometer and the 2K output to the cruise control. The sensor -> 2K and 4K outputs need more math than just that 1 of 7 divider to produce these outputs correctly.

I have not used a '7730 PCM but your explanation of how the PCM creates the 2k and 4k outputs just doesn't make sense. There has to be more to it than a simple divider.

[1project2many]

It doesn't matter what vehicle or PCM since all RWD vehicles with 40 tooth wheels on the trans output produce the same signal where pulses per mile = tire revolutions per mile x rear gear ratio x 40. So, if you use a 40 tooth wheel and GM sensor connected to terminals B9 & B10 then those are examples of the signal the PCM is starting with.

Ahh. I understand. All RWD vehicles do not use a 40 tooth reluctor. Corvettes and Camaros with the 7730/7727 use a gear driven speed sensor like this one from Eckler's:
http://htsmall.ecklerscorvette.com/assets/corvette/images/size/265x265/sku/46893.jpg
Correcting a tire or gear ratio change on these vehicles means changing the speedo drive to driven gear ratio, just like a speedometer. This sensor produces a 4kppm signal. You can see more sensors here:
http://www.jagsthatrun.com/Pages/SpeedSensors_Speedometer.html
GM sensors are about 1/2 way down the page.

[lionelhutz]

Well then you really can't expect to feed a 40 tooth drive shaft derived speed signal into the 7730.

This means you must feed 8000 pulses per mile if you expect the 4k output to drive the Jeep speedometer. You could also feed it 72kppm, 56kppm, 88kppm, 48kppm, 80kppm or 64kppm assuming those dividers work. You'd have to see if the driveshaft rotations per mile divides into one of those evenly, or at least close to evenly then you could build a wheel with that number of teeth.

This also means that 40 pulses per drive shaft revolution won't work unless you can set the calibration constant from 4000 to something in the 100,000 range. My guess would be that the PCM just uses a 16 bit constant which can only hold up to 65k since it only had to hold 4k.

[JeepsAndGuns]

Ok, just to get this straight...
You had the 92+ speed sensor connected to B21 and the "magnetic VSS installed" box was unchecked, you changed the values in the bin, and the vehicle speed reading in the TP dash maxed out?


Correct.
I could max it out by spinning it with my fingers. It did not seem to be any different from a bin without those values changed.

[1project2many]

Is this easy to set up and try again? If I post a few modified .bins, would you mind trying them?

[lionelhutz]

The internal PCM speed maxing out any time that Jeep VSS puts an output into C6 is telling me that no amount of code changes will fix it. Something is causing the PCM to detect many, many more pulses than what the sensor is really producing.

Do you have 2 x 1k ohm, 1/4 Watt resistors? I so, connect one resistor from 5V to B10 and the other resistor from B10 to ground. Then connect your optical VSS output to B9. Try that and see if it works. This way, you can just use the road speed constant as 8000 instead of 4000 to get the internal PCM speed correct and the 4K output simply becomes an 8k output.

If you don't have resistors, a TPS set to 1/2 the travel could be used to create 2.5V for B10 too.

[JeepsAndGuns]

Is this easy to set up and try again? If I post a few modified .bins, would you mind trying them?

Yes, very easy to try again. I keep the harness laid out and have a full set of sensors and injectors hooked to it ready to test run a ecm. I even have a ignition module, coil, spark plug and distributor so that I can do a full running test if need be.
Post up the bins and I will give them a try.

lionelhutz, I do not have any resistors. The jeep vss is not maxing it out as soon as it is spun, but it does read way way too fast for what I am spinning it by hand. (I am comparing it to how the jeep speedo reacts when the jeep vss is hooked to it) I can max out the speed in the ecm easy if I give it a quick spin by hand.
I do have a couple spare tps sensors though.
So your saying input 2.5v to B10, and the hall effect jeep vss to B9?

[1project2many]

The internal PCM speed maxing out any time that Jeep VSS puts an output into C6 is telling me that no amount of code changes will fix it. Something is causing the PCM to detect many, many more pulses than what the sensor is really producing.

I spent about an hour yesterday looking online for scope signals or OEM information showing pulse counts for the Jeep VSS with no luck. At this point I wouldn't assume the sensor produces 8 square pulses per revolution. Nor would I assume the ecm is detecting more pulses than the sensor is producing. However, the code as written could work with up to 142,000 pulses per mile so it would be beneficial to others to determine a good way to adjust for optical or square wave sensors of various pulse counts.


