Related
Recommended

Who remembers EPROM?

dougw

This is just an off the wall way to say thank you to element14 and Xilinx. They had a little give-away draw to win a Zynq book and they ended up sending me a huge book called Exploring Zynq MPSoC and a screw driver set. I also recently completed a Project14 so I threw in some images from that project.

This is a digital microscope image of a 27512 EPROM. It took a lot of digging in my archives to find this chip, but I wanted a chip with a ready-made window so the microscope could image the actual silicon.

Click on the image for a full size view.

Here is what a typical EPROM looks like with its window to allow UV light to erase the memory:

I also took a a closer image just to see what this microscope can do....

I'm having a lot of fun with this microscope.

 

Here is a video of the FPGA Technology Quiz prizes and the Pi HQ Camera microscope in action....

 

I once used EPROM without a CPU as a look up table to convert 18 bit gray code to binary.

Have you ever tried to use sunlight to erase an EPROM?

Feel free to add your EPROM stories below, I'm sure there are lots of them.

 

Relevant Links:

FPGA Technology Quiz

I Spy With My Digital Eye .... a Pico PI

Anonymous
  • in reply to neilk

    Those one-time-programmable devices were a great incentive to write bug-free code, it seems there is no longer any stigma attached to writing buggy software.

    Floppy disks are a whole other trip down memory lane. I still have some 8" floppies, as well as 51/4" and 3.5" floppies. There is an outside chance I could rig something to read 51/4" floppies and a slightly better chance of reading 3.5" floppies, but no chance of reading 8" floppies.

  • I have been searching my disorganised archive, on and off since you first posted and finally located some chips. hence my late comment.

     

    In the early 80s through to 1990, I was a external consultant to a small company that manufactured echo sounders, logs and wind speed indicators for (mostly) pleasure boats. We developed the first echo sounder using the MCS- 8048 microcontroller which was one time programmable. A pain during development - if it had a bug, you threw it away!

     

    We moved on to 8748 chips which were UV eraseable. I had all the kit at my house and I used to program the production chips in batches of 100, cover the window with an identification label, and post them to the factory (200 miles away).

     

    Eventually, we developed a system whereby I burnt the code into a 2764 EPROM, with an identification label covering the window, and posted that to the factory. They then copied the code into the target micro controller - 8031, I think,  by then.

     

    I have samples of most of the chips we used, including a set for the last instrument I worked on, as follows:

     

    ver 1.0 12 July 1990

    ver 1.1 13 August 1990

    ver 1.2 12 October 1990

    ver 1.3 17 November 1990

    ver 1.4 26 November 1990

     

    I assume version 1.4 went into production, but I haven't located any documentation to support that. I still have two 5 1/4" floppy disk which purport to carry the code, but no way of reading them. Somewhere, I also have a listing of the code.....somewhere!

     

    Neil

  • in reply to dougw

    This is an advertisement of the time reproduced in the handbook. I like the cyclists (a play on the idea of the devices being recyclable).

     

  • in reply to dougw

    This is an explanation of the gate from the 1975 Intel Memory Design Handbook. This applies to the early parts like the 1702 and the 2704.

     

     

    I was wrong to talk of a 'well' as it's really a floating gate formed above the substate within a grown insulating oxide layer, not sunk into it. Charge gets onto the gate by avalanching, so it's different to a Flash memory where the mechanism is tunneling. The diagram shows that the space above the gate is kept clear of metal to allow UV light to get to the charge on the floating gate.

  • in reply to dougw

    My guess would be that the metallization on the surface is too dense with a Flash part (presumably the UV parts were deliberately designed with a clear area above much of the storage well for the light to get in), but I could be wrong.

     

    BTW thanks for the pictures. The one that shows clearly the welds of the bond wires is particularly nice. You can see, from the variation in the placement of the tack point, why they needed the landing pads the size they were.