W27c512 12v programmer/14v eraser advice?

[AndersNielsen]

When teaching people to make < 10$ sbc’s it seems a bit off to have them buy a 50-70$ TL866II to program 1$ ROMs instead of 10$ ROM’s.. so I’d like to make a cheap programmer specifically for the w27c512s since they are so cheap and readily available.

For now I have 1 x 8bit port in mind for data, two ‘595s for address lines, a 5V to 14V boost regulator and a LM339 to control the 5V to 14V for erase voltage on A9 and OE/VPE

One hurdle though - what’s an easy way to drop the 14V on OE/VPP down to 12V when programming after erasing? Can I use an LM339 output to pull down?

I also need 5V on A9 and OE but I’m thinking a 6522 probably won’t fry if it has 1k5 resistors on the OE pin and the same for the ‘595 pin(?)

[AndersNielsen]

Thinking it over the ‘339 might not be ideal since VPE/VPP current can go as high as 30mA so I need an output buffer that can source that.. 339 still might be needed. Or maybe discrete mosfets are better.

[GARTHWILSON]

This is the first time I've seen a 27Cxxx part that was electrically erasable, and it looks like it has been out at least since 1999.  Now I need to edit a recent post of mine.  I'll have to study the datasheet.

[Michael]

Loading up a pair of 8-bit latches in cascade could be faster than loading up a pair of '595s. Also, please note the cheap MT3608 'boost' regulator modules will pass their input voltage when /enable is off which is perfect for supplying 6.5v or 5.0v on VCC and 12.5v or 5.0v on VPP for the 27C64, 27C128, and 27C256 12 volt EPROMs. Unfortunately, cutting a trace on the MT3608 modules to expose the /enable input can test your mettle, and, I'm not sure what you'd need for the various voltages required for the 27C512.

Have you looked at the relatively inexpensive single-voltage SST 39SF series flash ROM's?

[gfoot]

If the EEPROM has a "page write" mode then you might need to be careful to ensure that the top address bits stay constant during the operation. Atmel's, at least, seem to flush the page if the top bits change between writes, even if they're back to the old values by the time the next write begins. So two latches with independent clock pins may work better than daisychaining them.

If you only need to program the addresses in sequence you could also consider just using 74HC590 counters to generate the addresses.

You might also find https://github.com/gfoot/nanoprom interesting, although it is quite tailored to the Atmel EEPROMs which latch the address during write operations.

[Michael]

gfoot wrote:

You might also find https://github.com/gfoot/nanoprom interesting, although it is quite tailored to the Atmel EEPROMs which latch the address during write operations.

That's a clever and elegant method. Bravo! I wonder if it could be extended to include the 32-pin SST 39SF series flash ROMs?

I have to say I really like the SST flash ROM's but a design could support either AT28C256 (32k) or 39SF010A (128k) devices by including a 32-pin socket and simple set of jumpers (see below).

[AndersNielsen]

Michael wrote:

Loading up a pair of 8-bit latches in cascade could be faster than loading up a pair of '595s. Also, please note the cheap MT3608 'boost' regulator modules will pass their input voltage when /enable is off which is perfect for supplying 6.5v or 5.0v on VCC and 12.5v or 5.0v on VPP for the 27C64, 27C128, and 27C256 12 volt EPROMs. Unfortunately, cutting a trace on the MT3608 modules to expose the /enable input can test your mettle, and, I'm not sure what you'd need for the various voltages required for the 27C512.

Have you looked at the relatively inexpensive single-voltage SST 39SF series flash ROM's?

Indeed the 39SF010's are useful but they are about 3 x more expensive - and sorely fallen victim to the chip shortage. Farnell has 0 in stock. LCSC has 2 of them.. But for over 8$ each!

gfoot wrote:

If you only need to program the addresses in sequence you could also consider just using 74HC590 counters to generate the addresses.

I thought about it but the '595s are cheaper and slightly more available.

I've attached a schematic to make things clearer. I think only VEP/VPP (~OE) needs to source (30mA max) current so I've added ... what I think is. ... a current buffer BJT. A9 probably just needs the voltage. Will it pass on the voltage and add a current source?

Will it work and will it successfully switch between 12V and 14V depending on VPE_TO_VPP?