Do you have 2 x 1k ohm, 1/4 Watt resistors? I so, connect one resistor from 5V to B10 and the other resistor from B10 to ground. Then connect your optical VSS output to B9. Try that and see if it works. This way, you can just use the road speed constant as 8000 instead of 4000 to get the internal PCM speed correct and the 4K output simply becomes an 8k output.

What are you trying to achieve with this? This provides a constant 6V signal to a circuit that won't detect a pulse without a zero voltage crossing. Maybe we can design a circuit to achieve what you want to try.

[lionelhutz]

So your saying input 2.5v to B10, and the hall effect jeep vss to B9?

Exactly.

[JeepsAndGuns]

I spent about an hour yesterday looking online for scope signals or OEM information showing pulse counts for the Jeep VSS with no luck. At this point I wouldn't assume the sensor produces 8 square pulses per revolution.

Only thing I can find in my FSM just says it makes 8 pulses. There is a small chart under speedo calibration that says at 20mph it should be 44.4HZ, at 55mph it should be 122.2HZ, and at 75mph it should be 166.7HZ. Not sure if that helps you or not.
If you have the equipment that you can measure it with, I wouldn't have any problem dropping this sensor and pigtail in the mail to let you test it.

[lionelhutz]

that works out to 8000ppm by my math. Find the frequency at 60mph and multiply by 60 to get ppm. 166.7*60/75 = 133.3 pulses or cycles per second x 60 = 8000 pulses per minute. At 60mph the vehicle is moving 1 mile a minute. So this is 8000 pulses per mile.

So, the only question is what the output looks like. I could easily scope it but I doubt you'd want to ship it to Canada.

Try the resistors and the B9 & B10 inputs. That will put B10 at 2.5V and then hopefully this sensor should cause B9 to switch back and forth between 0V and 5V. The PCM then hopefully sees the difference voltage as AC and counts the crossings.

[sturgillbd]

If you guys can get the ecm and speedo to see the 8k pulse vss and the speed the ecm sees is doubled, as simple fix may be to use a flip-flop to divide the 8k to a 4k signal to feed the ecm. I used to use them as clock dividers. If it accepts the square wave signal, the flip-flop can divide it pretty easily. Feed the speedo with vss and tee signal to flip-flop circuit. Output of it will be half frequency. The flip-flop clock signal will be whatever the vss frequency is and the output will be half. There are probably cmos versions of them that will work at other than 5v. I'll attach a circuit so you can see how the pulsetrain outputs. http://www.electronics-tutorials.ws/counter/count_1.html

[1project2many]

Only thing I can find in my FSM just says it makes 8 pulses. There is a small chart under speedo calibration that says at 20mph it should be 44.4HZ, at 55mph it should be 122.2HZ, and at 75mph it should be 166.7HZ. Not sure if that helps you or not.
It does, thank you.


Try the resistors and the B9 & B10 inputs. That will put B10 at 2.5V and then hopefully this sensor should cause B9 to switch back and forth between 0V and 5V.
I didn't realize that you were calling out pins for a 7730 this morning. JeepsAndGuns is using a 7727. B10 is the 12V ignition feed on that ecm. So the correct pins for the 7727 would be C2 and C8.

The idea is to use the voltage divider resistors across C2 to bring VSSHi up, then use the speed signal at C8 to trick the detection circuitry into thinking it's swinging from a high on one pin to a high on the other. Does anything change in your circuit design if the A/C VSS input circuit is not isolated from the case? The pulse detection voltage is obtained at an R/C junction on VSSHi. C8 (VSSLo) is connected directly to ground. The OEM A/C VSS sensor signal works because it's a separate current generator independent of the 12V automotive circuit. To the A/C VSS, the ground connection through the ecm case is not the VSS ground, it's simply part of the VSS' circuit which allows voltage measured at the pulse detector to be positive, negative, or 0V relative to the battery ground. Connecting the digital VSS, which shares ground with the ecm, to C8 will ground the VSS output and pull the signal to 0V. The VSSHi signal will remain at a constant 2.5V once the cap to ground in the R/C filter charges. I predict the detection circuit will see no voltage transitions.


If you guys can get the ecm and speedo to see the 8k pulse vss
The ecm does see the 8k pulse VSS on the optical input. And the speedo sees the signal from the sensor as well. Many of the early GM vehicles are wired such that the vehicle speed sensor signal is delivered to both speedometer and ecm so wiring the Jeep like this would mimic an OEM GM installation.


as simple fix may be to use a flip-flop to divide the 8k to a 4k signal to feed the ecm.
True, but it would need to be further reduced to 2k to fall within the ecm's expected parameters at the optical input.

