What's up with the 84 ecus?

Stinky · Post by **Stinky** » Wed Sep 06, 2006 5:46 am

A friend of mine is working on chipping an 84. So far he has pulled a little information off of each 24 pin chip. The plan is to follow the 87 ecu instructions and modify the ecu to take the 28 pin chip. My question is what bin file would we need to use. It doesnt seem like the stuff he's got off the two chips is enough. Comments, suggestions, tips?

Matt · Post by **Matt** » Mon Sep 11, 2006 2:37 pm

I had to make two adaptors up to read the information from the 1K and 8K chips. right now they are makeshift, but I might get a proper adaptor PCB made up to do this

Using the adaptors I read them as a 27C128 image and then use 0000-0400 of the 1K chip and 0000-2000 of the 8K chip

these then are combined together into a 16K chip with the 8K moved to 0000-2000 at the start and 1K are moved to 3B00-3FFF in the image

however from what i can see the 84 ecus do not have the address decoding in place to access a single 16K chip using the /OE and /CE lines

the NIStune boards use internal address decoding rather than the ECU decoding which is why my boards work in 84 ecus using the 16K image decribed above.

These are the most difficult ecus to modify as the jumper of changes involved. The PCB tracks and throughh oles also damage very easily due to the technology of the actual PCB material back in those days

Stinky · Post by **Stinky** » Tue Sep 12, 2006 12:03 am

Ah ok. thanks. So basically he's going to have to buy nistune in order to get it working?

Matt · Post by **Matt** » Tue Sep 12, 2006 6:53 pm

Yeah thats correct, unless someone can work out how to get the address decoding working from two chips, to one chip only on these ECUs. But I doubt you can

I dont see any jumpers for doing it on the board, only the ones for putting a 16K chip in place of the 8K chip. seems like you still have to have the 1K chip there which has most of the maps on it with the changed jumper settings

This is prob why JWT has never bothered with the 84 ecus. There doesn't appear to be a quick fix around this. The NIStune revc+ boards had to work with 6802 and 6303 ecus which have all sorts of EPROM layouts, which is why I connect directly to the CPU lines and decode my own addresses.

bemis · Post by **bemis** » Thu Nov 29, 2007 2:45 pm

Can't I install a SST27SF512 EEPROM in the 1k PROM spot using a 28-24 pin adapter and at the very least adjust the timing and fuel maps that reside on that chip? I could probably do the same for the 8k PROM, if I found the pinout on it. It would be a bit unorthadox and cumbersome, using two EEPROMS on one ECCS, but it would technically make it programmable.

Matt · Post by **Matt** » Thu Nov 29, 2007 4:26 pm

Old thread. Things have progressed too

Techincally, yes you can

I have made 1K and 8K adaptor PCBs, which work two ways, depending which side up you install the pins/socket

They take a 27C256 chip and translate the lower 1K / 8K

Or you can read a 1K/8K chip into the Willem programmer and grab the ROM image

So with an emulator on the second socket you can realtime tune those ECUs that way and also rechip using an adaptor. However the scalings (and VQ map?) are located on the 8K chip

I'm going to get some more adaptor sockets made up along with my Type 5 prototype panel next week if interested.

bemis · Post by **bemis** » Mon Dec 03, 2007 6:37 am

I would be interested in adapters that allow me to utilize SST27SF512 EEPROMs in the 1k/8k chip slots because it would save me alot of time from making my own adapters. I'm also interested in where you get the Vpp from, but I assume that would be included of the installation instructions for the adapters.

Matt · Post by **Matt** » Mon Dec 03, 2007 8:40 am

Those adaptors were made primarily for me for testing and reading chips on the bench (ie dont normally sell them commercially so only have a single spare 1K on hand and no installation instructions).

However I can get some more 8K made up soon and help you get them up and running

They take about a month to get PCBs back and built up when I get my next PCB prototype batch made. Type 5 boards are getting prototyped soon so I can get some more included with that

There is no VPP line mapping out to those adaptors. With the EPROM chips I used a regular EPROM programmer on 27C256 and then plug them into the adaptor socket

To get NIStune realtime tuning running an the emulator in to try it out, but I would have to customise the address file (rom offset, length etc) to get it running

bemis · Post by **bemis** » Sat Feb 16, 2008 5:46 am

I'm having some trouble reading some of the 1k PROM chips. Apparently, the ECCS manufacturer decided to use quite a few different chips to fill the 1K PROM slot. Though some are from different manufactures, they all have the same pinouts and similar ratings. I have a HN25089LGJ that reads perfectly with my adapter. With the same adapter, I have a HN25089GJ and a MB7132E that don't read, no matter what I try, ie. read timing, Vcc voltages, grounding. I haven't burnt the chips, thus resulting in failed reads because I can put the chips back into the computer and have the engine run just fine. Any suggestions? Thanks.

Matt · Post by **Matt** » Mon Feb 18, 2008 12:22 am

those two chips have different pinouts.

Fujitsu
-------
MB7132

1024*8 (latched)
+------+
A7 |1 24| Vcc
A6 |2 23| A8
A5 |3 22| A9
A4 |4 21| CE1/
A3 |5 20| CE2/
A2 |6 19| CE3
A1 |7 18| STROBE
A0 |8 17| O8
O1 |9 16| O7
O2 |10 15| O6
O3 |11 14| O5
GND|12 13| O4
+------+

a little different from data sheet with the naming but same functionality

you must have /CE1 + /CE2 low and CE3 + STROBE high to read this. I think i read this straight out with the willem as 2732

the Z31 PROM chips are different again and that is where i used the adaptor (sorry no pinout on hand for that one at the moment)

bemis · Post by **bemis** » Wed Feb 20, 2008 6:17 am

Ok, the Fujitsu MB7132 chip just read out of the blue. But the HN25089GJ chip still doesn't want to give it up. I'm thinking that I may not be supplying enough current through Vcc. I've tried combining all of the Vcc inputs together on the pins and the chip acts like it wants to work, but it still isn't enough to keep it alive. I have to do this with GM 2532 PROMs to read them. I guess I will have to use Stinky's Willem programmer since I believe he as the upgraded Dual Power model with fixes this current problem.

Matt · Post by **Matt** » Wed Feb 20, 2008 2:09 pm

current would not be the issue. ROMs use stuff all current... its the voltage level that has to be correct when programming it

bemis · Post by **bemis** » Sat Feb 23, 2008 1:02 am

Instead of using the USB power source for the Willem Programmer, I used a spare PC power supply to source the 5v. I can now read the chip, since I have a power supply that supplies a constant 5v to the chip when I read it.

bemis · Post by **bemis** » Wed Sep 17, 2008 5:35 am

Sorry to revive an old post, but this is still related to the early ECCSs.

I've finally built an adapter for the 84 ECCS that will allow a SST27SF512 EEPROM to be installed in the slot designed for the 8K ROM. I've already made another adapter to do the same thing for the 1K ROM with great success. I based the pinout of the 8K ROM using the specifications available from a release of Nistune's Hardware Installation Guide. I'm limited to only this reference because I cannot find datasheet on this chip. With everything completed, checked, and double-checked, I do not get any read from the SST. The computer acts as if there is nothing there. When I return the original chip into the slot, everything works. This leads me to believe that the address and data lines do not match. Below, I have a diagram of the adapter that I made.

Just an observation, when I was probbing the 8K ROM chip will it was ON, I noticed that the /OE line read ~3V. This would indicate that it is in fact not an /OE line.

I've written the original software from the 8K ROM onto the latter part of the SST, because that is how I have it configured to read.