[Michael]

AndersNielsen wrote:

Indeed the 39SF010's are useful but they are about 3 x more expensive - and sorely fallen victim to the chip shortage. Farnell has 0 in stock. LCSC has 2 of them.. But for over 8$ each!

Mouser has over 14000 39SF010A's in DIP-32 package at $3.06 each. What's the source for the $1 W27C512's?

[Proxy]

AndersNielsen wrote:

Indeed the 39SF010's are useful but they are about 3 x more expensive - and sorely fallen victim to the chip shortage. Farnell has 0 in stock. LCSC has 2 of them.. But for over 8$ each!

mouser has plenty of stock, especially for the PLCC versions, all of them are around 2-3 USD per chip (with the DIP variant being around 1 USD more expensive than the PLCC one) so i would stay away from whatever site shows you 8 USD
example, 512kB PLCC-32
only one not in stock right now is the DIP version of the 512kB Chip

also where do you even get the W27C512 from? i was unable to find on Mouser or Digikey, and is the price difference that huge that it would be worth using old stock over new chips and dealing with the 12/14V when programming?
i honestly doubt it...

[gfoot]

Hmm I'm not sure I follow what's going on with the comparator. I'd be more inclined to take the 14V divided down to 12V, pass it through an op amp to buffer it (if you can get one that supports the output current you need), and use a P-channel MOSFET or PNP transistor to optionally bypass the top resistor of the divider - so that if the base/gate of this transistor is grounded then it will bypass the top resistor and the output of the divider - and hence op amp - will be 14V instead of 12V.

I don't think your "Q?" transistor is correct, you have the base grounded. An emitter-follower current amplifier usually has the input going into the base, output coming from the emitter, and collector connected to the high supply voltage. The emitter output is then about a diode drop lower than the base input. I'd prefer an op amp as it won't have that drop over it, but you'd still need one that pulls all the way to the high voltage input (or boost to an even higher voltage to power the op amp). I'm sure others here have better suggestions too, my analogue experience is very limited.

[AndersNielsen]

gfoot wrote:

Hmm I'm not sure I follow what's going on with the comparator.

It's just level shifting from 5V -> 14V. @GARTHWILSON has an explainer here: http://wilsonminesco.com/6502primer/potpourri.html

gfoot wrote:

I don't think your "Q?" transistor is correct, you have the base grounded.

That's usually how a common base amp works - still not sure it works as I expect though https://www.electronics-tutorials.ws/am ... ifier.html

Michael wrote:

What's the source for the $1 W27C512's?

AliExpress - they're all "refurbished". Actually they usually come with the previous owner's data. They also take about 30 seconds to program opposed to the 5 seconds for the 39sf010.

[Michael]

AndersNielsen wrote:

Michael wrote:

What's the source for the $1 W27C512's?

AliExpress - they're all "refurbished". Actually they usually come with the previous owner's data. They also take about 30 seconds to program opposed to the 5 seconds for the 39sf010.

You can get 39SF010A 'pulls' or 'refurbs' on AliExpress, too. I purchased ten from this vendor a few weeks ago (nice markings & all tested good). So if I were to 'recycle' parts, like you, it seems I could produce an SBC + Programmer for about $10 in parts (including a $3 Arduino Nano clone and a PCB).

[AndersNielsen]

Michael wrote:

AndersNielsen wrote:

Michael wrote:

What's the source for the $1 W27C512's?

AliExpress - they're all "refurbished". Actually they usually come with the previous owner's data. They also take about 30 seconds to program opposed to the 5 seconds for the 39sf010.

You can get 39SF010A 'pulls' or 'refurbs' on AliExpress, too. I purchased ten from this vendor a few weeks ago (nice markings & all tested good). So if I were to 'recycle' parts, like you, it seems I could produce an SBC + Programmer for about $10 in parts (including a $3 Arduino Nano clone and a PCB).

Amazing news! Gotta grab them while we can!
Then this’ll just be a fun project - even if it’s not saving a bunch of money

[Michael]

I'm anxious to see how your W27C512 programmer turns out...

[AndersNielsen]

Michael wrote:

I'm anxious to see how your W27C512 programmer turns out...

Me too!

I guess I can also just make a super cheap 14V “eraser” on a breadboard first and see if I can generate the “high” voltages.