Help with programming 27xxx series eproms

By: Racer5 - 2009-03-28 20:56:45

I bought a v5.0c dual power programmer from ebay to program some 27128 and 27256 chips.

I can read the chips just fine, but if I try to write, it fails instantly.  From what I have read its either a voltage problem or bad chip.   I have tried maybe 20 chips, so I really doubt they are all bad

I have also tried every jumper combination I can think of.  Changed the vcc from 12.5 up to 21.  Some chips say 12.5 so that should work.

Its almost like it is write protected.

I have it powered via 10v 1200ma supply.  I figured that should be enough power to write a chip correct?

If there is any more info you need from me please ask.  I am not sure where to test the voltages or I would give you my readings on that.

I have tried it on 3 different computers using different cables and printer port settings and same result.

I am beginning to think I might have a bad programmer.

Thanks for any help you can provide.

-Jeff

Re: Help with programming 27xxx series eproms

By: DL - 2009-03-29 00:42:16

Hi Jeff,
Jeff from Mooresville wrote on 03/28/09 at 20:56:45:
I bought a v5.0c dual power programmer from ebay to program some 27128 and 27256 chips.
I think you're going to have to make some mods on your programmer.   It's a PCB3 variant and has some power problems that need to be fixed if you want to be able to program NMOS EPROMs reliably.

I think the first problem you'll need to address is VPP drop.   27128 require VPP=21V and I suspect your programmer won't be able to provide 20.5-21.5V @ 50mA long enough.   You can check this by entering the Test H/W tab, set program as indicated and set VPP=21V.   Put a 420 ohm, 1W resistor between pins 1 and 16 on the socket.   Click the VPP box on the screen and measure the voltage on pin 1 of the socket.   It should be 20.5-21.5V.   Wait 30 seconds and measure again.   If you were programming 2732, it would need to stay between 24-26V for over 2 minutes.

If the voltage is low as I expect, you can try replacing the inductor with one that has a higher saturation current.   My dual powered PCB3 drops to 18.0V at the transistor output with a 412 ohm load (using an external 5V supply connected to the USB connector).   I didn't measure the drop from the regulator and transistor separately; it's the total drop that matters.

You may still have voltage drop problems due to the transistor used to switch VPP.   The typical transistor is rated at 100mA and has a significant voltage drop.   I usually use a 2N2907 to get less drop.

You'll also have problems with VCC voltage drop.   27128 program at VCC=6.25V +/- 0.25V.   Your programmer probably has a 6.2V setting, but the transistor switch will have significant drop at the 100mA Icc specified for 27128.

To check this, put a 62 ohm 1/2W resistor between pins 32 and 16 on the ZIF socket.   Enable VCC and measure the voltage on pin 32 of the socket.   It should be 6.0-6.5V; but at that current, no transistor typically used will have less than a 0.25V drop.   A quick and dirty fix is to put a relay in parallel with the transistor.   This will limit the voltage drop to significantly less than 0.1V.
Quote:
I can read the chips just fine, but if I try to write, it fails instantly.
Unless you're comparing the data against known good data, reading with no error doesn't mean much.   The programmer will read an empty socket...
Quote:
I have also tried every jumper combination I can think of.   Changed the vcc from 12.5 up to 21.   Some chips say 12.5 so that should work.
Not a good idea to use the wrong voltages.   27128 are VPP=21V, 27128A are VPP=12.5V, and 27256 are VPP=12.5V.   Using higher voltages is the easiest way to damage devices.
Quote:
Its almost like it is write protected.
That feature could come in handy if it existed...
Quote:
I have it powered via 10v 1200ma supply.   I figured that should be enough power to write a chip correct?
The problem is most likely your programmer (design), not the power supply.
Quote:
If there is any more info you need from me please ask.   I am not sure where to test the voltages or I would give you my readings on that.
If you're uisng WinXP, make sure legacy plug/play is disabled on your printer port to prevent inadvertent writes (after the power problems are addressed).
Quote:
I have tried it on 3 different computers using different cables and printer port settings and same result.
Good thinking; it's not likely that a defective programmer could damage a printer port.   A problem with the STB signal (pin 1) can prevent programming.   I had this problem when I got my first Willem.
Quote:
I am beginning to think I might have a bad programmer.
Probably just bad VCC and VPP circuit design.

You should blank check EPROMs at VCC=4.2 (STM spec) to 4.45V to insure sufficient erase margin.   Don't operate your programmer much lower than 4.2V because it's interfacing to the parallel port operating at 5V.

You should program most EPROMs at VCC=6.25V.   You should be cautious here too because you'll be overdriving the parallel port inputs by over a diode drop so you'll probably be forward biasing the clamp diodes on the data line inputs.   This could damage the parallel port.

Only the Willem 4.1 does this mostly right.   It includes a level shifter for interfacing with the parallel port.

HTH,
Dennis

Re: Help with programming 27xxx series eproms

By: Racer5 - 2009-03-29 02:10:09

Well that sure sounds like what is going on.  I would have no way in replacing those parts.  I bought it from Kee Electronics and he said to send it back and he will check it out.

