911 3.2 Chip question

[johnman001]

I have a question regarding chips in a 911 3.2.

Are the chips interchangeable from year to year?

Specifically, can a 1986 24-pin Eprom chip be used in a 1987 24-pin DME unit? Or will the mapping be wonky?

[Elombard]

At some point the chips changed from 24 to 28??? but you can have the 24 pin holder changed in the DME and run the later chips which are supposed to be better in some way (I cant rememeber why).

Not really answering your question. I have read that the 89 chips had mapping that is considered slightly better than earlier. Of course you can go to a custom mapped chip to eliminate some of the compromise that a manufacturer has to build in (mileage, emissions, fuel grades, lessons learned over 30 years of running these cars) and get slightly better performance.

Maybe one of the chip guys will chime in. I think there would not be much difference.

[johnman001]

Thanks for the reply. I do understand that in later 1987, the DME went from 24-pin (4k) to 28-pin (8k) Eprom chips.

I know that the mapping is different on the 1984-1986 24-pin (4k) chip as opposed with the 1987 24-pin (4k) chips.


I'm questioning if one can replace a 1987 24-pin (4k) chip with a 1984-1986 24-pin (4k) chip.


Thanks,

John

[BER]

I have an '84 3.2 Carrera. I sent my Motronic box to Steve Wong a couple of years ago to be updated.

Steve removed the original 24-pin connector in my '84 Motronic unit and replaced it with a 28-pin connector. He then installed a new 28-pin (8k) performance chip with the '88-89 fuel mapping. I got Steve's chip for a Carrera with a cat bypass and that uses 93 octane fuel.

The new chip made a significant improvement in the drivability of the car. For example, the car used to surge occasionally when driving in the 2-3k RPM range. I have not had that problem since installing the new eprom chip. The car cars pull stronger over the rpm range from 1500 rpm up to the rev limit.

The car has been dyno'ed and it had the following results (averages from 6 pulls on the Dynojet:

205.87 HP at rear wheels
187.90 ft lbs Torque at rear wheels

Peak values were reached at approximately 6200 rpm.

My car is a stock 3.2 Carrera with cat bypass, OEM 1-in, 1-out muffler and Steve's 28-pin performance eprom chip.

There is a lot of debate as to the percentage of power lost in the drive-train, so use whatever % factor you desire. Whatever factor you use, you can see that there is a performance improvement which I mostly attribute to Steve's performance chip.

Hope this helps. BER

[johnman001]

Thanks for the reply.

I know that the DME's 1984-1986 and the earlier 1987 3.2's can be upgraded to 28-pin E-Proms.

I know that the later software (8k) 1988-1989 has some improvements over the earlier (4k) software. The later (8k) chips also included the DME program. (A jumper {B700} was added to the circuit board to let the CPU ignore the program internal to the 8051 processor.)


The specific question is: Can EARLIER software (4k) 1984-1986 be used on an LATER unmodified (4k) 1987 DME?

[Quicksilver]

Originally Posted by johnman001

. . .

The specific question is: Can EARLIER software (4k) 1984-1986 be used on an LATER unmodified (4k) 1987 DME?

Yes. The chips are interchangeable as long as they have the same pin count.

The question is why would you swap in an earlier chip?
Porsche improved the chip as time went on so why make a known step backwards?

In fact the later 28 pin chips made a noticeable improvement as they added the ability to change the actual program as opposed to just changing the look-up tables that the 24 pin chips offered. Updating the chip to a 28 pin socket and a corresponding chip gives an improvement in idle, smoothness, and driveability.

[BER]

Wayne,

I was thinking the exact same thing. Why would you want to put a stock '84 chip in a '87 Carrera? I'm sure there is a reason that will hopefully be shared with us.

Only thing I can think of is he has a '84 Motronic ECU and wants to put the matching chip into the ECU.

BER

[johnman001]

Are you guys certain that an earlier chip mapping (pre 1987) is compatible with the later (1987) chip mapping? I would prefer to hear back from someone who has done these chips before chime in.

The software used to program these chips seem to make a definite distinction between the 1984-1986 (4k) chips and the 1987 (4k) chips. It seems to stem from the map locations in the software.

I am not asking if you can run a 1984 (2k) chip in a 1987 unit. However, it appears that the map locations are common between the 1984-1986 chips and for whatever reason, the map locations seems to have changed in (4k) 1987 chips before they made the switch to the 28-pin (8k) chip in mid-1987.

I am getting no responses from the one whom I purchased a chip from. (I have been waiting for a couple of weeks now without any response.) I'm trying to determine if the custom chip is simply mapped incorrectly.

