6

I teach a hands-on computer architecture course in which the students burn their programs onto an EPROM with a commerical PROM burner. The instructions are read by a low-speed TTL computer they build. I started with the 27C256 EPROM but switched to the 29C256 EEPROM, which has now gotten hard to find.

Can anyone recommend an EEPROM (I'm sick of UV erasers) that would be a good replacement? Specifically, it should be reasonably priced, reliable, easily available and likely to remain so. It doesn't need to be pin-compatible, although that would be nice. I'd be happy with as few as 8 address bits (we use two LS163s for the PC), although I doubt anyone makes something that small. I also don't care about speed. Robustness to student wiring mistakes would be great. (All of the EPROMs have seemed pretty good for this.) Would the 28C family be a good place to look?

We have an Andromeda Labs PROM burner, which supports lots of devices.

Added detail: I've been using the 28-pin DIP package.

Ellen Spertus
  • 1,609
  • 4
  • 20
  • 23
  • Not an 'answer' per se, just an opinion: you might consider switching to an inexpensive FPGA trainer. I don't think you're doing your students any favors by getting them to build stuff out of 7400-series TTL in this day and age. – John Miles Dec 04 '11 at 00:18
  • 2
    @JohnMiles, could you elaborate? The goal of my course, which is the introductory (and only) architecture course at a liberal arts college, is for students to understand all the way down (to the gate level) how computers work, not to directly prepare them for industry. (I hope to inspire some to grad school.) I find the hands-on work, including the debugging, to really get this information across to all levels of students. Could you let me know why you think a FPGA approach would be better and point me to any educational material? I don't have experience with FPGAs. Thanks. – Ellen Spertus Dec 04 '11 at 07:33
  • I do like the hands-on nature of the TTL/EPROM-based curriculum. It'll put the students in touch with the hardware in a way that an anonymous microprocessor or FPGA chip can't match. However, it struck me that you must be spending a lot of time getting them to the point where they understand what they're hooking up, and why. Perhaps the limited available time could be used more productively with modern hardware. – John Miles Dec 04 '11 at 09:00
  • (Broke into two comments since the edit function doesn't seem to work) Have you checked out some of the boards available at www.digilentinc.com and similar education-oriented sites? I like the Nexys2, personally, and the books by Haskell and Hanna look pretty good for bringing newbies up to speed. – John Miles Dec 04 '11 at 09:03
  • I strongly second the suggestion to switch to FPGAs. You can write VHDL or Verilog in a very explicit gate way if you want... and you can do a universe worth of additional stuff with it. The concept of building up from gates can be maintained... without all the grunt work of trying to wire up a bunch of TTL correctly on breadboards. Is it a computer architecture class or a class on building prototypes? – darron Dec 04 '11 at 17:26
  • You could provide a VHDL/Verilog module for loading code from USB (using an FTDI serial to USB chip) or RS232 into FPGA block RAM. That'd additionally teach bootloaders, if desired... would would be much more useful today than UV burners. – darron Dec 04 '11 at 17:32
  • It's also possible for students to append user data (program code & data) on the end of an FPGA configuration datastream. You could then have the FPGA design load that data into block RAM on boot. – darron Dec 04 '11 at 17:35
  • You could also likely do what you're trying to do with CPLDs, which would need almost no external components, contain the nonvolatile data internally, and come in more friendly packages. They're possibly better direct matches, but I'd still suggest an FPGA platform for versatility. – darron Dec 04 '11 at 17:38
  • 1
    I appreciate all of the advice and will look at the referenced technology, although I find it's when debugging that the students gain an understanding of what's really going on. Do you think I should start another question on how best to teach intro architecture so we can continue this discussion? – Ellen Spertus Dec 04 '11 at 22:37
  • 1
    FWIW, we use the textbook Computer Organization and Design by Hennessy and Patterson. – Ellen Spertus Dec 04 '11 at 22:38
  • I created another question for the issues in these comments ("http://electronics.stackexchange.com/questions/23231/what-are-the-advantages-of-using-fpgas-over-ttl-in-intro-computer-architecture") and am hoping it doesn't get closed for being too discussion-y. Thanks again for all of the advice. – Ellen Spertus Dec 04 '11 at 23:07
  • Sst 27SF256 is a direct EEPROM replacement for the 27C256 –  Sep 14 '15 at 17:34
  • Yeah, let's never teach 74 series logic any more. Goodness knows everything is FPGA and no companies manufacture devices with 74 logic anymore. – cbmeeks Apr 20 '21 at 17:51

3 Answers3

5

I believe the 29Cxxx series is now mostly flash memory.

Digi-Key appears to have several thousand of the 32Kx8 AT28C256 EEPROM's, which are almost the same pinout as the 27C256, and appear to be supported by your programmer. You didn't say which package you are using, I assuming it is a 28-pin DIP. There is also a 8Kx8 version of the same chip with a compatible pinout (unused address lines are no connects). It is a little cheaper in price, and also in stock.

tcrosley
  • 47,708
  • 5
  • 97
  • 161
  • I've never used the 28C family. Can you tell me anything about it? – Ellen Spertus Dec 04 '11 at 07:26
  • @espertus, I have not used these devices myself, but from the datasheet they appear to be same as the 27C and 29C families in terms of reading out the data, and since you are using a PROM burner to program them any differences re their erasing and programming wouldn't matter. – tcrosley Dec 04 '11 at 08:09
  • 1
    FYI, the 28C256 is *ALMOST* pin compatible with the 27C256. IIRC, there are two pins that are different. A14 and /WE. If you made a custom socket or PCB and swapped those two pins appropriately, then that should work if you need to retrofit the 28C256 to a 27C256. Otherwise, the AT28C256 is still manufactured and easy to work with. The only negative is that they have become expensive in recent years. Also, there are other flash-based options out there too if you don't need the DIP28/28C256 package. – cbmeeks Apr 20 '21 at 17:54
2

One more comment in the form of an actual answer: whenever I've needed to buy 27(C)256 EPROMs, I've always found eBay to be a good source. You could also try Jameco.

John Miles
  • 512
  • 3
  • 5
1

There are still a few parallel flash chips available in DIP, e.g. SST39SF010A from Microchip. They tend to be higher capacity than eeprom devices, so more address lines to tie but otherwise equivalent to eeprom for your purposes.

mikeselectricstuff
  • 10,655
  • 33
  • 34