7

Using Microchip EREPROM 24AA02E48. Datasheet says it is a "2-Kbit I2C Serial EEPROMs". In my eyes, 2KBits/8 = 256 Bytes.

Storage is addressed in what they call "words", addressable range is 256 bytes wide, ok so far. Storage only allows R/W access to address below 0x80, which means there is a true data space of 0x80=128 bytes = 1KBits, inside an address area of 0x100=256 bytes.

Can anybody confirm this (which means the "2K" size is a lie), or did I miss the magic wand command that allows full 256 bytes R/W access?

toolic
  • 8,262
  • 7
  • 24
  • 35
Michael Schmelz
  • 73
  • 4
  • 3
    "word" is ambiguous. A word can be anything, depending on context and individual definition, from 1 bit (though that would be indeed surprising) to tens of thousands of bits (in context of SSD-grady forward error correction words). – Marcus Müller Nov 02 '22 at 10:10
  • 2
    Before ordering any part it is wise to read the the part number spec, in this case found at page 29 of the datasheet. All parts come with some flavour of this EUI node id. – Lundin Nov 02 '22 at 10:53
  • 1
    Michael Schmelz - Hi, I notice that you commented to say "thank you" below an answer. If your question has been solved, please consider "áccepting" an answer (i.e. click the "tick mark" next to that answer, to turn the relevant tick mark green). This shows that you don't need more help and future readers can quickly see which was your choice of answer / solution. Thanks. – SamGibson Nov 03 '22 at 12:57

2 Answers2

17

The device is unambiguously described as a 128 x 8 "standard" EEPROM + Unique ID EEPROM.

You might accuse Microchip of being confusing, but I don't see anything resembling a lie. It's 1 Kbit of EEPROM for you, and 1 Kbit used for the unique ID, a thing which is extremely useful for many devices. It's a 2 Kbit EEPROM with half preprogrammed with something extremely useful, which very often goes into EEPROM. It could save all the per-unit work of programming IDs. Being write protected at the factory prevents it ever getting changed even if desoldered from your board.

If you want a plain, empty, EEPROM, consider looking at 24AA02 without suffix, which at first glance appears identical except it's all writable, and has a write protect pin.

The moral of the story is always read the datasheet, including the part about the suffixes.

From -E48 datasheet, p14:

enter image description here

Plain Version

In fact, my guess is that the -E48 and -E64 is the same chip but with the WP connected to A7 internally, after they've programmed the unique ID. Microchip to classifies it as 2 Kbit EEPROM because it is a 2 Kbit EEPROM but a) conveniently pre-programmed with 6 (or 8) bytes of the ID, b) write protected ID. And, certainly, c) 250 (or 248) bytes "wasted" in the top write-protected half.

They also make 24AA024H which is the same but the write-protect pin protects the upper half only. Which would be useful for putting your own unique IDs or serial numbers in. Obviously you win the "wasted" portion of the top 256 bytes, but also it would be possible for it to be reprogrammed under some conditions.

