Programming old 27xxx EPROMS (TL866II Plus)

[sequoia]

While restoring an old PC/XT clone, I needed new BIOS (to fix buggy clone bios from 1987 on the computer).

I compiled latest Super PC/Turbo XT BIOS (https://www.phatcode.net/downloads.php?id=101), and found some old Intel D27128A chips (to replace MBM27128-25 that was on the motherboard), and thought I'd simply program one of these with a TL866II Plus programmer, but that's where the problems started...

Chip verified as blank ok, but programming kept failing about half way through with  "OverCurrent Protection ! External Short or IC reverse or incorrect package!" error.
Software seemed to default: VPP Voltage = 12.00V, VDD Write = 5.00V, VCC Verify = 5.00V.   I looked up the datasheet and seems like programming voltage should be "Vcc = 6.0V and Vpp = 12.5V"...

So, I tried changing settings in the (Xgpro v9.16) programing software to: VPP Voltage = 12.50V and VCC Write = 5.5V  (no 6.0V option, next closest is 6.5V), but programming failed immediately...

After some experimenting with voltages I noticed dropping VPP to 11.5V allowed programming to proceed about 75% before erroring out.... and with VPP set to 11.0V programming was successful! ...go figure.


Chip seems to now work ok (nice that you can "continue" if programming ever fails on these UV EPROMs...), I guess only time will tell if using lower than specified Vpp has any effect on longevity of the programming?


Btw, anyone run into similar issues with cheap "TLL866" programmers?   I was wondering if mine is just defective unit or do all of these behave the same?

[Alex Eisenhut]

Don't know but I've found that USB cables are often crap, what are you using for your programmer?
I had a cable that looked fine but the programmer would not power up properly , the RUN light would flash all the time.
Tried another cable and it works fine. Maybe some cables are so bad they can't carry the current.

[sequoia]

Thanks, didn't think about the cable.

I tried with known good cable; a 1.5ft cable from Monoprice (UL listed cable labeled as 28/24 AWG), no difference versus the 3ft cable that came with the unit (this cable has nice feel and labeled 28/24 AWG , but no UL listing/E number...).

Also checked voltage/current during write and its 5.00V/0.24A (measured using UNI-T U658), so doesn't seem like its issue with too much current draw from the host port either...
 

[Alex Eisenhut]

Well I tried a few different EEPROMs of the 28C variety. An Atmel 28C256 can't be erased with my TL866ii plus. Strangely when I program it it erases just fine and I can verify the data is programmed correctly.

There seem to be some weird quirks here and there with the unit. It's still a good programmer IMO.

How long did you erase the EPROMs?

[GeorgeOfTheJungle]

Yes, back in the day, many EPROMs needed 21 or even 24V. And some chips needed longer program pulses than others, usually ~50 ms, sometimes more than that.

In the 70s, you were nobody unless you had made your own EPROM programmer with a bunch of TTLs :-)
Bonus points for
-Using a 6820 or 6522 PIA
-Boost converter to generate VPROG from 5V

[techman-001]

Yes, back in the day, many EPROMs needed 21 or even 24V. And some chips needed longer program pulses than others, usually ~50 ms, sometimes more than that.

In the 70s, you were nobody unless you had made your own EPROM programmer with a bunch of TTLs :-)
Bonus points for
-Using a 6820 or 6522 PIA
-Boost converter to generate VPROG from 5V

Yep, making a eprom burner back then was about as special as a postman knowing how to ride a bike.

Bonus points for
-Boost converter to generate VPROG from 5V
CHECK    Small potcore, pushpull with feedback and linear regulator on output. Takes about an hour to make inc winding the bobbin.

-Using a 6820 or 6522 PIA
CHECK 
 I still have about 200x 6821's in 40 pin DIP in my stock. I've interfaced them with 8085's and PC parallel ports in the past.

Here is a GPL'd general purpose mcu and prom burner (using a 6821) I designed and built 20 years ago. In this configuration it programs a 89C2051 in circuit using relays to switch the chip between target and programmer as orchestrated via a Makefile. The program was written in C using Gcc on X86 with a emphasis on general purpose configuration so it was easy to adapt new targets. One person rewrote my crappy direct hardware access and sent me a driver for it 

[GeorgeOfTheJungle]

Yep, making a eprom burner back then was about as special as a postman knowing how to ride a bike.