[BER]

Hopefully you will hear from the chip manufacturer. Sounds like they are the logical people to address your question.

I obviously do not have the technical expertise to answer your question. I do know the end result of a chip upgrade.

Good luck. BER

[ischmitz]

There are two things to consider:

1. ) A DME with a 24 pin EPROM relies on a section of "base code" (aka: software) stored on the MCU and the EPROM then adds the fuel maps, ignition maps (table) and some more software to "personalize" the box to become a Euro or U.S version. The base code is the same for all DME regardless of vintage for the 3.2

Once the change was made to the 28 pin EPROM Bosch placed the entire software and all maps onto the EPROM and disabled the base code in the MCU.

Any DME can be converted to accept and use the 28-pin EPROM by changing the socket and placing a jumper. I can program a 28-pin EPROM with the earlier code, maps and early base code so you could down-grade a late DME to behave exactly like an early version.


2.) When it comes to the question is one mapping (fuel, ignition) compatible with another engine configuration you need to consider fuel quality and compression and then finer details such as exhaust configuration, O2 sensor, idle target RPM, etc. This is completely independent of the number of pins on the EPROM.

A chip tuner grabs a copy of the base code with its stock maps and then changes fueling and timing constants in these tables. Further, some change certain function calls e.g. to disable O2 sensing or detecting a region coding plug. But only in the rarest instances is the software itself changed. To do so you would need to disassemble and translate the machine code into a readable language, make changes and them compile it again. Not something for the faint-hearted....

Cheers,
Ingo

[johnman001]

Thanks Ingo, that helps to eliminate that as a potential issue.

I was getting concerned due to the information available on the internet that perhaps there is something odd about 1987 DME units that came with 24-pin E-proms.

Somehow my e-mails to the chip programmer either fell through cracks or were in some other way lost. In any event, they are having me check some more items out to try to determine where the problem may be.

The perceived problem is that the car is running lean during the Dyno testing. The things we checked so far include:

Throttle opening fully - OK
Full Throttle switch functioning 3/4 throttle + - OK
Fuel Quality Switch set at Full CCW - OK
Fuel Pressure - OK (40 psi at the fuel rail, vacuum hose disconnected)
Tachometer from a Motronic 911 - OK (unit from 1987 3.2 911)
AFM ok? - OK (replaced about 3 years ago, never been opened)
Brown "California Wires" disconnected at DME - Need to check this, but believe them to be taped into harness
Fuel Pump - OK (using standard Bosch 911 3.2 Fuel Pump)

I don't own a Wideband O2 meter, but am considering the purchase of one to verify the results from the DYNO test.

This was our first and only Dyno test ever and I wasn't present at it. The car is not in my shop, and I have to have someone else check some of these items for me remotely.

Just to re-iterate, the chip programmer did respond and we did start communicating again. I am open for suggestions as to what items I should be checking into for a potentially lean running 3.2.

The car has no CAT, is in a mid-engine configuration, with headers and 964 cams, and I strongly believe that we don't have any exhaust leaks.

Our first Dyno tests were a little disappointing as the data provided was less than expected. Two of the Four pulls did not get to 6,000 RPM, One pull had no A/F data, and I had to give the raw file to the manufacturer of the Dynamometer in order to get the raw data or decent graphs.

The results were: 207 HP at the rear wheels (5950 RPM), 187 lb-ft Torque (5650 RPM).

The bigger concern was the A/F ratio is always quite a bit higher than the 13:1 the programmer was shooting for.

[johnman001]

Forgot the compilation of the 4 dyno pulls made.

[ischmitz]

John,

the only hardware difference between some of the early DME boxes is that some early U.S. boxes came without the electronics for the O2 sensor while all R.O.W. boxes had the electronics. Remember that early U.S. did not have the CAT and didn't require the O2 sensor. Later all DME boxes had the electronics and the chip would determine whether the O2 sensor signal gets processed or not.

In your case on the dyno you really only test/employ the full-throttle maps and thus the O2 sensor is a mute point. I would start with the basics:

- chip: I am assuming you are running a stock chip and not something "tuned" with changed fueling. If not get a stock chip and repeat

- WOT switch: do you get the WOT (wide-open-throttle) switch to engage telling the DME to switch to the WOT maps. If not this could cause lean-running. Also, if for any reason the DME doesn't recognize this input (needs bench test) the same would happen.