jonathanjo
  • 14,034
  • 3
  • 34
  • 66
  • 2
    That is definitely not "clearly" marked. I dont normally look at datasheets for memory, but that is terrible naming convention. Good catch though. – Linkyyy Nov 02 '22 at 10:39
  • 7
    @Linkyyy Well, both the datasheet title and the first listed feature prominently mention pre-programmed node ID. You won't find that on any generic EEPROM. That E48 or E64 suffix says "this part is special". – TooTea Nov 02 '22 at 10:42
  • 1
    I see. Could have fooled me aswell. You could argue that the ID data would be in an additional 1kbit of memory. – Linkyyy Nov 02 '22 at 10:48
  • 6
    @Linkyyy, it is very clearly marked - it's in the title. And it's detailed in the datasheet. Best to check those thoroughly before using terms like 'lie'. I've designed in these parts before and found them very clear. It uses 2 Kb of EEPROM, half of which is set aside for the smaller EUI data. When you buy a PC, manfs aren't 'lying' to say you've a 256 GB HDD when there's only 200 GB free after Windows etc. – TonyM Nov 02 '22 at 10:50
  • @Linkyyy I see your point, changed it to "unambiguous" and wrote about the plain part. Microchip thinks this is a 2kbit EEPROM with the ID in it and the write protect used to protect it. We might think of it as a 1kbit EEPROM + ID ROM. – jonathanjo Nov 02 '22 at 10:50
  • 7
    The "error" might be in a vendor's selection tool if you sort by capacity. I can see how it can happen, but it's a good lesson in the need to read the datasheet – Scott Seidman Nov 02 '22 at 10:52
  • 1
    @TonyM, that doesnt make sense at all. When you buy a 256GB drive, the full capacity is available for you. In this case you buy storage, and only half of it is available, because it is literally read-only. – Linkyyy Nov 02 '22 at 19:01
  • 1
    @Linkyyy, that doesnt make sense at all Maybe because you're not reading what I wrote, which starts 'When you buy a PC...', not 'when you buy a drive'. – TonyM Nov 02 '22 at 19:20
  • 1
    @ScottSeidman I've seen enough data entry errors at online retailers that, frankly, I always double check the P/N with the datasheet or at least manufacturer website. Frankly, those two, are the only sources of truth for a part. – jaskij Nov 02 '22 at 19:21
  • 1
    @jaskij The manufacturer Microchip classifies it as 2 Kbit in the part selector and datasheets. They think of it as "ready-programmed with something useful in the top half". – jonathanjo Nov 02 '22 at 19:46
  • 1
    @jonathanjo ah, fair, I'm always unsure what "vendor" means in these contexts - might be the manufacturer, might be the seller. Yeah, that could surprise someone. That said, is this even MCP's fault? I believe this is an inherited Atmel part, and they probably just didn't change the info on it, which has it's pros and cons. – jaskij Nov 02 '22 at 19:49
  • 1
    @jaskij Honestly I think I accept Microchip's description. They make a whole range of these things with various write-protect facilities. Some of them come with a suffix eg -E48 showing they are part-programmed with something extremely useful which EEPROMS very usually contain, programmed by the board manufacturer. This -E48 saves time, money, and per-unit work by the board mfr. Caveat emptor, surely? – jonathanjo Nov 02 '22 at 19:57
  • @jonathanjo oh, caveat emptor absolutely, as this are items meant for use by professionals. As to the usefulness... I'm currently lamenting the lack of stock on these ICs, as the E48 parts actually contains EUI48 addresses, and our company, being as small as it is, doesn't really have the infra to hold a MAC address database. Although I think there are some available (Q)SPI flash parts in the series. – jaskij Nov 02 '22 at 20:00
6

From the Webpage of the Part enter image description here

Let's check the datasheet enter image description here

So, the first 128 Bytes are usable. The Rest is read only and some bytes at the start of the high block are used for unique IDs...

So if you want 256 Bytes of EEPROM and don't need the unique IDs, this is not the chip you're looking for.

kruemi
  • 3,054
  • 17
  • 29
  • THANKS HEAPS everybody for the explanations. All of you feedback gave me a good portion of insight !

    The datasheet says with BIG LETTERS that I can expect 256 bytes, with a bit smaller letters that there are 2 pages of 128 Bytes available, and finally that only half of it is accessible.

    In my opinion, it is still simply misleading. They can write a header saying "2k EEPROM", but in this case , the very next line should say "1k free for programming only". Otherwise it is still a kind of cheat to me.

     – Michael Schmelz Nov 03 '22 at 12:14
  • Sure, Microchip would have done you a service in making this more clear. Finding the right part for our needs is one of the biggest challenges. Because there are often dozens of parts with only minute differences. Ordering the wrong part and finding out the hard way is part of the process. It happens to everyone. Complaining doesn't help. Learn and move on, everything else is a waste of energy. – kruemi Nov 03 '22 at 12:48