[1project2many]

Does the .adx you're using allow you to see promid? If so, would you mind trying this .bin and watching the promid. If I understand the code properly then the first number displayed will be fairly low, less than 10, when the VSS is spinning. The second number will increase or decrease with speed sensor speed.

Thanks.

[lionelhutz]

Hmmm, if either B9 or B10 is at ground potential then you can't apply 2.5V to that terminal or the 5V optical VSS signal to that terminal, which means that idea won't work. I was hoping both pins had input circuits on them and took the difference, similar to how the O2 inputs are. Got a schematic of the VSS input circuit?

[sturgillbd]

I have downloaded the bin that 1project modified and have it running on a 7727 on the desk. I have a signal generator set up feeding a square wave into the C2 and C8 pins. I have an oscilloscope connected to the output of the generator and can measure the frequency fairly accurately. I can output a sine, or square wave output and also induce a dc offset. I took a few pics of the screens and here are the results. 1project, you asked about the prom ID. It does change with speed. The whole prom ID number gets lower with frequency decreasing. It gets larger with increasing frequency. Attached are some pictures of the screen. I do not have any other sensors connected to the ecm but speed does display in the dash when connected in tunerpro. It is showing pretty much double what the speed should be.

[1project2many]

Got a schematic of the VSS input circuit?
http://www.exatorq.com/ludis_obd1/
Follow link for ecm schematics.


The whole prom ID number gets lower with frequency decreasing.

Does the first byte stay below ten? Maybe even three or less? It appears to be a pulse counter in code. IIRC (code and notes at home right now) the second byte should be the VSS.

[sturgillbd]

Yes, for any normal speed. I'll give you a ten sample chart if you want. It is normally 2, 3 or 4

[lionelhutz]

Good work. I see 133kHz which is 8k ppm so the PCM is expecting 4k ppm.

If I'm interpreting that correctly, your waveform in the picture has an amplitude of about 4.5V and it has an offset that makes it go 1.5V negative. Correct?

The issue seems to be getting the PCM to recognize the 5V Jeep VSS signal.

[sturgillbd]

It is 133 HZ and yes it seems to be expecting 4K/mile. The offset doesn't change any reading in the ECM unless you go above or below the 0V reference with the whole waveform. I did try to see how the offset would affect what the ECM reads. and that was my observation. I can turn off the offset but I do not have the VSS low tied to ECM gnd. Anything over 2V P/P, the ecm is rock solid on speed reading. I think if upper voltage is an issue, a zener clamp on the vss lines will solve that. I can capacitive couple the signal into the ECM and that will do away with any offset or bias and I can tell the results of that if wanted.

[lionelhutz]

Ok, but the last tests were trying to use the PCM optical input which is C6 on the 7730 or B21 on a 7727. Can you try your tests with 0-5V signals of different types on that input.

^ Jeep VSS tests is what I was referring to.

Don't use an offset. The Jeep VSS is powered by 5V and ground so I doubt it's producing anything besides a signal that goes from around 0V to 5V.

My first idea was bad so I've been trying to think of a way to introduce an offset to the Jeep sensor so it goes -2.5V to +2.5V and can be fed to the B9 & B10 (or C2 & C8) VSS input. A simple high pass C-R filter would work except for the transition between stopped and moving where it might take a little time for the capacitor to charge to 2.5V so it's giving the correct offset.

[sturgillbd]

Yes. It will take me a few mins to move some cables. I have to make sure my V6 pigtails have B21. If not, I will have to drill a hole to put a pin in. I will post back with results in a bit. I lost track in the thread as to which input you guys were trying to use. Is the signal between B21 and ground? Also do I need to change the flag for optical? Ill have to look at the xdf I have for 8D. I just tried it and I got no vss data in TunerPro. It will be a bit before I get back. Supper is almost ready :)

[1project2many]

Yes, for any normal speed. I'll give you a ten sample chart if you want. It is normally 2, 3 or 4

A chart would be great, thanks. The code tests for a count of 4, 3,or 2 then selects a pulse divisor based on the count. As long as the count stays below 4 it looks like all I need to do is alter the divisor values.


I lost track in the thread as to which input you guys were trying to use.
Well... we've got two different tests going on. lionelhutz is working on using the a/c input on C2 with the digital sensor. The ecm should be configured for a magnetic sensor to use this pin.

I'm asking about using the optical input at pin B21. The ecm should be configured for the optical sensor.


