EEPROM "self timed write cycle" question

[Jay112]

I'm comparing the datasheets of the AT24C256 and AT24C32 EEPROMs. The C256 says it has a "Self Timed Write Cycle" of "5ms max", whereas the C32 says the self timed write cycle is "10ms max".

Does that mean that writes to the C256 will generally be twice as fast as compared to the C32? Sorry if I'm way off base. Also if the writes are faster in one, does that imply that the reads might also be faster, or would you expect the read times to be the same?

Thanks!

[ataradov]

Does that mean that writes to the C256 will generally be twice as fast as compared to the C32?

Yes, it does.

Also if the writes are faster in one, does that imply that the reads might also be faster, or would you expect the read times to be the same?

Nope. Typically read time is as fast as your interface can go. Which is not that fast for I2C.

[uncle_bob]

Hi

If you want fast reads, switch over to an SPI based part.

Bob

[Jay112]

Thanks guys!

That's a good suggestion, uncle_bob. I actually just ran into information about FRAM for my first time. I was just about to buy an FM25V02 from digikey, along with a SOIC8 to DIP8 adapter.

Do you think the FM24V02 is a good choice? I'll be using it with an Arduino Due (SAM 32), with the SPI running at 40MHz if possible.

Also I've never dealt with SOIC before. Do you think a beginner (who's comfortable with soldering everything else) will be able to handle soldering a SOIC for his first time, or is it usually very difficult the first time?

[ataradov]

Do you think the FM24V02 is a good choice?

What are your actual requirements?

Do you think a beginner (who's comfortable with soldering everything else) will be able to handle soldering a SOIC for his first time, or is it usually very difficult the first time?

The easiest way to find out is to get some cheap ICs in that package and cheap breakout boards from eBay and practice a bit. In general, it is not that hard, especially for SOIC.

[Jay112]

Thanks Alex!

I don't actually have any "requirements", since it's mostly just a hobby. It's just for a general "remote controller" system I made that can control lots of things around me via a web page. I have hundreds of variables stored in EEPROM, and most of the information stored is just the positions of different switches/toggles. I try to make the system as fast as possible so that the web page is very responsive.

Even my current EEPROM is working well enough (it's the one that has the 10ms write cycle), but I was considering swapping it out for an EEPROM I already have on hand (the one that has a 5ms write cycle). But then if I was going to do the work of swapping the EEPROMs and changing the code, I figured I might as well just make it a big upgrade so that storing variables in non-volatile isn't slowing down anything anymore, and also so that I don't have to worry about burning out the chip from too many writes.

And thanks for the encouragement about the SOIC soldering! I guess I'll just take it really slow and see how it goes. I'm guessing it'll be easier than I thought it would be.

[uncle_bob]

Thanks Alex!

I don't actually have any "requirements", since it's mostly just a hobby. It's just for a general "remote controller" system I made that can control lots of things around me via a web page. I have hundreds of variables stored in EEPROM, and most of the information stored is just the positions of different switches/toggles. I try to make the system as fast as possible so that the web page is very responsive.

Even my current EEPROM is working well enough (it's the one that has the 10ms write cycle), but I was considering swapping it out for an EEPROM I already have on hand (the one that has a 5ms write cycle). But then if I was going to do the work of swapping the EEPROMs and changing the code, I figured I might as well just make it a big upgrade so that storing variables in non-volatile isn't slowing down anything anymore, and also so that I don't have to worry about burning out the chip from too many writes.

And thanks for the encouragement about the SOIC soldering! I guess I'll just take it really slow and see how it goes. I'm guessing it'll be easier than I thought it would be.

Hi

If you are serving web pages fast, you will have to have a *lot* of RAM for the buffers associated with the TCP-IP side of things. Doing a few thousand variables and moving them into RAM for fast access will not significantly impact the total RAM budget on the device. I'd stick with normal (cheap) eeprom and spend the extra two or three dollars on a bigger MCU. If you are serving the web pages out of one of these dedicated chip gizmos, the i/o to that chip will be a big enough bottleneck that eeprom access time on an SPI eeprom will likely be a non-issue.

Bob

[Jay112]

Thanks for the help, Bob!

A big reason why I try new hardware like this is just for learning, so that I can do more advanced things in the future. So even if there's no noticeable speed improvement by switching from EEPROM to FRAM, I'll still benefit from the project because I'm learning about FRAM, more about SPI, and also I'll be able to solder my first soic.

For my hardware, I have the webserver running on an Arduino Due (SAM32 80Mhz), with a standard Arduino ethernet shield on top. I'm not knowledgeable enough to make my own TCP/IP stuff yet.