I had one before we moved and it worked just great.  I even powered it from usb and could burn any of those chips.  Did I just get lucky?  The thing is that I dont remember what version it was.  I am thinking 4.5 but not positive on that.  I think it came from Savia(sp?) in HongKong but like i said, it worked nicely.  Somewhere in the move a whole box of electronic stuff went missing  

Kee has an advanced willem programmer for a little more at this link here http://www.keeelectronics.com/catalog/product_info.php?products_id=53

Do you think this is better than what I have?  I can probably upgrade for the difference in price if you think its worth while.  The only thing I program is the 27 series chips, so if there is a better solution for about the same price that will do these chips, I am all for that.

Re: Help with programming 27xxx series eproms

By: DL - 2009-03-29 02:57:49

Hi Jeff,
Jeff from Mooresville wrote on 03/29/09 at 02:10:09:
Well that sure sounds like what is going on.   I would have no way in replacing those parts.   I bought it from Kee Electronics and he said to send it back and he will check it out.
I think all of their programmers are PCB3 variants.  If this is the case, all will have problems with 27128.  They also have an enhanced universal Willem that's not worth any price.
Quote:
I had one before we moved and it worked just great.   I even powered it from usb and could burn any of those chips.   Did I just get lucky?
It's unlikely that any EPROMs you programmed with USB power were programmed reliably.
Quote:
The thing is that I dont remember what version it was.   I am thinking 4.5 but not positive on that.   I think it came from Savia(sp?) in HongKong but like i said, it worked nicely.   Somewhere in the move a whole box of electronic stuff went missing   
Sivava is the primary innovator when it comes to Willem type programmers.  Unfortunately, they continue to proliferate the questionable VCC circuit from the original Willem and the weak step-up regulator design introduced on the PCB3 variant.
Quote:
Kee has an advanced willem programmer for a little more at this link here http://www.keeelectronics.com/catalog/product_info.php?products_id=53
I wouldn't buy any PCB3 variant unless I intended to fix the power problems.  I've fixed a dozen programmers for others...
Quote:
Do you think this is better than what I have?   I can probably upgrade for the difference in price if you think its worth while.   The only thing I program is the 27 series chips, so if there is a better solution for about the same price that will do these chips, I am all for that.
The only Willem type programmer I'd recommend for EPROMs is the Willem 4.1.  Willem has discontinued that version, but I think someone is still selling a copy on eBay.  The relay is mandatory for NMOS EPROMs and a simple modification will give you all of the VCC options you need to blank verify and program reliably.

HTH,
Dennis

Re: Help with programming 27xxx series eproms

By: Racer5 - 2009-03-29 04:12:12

would you be interested in fixing the programmer I have?  if so, how much and how long?  I can get it shipped out on Monday.  I am located in North Carolina


Yes, my old programmer worked reliably.  I basically use it to make replacement roms for commodore computers.  I have 2 in the computer sitting next to me and its been 2 years and still going strong.  Maybe I just got lucky.  dunno.

Re: Help with programming 27xxx series eproms

By: DL - 2009-03-29 15:21:11

Hi,
Jeff from Mooresville wrote on 03/29/09 at 04:12:12:
would you be interested in fixing the programmer I have?   if so, how much and how long?   I can get it shipped out on Monday.   I am located in North Carolina
It would cost you more than you paid for your programmer and I don't have the right equipment for working on SMT designs.   From the picture on their website, it looks like they continue to use a molded inductor.

If you're still interested in pursuing this, PM me your email address and we can work out details.

Whatever you do, don't buy the one they call Pro.   It's based on the enhanced universal design that has too many flaws.   If they didn't fix them, that variant shouldn't even be sold.   I tried telling them about the flaws with the enhanced universal design they used to sell but wasn't able to make them understand or didn't get a response.
Quote:
Yes, my old programmer worked reliably.   I basically use it to make replacement roms for commodore computers.   I have 2 in the computer sitting next to me and its been 2 years and still going strong.   Maybe I just got lucky.   dunno.
If they were NMOS EPROMs, I suspect they weren't programmed reliably.   Until a few years ago, it was generally accepted by people designing and using Willem type programmers that any EPROMs could be programmed reliably at VCC=5V (ignoring the fact that no Willem type programmer could provide 4.75-5.25V to an NMOS EPROM).   This was only true for 2716/2732.   After that generation, faster algorithms were introduced and they required explicit program margin checking (generally to 6.25V).   If you program using the faster algorithms with VCC=5V, you were only checking margin to 5V (again, ignoring voltage drop in the switch).   This is likely to have a significant affect on long term reliability (10-20 years, not 2).

The original algorithm was open loop.   They just programmed with a 50mS pulse that was guaranteed to program every bit of every EPROM that left the factory (they screened 100% of the parts).   The newer closed loop algorithms program in as little as 100uS (up to 2500uS worst case, but the programs we use don't seem to implement the algorithms correctly).   If you assume that the charging profile is exponential (like a capacitor), you can see that significantly fewer electrons are being deposited on the floating gate.   Fewer electrons means less long term reliability.

As a test, you could re-verify the EPROMs from your Commodore computers at VCC=6.25V; after you get a programmer that addresses the VCC voltage drop I mentioned earlier.   With a 100mA load at VCC=6.25V, I think most Willem designs will drop to around 5.5V.   So you'd still not be checking margin to the correct voltage on an unmodified programmer; unless you were using a 4.1 with the optional relay.

I've done experiments to simulate unreliable programming.   I programmed devices properly, then erased them for a short time.   I was able to make devices that read correctly at low VCC, but were read as mostly blank at 5.5V (which is still in the normal operating voltage).

HTH,
Dennis