Erasing PIC16C54C

[rasmus]

Hi

With my limited experience/info with microcontrollers an EPROM is erasable by exposing it to UV light, the chip has a window on top of it so the UV light can reach the chip inside the package. My question is: i have this PIC16C54C which as per the data sheet says it is an EPROM but, it don't have any window or opening to allow the UV light to reach the chip, then how is this erasable if its an EPROM ?


Thanks in advance

 

[kellys_eye]

Most EPROMs are programmed and erased using the program voltage pin (/MCLR/Vpp) these days. Windowed EPROMs are still available but hardly ever used.

Erasing would simply be writing all 1's or 0's to the relevant memory locations.

 

[rasmus]

if they are erased electrically then why are they called EPROM, shouldn't they be called EEPROM ?

 

[kellys_eye]

'E' meaning electrically (in EPROM) and 'EE' meaning electrically erasable - in the first instance you can erase individual memory locations. In the second you usually have to erase the whole device. 'EE' PROMs are 'old' technology nowadays introduced, I believe, as a simpler way to erase ROM other than the UV route which was, let's face it, a real PITA procedure. The further advancement to permit selective memory location erasure was the next logical step I suppose?

Q. Anyone else have any historical knowledge of E/EEPROM development?

It's all semantics really - the datasheet reveals precisely what you can and can't do with the on-chip memory.

 

[Harald Kapp]

This PIC seemingly comes in 2 flavors: UV-eraseable and OTP (Datasheet chapters 2.1 and 2.2).
The UV eraseable variant obviously has a window for erasing the EPROM during development.
The OTP variant has an EPROM but no windows. You can't erase the EPROM once it has been programmed. This variant is for production purposes.

 

[dorke]

'E' meaning electrically (in EPROM) and 'EE' meaning electrically erasable - in the first instance you can erase individual memory locations. In the second you usually have to erase the whole device. 'EE' PROMs are 'old' technology nowadays introduced, I believe, as a simpler way to erase ROM other than the UV route which was, let's face it, a real PITA procedure. The further advancement to permit selective memory location erasure was the next logical step I suppose?

Q. Anyone else have any historical knowledge of E/EEPROM development?

It's all semantics really - the datasheet reveals precisely what you can and can't do with the on-chip memory.

The history of non-volatile memory actually goes back to magnetic media.
Like drums,tapes etc.

The first EPROM was invented by Dov Frohman ,an Israeli Engineer working for Intel in 1971(later on was vice president and head of Intel Israel,the little place they invent and design Intel's top of the top CPUs.)

The 1702 was the name of the first device from intel,
it stored 256 Bytes allowing it the store the full EBCDIC code
It Not only had a "huge" storage size ,it was also "lighting fast" at 1.5uSEC access time
That was the humble binging of the multi-billions of billions Erasable memory IC market.

 

[kellys_eye]

it stored 256 Bytes

Wow, the same as my first ever computer!

Anyone recall the ETI 'Triton' single-board PC project? A kit of parts to build it cost me £300 in 1978(?) and it came with a 2k BASIC in ROM and, yes, a whole 256 BYTES or RAM! Expandable of course

 

[hevans1944]

This PIC seemingly comes in 2 flavors

Actually, 3 flavors, if you have a really high volume application and mature (bullet-proof) code. There is a masked ROM version available, but it is obviously not for the faint of heart. Still, if your market volume is measured in the millions of units, there could be considerable savings, not to mention inherent security, in purchasing the ROM version.

I believe the OTP (one-time programmable) version uses the same die as the EPROM version, but verity with Microchip before making a commitment. If so, that means you don't need a UV window to erase it. Any old dental x-ray machine should erase it. It would be interesting if any dentists on this forum could verify that their office x-ray machines can erase the windowless packages. You would need to program them first, before erasing, since they are shipped from the factory in erased condition, ready to be programmed.

 

[Harald Kapp]

Any old dental x-ray machine should erase it.

None in my backyard,,,

 

[Arouse1973]

OTP is slightly misleading. You can programme them more than once if you have enough space left, you just can't erase them.
Adam

 

[dorke]

Actually, 3 flavors, if you have a really high volume application and mature (bullet-proof) code. There is a masked ROM version available, but it is obviously not for the faint of heart. Still, if your market volume is measured in the millions of units, there could be considerable savings, not to mention inherent security, in purchasing the ROM version.

I believe the OTP (one-time programmable) version uses the same die as the EPROM version, but verity with Microchip before making a commitment. If so, that means you don't need a UV window to erase it. Any old dental x-ray machine should erase it. It would be interesting if any dentists on this forum could verify that their office x-ray machines can erase the windowless packages. You would need to program them first, before erasing, since they are shipped from the factory in erased condition, ready to be programmed.

Dental X-ray machine?
Senior,that can mess-up your Kahonas,not recommended...

There is also the gentle way ,
exposure to direct sun-light,1 week should do it

 

[Harald Kapp]

exposure to direct sun-light,1 week should do it

How's that going to work for an OTP without glass window?

OTP is slightly misleading. You can programme them more than once if you have enough space left, you just can't erase them.

You can progam every cell only once. And only from '1' to '0'.
If you want to put stuff into previously unused memory locations and want to make use of that stuff, you need a way to tell the controller to access those new locations instead of the old ones. One way could be by havin a kind of bootloader routine (which must be fixed and cannot be overwritten or replaced, of course) which scans the EPROM for the first valid address entry and jumps to that address to start the program. You'd provide a reserved memory array to hold these start addresses. When writing new code to the EPROM you simply set the previous start adddress to 0000 and write the new start address to the next free location within that array.
But this is not how OTPs are meant to be used, it's probably a poor man's solution.

Anyway: modern processors use FLASH and the whole problem is moot.

 

[Arouse1973]

How's that going to work for an OTP without glass window?


You can progam every cell only once. And only from '1' to '0'.
If you want to put stuff into previously unused memory locations and want to make use of that stuff, you need a way to tell the controller to access those new locations instead of the old ones. One way could be by havin a kind of bootloader routine (which must be fixed and cannot be overwritten or replaced, of course) which scans the EPROM for the first valid address entry and jumps to that address to start the program. You'd provide a reserved memory array to hold these start addresses. When writing new code to the EPROM you simply set the previous start adddress to 0000 and write the new start address to the next free location within that array.
But this is not how OTPs are meant to be used, it's probably a poor man's solution.

Anyway: modern processors use FLASH and the whole problem is moot.

Yes the way I have done it is to use a new start address.
Adam

 

[hevans1944]

I have some EPROMs with the quartz window. I have the ultraviolet eraser thingy. I had a programmer that ran under DOS 5.1 and required one slot in my PC. No idea where the software for the programmer is, nor do I want to install an ancient I/O card in one of my existing PCs. Sooo... if I ever decide I need to program an EPROM for another Intel 8085 project, looks like I will need to purchase another EPROM programmer. I will just add on that cost to the customer requesting such a project. So, yeah, the problem is moot as far as I am concerned.