Good work. I see 133kHz which is 8k ppm so the PCM is expecting 4k ppm.
Yep, so doubling the current divisor values looks like the correct answer for the optical input.

FWIW, I'll be away from this for a bit, too. If the table looks as I expect, I should have another .bin to try.

[JeepsAndGuns]

Does the .adx you're using allow you to see promid? If so, would you mind trying this .bin and watching the promid. If I understand the code properly then the first number displayed will be fairly low, less than 10, when the VSS is spinning. The second number will increase or decrease with speed sensor speed.

Thanks.

Yes I do have a adx that can display that. It will probably saturday afternoon or sunday afternoon before I can run the test. Not sure if sturgillbd's test gave you what your needing or not. I can still run the test though.

[JeepsAndGuns]

Yes. It will take me a few mins to move some cables. I have to make sure my V6 pigtails have B21. If not, I will have to drill a hole to put a pin in. I will post back with results in a bit. I lost track in the thread as to which input you guys were trying to use. Is the signal between B21 and ground? Also do I need to change the flag for optical? Ill have to look at the xdf I have for 8D. I just tried it and I got no vss data in TunerPro. It will be a bit before I get back. Supper is almost ready :)

I had to drill a hole in the V6 car connectors I am using to be able to input the correct pin. And yes, you will need to uncheck the box for "magnetic vss installed" before that input will work.

[sturgillbd]

Ok, Breaking it down in 15Hz increments shows as follows

0Hz Prom ID 119 Speed 0mph
15Hz Prom ID 111 Speed 13mph
30Hz Prom ID 108 Speed 27mph
45Hz Prom ID 105 Speed 40mph
60Hz Prom ID 104/208 Speed 54mph
75Hz Prom ID 103/206 Speed 67mph
90Hz Prom ID 205/308 Speed 81mph
105Hz Prom ID 204/307 Speed 94mph
120Hz Prom ID 306 Speed 108mph
135 Prom ID 305/407 Speed 122mph

The two numbered Prom ID's are because the prom ID cycles back and forth between those two numbers even though the frequency is stable and the MPH reading is stable. This is with a 5v p/p square wave signal with no DC offset.

[sturgillbd]

I'll be back on a little later. I'm heading to a buddy's and I will take some measurements on that 4.0 Jeep TPS sensor while I am there. I should be back by 10:00pm. If needed, I'll do more testing when I get back and post results then.
Brian

[lionelhutz]

1project2many - I just want to point out that all the data sturgillbd has posted so far is for the magnetic AC input. Your posted code was intended to test at the optical input, was it not?

JeepsAndGuns - I'm not convinced the Jeep hall-effect VSS is giving a signal compatible with the PCM optical input based on the way you describe turning the sensor and getting a high speed. If your feed that Jeep VSS into both the speedometer and PCM optical input and turn it at a consistent rate, you should see the PCM showing a speed that stable and a multiple of the speedometer (it'd be 4X the speedometer I think). If you hook that VSS back to power, connect a voltmeter to it's output and turn it slowly and see what the meter reads. I am expecting that it switches between 0V and 5V but who knows what it really does.

sturgillbd - See if the PCM is stable at reading the speed when feeding the different types of 0-5V signals into pin B21.

[1project2many]

The two numbered Prom ID's are because the prom ID cycles back and forth between those two numbers even though the frequency is stable and the MPH reading is stable.
I had a hunch that was going to happen.


1project2many - I just want to point out that all the data sturgillbd has posted so far is for the magnetic AC input. Your posted code was intended to test at the optical input, was it not?

It was. If I'd remembered to select optical VSS the ecm would have ignored the magnetic input anyway. The results are still valid though. The variable displayed in the high byte of promid was the main concern. It's in play with either speed sensor type. With magnetic VSS it's value is multiplied by the road speed sensor constant to create a divisor. It's still at 4006 in the test cal so displayed speed is 2x measured speed. If that test is run again supplying signal at C2, magnetic VSS selected, and the road speed constant set to approximately 8000, the displayed speed will be correct.

With optical selected the code checks this byte and determines which hard coded divisor is selected. The divisor is a multiple of 461. Since the original optical VSS is 4k then it makes sense the divisor should be doubled. We haven't tested it though.

So if this is simple as it seems on the face, the attached calibration will display correct speed if an 8 pulse square wave at about 5V is supplied to B21. The optical VSS is already selected and the Promid is still modified to show two raw speed related bytes. I am expecting problems with calculated speed when the high byte of promid shows 04 but we'll see.


