andrew96

Patching of the 128k rom is much the same on the later machines, you just have to do it twice for chr and checksum before its happy, and you have to add in 2 identical lamp tables and change the #802 / #803 references to the chr chip 4 times. I remember you saying some you did you found 8 occurrences to have to patch, so this is not that different to yours, It just took me most of a day to do it

he is going to find out tomorrow what version of cloud 999 he had

why? I still don't understand.....that's a bold statement when the figures used for the table bear no relation to any of the figures in a actual real chip, I have proved that basically the keys are different one table against three in a real chip. I can understand its great for mfme as that's designed around 1 table, a real machine is designed around 3 whatever machine is tested we are still comparing one lamp table against 3 so whatever the age of machine is not a factor

Now patched with the MFME lamp table and works fine in MFME with all chr data cleared And its using John Parkers code for the lamp table in the rom so am getting his roms to modify hopefully tomorrow night, so should now be a easy task to do the 2 roms

just wait till I have done cash attack first, I am all for finding out! the niggling thing in my mind is barcrest used 3 lamp tables and didn't always read them out in sequence, and this happens through out the game very very regularly, why would they do that? it only needs to be checked occasionally like the character check does. this is my way of thinking

all pointers point to the answer being YES at this time once cash attack is tested we will know more

well I know enough 6809 code to get round checksums and character checking and the little tricks Barcrest do to fool ya!

lol give us a chance!! I am working on the answer!! so far.... its taken a bit longer than I thought it would but I have hacked the character checksum and the characteiser check out of the rom.... so now when I get his version of the rom I can hack that knowing it CAN be done! so the plan now is to hack his for 2 copies... one just with the chr and checksum hack, this will then still use the 3 lamp tables in the origional chr chip..... then another copy with a SINGLE lamp table in it, so the only difference between the roms WILL BE THE LAMP TABLES... this should then be able to prove it once and forall

why does barcrest use 3 lamp tables? all 7 figures are firstly read out in sequence for lots of passes, then some odd reading of just the last 2 figures over and over in all 3 tables......why the bizarre read patterns?.... so it stands to reason it is not just for lamps, barcrest must be doing something else with the codes. I nether have the 6809 code experience or the inclination to decode and go through the rom code to see exactly what it does, but I really do think something more must be going on as why go to the bother of having to get custom PAL chips made to contain the tables it needs when you could do it all with a much cheaper 'off the shelf' PAL chip with just 1 table inside...

we are just going round in circles, I can only tell you what I have been told and given all the info I know! the fact is the real machine has 3 lamp tables and one character look up table in, the hacked version of the rom bypasses the character check (which is fine and not the problem), BUT only has 1 lamp table in..... so are they the same... nope...

yeah he knows all this, he has been in 'the game' a long time. he is very much aware what versions of games and empty tricks etc

thankyou for the ayout all I can tell you is what I know so far which I have done. No meticulous writing down of figures was done, he knew it was wrong by the way it played, again he knows how they play like the back of his hand. all I know is what I have said with screenshots. Real machines are different to emulation for sure when it comes to the chr chip issue that I have proved by my reading out a real chip, If mfme was perfect then why does a real chips tables NOT appear in mfme anywhere? again its a emulator not real hardware

removing the character check has been known for years, its just a table it runs through and if it gets the correct figure it goes onto the next until complete, it if gets a wrong answer from the chr chip it drops out with a characteriser error, once this check has been removed then the checksum check has to be disabled, also that's been known for years..... these are not the problem, it is the lamp tables within the character that cannot be reproduced in a rom hack at the moment. its not the case of doing it different ways, it is just one lamp table is used in the hack when the real character chip has 3 tables in it

Perhaps I should continue this over on my mpu4 character hacking thread perhaps...

Thanks John the plan is..... he is going to dump the rom, its a version he knows like the back of his hand!! I am then going to remove the character check so it runs... (is there a link on here to cash attack mpu4 mfme?) I will then either patch in the lamp table into the rom, OR make a lamp table CHR chip with just the one mfme table in (not sure at this stage which option... then he can test it!! If I cant remove the chr check, then I can just divert the #800 references toa new lamp table within the rom... I just need his copy of the rom and the mfme version to put the plan into action!!

well this is why he thought cash attack would be a good machine to test it on, I just need the lamp table from mfme for it if someone can tell me it then we can prove it.

all I can say is 1) there is a percentage issue on a real machine when using patched roms. 2) a real machine uses 3 completely different lamp tables in the CHR to what mfme uses.

well its in a real arcade so thats not possible, he proved it some ago

so the mfme figures are somehow "fudged" to work by some manuplitation, I guess its like saying you can use one road or another to get to the same destination

perhaps this might help to see the problem, top is the patched figures for hyper viper club in MFME below are the 3 lamp table values held in the CHR chip, as you can see not one figure matches the ones used in MFME!! but the mfme and CHR produce the right lamp sequencing on a real machine..... but using the MFME table on a real machine causes problems... as you can see the real 3 tables are correct for the machine, these I read out using my fluke 9010 so I know they are correct, but remember the last 2 bits of each hex number are discarded in all cases which is not been applied to these raw figures

forget mfme, it uses ONE LAMP TABLE, a real machine uses 3 COMPLETELY DIFFERENT LAMP TABLES, MFME wont prove or disprove anything!! your trying to compare oranges with pineapples!

running patched or non patched roms in MFME wont do any good at all as MFME only uses ONE lamp table, so results will be the same patched or unpatched, mfme will not prove or disprove a thing!!! real machines have THREE completely different lamp tables in the chr chip, if a rom is patched with the one from MFME yes it works but screws over the machine, It appears to me there is more going on with the lamp tables than just lamps...

a real machine uses 3 different lamp tables, mfme uses 1 and the 1 table mfme uses WILL work in a real machine but causes percentage problems in a real machine..... that's my findings so far backed up by a arcade operator that's used chr hacked roms. what more can I do!!! I patched out the chr years ago, its not difficult to do by hand

no,the patched rom will overpay in a real machine as it won't have the 3 correct lamp tables in

Its almost like Barcrest thought about having one lamp table which hackers could do in the future, so they opted for 3 and put things in place so it plays wrong if 1 table is used