7060 PCM - - DRAC is Current Sensor

[b65cuda]

I have an intermittent issue that come and goes with ignition cycle. When reading live data ($85, A138) if the "DRAC is Current Sensor" = No/Magnetic, then my PCM reads the vehicle speed and the TOSS as about 1/2 causing it to

1)Shift at very high RPM and
2)Think the TISS/TOSS ratio is >4 and maxes out line pressure causing hard shifts.

On the rare occasion that DRAC is Current Sensor = Yes, TISS/TOSS and ratio read correct, pcm vehicle speed seems correct (a little lower than speedo like 40 vs 45).

The speedo always looks good.

Looking at o-scope signals
The AC signals from TISS, TOSS, and VSS look good and freq/amplitude tracks with speed.
The DRAC to PCM signal look like a clean sq wave at 5v that freqency tracs with speed.

So how does the PCM decide to use the drac or not, and when it decides no and uses magnetic, why does it seem to half all of the values of the TOSS and vehicle speed?

Appreciate any insight anyone might have, or any other troubleshooting steps I might be able to take.

Thanks,

Tony

[dave w]

I'm unable to answer the question; "how does the PCM decide to use the drac or not, and when it decides no and uses magnetic".

One would expect the .bin file "Switches" to define DRAC or Magnetic. Possibly the DRAC or Magnetic option is an undefined parameter?

The 7060 uses two different definition files, $85 & $D8.

http://www.gearhead-efi.com/Fuel-Inj...Information-85

http://www.gearhead-efi.com/Fuel-Inj...erceeded-to-D8

I've noticed $D8 .bin files error codes 24 (Vehicle Speed Diagonostic), 72 (Output Speed Loss Diagonostic), and 74 (Input Speed Sensor Diagonostic) are not selected or enabled.

It appears $D8 definitions start with Letter "B" and $85 definitions start with "A".
Example $85 "AWYN"
Example $D8 "BAJW"

I'm left wondering why $85 was superseded with $D8

[b65cuda]

I decided to pull the PCM out. Once the cover was off, i saw a memcal adapter and chip labeled AXXZ from The Turbo Shop, INC. Removed it and just installed the original AWBP memcal alone and so far so good. It was an intermittent issue, so I won't know for sure until I get a few drives on her.

[dave w]

I decided to pull the PCM out. Once the cover was off, i saw a memcal adapter and chip labeled AXXZ from The Turbo Shop, INC. Removed it and just installed the original AWBP memcal alone and so far so good. It was an intermittent issue, so I won't know for sure until I get a few drives on her.

It appears AWBP has been superseded AWBP --> AXXZ ---> BAJW ----> BDPN

I attached a Read Original BDPN .bin ($D8) from my library.

AWBP 01.JPG

AXXZ.JPG

BAJW.JPG

BDPN.JPG

[b65cuda]

Well, it came back so it is still intermittent. The "DRAC is current sensor" isn't set in the bin, but it sets right after the vehicle starts moving. It's like it looks for a signal and sets it if it exists. It doesn't check again until you cycle the ignition and drive again.

The PCM supplies 5v to the drac and the drac switches to ground at the same frequency as the VSS. The VSS being a sine wave that increases in amplitude and frequency with speed and the drac being a 5v square wave that matches the VSS frequency.

I've hooked the o-scope up to both VSS and drac and always see the expected results. I hate just replacing parts, but without being able to see or know what the computers logic is for setting that flag, I don't know what else I can do. Everything else seems to be working.

I'm studying a hacked bin and it looks like it has a timer to switch between magnetic and drac and look for valid input. Once it finds a valid input, it sets that as the vss to use.

So now I am thinking that sometime in its past someone hooked up too many speed sensors. I have found a lot of conflicting information for this model suburban. I have both TISS and TOSS as well as a Transfer case VSS. Maybe the TOSS was never hooked up from the factory? I'll have to try and unplug it and see what happens.

[dave w]

The $D8 definition file I use is from TunerCats, which does not have a defined Constant for DRAC vs. Magnetic.

Possibly TunerCats can provide an updated $D8 definition file with a defined Constant for DRAC vs. Magnetic?

Recently TunerCats helped me with an updated $4E definition which now defines parameters for Air Diverter.

TunerCats is not freeware.

[1project2many]

Interesting... $85 / $D8 have the ability to guess at which sensor type is installed. The data word at $0043 contains a status bit for the type of sensor currently selected. If code is set to guess at which type of sensor is present, and the code does not see one type of speed sensor signal present within a preset time after selecting one type of sensor it will toggle this bit to select the other type.