It will probably saturday afternoon or sunday afternoon before I can run the test. Not sure if sturgillbd's test gave you what your needing or not. I can still run the test though.

Sturgillbd's testing with the signal generator is helping this move along well. Eventually, with either solution, actual testing with a VSS will be necessary.

Again, a I'd like to give a big thanks to guys willing to take time and run tests, especially with two different solutions being worked on. It's really great to do this.

[JeepsAndGuns]

Again, a I'd like to give a big thanks to guys willing to take time and run tests, especially with two different solutions being worked on. It's really great to do this.

Its just as much to help me as it is to help the OP. We both had the same desire. Being able to keep the stock jeep speedo and vss, but also be able to run a GM ecm.
I will try and run the tests when I get home from work today and will report back when I do.

[sturgillbd]

With signal generator on B21 and square wave set for 0-5v, the speed is way too fast. With the other bin 133 Hz would cause 120 MPH. This new bin it is maxed out at 255. At 30Hz, it shows 108MPH and at 60Hz it shows 218Mph

[1project2many]

Thanks.

How about the high byte in the promid? 2,3,4 again or higher values?

[sturgillbd]

Prom ID's appear to be the same as the charted values posted last night. Only the speed displayed has changed
I can post the results every 15Hz again if needed

[1project2many]

No, that's good, thanks.

I looked at the code this morning and I have a preliminary rewrite that allows the road speed constant at for the magnetic VSS to be used for both speed sensor types. That would save time with trial and error and would allow one to determine pretty readily whether or not the optical input is viable.

If running this code and doing trial and error with the road speed constant seems like too much hassle, I can try and dig out my ecm bench and my speed signal generator.

[sturgillbd]

I guess that choice is up to the OP and JeepsandGuns. I'll test any code changes with the bench setup you guys want. I'm going to be around home pretty much all day. Its pouring rain here so outside work is on hold. I'll check the thread for updates throughout today. I just set up the signal generator so we could input a known frequency and test changes. Thats a lot easier than trying to spin a sensor with a drill etc. I'll load up the bins later and do a compare to see what you are changing.

[lionelhutz]

Hmm, well if I'm reading the schematic right then B21 goes into a logic gate and the gate is CMOS and powered by battery voltage, so the input signal must switch from 0V to Vbat-3V. So, 0V to 9V if you're using a 12V supply, or 0V to about 11.5V if it's in vehicle. This could be tested by increasing the signal voltage and see what voltage is required for the PCM to read the correct speed.

There are 3 ways to fix this problem.

#1 - Power the sensor by battery voltage.
- the sensor could just output at 12V switched signal instead and everything would work correctly.
- the higher voltage might also kill the sensor
- feeding the higher signal voltage to the speedometer might hurt the speedometer

#2 - build a circuit to take the 0-5V signal and convert it to a 0-12V signal
- this is fairly easy using a transistor and a couple of resistors.

#3 - build a circuit to take the 0-5V signal and convert it to a -2.5V to 2.5V signal and use the magnetic input.
- I think a capacitor and resistor of the correct value would work. Basically a high pass filter which removed the DC offset
- this could have issues measuring speed when transitioning from a stop to moving

[1project2many]

I guess that choice is up to the OP and JeepsandGuns. I'll test any code changes with the bench setup you guys want. I'm going to be around home pretty much all day. Its pouring rain here so outside work is on hold. I'll check the thread for updates throughout today. I just set up the signal generator so we could input a known frequency and test changes. Thats a lot easier than trying to spin a sensor with a drill etc. I'll load up the bins later and do a compare to see what you are changing.

Once I get back from the shop I'll post the new file. Prolly 2:30-3:00 eastern time.

[sturgillbd]

I just checked the lowest input signal it can accept and it is about 2.6 Vp/p. I just now measured the input and it is internally pulled up to battery voltage. When I connect the signal generator directly to the input, it lets signal bounce betwen 0 and 5v or whatever I set the amplitude of the output. If this is a concern, all that would have to be done, is to use a 2n3904 transistor with about a 1k series resistor to the base from the sensor. Tie the collector to the B21 input and the emitter to ground. This input is just switched to ground through the sensor. It works fine if the input is pulled to 5v on high and near zero on low so the direct connection of the sensor to the input will probably work.

Edit: I can breadboard my solution if needed

[sturgillbd]

I went ahead and breadboarded the transistor circuit I described. It made absolutely no difference in what speed the ECM shows in the datastream. So a 0-5v signal is no different than a 0 to battery voltage signal. The 0 -5 volt VSS will supply the proper signal if the code can be worked out.

