Hey Guys,

I have a memory DIP (one of the HM62256B series chips) and the datasheet says that the memory is organized in a 32-kword x 8 bit grid but the chip only has 15 address inputs. How can the datasheet claim that the chip is organized in words when the packet only has 15-bits of address access?

http://www.datasheetarchive.com/Indexer/Datasheet-028/DSA00489713.html

 

A "word" is arbitrary; on an 8-bit processor, it could be 8 bits (synonymous with byte), and on a 16-bit processor, it could be 16 bits. In this case, your memory has an 8 bit word. It only has 32,768 locations, which requires exactly 15 address lines.

http://en.wikipedia.org/wiki/Word_(computing)

 

Thank you; I understand now.

 

According to this datasheet, is the memory organized as 32Kilobits x 8 or 32Kilobytes x 8?

http://www.datasheetcatalog.org/datasheet2/7/0yuw0yc3x0278cpz78zge08qi13y.pdf

 

The device in the datasheet is 256K bits organized as 32Kbytes.

hgmjr

 

Some memory capacity is given in bits, so the word size doesn't matter.

Normally, when memory is specified in bytes, an 8 bit byte is assumed. Otherwise, 4GB of memory would be "twice as much" in a 32 bit system compared to a 64 bit system.

 

The device in the datasheet is 256K bits organized as 32Kbytes.

hgmjr

256 Kbits = 256,000 bits
32 Kbytes = 262,144 bits

How can this configuration be possible?

I believe that the confusion I am having arose from the ambiguity of whether a kilobit is equal to 1,000 bits or 1,024 bits. I have seen both definitions.

 

The use of the prefix kilo is used loosely when specifying bytes. Your bit count calculation is correct.

1 Kilobyte = 1024 Bytes

hgmjr

 

256 Kbits = 256,000 bits
32 Kbytes = 262,144 bits

How can this configuration be possible?

I believe that the confusion I am having arose from the ambiguity of whether a kilobit is equal to 1,000 bits or 1,024 bits. I have seen both definitions.

In electronics a Kbit is 1024 bits, a Mbit is 1048576 bits and so on, binary prefixes are mixed with metric and it all gets confusing.

 

Ah, I see. Thanks guys for the quick responses.

 

256 Kbits = 256,000 bits
32 Kbytes = 262,144 bits

How can this configuration be possible?

I believe that the confusion I am having arose from the ambiguity of whether a kilobit is equal to 1,000 bits or 1,024 bits. I have seen both definitions.

It's frustating and confusing. This practice carries over to hard drive and other storage devices as well. Sometimes kilo/mega/etc is a 1024 based number, sometimes kilo/mega/gig/etc is a 1000 based number, and most annoying of all is when they mix the two, such as 1 million bytes. First is the 1000 system, second is the 1024 system. Until you work with it, you can't exactly figure out the number of usable bits, especially with hidden bad sectors, etc.

 

It's frustating and confusing. This practice carries over to hard drive and other storage devices as well. Sometimes kilo/mega/etc is a 1024 based number, sometimes kilo/mega/gig/etc is a 1000 based number, and most annoying of all is when they mix the two, such as 1 million bytes. First is the 1000 system, second is the 1024 system. Until you work with it, you can't exactly figure out the number of usable bits, especially with hidden bad sectors, etc.

As an experienced person in electronics, would you conclude that the potential loss of 24 bits would affect larger-scaled systems badly? Can they be safely ignored?

 

As an experienced person in electronics, would you conclude that the potential loss of 24 bits would affect larger-scaled systems badly? Can they be safely ignored?

One potential problem: An operating system computes drive space in megabytes by using integer division. It divides say 1,000,000,000 bytes by 1024 to get the number of kilobytes = 976562 KB. In this case, it rounds down - but then another operating system decides to round up to 976563 KB. This creates a size incompatibility - for example, duplicating a backup. Now if the sizes were multiples of 1024, there'd be no problem.

It's not the loss of just 24 bits, either. On a large scale, a 1 GB drive is 1,000,000,000 bytes instead of 1,073,741,824 bytes - a whole 7.3%. A terabyte drive is usually only 931 GB. As you can see, a considerable amount of space is not actually there.

 

As an experienced person in electronics, would you conclude that the potential loss of 24 bits would affect larger-scaled systems badly? Can they be safely ignored?

It depends on what the 24 bits are.

If they are a platter on a Terabyte hard drive, it wouldn't matter at all.

If it is a defective area of RAM in a microcontroller that only has 128 bits of RAM to start with, it can be very bad.