- fuel pressure: (make sure the regulator properly reacts to the MAP and raises fuel pressure when the throttle is wide-open). I believe one simple test is to pull the vacuum hose when idling. This should cause the fuel pressure to rise by about a bar and you'll hear the engine note changing due to going rich. If that is not happening you can deduce the pressure regulator has issues.

- fuel starvation: maybe a partially blocked fuel filter or a pump that can not keep up pressure under load and has the fuel pressure sag when fuel flow goes up.

- DME config: check the rotary switch inside the DME is all the way counter-clockwise for stock operation.

- DME: The DME could have issues (not likely but possible). If e.g. the IAT or the CHT signal isn't properly recognized or said WOT input it could lean out the mixture too much. This would have to be checked with a bench test (comparing fuel signal length at a set RPM/input signals to another DME)

Best,
Ingo

[johnman001]

Ingo,

Thanks! I will try your suggestions to see what the outcome is.

1. These were with the tuned chip. I may need to attend the next dyno testing session so that I could swap chips on-site to re-run tests back to back on the same equipment. Unfortunately, I was not at the first session.

2. The WOT switch was tested and found to be functional. The ground makes it's way back to the DME connector at 75%+ throttle pedal depression, so we know the wiring back to the DME is correct. Also, the DME does react when the test connector is jumpered in order to set idle, so I am assuming that the DME does recognize the WOT signal.

3. Fuel pressure. It tested at 40 psi with the vacuum line removed and it was lower with the line attached, but I did not record what it was at idle with vacuum hooked up.

4. I am suggesting changing the fuel filter and need to perform a flow rate test. This will have to wait until I have access to the car (currently in my dad's shop in another state).

5. Yes, the rotary switch was set full counter clockwise. Would setting this 2 positions CW enrichen the fuel mixture without changing timing? (just a question)

6. The dynomometer test was done on a very cold day in the Chicago area back in November (11-21-2014). The oil was somewhat warmed up prior to making the runs, but I don't know how warm it really got.


I will have to do some checking/verification with fuel pressure and flow next time I have the car available to me, but I can get someone to check the brown "California" to make sure Pin 10 on the DME plug is not grounded.

Thanks again. I have some items to check and hopefully will get to the bottom of what is going on. I still am wondering if the AF ratios recorded were accurate.


What do you guys think about the LM-2 Wideband Meters?

[Quicksilver]

Originally Posted by johnman001

Are you guys certain that an earlier chip mapping (pre 1987) is compatible with the later (1987) chip mapping? I would prefer to hear back from someone who has done these chips before chime in.

The software used to program these chips seem to make a definite distinction between the 1984-1986 (4k) chips and the 1987 (4k) chips. It seems to stem from the map locations in the software.
. . .

In the 24 pin chips the actual program is hardcoded in the processor. When you change the chip you are only changing the data in the look up tables on the chip.
A "look up table" is a pretty simple arrangement. The DME sees specific conditions and it doesn't calculate what to do about it. It just looks it up: I'm at this RPM, at this quantity of air flow, at this temperature... Look at the table... It says use this much gas with this much advance.

In the 28 pin chip the actual program has been moved to the chip. One of the jumpers that you change to do the 28 upgrade specifically tells the processor to ignore its internal program and to run the program on the chip. The ability to change the actual program that is running is huge. That is why so much improvement was possible with the 28 pin chip.

The reason that there is a question about 87 DMEs is you can't be sure if it is a 24 pin or 28 pin. (Mine was 24 pin until I had it repaired/reflowed by Systems Consulting. I had it changed to a 28 pin at that time.) For a description you can look here: http://www.911chips.com/dmeconvr.htm Note that "2k" is a 24 pin and "4K" is a 28 pin.

[GaryR]

I suggest anyone with an original Motronics ECU send it to Ingo for a comprehensive check, re-solder of bad/cracked connections and have him change the socket/chip to 28 pin. It will not only update you but eliminate the chance of those next to impossible to figure out issues that pop up. Can't recommend him highly enough.

[johnman001]

Originally Posted by Quicksilver

In the 24 pin chips the actual program is hardcoded in the processor. When you change the chip you are only changing the data in the look up tables on the chip.
A "look up table" is a pretty simple arrangement. The DME sees specific conditions and it doesn't calculate what to do about it. It just looks it up: I'm at this RPM, at this quantity of air flow, at this temperature... Look at the table... It says use this much gas with this much advance.

In the 28 pin chip the actual program has been moved to the chip. One of the jumpers that you change to do the 28 upgrade specifically tells the processor to ignore its internal program and to run the program on the chip. The ability to change the actual program that is running is huge. That is why so much improvement was possible with the 28 pin chip.

