I know that ECU for 2002 A4 is very different from the 2003-2005 A4. I dont know if there are similar differences for cluster but it might be problematic trying to use a later cluster on a 2002, is your new cluster also from a 2002?

Maybe try opening the decrypted cluster bins in tunerpro, compare both the old and new cluster and try to figure out where the old cluster might have become corrupted and fix. Some of the locations will be obviously different for data such as the VIN and immobilizer ID but if you can find areas where the old cluster should match the new but is different it might be possible to edit and fix, then re-calc the checksum. It would be especially helpful if you can find or develop a tunerpro xdf map or winOLS file for the RB4 cluster but thus far I have not found such a map of the RB4 data locations readily available. Studying cluster maps and matching .bin for the other available clusters with defined maps might help to understand overall structure even if locations are different.   

Another tool you might look into using is KW1281test;
https://forums.tdiclub.com/index.php?threads/kw1281test-a-free-vds-pro-vagtacho-alternative.509151/

This program is more specifically developed for EDC15 (TDI diesel) ECUs but the cluster functions do work with RB4, I was able to use it to pair new keys to cluster & ECU immobilizer on a 2003 B6. KW1281test does have a command for adapting cluster to ecu, might be worth trying in place of VCDS to see if you get better result. Be sure that ECU immobilizer is turned back on while trying any cluster or key adaptation!


 

Might be helpful to review some of the cluster location info available at;
https://github.com/gmenounos/vwcluster/tree/main
The EEPROM.mb section listing the 10 items that need to be copied over to clone the data from one cluster to another without going through adaptation.

Problem is that all the address shown are for other VW clusters, not specific to audi RB4 cluster and thus far Ive not located any similar complete listing of address for the RB4. Some of them like the VIN and IMMO ID are easy enough to identify in hex editor as plain ASCI text but Ive not found them all and some of the addresses appear to be somewhat scrambled even in the decrypted EEPROM .bin.

Also, anyone know of a software tool that is able to perform checksum correction for edited RB4 clusters?   

post bins

ok so i've got all of the adaptations and coding done on the replacement cluster through VCDS and i've been modifying the encrypted data for the VIN and the Immo data and it hasn't thrown an error with the most recent modifications. i tried to modify the decrypted files and upload them but it bricks the cluster every time i try then i have to load in a good encrypted backup to get the cluster alive again. there has to be a way to figure this out! even if it just requires setting a few things to FF so it can relearn that data if it works that way. i know the lexus immo stuff i've done in the past works that way just reset the data to FF and do a relearn. i feel like i'm very close to getting this working correctly i just need the last couple items set right.

I have included multiple BIN files below showing the various stages ive gone through with this.