It mostly suits my needs, but I'm always keeping speed in mind because the web page is complex, since it has to check for so many different variables on each load. At first I was using one of the cheaper Arduinos that had an (I think) 16Mhz processor, but after a while the webpage became sluggish on that due to the complexity. But when I switched from that to the Due board, there was a very noticeable and pleasant speed improvement. The page currently takes about 900 ms to load, which is fine for now, but I wouldn't want it any slower than that, and I'm always adding new features.

[uncle_bob]

Thanks for the help, Bob!

A big reason why I try new hardware like this is just for learning, so that I can do more advanced things in the future. So even if there's no noticeable speed improvement by switching from EEPROM to FRAM, I'll still benefit from the project because I'm learning about FRAM, more about SPI, and also I'll be able to solder my first soic.

For my hardware, I have the webserver running on an Arduino Due (SAM32 80Mhz), with a standard Arduino ethernet shield on top. I'm not knowledgeable enough to make my own TCP/IP stuff yet.

It mostly suits my needs, but I'm always keeping speed in mind because the web page is complex, since it has to check for so many different variables on each load. At first I was using one of the cheaper Arduinos that had an (I think) 16Mhz processor, but after a while the webpage became sluggish on that due to the complexity. But when I switched from that to the Due board, there was a very noticeable and pleasant speed improvement. The page currently takes about 900 ms to load, which is fine for now, but I wouldn't want it any slower than that, and I'm always adding new features.

Hi

It's *always* worth looking at the bandwidth bottlenecks in any design. The "web server on a chip" ( = Arduino shield) gizmos have significant advantages (they have zero MCU load). They also have disadvantages, you can only get just so complex before they fall to pieces. The bottom line - this is why you move up by another $20 or so and run a setup with a real OS ( = Linux).

Bob

[Jay112]

Thanks for the help, Bob!

A big reason why I try new hardware like this is just for learning, so that I can do more advanced things in the future. So even if there's no noticeable speed improvement by switching from EEPROM to FRAM, I'll still benefit from the project because I'm learning about FRAM, more about SPI, and also I'll be able to solder my first soic.

For my hardware, I have the webserver running on an Arduino Due (SAM32 80Mhz), with a standard Arduino ethernet shield on top. I'm not knowledgeable enough to make my own TCP/IP stuff yet.

It mostly suits my needs, but I'm always keeping speed in mind because the web page is complex, since it has to check for so many different variables on each load. At first I was using one of the cheaper Arduinos that had an (I think) 16Mhz processor, but after a while the webpage became sluggish on that due to the complexity. But when I switched from that to the Due board, there was a very noticeable and pleasant speed improvement. The page currently takes about 900 ms to load, which is fine for now, but I wouldn't want it any slower than that, and I'm always adding new features.

Hi

It's *always* worth looking at the bandwidth bottlenecks in any design. The "web server on a chip" ( = Arduino shield) gizmos have significant advantages (they have zero MCU load). They also have disadvantages, you can only get just so complex before they fall to pieces. The bottom line - this is why you move up by another $20 or so and run a setup with a real OS ( = Linux).

Bob

I was actually considering Raspberry Pi as the next upgrade. I got the newest model which has I think a quad core 1.2GHz processor. I'll probably experiment with that over the winter time when there's less to do. It would be interesting to see how big the performance improvements might be.

[uncle_bob]

Thanks for the help, Bob!

A big reason why I try new hardware like this is just for learning, so that I can do more advanced things in the future. So even if there's no noticeable speed improvement by switching from EEPROM to FRAM, I'll still benefit from the project because I'm learning about FRAM, more about SPI, and also I'll be able to solder my first soic.

For my hardware, I have the webserver running on an Arduino Due (SAM32 80Mhz), with a standard Arduino ethernet shield on top. I'm not knowledgeable enough to make my own TCP/IP stuff yet.

It mostly suits my needs, but I'm always keeping speed in mind because the web page is complex, since it has to check for so many different variables on each load. At first I was using one of the cheaper Arduinos that had an (I think) 16Mhz processor, but after a while the webpage became sluggish on that due to the complexity. But when I switched from that to the Due board, there was a very noticeable and pleasant speed improvement. The page currently takes about 900 ms to load, which is fine for now, but I wouldn't want it any slower than that, and I'm always adding new features.

Hi

It's *always* worth looking at the bandwidth bottlenecks in any design. The "web server on a chip" ( = Arduino shield) gizmos have significant advantages (they have zero MCU load). They also have disadvantages, you can only get just so complex before they fall to pieces. The bottom line - this is why you move up by another $20 or so and run a setup with a real OS ( = Linux).

Bob

I was actually considering Raspberry Pi as the next upgrade. I got the newest model which has I think a quad core 1.2GHz processor. I'll probably experiment with that over the winter time when there's less to do. It would be interesting to see how big the performance improvements might be.

Hi

If you actually have a lot of dynamic web pages (lots of re-done graphics) you will find that the PI will do quite a bit better. That's not to say it will do well. A Dell server will blow it away. It all depends on your frame of reference.

Bob