Internal eeprom - erase block

[skimask]

I could use more memory blocks and increase the project's performance but i limited it to 2 blocks because the erasing process takes too long ( 2/3 minutes for both blocks ).
I know that i could use a faster clock to increase the instruction's speed and erasing process, but i'm working with the internal clock of a 16F88.
Is there a way i can speed a bit more via software ? Maybe a diferent code/set of instructions that work a bit better ?
Thanks

Post your code. 2-3 minutes is entirely TOO LONG! There has to be something silly going on that's causing it to take sooooo long.

[sayzer]

By "block" do you mean the "page" feature of external eeproms?

If so, page write takes about the same time as writing a single byte. Should be somewhere between 2ms to 10ms depending on the eeprom you are using.

It seems that your block writing routine is not really accuate.

-------------------

[ruijc]

Hello sayzer,

i dont mean page.

When i say block it means that i'm using an area of memory for one record session and other area for a diferent record session.

For example...

Using memory address from 0 to 3000 will be used as "block" 1 or record session #1 and memory address 3500 to 6500 will be used as "block" 2 or session #2.

The user will choose ( using a button ) where it will record ( either block 1 or 2 ) and the same concept is used when reading back data.

I'm using a 16F88 pic with internal osc and a 24LC512 external eeprom.
The program works very well except the erasing rotine ( note that i'm erasing all blocks at the same time ).

With 2 blocks it erases in an acceptable time frame...but i would like to use more blocks without loosing too much time erasing them all.

[Dave]

ruijc, But why are you required to erase all of the data? Why not just keep a couple of pointers? 1 for number of data bytes recorded and 1 for the next data location. The reason for 2 is to verify if either is corrupted. Just update the 2 pointers when writing the data or clear them when you require a memory clear function.

Dave Purola,
N8NTA

[ruijc]

Hello Dave,

Very interesting idea.

It makes perfect sense. It's a completly diferent logic this way.

Thanks

What i thought was:

I can reserve 4 bytes ( using 2 blocks of data...reserve more with more blocks ) of memory such as:

1 will record the starting address for first block
1 will record the ending point for the first block
1 will record the starting point of second block
1 will record the ending address of second block

This way i can have variable size of block with data

These can be stored either in the external or internal eeprom ( i guess it will be safer to store them in the external one )

For reading it will lookup these bytes for it to know where to start colecting data and where to stop.

To erase just simply put these to 0 ( reminding to check for 0 before start recording ).


I think it can work this way, dont you think ?



.

[Dave]

ruijc, Bygolly youv'e got it....... I have used this same method for about 3 years or so using 4 24LC1025's for data storage of depth, temperature, diveplane angle, and status at 1 second intervals for about 18 hours on my TempTracker.

Dave Purola,
N8NTA

[sayzer]

So what Dave is suggesting is like a partition table.

Instead of erasing all data, just erase or clear partition table.

If the partition table is empty, the data is not accessible or is useless.

--------------