Bonus points for
-Boost converter to generate VPROG from 5V
CHECK    Small potcore, pushpull with feedback and linear regulator on output. Takes about an hour to make inc winding the bobbin.

-Using a 6820 or 6522 PIA
CHECK 
 I still have about 200x 6821's in 40 pin DIP in my stock. I've interfaced them with 8085's and PC parallel ports in the past.

Here is a GPL'd general purpose mcu and prom burner (using a 6821) I designed and built 20 years ago.[...]

 

I used bit banging + (74164?) shift registers (invented SPI!) and the game connector of my Apple II. It worked so well on the breadboard that couldn't be arsed to make a proper pcb.

[dexters_lab]

While restoring an old PC/XT clone, I needed new BIOS (to fix buggy clone bios from 1987 on the computer).

I compiled latest Super PC/Turbo XT BIOS (https://www.phatcode.net/downloads.php?id=101), and found some old Intel D27128A chips (to replace MBM27128-25 that was on the motherboard), and thought I'd simply program one of these with a TL866II Plus programmer, but that's where the problems started...

Chip verified as blank ok, but programming kept failing about half way through with  "OverCurrent Protection ! External Short or IC reverse or incorrect package!" error.
Software seemed to default: VPP Voltage = 12.00V, VDD Write = 5.00V, VCC Verify = 5.00V.   I looked up the datasheet and seems like programming voltage should be "Vcc = 6.0V and Vpp = 12.5V"...

So, I tried changing settings in the (Xgpro v9.16) programing software to: VPP Voltage = 12.50V and VCC Write = 5.5V  (no 6.0V option, next closest is 6.5V), but programming failed immediately...

After some experimenting with voltages I noticed dropping VPP to 11.5V allowed programming to proceed about 75% before erroring out.... and with VPP set to 11.0V programming was successful! ...go figure.


Chip seems to now work ok (nice that you can "continue" if programming ever fails on these UV EPROMs...), I guess only time will tell if using lower than specified Vpp has any effect on longevity of the programming?


Btw, anyone run into similar issues with cheap "TLL866" programmers?   I was wondering if mine is just defective unit or do all of these behave the same?

check the exact part number of the EPROM, there's countless versions of the 27Cxxxx and it's easy to end up with an old one with high write voltages.

The older TL866 can program at 21v but the max voltage is lower in the new versions, there are also fake TL866 programmers out there...

[sequoia]

check the exact part number of the EPROM, there's countless versions of the 27Cxxxx and it's easy to end up with an old one with high write voltages.

The older TL866 can program at 21v but the max voltage is lower in the new versions, there are also fake TL866 programmers out there...

Chip I was using is Intel D27127A. All datasheets online for "D27128A" seem to actually be for Intel "M27128A".... (that has Vpp 12.5V)

Software for TL866 has entry for "INTEL D27128A @DIP28" that I used... Verify (chip) ID seems to work, so seems like chip matches what the programmer expects ("Read Device ID: 89 89").

[Alex Eisenhut]

Curious: did you try the tools->system self-check?

[rsjsouza]

Yes, back in the day, many EPROMs needed 21 or even 24V. And some chips needed longer program pulses than others, usually ~50 ms, sometimes more than that.

Yes, the TL866CS can't program 25V parts, as I found out with specific TI parts:
https://www.eevblog.com/forum/vintage-computing/programmer-for-tms2516/msg2699648/#msg2699648

Here is a GPL'd general purpose mcu and prom burner (using a 6821) I designed and built 20 years ago.

That is a very nice project! Thanks for sharing. I have several PIAs lingering around, as well as some 6502, 6507 and Z80s - I may take the spiderwebs out of them and put them to work some day.

[sequoia]

Curious: did you try the tools->system self-check?

Yes, self-check works, and unit works fine programming/reading other chips.

No issues writing Intel D27256A  or D2764A.  Maybe there is something "wrong" with this particula D27128A I have been testing with, I'll see if I can locat another chip and see if see same behaviour with the programmer....