[lionelhutz]

I was expecting the code would just use 2k instead if 4k meaning 133hz would produce 240mph. But, it's never that simple.

The magnetic VSS input goes VSS -> zero crossing detector -> divide by 2 flip flop -> logic to select which input -> processor

The optical VSS input goes VSS -> inverter -> logic to select which input -> processor

Here's the schematic of the circuit. Once processed, either VSS signal should produce the same the same signal at VSS on the right. Well the ZCD and flip flop likely have the magnetic still putting 4k at VSS while the optical puts 2k at VSS, assuming the ZCD produces 2 pulses per cycle as 1project2many posted earlier. If the code just switches out4 then the other input would work, but it has a 2x speed multiplication on it. You have pretty much proved the 2x pulse difference is there with your test giving 30hz producing 118mph instead of ~60mph. So, 1project2many is correct. If he can figure out how to count 8k pulses instead of 2k pulses it should work. So, routing the optical VSS to the same magnetic VSS code should make it work.

I'm just curious why the code suddenly starts counting the optical input as other than 2kppm when you increase the pulse input from 60hz to 133hz. How does it decide that's not a 2k ppm input?


8638

[sturgillbd]

This is the second bin file that 1project edited. I can flip the flag and tell you the results of connecting to the magnetic vss inputs too. I may just connect my emulator up so the changes can happen more easily.

[lionelhutz]

If you're bored, try slowing increasing the frequency into the optical input and see if there is a sudden jump in speed. It seems the PCM drops down to a 1000ppm input with a high number of VSS pulses. Just curious if it does it on the fly.

[sturgillbd]

It seems to be pretty linear on the speed. I noticed no sudden jump in speed. Lowest it will register is 4 mph then drops out. I just rechecked with optical vs magnetic. Magnetic seems to be expecting 4kppm and the optical is expecting 2kppm. That is just flipping the flag and moving the input. It does need the slightest offset to make it cross zero for the magnetic sensor to register. Emulator is making it much easier to test. I didn't zoom in very close to see that last night that the signal was just crossing zero to a negative voltage when I was connected to the magnetic inputs.

[lionelhutz]

If you're still bored, try a 10uf cap between your 0-5V signal and the magnetic input and a 100k resistor between the magnetic input and ground. See if that gives enough DC offset that it works.

[sturgillbd]

Been looking for a dual flip flop IC in my parts bins. The second bin file that 1project posted seems to want 1kppm for the optical. The first bin seems to be 2kppm. I had noticed that earlier and decided to try and find the ic so I could do a divide by 4 on the signal generator output. If it works, I can post up a diagram and pics to show how to do it without modifying the code. It would just be a flag switch and go. The chips are less than a dollar. A few parts, and a perf board and the divider could be built with hardware pretty simply. We'll see how all this stuff plays out. I'll still test bins etc but I am going to work on a hardware solution.

[JeepsAndGuns]

Looks like a lot has happened since I left for work this morning. I got home around 12:30, had to eat then run some errands and I was finally able to get some testing done.

I started out with test bin #1 and run it. I figured it would be easier to record it then try to describe it. I had already run it a tiny bit before I started recording. I tried to vary the drill speed a bit for different speeds. I also went ahead and hooked up the jeep speedo I have so I could compare the datastream speed to what the jeep speedo reads I show both so you can see.
About halfway through I paused it, switched to the 2nd bin and then tried to run about the same test. I did notice the 2nd bin read much faster than the first.
https://www.youtube.com/watch?v=-Svl0gAOVgo&feature=youtu.be

[1project2many]

The second bin file that 1project posted seems to want 1kppm for the optical. The first bin seems to be 2kppm.


I did notice the 2nd bin read much faster than the first.

Just to keep track, are both of these comparisons against the the TEST1_ANHT file using optical sensor input on B21?

Anyway, here's the latest file. The VSS road speed constant is set to approximately 8000 pulses but can be changed in your favorite cal editing program. lmk.

[sturgillbd]

I found a 74ls74 and wired it up on the breadboard. It works but the speed varies 6-8 mph at low speed. I uploaded a video of it and you can see why. I don't know exactly why it doesn't like the low speed but the pulse train doesn't clean up until the input frequency goes up. It was a test with an old chip. Could be the chip or just trying to use the chip at a really slow speed compared to what it is designed for. The video will show what I am describing. I'll take the divider out of the circuit and load up the new bin.

https://youtu.be/5dkDiUMUVbs

[lionelhutz]

Try the cap and resistor into the magnetic input. It'd be simpler and let the magnetic VSS pulse setting function.