The mode word constant at $900A has three bits related to speed sensing. Bit 0 selected indicates TISS input shaft speed sensor installed. Bit 4 selected indicates output speed comes from DRAC sensor. Bit 7 is used to indicate the code should guess at what type of sensor is being used. BAMW, AWAY, and AXXZ are both set to $80, "Guess at speed senor type." Setting this bit to $20 will deselect "guess at speed sensor type" and select "Output speed from DRAC sensor." This could solve the switching VSS issue. Optionally, changing the value at $9039 from $28 (one second) to $50 should effectively double the window before speed sensor type is changed.

Without the ability to burn or replace the chip, I would likely test VSS signal for consistency while driving by scoping the VSS signal line within the PCM. As noted, you should see no pulses for over one second before the signal type status changes.

[b65cuda]

Thanks for the info 1project2many! Where did you find this? What you said seems to check with what I could see in one of the bins I found, but I'm not sure how I make the changes just yet. I did order a programmer and a couple of chips, but I don't think it'll be arriving until next week. The problem is intermittent, but so far every time I have the scope hooked up it has been selecting the drac. VSS and drac output have so far always matched in frequency. The flag to set the source happens right after you start moving, so I need to catch it "glitching" in that short selection time.

What is interesting is that when it selects the VSS, why does the computer calculate the vehicle speed to be 1/2 and think the transmission is slipping? First I thought it was a bad setting in the aftermarket AXXZ chip, but after removing it and running the stock memcal with AWBP it does the same. I think I would be perfectly happy with programming it to just select the drac, but the curiosity in me still needs to figure out the 1/2 speed issue. One thing that is difficult is the speed sensor data is in the A150.adx, and the speed sensor selected and drac is current sensor are in the A138.adx, so I cannot monitor or log both that the same time. At least I don't know how if there is a way.

One other thing I noticed is there is another flag for loop exceeded 6.25mS that goes on and off constantly and I can't find any information on what that is.

[1project2many]

You're very welcome. There are disassemblies on the page linked above. here. As Dave W noted, the filenames that start with B are actually $D8.

DRAC produces a 2000 Pulse Per Mile signal. Possibly trucks using a magnetic sensor produce a 4000 ppm signal?

Hmm... when a sensor doesn't produce a signal at low speed but seems to work at high speed, the issue could be that the sensor is too far from the toothed wheel. Maybe you can jack up one rear wheel then use a scope or DVOM to watch speed signal while you turn tire by hand? VSS is usually able to produce a signal at very low speed.

[b65cuda]

My eeprom reader/burner came in. I read both the stock AWBP(in the memcal) and the aftermarket AXXZ on the adapter and got some interesting results. The AXXZ matched up pretty well with all of the other $85/$D8 bins with some table tweaks.... but that AWBP is 100% different. Mask ID says $8f instead of $85 and everywhere that should be 00 in the blank spaces is 0A instead. Anyone seen anything like that before? The suburban seems to run with it.

[dave w]

My eeprom reader/burner came in. I read both the stock AWBP(in the memcal) and the aftermarket AXXZ on the adapter and got some interesting results. The AXXZ matched up pretty well with all of the other $85/$D8 bins with some table tweaks.... but that AWBP is 100% different. Mask ID says $8f instead of $85 and everywhere that should be 00 in the blank spaces is 0A instead. Anyone seen anything like that before? The suburban seems to run with it.

The AWBP.bin read I have.

85.JPG

[1project2many]

Anyone seen anything like that before? The suburban seems to run with it.

Maybe, yes. Check the size / number on the chip. At times GM used larger chips and burned the file into the highest addresses. Try changing the start address for reading to $8000. The file you uploaded is not a valid calibration.

[b65cuda]

I did end up reading the chip again as a 64kb and getting the valid results in the other half.

When I attempt to change $900A from $80 to $10 the SES light doesn't do the quick flash on key on, TunerproRT wont communicate, and it save a 51 PROM error. I have confirmed I am burning correctly by burning two chips, one stock, one with just $900A modified. Is there something I am not doing correct? I am just opening the bin in a hex editor, changing the value, and saving.

[b65cuda]

I did finally manage to burn a bin that worked. Just editing the hex did not work. I had to create the flags in the tunerpro xdf, make the change, then save. This updated the checksum and then it worked.

[1project2many]

Interesting. Did you use $900A as the address in Tunerpro? You may not be changing the MASK ID. I believe the correct address to set the "Experimental" Mask ID would be $8000A?