memcal question:Multiple maps on one chip

[jamieaus]

my biggest problem was that the Led display was lagging I assumed that was dodgy connections.

Will try the switch you suggested and see if it makes any difference.
Thanks

Jamie.

 

[hakhawk]

yeh, mine doesnt lag, does well, except i have now killed 2 leds. 1st one i killed B, replaced it(i bought 3), and second one i killed F and A is dimmer than rest. i really need to make a decent board. but, it works, and it works well. although i tried with 2 vn maps and 1 vr map, but for me, the vr map comes up as an error on engine light, even when i have aldl running, its like it switches to limp home, but i need to do a memlca with no limp home chips to see if it starts on the vr one.

greenfoam said his never shows an error on vr. i dont know what anyone elses does though.

i tried one with 4 different vn maps and all worked fine. you have to stack them in tunerpro upside down though. or else it shows 0123, but switches 3210.

 

[jamieaus]

I read somewhere that the codes need to be all of the same base, I assume that means you can't swap VN, VP VR with this method.

 

[VL400]

Excellent work with the code switchers jamieaus and hakhawk, good to se they are working. Bring on the DIY i say!

Not sure how you both have the eproms wired or what size they are, but if you are using 27c512s be carefull with the A15 pin. Inside the ECM it is hardwired to +5v so remove that pin from the IC socket before you try switching it with your circuit.

hakhawk - If you can get ALDL data with a VR bin then the code is being executed, try extracting the errors. Limp home mode is most easily recognised by the engine fan turning on while the ignition is on and engine off. Make sure the Anti-theft is disabled, the speed PPK is adjusted and the checksum is correct if you are using a VR binary.

Now that i think about it using a switching device on the address pins may also be your problem. The VR code uses A14 during normal code execution, it is high during a reset to get to the vector table but to get to the start of the code it needs to be low so make sure you switching circuit does not interfere with it. I dont know how your circuit is wired so you may have thought of this already.

Using smaller caps on the 555 with it configured as a one shot just means the debounce time is lower. With 2nF caps it is only 4mS of debounce, it is just just enough.

jamieaus - If you want to switch maps on the fly make sure you have all the stacked binarys with the same broadcast code (the 4 letter code each binary has). If you stack different BCCs then have the engine off while switching, this way you could stack 1 VR and 2 VN binarys on one 27c512 at the same time, just do not switch with engine running - the ECU will most likely freak and reset either from an illegal operation or a watchdog timeout from the main processor or supervisory chip.

 

[VL400]

i've read that you can run 12 volts to pin c9 on the ecu and it will switch pin 27(a14) of the eprom with 5 volts. apparently they used it on lpg cars to switch between 2 different maps but i haven't been able to get it to work. can anyone confirm this
Ross

From what i understand grounding pin C9 puts the ECM into LPG mode which disables the injectors while the engine is running (above 300RPM, so it would appear the ECM controls cranking fuel then switches to LPG). I havnt gone right through the LPG code but it certainly does not switch a14 and run different maps/code. There are a few other little bits of code that are run/not run while in LPG mode but i havnt got the time to work through the BBMM dissassmbly as well. There is an extra option flag to say it is an lpg vehicle too, this needs to be set for pin C9 to do anything. BBMM is setup for LPG so try that one.

 

[hakhawk]

VL400, thanks for the info, A14 could be the problem, as other people greenfoam seems to have his vr map work fine, mine doesnt. i dont get any data with the vr map, its like limp home mode comes on, as fan turns on with engine off and i get solid engine light. i need to do a memcal with no limp home chips and see if it starts. might have to have a play around, need to be able to get a14 low while vr map is selected, or it might try and pull low but switch maps in the process. will it have to be hooked to the memcal pin? i have it and a15 seperated from memcal itself.

sice greenfoam and others that have the vr map working, i think thats my problem, cause they arent using switching as far as im aware.

oh, and just out of curiosity, when you say broadcast code, do you mean the promid? if so, i was using 4 vn maps today, all with different promid's on the run and there was only one map that didnt play, but it doesnt like my car by itself anyway.

 

[jamieaus]

Vl400 thanks for clearing that up for me could save me some trouble when I get around to testing.
Here is a protel drawn Schematic of the schematic above with hachawk's 555 timer switch incorporated I have not tested this yet however I see no reason why it won't work the oly thing I havn't put on here is the wires going to a14 and a15 I am sure you will work that out.

 

[michaelw]

Where are you current limiting resistors for the led, also have you tied all the cmos inputs either high or low?

Mike

 

[hakhawk]

yeh, you'll need those resistors, ive killed 2 displays already, even with resistors, i think it was during breadboard testing tho i didnt use them then

 

[VL400]

hakhawk - If you have A14 disconnected from the ECMs address bus a VR binary will not work. Disregarding A15 for the moment (leaving it high), with A14 low you can address only the $8000 to $BFFF range and with it high from $C000 to $FFFF. A VR binary spans $8000 to $FFFF so the ECM needs to control this line, you cannot just leave it high or low either. If the fan is on with engine off it is def in limp mode, the code prevents the fan from working until the engine is running, without the limp home resistors it will not start.

Broadcast code is like ASBY (VN), BLCF(VR) etc etc Using different BCCs, even with the same program ID (say $5D), could confuse the ECM. GM made a few changes to fix a couple of bugs add some new features. Some are very minor like adding an extra flag which means only a few lines of code but that is more than enough to cause the ECM to reset after it returns to the wrong spot or executes the wrong routine or even something bad like dialing in shitloads of spark......

jamieaus - From the quick look at the schematic it looks good except you need some resistors for the display, the counter lines need to go to A and B inputs of the 4511 and BI, LT need to be tied high and LE low.

Edit: beat me too it about the resistors!