home > tech tips >  chips and eproms

.... & Chips

Eprom Burner
From: Greg Powzun

Hi Eddy,

welcome back matey. It's been a while, but we still miss your technical input.

As for the EPROM in the 928, I think it's a 27C256, dual-in-line package. This type of EPROM is like your common garden variety and is used in heaps of electronic gear (I have a draw full of them at work for the fibre-optic terminal gear). Being a CMOS device, you'll want to make sure that the programmer is compatible not only with the device model, but the manufacturer as well as some manufacturers use a slightly different programming voltage to others even when the device model is the same. I will try and find out for sure.

Great to have you back!
Greg 9ine2eight Dankelblau

From: "Adrian 007"
Subject: RE: [landshark] Eprom Burner


A lot of work has been done on chip grades/comparisons by fellow US sharksters.

Bottom line was as follows:

- benefits of reprogramming will depend on the year of your car & spec. eg if you have a 1989 Euro spec GT without a catalyst, you are probably wasting your time - However you can get 10-15bhp more from a 1990 s4. There is a company in the US called Authority I think which sells upgraded chips for about US$500.

- easiest benefit/$ is to change the rear silencer & (if local regulations allow) replace the catalyst with a performance version/normal exhaust. This will impact other parts of the rev range though. Porsche did actually offer a 928 'Sport' stainless steel rear exhaust (without baffles!) as part of their 'Tequipment' range in the early 1990's. I picked up one a few years ago, but you will be lucky to get hold of one now (basically it looks more like a mortar launcher!)

If anyone needs any more info on the above, please let me know.

Regards, Adrian

From: Eddy Sain
Subject: Re: [landshark] Eprom Burner


I followed up some of the US guys and got information from knowledgeable US users of lots of chips including Authority. Lots of these users sent the chips back ( 30 day satisfaction guarantee in a lot of cases) the end result was as I said sometimes felt like it went harder but when checked on an accurate dyno back to back before and after not worth the money. On 90 and later cars (maybe even before) you need two chips one for fuel one for spark the US guys believe that custom chips for your individual car are worth the effort. OK you can
replace the whole computer with lots of good units on the market but you need to "bastardise" the car wiring etc. etc. and then you need to set it up for your individual car anyway.

I agree a good quick cheap way to get a few free Hp is the exhaust especially the muffler and even the air cleaner element but there is more available in the brain with out going into the motor or good old Nitrous Oxide.

Regards, Eddy

Subject: [landshark] S4+GTS = 27C256 Chip
From: "Greg Powzun"


found that site I remembered. It's at http://www.eproms.net

These French guys have 4 chips. two (identical checksum?) are for 928 S4, and two are for a 928 5.4 (GTS). The interesting part is that they detail the chip used for these models. They are: "E DIL27C256" chips. Therefore we could reasonably say that your programmer is suitable.

Darren, it is also possible that the instruction code is stored either on the processor (therefore it is strictly a microcomputer) or on a PROM chip elsewhere, in which case you would conclude that the EPROM chip then only contains operating parameters. In the case of two chips, it is plausable that one is program and one is data.

Greg 9ine2eight Dankelblau

Subject: [landshark] Re: S4+GTS = 27C256 Chip
From: Sterling Gee

Chips for 928's come in sets. One for each computer. One for ignition mapping, and the other for fuel mapping. I forget which one is which. I actaully made copies of a factory chip with a $30 chipreader and some blank chips, from the local electronics shop. I took the factory chips in, and they found suitable matches. The two chips are different specifications. The chip reader created a dos file for each chip, and I put in the new chips and wrote to them..... 
I also tested them and they worked. I did this for backup reasons and was trying to reverse engineer them for my project. 

This was 2+ years ago, and I decided the gains were minimal. On a stock car, the gains to be made by chips are minimal. The more modifications, the more gain possible by remapping the chips. The computers extrapolate operating parameters from sensors in the car, ie temp, timing, and airflow, and reference the table for the injector firing. 

What is interesting, is that the computer can change the timing on the injectors, not the pulse width. Certain parameters like the injector impedance and flow are not adjustable. This is why some people have worked around this limitation with stroker engines, and just put in larger injectors. It is an interesting topic though.

The best tuning device is the Dastek product, providing you stay with the factory computers. If you were to say, go with a MoTeC, you could realistically gain 25 hp on a stock engine.


From: Greg Powzun

Hi Eddy,

The task you have set yourself is by no means a small feat. For you to be successful, you will need access to some major information. I will briefly outline what I think you will have to do, not to discourage you, but to make you aware of the mountain you must climb. You already may know most of this so excuse me for painting your task as daunting.

The engine management system as a module is responsible for taking sensor readings, interpreting them and making adjustments to optimise the balance between power, emissions and economy. At the heart of the module is a microprocessor, microcomputer or PIC (depending on the car). The microprocessor is no magical device, it is simply an Algorithmic State Machine. It can only follow very very simple instructions (machine code or assembler code). Each instruction is made up of microinstructions that are carried out internal to the processor. For ease of understanding and use, programmers use a higher level language such as C in which each command is made up of a huge collection of machine code.

Okay now what is on the EPROM is just machine code in binary form. The code details to the processor what processor instruction it must carry out and with what parameters. The original designer would have started at the highest level with something like a flow chart (unless he used Object Orientated Programming, but I doubt it). From the flowchart, he writes the program in C. Then an assembler (a program dedicated to that processor) converts the high level C code to machine code (binary) which is then burnt onto the EPROM.