The reason that there is a question about 87 DMEs is you can't be sure if it is a 24 pin or 28 pin. (Mine was 24 pin until I had it repaired/reflowed by Systems Consulting. I had it changed to a 28 pin at that time.) For a description you can look here: 911Chips.com - Motronic DME Update Note that "2k" is a 24 pin and "4K" is a 28 pin.

I thought this was true until I looked at some of the DME editing software that shows that there are three (at least) distinct sets of map locations all on the 1987 24-pin EPROM chip. Can you help explain how that works?

I had thought that the 24-pin Motronic DME's were the same and had the same map locations and lookup tables until I saw references that there were multiple maps on the 1987 24-pin chips. (I suppose there could be multiple maps located on 1984-1986 24-pin chips, but I had not seen references to support that except for an odd resistor array {911.612.421.00} that could be inserted between the Altitude Correction Sensor and the harness that applied 1.8k Ohm resistance to Pin 28 of the DME normally, and 525 Ohm at above 1000m. Cars with CAT converters did not have these resistors and thus Pin 28 would see an open circuit unless above 1000m, then it would be shorted to ground. I am unclear if this is referring to the later 28-pin chips only, or if it applied to 24-pin chips as well, but it shows up in both 1986 and 1987 DME wiring diagrams.) The original chip was 1267355302 24-pin, (4k) {8732 = 4x8} chip. I understood that only the very early 1984 units had 24-pin (2k) chips {8716 = 2x8}. I also understood that mid-late 1987 they went to the 28-pin (8k) chip {8764 = 8x8} and located the program code to the EPROM and placed a jumper at b700 to disable the program on the 8051 CPU. The chip programmer went over some of this with me already.


The internet, while a great resource for information, can at the same time, be a great source of confusion and misinformation.


I was taking Ingo's word for it, and to some extent still do and I will remove the chip mapping from my list of root causes of very lean A/F ratios.



I plan to do the mechanical verifications of fuel pressure, fuel flow, brown "California" wire (pin 10), etc. before moving too much further forward.

I am also thinking that the chip provider will help me get to the bottom of what is going on here.

[ischmitz]

See my previous statements about code location - there is no change. The code on the MCU is hard-coded (hard mask, defined when the chip is built by Intel) and thus all MCU are the same. The customer (Bosch) specifies what code goes on and then Intel makes these chips. Think of it as the operating system of sorts. It is 4k in size (4096 bytes) and has no personalization so the MCU can be used very universal on many different engines.

The first EPROM actually only had a very tiny amount of space (2716, 2kB, 2048 bytes, 24 pin) and thus there wasn't room for multiple maps. As the EPROM devices grew with Moore's law space became cheaper and that allowed to make the code more complex and have the region coding based on multiple maps. The most common EPROM is the 24 pin 27C32 with 4k size. It makes for a total of 8k code size. Eventually the entire code moved to the EPROM (8192 bytes, 8kB, 27C64, 28pin)

At the same time the Bosch Motronic software evolved over the years where region coding is done by jumpers. Initially the region coding would be a simple "richen up" or "retard timing" kind of deal. In later versions of the code the region coding references different map sets located in the code based on what the region input line "sees". You need to refer to code with a Bosch part numbers. There are at least 10 different versions I am aware off over the years for the 3.2 and many more for other DME of similar vintage.

Whether they have one or multiple maps can be checked when disassembling and looking at the code. A 24-pin DME with a 27C32 4k EPROM holds exactly the same amount of code as a late DME with a 27C64 28 pin 8k EPROM. And with the provision that the lower 4k doesn't have any personalization both can have multiple maps.

Tuners more often than not bend the jump table to the one and only map set they tune and thereby effectively disable the region jumper.

Hope that help clearing it up,
Ingo

[ischmitz]

One more thing to answer John's initial question. If a chip wasn't compatible to the base code the engine wouldn't start at all. The DME would literally be brain-dead to the point that not event the ICV would vibrate.

A vibrating ICV is a sign that the DME processes code and is alive.

[johnman001]

Thanks guys! (especially ischmitz, and our chip supplier)



I believe that Ingo finally understood and answered my original question.

I believe that we are proceeding in the right direction. (and MAN, I have learned a huge amount about these Motronic units used in 3.2's in these last weeks after being around them for the last 25 years. Oh crap, I'm old..... and it seems like only yesterday)

I apologize if this thread had some vague areas, but I did want unbiased opinions. (and this seemed to go successfully)