[sturgillbd]

I found my issue with the divider circuit. I needed to tie the preset and clear pins high. They were in the gray area and were getting triggered sporadically causing the irregular output. The divide by 4 now works great. It has a smooth transition from 2mph through 190mph in the datastream. It also shows 60MPH with 133Hz feeding the divider.

[sturgillbd]

If anyone is interested, here is what the circuit looks like on a breadboard. This breadboard has seen better days. First pic is the circuit. Second is of dash in tunerpro with ecm running S-AUJP with signal of 133 hz feeding divider. Next is of signal gen at 133 hz and output of divider on the scope. Last is of signal gen at 6hz and what the divided signal looks like. All stable.

[JeepsAndGuns]

Just to keep track, are both of these comparisons against the the TEST1_ANHT file using optical sensor input on B21?

Anyway, here's the latest file. The VSS road speed constant is set to approximately 8000 pulses but can be changed in your favorite cal editing program. lmk.


My tests were with test1 and test2 bin files you made. The test2 bin read much faster than test1. And yes, it was inputted to B21.
I will give the new bin a try and report back this afternoon.


Another thing, we don't have to be dead set on making the jeep vss work. Remember I tested the gm truck magnetic vss and 40 tooth wheel and it worked fine on the magnetic inputs, and the output of the ecm did run the jeep speedo, it just read wrong. If it could be easier to change the output of the ecm, we could try that too. It is not really difficult to change the tailhousing on the transfer case. In fact I already have one I pulled from the JY.

[sturgillbd]

I ran a quick test on the test3 bin. It appears to be wanting a 4kppm signal on the optical input (B21 in my case 7727) . I changed the vss pulse parameter to 16024 (which Tunerpro saved as 15890) and it appears to accept the 8kppm signal. Tunerpro dash data follows my calculated input frequencies for speeds.

[lionelhutz]

Remember I tested the gm truck magnetic vss and 40 tooth wheel and it worked fine on the magnetic inputs, and the output of the ecm did run the jeep speedo, it just read wrong. If it could be easier to change the output of the ecm, we could try that too. It is not really difficult to change the tailhousing on the transfer case. In fact I already have one I pulled from the JY.

That combination of 40 tooth sensor and 4000ppm output will only work to produce a 8k output if the tire revolutions per mile times your gear ratio times 40 is equal to 2x one of those 7 values 1project2many posted way back in post 13.

[1project2many]

ran a quick test on the test3 bin. It appears to be wanting a 4kppm signal on the optical input (B21 in my case 7727) . I changed the vss pulse parameter to 16024 (which Tunerpro saved as 15890) and it appears to accept the 8kppm signal. Tunerpro dash data follows my calculated input frequencies for speeds.

Excellent! I suspected you'd end up near 16k. In the original code the optical divisor value for a 2k input is double the value used for a magnetic 4k input. I've added one more operation to double the divisor during calculations for the optical sensor. This will allow the road speed value to match the sensor output. If this file works as expected, I'll go back and remove the other files from this thread.


Another thing, we don't have to be dead set on making the jeep vss work.

It wasn't just about the Jeep VSS. This problem has manifested itself in various ways over the years and the answer most often is to give up and buy an aftermarket part which is fairly expensive. It's very cool to have two reasonable solutions available. Having an option to easily use a square wave signal on the magnetic input is very applicable to later generations of pcm which do not have the circuitry or code for a square wave input. The second solution is code based and while it can be patched into $8D easily, other code masks will require different patches to achieve the same result.


Remember I tested the gm truck magnetic vss and 40 tooth wheel and it worked fine on the magnetic inputs, and the output of the ecm did run the jeep speedo, it just read wrong.

That combination of 40 tooth sensor and 4000ppm output will only work to produce a 8k output if the tire revolutions per mile times your gear ratio times 40 is equal to 2x one of those 7 values

There's also a 2k output line available but it may take removing teeth from the reluctor wheel to get a frequency that works.

[JeepsAndGuns]

Ok, I loaded up test3 bin and gave it a go. Speed looked ok in the bin and read a little more normally. However it was still off from the jeep speedo, still too fast. I checked this thread real quick and saw sturgillbd's post and changed the vss parameter to 16024 (which it saved as 15889.66) and it worked! The data stream dash reads about 2-3 mph slower than the jeep speedo, but that's close enough to work with! Nice job :thumbsup:
What is test4?



