@HAZArD
There are 3 possible reasons. Biostar mixed the components and added a wrong MISC.SIG in the BIOS file that provided this 3.3.1540.11 (RAID + MISC.SIG). But I checked with other 3.3.1540.11 modules from 1002-4393 and they are not the same as these ones. The second reason could be that these are only for AMI, since they were found in AMI BIOS and they were compressed. But I decompressed the modules before inserting them and added the decompressed MISC.SIG as it is, without the extra padding, which would not have messed the checksum in the first place.
The third reason could be that I haven’t cleared the area before adding this (smaller than original) UI750.BIN and some of the extra bytes got counted in this file. Please test the files attached, which have no extra bits of information in the area of misc.bin/MISC.SIG. If you still have the same error, then I will add the previous working ROM - 3.2.1540.15.
Edited: files removed.
I’m feel really bad my friend I’ve tested the two roms and even this time they are broken but this time if can be usefull the message error is only checksum is error for misc.bin 
may u update the previous raid rom as u said?
I have attached more test files. Flash them in order, test RAID and post your findings. The first two are still using 3.3.x, but with a few more bytes, to check some things. If it still fails, then either the MISC.SIG is corrupted or only works for AMI BIOS. The last three are with 3.2.x, but are using different approach to handle the bigger sized MISC.SIG. I will write more after your tests, because I have found some new things that maybe are worth mentioning, but could also be meaningless.
Edited: uploaded final working mod with 3.2.1540.11
ga-ma785gm-us2h - modfin.rar (578 KB)
Ok here’s: flashed all the five bioses and: mod4a and mod4b doesn’t work always the same message error checksum is error bla bla… I’ve tried mod5a, mod5b and mod5c and each one of them work I mean raid software and installation of OS. Right now I asked me wich is this different approach? anyway thanks very very much friend for your effort I appreciate it a lot
@HAZArD
I have found the checksum control byte and uploaded 2 more test files, the last ones. If this still fails, then it needs the misc.bin to be compressed and this only works for AMI, not Award. Test mod6b only if mod6a fails.
I will write tomorrow all the things related to this modding. Please test them as soon as possible, so I can remove all the test files and keep only the working ones, in case someone comes here with the same board. I would keep mod5a or mod5c for 3.2.1540.15 and mod6a for 3.3.1540.11.
Edited: files removed
Hi lordkag I’ve also tried these roms and no way U’re right module 3.3.1540.11 is for AMI bios 
It is too bad that this hasn’t worked. It appears series 3.3.x was only released for AMI. What I have found working on your BIOS, that could be used with this compelling guide.
The values from the Hole structure table are the base that provide the real value when multiplied with 1000. Example: 0C = C000, 04 = 4000, 10 = 10000, 0E = E000
The BIOS and cbrom don’t care about the size of the hole, as long as it fits the module. CBROM will not relocate any bytes if the new module/hole is bigger than original. For instance, in your case you are restricted to F000 bytes for hole2, otherwise you will override the information about hole used space, which starts at 1F000. So the hole2 can only fill 10000-1F000.
Maybe this can be worked around by altering the starting address of the holes, but I wouldn’t test this without backup measures. The hole seems to have a checksum byte that is ignored in most cases. So, if you want to do a proper inserting in hole area, do this: release, make sure it fits the hole area - otherwise patch with the new size, double check that the new module will not override other information, fill the hole area with FF (not 00), add the new module with cbrom, correct the checksum as bellow, re-add the new module with cbrom to have all the checksums in place.
Starting with 3.3.x RAID modules, AMD has also added a checksum control for MISC.SIG. For MISC.SIG the 0A offset is used to get the checksum to 00, for main RAID module is the 08 offset. This hasn’t been tested yet. The 3.2.1540.15 module seems to have a bad checksum, but I haven’t seen anyone complain about that.