jeeping68

I am running an 87 5.7 TPI in my Jeep CJ7. I have upgraded to a $6E mask and modified it to accomodate the engine upgrades I have made. However, I am running a different gear ratio in the axles with larger tires. What changes do I need to make so the speed the ECM is getting from the VSS matches my actual rate of speed?

tequilaboy

You can either correct the vss signal upstream of the ecm with appropriate drive and driven gears, an electronic ratio converter, or scale it internally if you're not also using the signal to run the speedometer.

I'll have to dig for the address.

I think its this one, but its been awhile since I've played with it:
E0B4: LDD #$01CC

jeeping68

I would like to scale the ECM to read it correctly. I am running an aftermarket speedo that is fully programmable. In addition, the speedo feed comes out of a Dana 18 Transfer case and I don't have any options for correcting by way of a gear. Does anyone know what formula I can use for calculating the adjustment ?

jeeping68

I

tequilaboy

Take the value 01CC, convert to decimal = 460. Scale the value proportionately based on your current error, then convert back to hex.

jeeping68

Can I make this edit in TunerPro or TunerCat, or does it require me to edit the binary file directly? I have not edited the binary file outside of TunerPro and TunerCat yet. Definitely going to require some research on my part.

RBob

iTrader: (1)

I was wondering what the VSS in use was. This can make a difference. If the VSS provides a value of about 40 pulse per driveshaft revolution then a DRAC module is the way to go. A DRAC is a stock GM part that they used on many trucks & SUVs (IOW: real cheap and readily available at a JY).

They are strap programmable for various tire diameters and rear ratios.

Let us know what sort of pulse count is provided and we can go from there.

RBob.

tequilaboy

You can easily add the value to your editor of choice. For the above example, just use the address: 20B5 and size: 16 bit.

You can choose to display the output value as hex digits or decimal values according to preference.

Update: I went back and checked the original bin that I tested this with back in 2003. I had also changed the value below. To be honest I don't remember which change actually worked. If all else fails you can try them one at a time.

E0D4: LDD #$3000

For this item you would want to edit the 2 bytes starting at address 20D5

jeeping68

I'm using a 2000 pulse Painless Performance retrofit VSS. At the time, it seemed simpler than the DRAC, but looking back now, probably not. How do you go about adjusting the DRAC?

tequilaboy

How far off is the internal vehicle speed signal compared with actual?

RBob

iTrader: (1)

The DRAC has a series of jumpers soldered onto the board. By adding/removing them the input to output ratio is changed. They are all in a row which makes it easy to remove them and use a DIP switch in there place.
Look on the diy-efi.org site for a file called DRAC2.PDF (or close to that). It includes the info on the ratios and which straps to set.

RBob.

jeeping68

I am still a couple of weeks off from getting the wiring finished and firing it up. Before I tore it apart to install the different engine/trans combo, the displayed speed was about 5 mph off from actual at highway speeds. Once I get it back on the street, I can give you a more precise measurement. RBob, thanks for the info on the DRAC. I will take a look at the file and see what I can find.

tequilaboy

When I originally tested the internal scaling changes, I was trying to simulate correcting the vss for 3.54 gears, in place of 3.07 gears.

I just scaled the constants already discussed multilpying by 3.07/3.54.

The result was that the internal velocity signal was reduced appropriately.

Unfortunately the IP displayed mph was unchanged since the VSS signal in my car goes first to the cluster and then on to the ecm. The ecm change therefore only corrected the internal velocity.

A reasonably small error like 5 mph (or maybe 5-10%) should be easily corrected by this method as long as you're only trying to correct the ecm internal signal as previously discussed.

jeeping68

RBob, my VSS, much like the stock GM unit puts out 2000 pulses per mile. Later models use the reluctor wheel and it puts out 40 pulses per revolution. What is the ratio of pulses per revolution per pulses per mile?

jeeping68

Tequilaboy, I believe I understand now how to do the edit to the ECM. This is going into a portion of ECM mods that is new to me, so I don't want to burn my chip until I have an actual value to compare vs. the estimated one that I'm working with right now. Just to make sure I'm understanding, my actual speed is 55, but my displayed speed is 60. I would take the value in my ECM and multiply it by 91.66% (55mph / 60mph). Does that sound correct?

jeeping68

TequilaBoy,

I know your memory is a little rusty, but I thought I would run my values by you anyways. In location 20B5, I have a value of 56. In location 20D5, I have a value of 49920. Do those value sound similar to what you ran into?

tequilaboy

With a stock 6E bin, you should expect to see the following values:
At address 20B5: (16 bit) 460 (floating point)
At address 20D5 (16 bit) 12288 (floating point)

When scaled by 0.9166 or 91.66%, you should expect to see the following values (rounded down):
At address 20B5: (16 bit) 421
At address 20D5 (16 bit) 11263

Make sure you're not adding in some sort of conversion to the displayed value.

RBob

iTrader: (1)

I don't recall the actual ratio although I do know that they are too far apart. A 2K PPM VSS won't work with a DRAC. The reason I brought up the DRAC is when you mentioned using a transfer case. This opened the possibility that you were using a truck 700R4 with the 40T wheel.

RBob.

jeeping68

TequilaBoy,

I opened up several stock bin files (ARAP, ANYJ) I had and the numbers match up to your 460 at 20B5 and 12288 at 20D5. I opened up several AUJM bin files (what mine is based off of) and they all had the 56 and 49920 values. Looks like I will just have to experiment a bit and find what works for my setup. I appreciate the help. If nothing else, I've started to get into some new areas of the bin that I haven't touched before.

jeeping68

RBob,

Thanks for the input. I'm running an older 700R4 without the ring. I also have an 8000 pulse per mile VSS for my cruise control. I might have to do some calculations and see how that converts to revolutions. Now I know why they made me take all that math in school.

tequilaboy

Interesting, could be that the F-bodies and Y-body, used a different VSS sensor scaling.

Check the hex code for the preceding LDD opcode "CC" to verify that these addresses are used in the same manner as in the Y-body bins. If that's the case, then scale the values that you have as already described.

tequilaboy

I just checked a copy of AUJM, The addresses are shifted by a few bytes to the following values:
20BD
20DD

Update: I checked the road speed calc in the source code that I have, the second value at 20DD (#3000) is only a max value that is set if ACCA overflows. You should only have to modify the first value at 20BD (#01CC).

This value appears to represent .0005 miles/pulse or the inverse of 2,000 ppm.

jeeping68

Ahh, that looks better now. On the LDD, I found what I thought was the address of E0D4 on one of your previous posts, but my bin doesn't go that high. What is the address for checking the LDD code?

tequilaboy

For AUJM, the LDD should be at address 20BC.

The commented source hac's addresses start with C instead of 0. Therefore D would represent 1, E would represent 2, etc. An address that appears as E0BC in the source hac would appear in the editor as 20BC.

jeeping68

I am starting to understand the concept better now. I made the correction to the address and it is matching up now. Thanks for all your help. Now I just have to test it in the real world.