[sequoia]

If the auto ID says it's 89 89 then it *should* be a 12.5V part, the older D parts should be 89 83 ID so I suspect you've a dodgy chip, either it's faulty or it's not erased quite as well as it could be.

I've had chips that misbehaved during programming and another ten mins in the eraser sorted them, similarly I've had chips which just failed in odd ways.

Would be interesting to know how it works with another chip.

Do you hav e an eraser?

"Read Device ID: 89 89"  so seems to be correct 12.5V part....

I found another chip and this one fails immediately, even if dropping Vpp voltage.

Seems like these need more current that TL866 can supply, since when measuring Vpp while programming with Vpp 11.0V, I see 8.25V across the pins Vpp and GND pins... If programming with (default 12.0V) Vpp is about 9V until it suddenly drops in the middle of programming (presumably over current protection kicking in...) and programming fails...


I have an eraser, I'll try erasing these longer than usual to see if it makes any difference....

[Bicurico]

I have a Genius G540 programmer. It fails programming a Winbond 27C512-45 on my main PC running Windows 10.
Using the same device and IC on my old Pentium III laptop running Windows XP I can program the IC without any issue - Read, Write, Erase and Verify work without flaws.
I think the cheap chinese programmers are not meant to use Windows 10 and very fast machines.
I keep that laptop specifically for these tasks: LPT JTAG, serial port communications, X1541 transfer, use of G540, use of smartcard programmer, etc.
While in theory I could or should be able to do all of this on my main PC, the truth is that I get all sorts of errors, need to use unsigned drivers in special mode, etc.

Conclusion: it might be worth to try and use an older computer.

Regards,
Vitor

[gslick]

"Read Device ID: 89 89"  so seems to be correct 12.5V part....

Seems like these need more current that TL866 can supply, since when measuring Vpp while programming with Vpp 11.0V, I see 8.25V across the pins Vpp and GND pins... If programming with (default 12.0V) Vpp is about 9V until it suddenly drops in the middle of programming (presumably over current protection kicking in...) and programming fails...

The Intel 27128A datasheet says for programming VCC = 6.0V with a minimum of 5.75V, and VPP = 12.5V with a minimum of 12.0V. The max VCC current during programming is 100mA, and the max VPP current during programming is 50mA.

Intel 1991 Memory Products
http://www.bitsavers.org/components/intel/_dataBooks/1991_Intel_Memory_Products.pdf

See pages 304 and 305 of the PDF (pages 5-35 and 5-36) for the 27128A programming flowchart and DC and AC programming characteristics.

[SilverSolder]

I have a Genius G540 programmer. It fails programming a Winbond 27C512-45 on my main PC running Windows 10.
Using the same device and IC on my old Pentium III laptop running Windows XP I can program the IC without any issue - Read, Write, Erase and Verify work without flaws.
I think the cheap chinese programmers are not meant to use Windows 10 and very fast machines.
I keep that laptop specifically for these tasks: LPT JTAG, serial port communications, X1541 transfer, use of G540, use of smartcard programmer, etc.
While in theory I could or should be able to do all of this on my main PC, the truth is that I get all sorts of errors, need to use unsigned drivers in special mode, etc.

Conclusion: it might be worth to try and use an older computer.

Regards,
Vitor

It is sometimes possible to use an older OS (e.g. Win 7) in a virtual machine (e.g. Virtual Box) which can then "take over" a USB port or two.   Magically, everything works now,  that Windows 10 couldn't do.

[sequoia]

It sounds like you've a problem with the programmer if VPP is bouncing around.

Is there anyone who has a different programmer who can test a chip by programming it for you?

Didn't seem like erasing these chips helped either. Normally about 10-15minutes works with the eraser I have, this time I had chips in there for over an hour, but that didn't help...

If someone is interested to test these on their own programmer, just PM me for address and I'll gladly mail you these couple chips (US only please, so shipping stays reasonable and no customs declaration forms to fill...)

[scan2005]

The programming of 27xxx EPROMS is perfectly handled by the programmer "Master" http://mprog.ru