Your first step would be to determine what processor is being used. Being about the ten year vintage, I imagine a commercially available processor like a 6800 series. For newer cars, a custom processor is more likely. At Uni, I managed to design my own dedicated microprocessor using one of the new Large Scale Programmable Algorithmic Logic chips. With the processor type, you can obtain the full instruction set. They are usually about 120 instructions.

BTW, this is called reverse engineering. Then you take your EPROM, put it into your EPROM reader and extract all the binary information. There will be 256,000 ones and zeros (arranged in 32,000 words of 8 bits each) in a 27C256 chip. Each instruction is made up of 1 to 4 words (generally), so we're looking at about 10,000 instructions. There are programs called disassembles that make good attempts at converting the machine code back to the discrete instructions, but from experience, unless you have a good one, it can end up being a mess. For example, a data value may have the same sequence as an instruction, which confuses the hell out of them. Assuming you convert the binary code back to 10,000 microprocessor instructions and data, it is nearly impossible to convert the assembly code back to an understandable high level code. You will grow old trying to plot out a flow chart using just the machine code.

Lets look at it from another perspective. Having two cars with the same engine but differing performance is a great plus. Take for example the S4 and the GT. IF and I stress IF they both used the same program, then a comparison between their EPROMs would show differences only in the parameter values (buried in the machine code). These values may be timing of ignition and injector on-times etc. The parameters you want to experiment with.

Simply take one differing parameter at a time and take it to an extreme to determine what it does (maybe to the determent of the development engine). With the after market engine management computers, they come with programs that access and alter any of the parameters at their known locations. These guys know where the values are because they wrote or have access to the original high level program (called source code). If, however, you could get the original source code, then we can do the assembly ourselves, but where could you get that top-secret information.

As for the programmer you detailed, although it didn't directly detail a 27C256, I think it would take them and notice the ability to input the vendor ... very important. Should do the job, but don't pay too much for it. 

Also, you should confirm the chip is a 27C256. I seem to recall reading it somewhere on the net, an aftermarket chip place, but I can't remember just where. I will look again soon. 

Eddy, when you finish this project, maybe we can put our heads together and try and crack the Human DNA code. It's only 30,000 something's long ;-) 

Greg 9ine2eight Dankelblau

Subject: Re: [landshark] EPROMs what I think happens
From: Eddy Sain


I knew most of the stuff you mentioned in your email but boy I could not have expressed it so well.

What you are saying is correct but there are some existing programs that do a lot of the work already available I found a heap of sites in the past and some more tonight

The source codes are available off the net. With regard to the programming all that I am trying to do is to rewrite the look up table for the base program not rewriting the whole thing so by modifying the look up table we change the  parameters and hence the characteristics.

I still need to get the chip manufacturer and number but am reluctant to pull apart my ECU units until I have a spare. I can probably borrow GT ECU units so i may start from there.

Check out the following link and the other links under chips and the sample software that you can download I haven't tried anything yet as this is a long term project and I need to learn before I crawl let alone walk or run still good stuff to get your head around and if others do it why not me.




DNA code can we engineer more horsepower into it?

Regards, Eddy

From: Greg Powzun


one of your links goes to this mob in Melbourne: http://www.powerchip.com.au/chips/index.asp?make=Porsche

They say

928 GT 89-91 5.0L uses 2 plug-in pull-out chips
928 GTS 92-95 5.4L uses 2 plug-in pull-out chips
928 S 80-84 4.7L uses a soldered chip
928 S 85 4.7L uses a soldered chip
928 S 86 5.0L uses a soldered chip
928 S4 87-91 5.0L uses 2 plug-in pull-out chips

But no-ones giving away what they are.

Greg 9ine2eight Dankelblau

Subject: [landshark] Chips etc
From: "Pitman"

I have opened and chipped a hot wire late Bosch (but badged Lucas) ECU in a
Discovery, and I suspect the mods are only to data tables. The replacement chip needed to be matched to the 'family' of the surrounding chips (several bigger ones) based on codes on one of them. Most of the rest of the ECU is A->D and D->A hardware I think (analogue to digital and back again).

You cant learn a lot from looking at much of the hardware here, but you can in a flapper type air flow meter- there are only 6 or so steps in these units, based on a few I have opened, so there is no continuous smooth curve of signal to the ECU for air flow - just 6 or so discrete values.

On Bosch EFI, the throttle pot sends a continuously varying voltage, but on Mazda eg, the throttle pot only sends fully shut, or not, and maybe fully open. From some reading I have done it seems the Bosch EFI controls fuel delivery by changing pulse length of the injector opening, which means that at 6000rpm, you only have 10msec of open available to you and you can end up with continuously open injectors - time for larger capacity units, with less open time required.

Also, I have found that output on this LR system can be improved by checking the static volts at the throttle pot are in spec - move it on its screws, and that of the air flow meter - which has an adjustment screw in the top side, under a tamper proof cover.

I suspect for most non racing uses, more power can be had by normal improvements to inlet, exhaust, and advancing timing to just below pinging. 

EG, MX5s , normally timed at 10 BTDC, will run as high as 18 BTDC. Advancing timing 2 to 4 degrees was part of the chipping instructions on the Discovery kit. If it ain't pinging, it cant do much harm?


John Pitman