strange 24lc512 eeprom problem

[cluster]

Hi all,

i've very strange problem while writing to external eeprom 24lc512. its connected to 18f4550 with 4.7k pullups. eeprom data is sent through USB from PC

i am writing eeprom in chunk of 28bytes, writing to first 28 location is ok but when i write to 29th location the 28th location byte gets corrupt. when i try to correct it by writing a single byte at 28th location then the 27th location byte gets corrupt.
i have no idea why eeprom is not working as expected. below is the code:

              for aa = 3 to 62 step 2
                temp_add = USBRXBuffer[aa]
                gosub hexconvert    
                edata[0] = temp_add * 16
                
                temp_add = USBRXBuffer[aa+1]
                gosub hexconvert
                edata[1] = temp_add
                
                edata[2] = edata[1] + edata[0]
                prg_block[cc] = edata[2]
                cc = cc + 1                             
            next aa                 
            I2cwrite SDA,SCL,CONTROL,address[4],[str prg_block\cc]
            pause 10

eeprom dump, see at last 28th byte should be FF but its D9:

FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
FF
D9
FF
FF
FF

when i specifically try to correct 28th byte then the 27th byte gets corrupted
can you please help me solving this issue

Thank You

[kenif]

If you're on 3V3 try a smaller value pullup, which wouldn't hurt on 5V.
Check the leading edge with a scope.
If you have a lot of capacitance on those tracks pullup should be smaller, as low as 1K5.

It sounds like a timing issue.
Have you tried writing in small chunks with a pause in between?

[Demon]

Try writing fixed data to the eeprom first; writing a loop value in binary. That way you will be sure the proper data gets written to the proper address. Using FF isn't giving you a fair test of your logic, something like:

0000 0001 pos. 1
0000 0010 2
0000 0011 3
0000 0100 4
0000 0101 5
0000 0110 6
0000 0111 7
0000 1000 8
0000 1001 9
0000 1010 10
0000 1011 11
0000 1100 12
0000 1101 13
0000 1110 14
0000 1111 15
0001 0000 16
0001 0001 17
0001 0010 18
0001 0011 19
0001 0100 20
0001 0101 21
0001 0110 22
0001 0111 23
0001 1000 24
0001 1001 25
0001 1010 26
0001 1011 27
0001 1100 28
0000 0001 pos. 29
0000 0010 30
0000 0011 31
0000 0100 32
0000 0101 33
0000 0110 34
0000 0111 35
0000 1000 36
0000 1001 37
0000 1010 38
0000 1011 39
0000 1100 40
0000 1101 41
0000 1110 42
0000 1111 43
0001 0000 44
0001 0001 45
0001 0010 46
0001 0011 47
0001 0100 48
0001 0101 49
0001 0110 50
0001 0111 51
0001 1000 52
0001 1001 53
0001 1010 54
0001 1011 55
0001 1100 56
etc

Robert

[sayzer]

Writing to LC eeproms requires 2ms to 5ms delay after each write.

You are sending a block of bytes at one shot.
Try sending(writing) them single by single and have 5ms pause after each write.
Or, you may want to use page write feature for one whole 128-byte page (for LC512) and have one pause after each page.
check this link :
http://www.picbasic.co.uk/forum/show...8811&highlight=

[cluster]

hi
thanks all for your reply
problem solved... i did two things:
1. started to write 128 locations in one shot instead of 28 as suggested by sayzer
2. i was using usb port for power supply so i changed it as suggested by kenif
and then i tested as suggested by demon... issue vanished..
once again thank you all.