That combination of 40 tooth sensor and 4000ppm output will only work to produce a 8k output if the tire revolutions per mile times your gear ratio times 40 is equal to 2x one of those 7 values 1project2many posted way back in post 13.


There's also a 2k output line available but it may take removing teeth from the reluctor wheel to get a frequency that works

Would adding a drac/4 out buffer/whatever they are called, like used on mid 90's trucks help any?

[sturgillbd]

Having an option to easily use a square wave signal on the magnetic input is very applicable to later generations of pcm which do not have the circuitry or code for a square wave input.
I have to back up and punt on that one because the original test was flawed on my setup. I used the normal square wave output on the generator but had offset the signal to the low side being what looked like 0v. When I zoomed in with the scope, the low was actually at -.089V. The low was actually crossing zero. I may have a solution for feeding it with a 0-5V signal but I have to use another chip. TTL is 0-5v DC. I will set up feeding a max232 chip with the 0-5v signal and the max232 output will be approximately +13 to -13 v. I will set up a test circuit and report the results in a new thread.


What is test4?

I think 1project2many was trying to correct where we were having to double the pulses per mile setting in the bin to get the proper output. I tested the new bin and had to quadruple the ppm (32000) to get the speed in the datastream to show correctly. One interesting note is now I can display 1Mph where with the other bins, I could only make it display 2Mph and above.

[1project2many]

I think 1project2many was trying to correct where we were having to double the pulses per mile setting in the bin to get the proper output. I tested the new bin and had to quadruple the ppm (32000) to get the speed in the datastream to show correctly. One interesting note is now I can display 1Mph where with the other bins, I could only make it display 2Mph and above.

Interesting. It seems like using the pulse per mile count with no doubling is the answer in this case. It doesn't make sense though. GM wouldn't write code with a dedicated divisor that's 4X larger than it needs to be.

New code posted for trial and error. I'll have to see if I can figure this out tomorrow, though.

[lionelhutz]

Would adding a drac/4 out buffer/whatever they are called, like used on mid 90's trucks help any?

Not without modified code. The DRAC has a 2k & 4k output. You might be able to re-purpose the outputs to be 8k and 4k, but the PCM still expects 2k on the optical input so you have a mismatch and need the modified code to turn the 2k PCM optical input into a 4k input. You could feed it into the magnetic input but that would require something between the DRAC and PCM to put an offset on the signal so it crosses below 0V.

The signal must be converted from 0-5V or 0-12V to go below 0V and then fed to the magnetic PCM input to use either the 40-tooth wheel and DRAC (so PCM is also feeding the Jeep speedometer) or the Jeep VSS without modifying code as-is. I suspect it'd work OK just using a capacitor and resistor low pass filter, but it might require a more complex circuit to level shift. Besides the MAX232 chip, an simple op-amp differentiator circuit would likely also work.

[JeepsAndGuns]

I should have time in the next couple days to try test5. Anything specific to look for or do?

[JeepsAndGuns]

Ok, sorry it took so long to get back with test results, I have been really busy this week.

I finally had a little time this afternoon and tested test4 and test5. Here are my results.

Test4. Read way too fast. Ended up changing vss parameter to 30000 (which tp saved as 29729.03) to make it read correct. With that value, the tp dash read correct to 1 mph slower than the jeep speedo. I did find that after the jeep speedo read 65 and up, the tp speed started reading 5 mph too fast.

Test5. It read much closer, but I had to decrease the vss parameter by 1000 (which tp saved as 7035.11) With that value changed, it read about right at 1 mph faster than the jeep speedo. However, it started reading 5 mph to fast after the jeep speedo got to 30 mph and above.

Hope this helps.

[misterpat]

I guess no one thought it would prove to be this difficult, huh? Thanks for all this work!

[1project2many]

I don't know if you noticed but there is a solution thanks to sturgillbd.
http://www.gearhead-efi.com/Fuel-Injection/showthread.php?4275-VSS-signal-converter

I'm currently working on cleaning out some space to set up my ecm test bench to further work on the possibility of a code based solution.

[JeepsAndGuns]

Yea I saw that, but I was not aware it was taking the place of this discussion.

There are little boxes sold in the aftermarket that can be used to correct the jeep speedos. But they does not change the output signal. They are a little expensive and that's what has kept me from buying one.

http://www.summitracing.com/parts/jet-50110

http://www.bluemonkeymotorsports.com/products_ht.html

[1project2many]

If I get that ecm bench set up to do more testing I'll post back here. The handful of electronics required for the converter in the other thread is much less than any of those aftermarket parts. You could buy enough pieces to build two converters and use the first one for practice.