Info on 1702 eprom programmer?

Post by Kevin
I'm trying to get a 8080 S-100 system I built in the 70's working
again, but the cpu card is missing one of the 1702 eproms that hold
the bios. I have the code and several blank 1702s, but no programmer
that can handle a 1702 (multiple voltages, relatively large
programming voltage). I have been unable to find any information on
making a 1702 programmer and don't want to spend several hundred
dollars to buy one that can. Does anybody have a schematic or link to
a site that has information on building a 1702 programmer? I'm only
going to program a few of them, so I am willing to set the
address/data manually with switches.

Those are so obsolete, small, and require such horrendous drivers
to program that nobody has used them since the 2708 etc. came out.
They were obsolete by the time the 8080 appeared, and there were
even various flavors (1702 and 1702A) with different programming
requirements. I built one about 30 years ago, and never wanted to
see it again. Good luck in finding somebody with one.

The programming info would be on the 1702 (A) data sheets, if you
can find those.

--
Chuck F (***@yahoo.com) (***@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

Kevin

2004-10-04 20:44:29 UTC

Actually, what I have is a 1702A, and I have data sheets for it. I was
hoping to find a schematic already made so I wouldn't have to figure
it out.

Those are so obsolete, small, and require such horrendous drivers
to program that nobody has used them since the 2708 etc. came out.
They were obsolete by the time the 8080 appeared, and there were
even various flavors (1702 and 1702A) with different programming
requirements. I built one about 30 years ago, and never wanted to
see it again. Good luck in finding somebody with one.
The programming info would be on the 1702 (A) data sheets, if you
can find those.

Al Kossow

2004-10-08 01:03:16 UTC

Post by Kevin
Actually, what I have is a 1702A, and I have data sheets for it. I was
hoping to find a schematic already made so I wouldn't have to figure
it out.

look in the MCS 8 documentation at www.bitsavers.org/pdf/intel

there was a 1702 programming board for the MCS

CBFalconer

2004-10-08 03:12:52 UTC

Post by Al Kossow

Post by Kevin
Actually, what I have is a 1702A, and I have data sheets for it.
I was hoping to find a schematic already made so I wouldn't have
to figure it out.

look in the MCS 8 documentation at www.bitsavers.org/pdf/intel
there was a 1702 programming board for the MCS

If that was the one board development system for the 8008, it had a
1702 programmer built into it.

--
Chuck F (***@yahoo.com) (***@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

Barry Watzman

2004-10-04 19:34:19 UTC

The 1702 is a bitch to program, and there are very few programmers that
handle it. Any chance of using a 2708 instead?

I'm trying to get a 8080 S-100 system I built in the 70's working again,
but the cpu card is missing one of the 1702 eproms that hold the bios. I
have the code and several blank 1702s, but no programmer that can handle
a 1702 (multiple voltages, relatively large programming voltage). I have
been unable to find any information on making a 1702 programmer and
don't want to spend several hundred dollars to buy one that can. Does
anybody have a schematic or link to a site that has information on
building a 1702 programmer? I'm only going to program a few of them, so
I am willing to set the address/data manually with switches.
Kevin

Kevin

2004-10-04 20:48:36 UTC

My 'last resort effort' would be to do that; I have a simple schematic
for a 2708 programmer. However, I haven't been able to find any
locally. I would also have to make an adapter board for the eprom
sockets on the motherboard to replace the 1702As and handle the 2708.
In that case, I would probably go with a 2716, as I have those and a
programmer for them. Anything I do (short of finding a 1702 programmer
cheap) will require wiring. I'm looking for the easiest solution.

Kevin

Post by Barry Watzman
The 1702 is a bitch to program, and there are very few programmers that
handle it. Any chance of using a 2708 instead?

Barry Watzman

2004-10-04 23:57:14 UTC

The 2716 might make a lot of sense. But for 2708's, Cromemco made the
"Bytesaver" board, an EPROM board that has on-board programming.
Schematics are in the manuals which are on the Harte site.

Post by Kevin
My 'last resort effort' would be to do that; I have a simple schematic
for a 2708 programmer. However, I haven't been able to find any locally.
I would also have to make an adapter board for the eprom sockets on the
motherboard to replace the 1702As and handle the 2708. In that case, I
would probably go with a 2716, as I have those and a programmer for
them. Anything I do (short of finding a 1702 programmer cheap) will
require wiring. I'm looking for the easiest solution.
Kevin

Post by Barry Watzman
The 1702 is a bitch to program, and there are very few programmers
that handle it. Any chance of using a 2708 instead?

Edmund Cramp

2004-10-04 21:38:59 UTC

Post by Kevin
information on building a 1702 programmer?

I think that there was a JADE S-100 card that programmed 1702A's - I
should still have the card and manual around somewhere...

--
Remove nuts to reply.

Edmund Cramp

2004-10-04 22:57:49 UTC

Post by Edmund Cramp

Post by Kevin
information on building a 1702 programmer?

I think that there was a JADE S-100 card that programmed 1702A's - I
should still have the card and manual around somewhere...

Rats! It's an SD-SALES PROM-100 ... and does not support 1702's.
Sorry.

Edmund Cramp

--
Remove nuts to reply.

Fred J. Scipione

2004-10-05 01:43:46 UTC

Any possibility that you can get access to a library copies of Byte
magazine's first year? Carl Helmer's brother Pete wrote such an
article in the first few issues, and re-buffed Carl's suggestion to
automate the circuit with relays in place of switches.

primo

2004-10-07 22:07:14 UTC

On Tue, 05 Oct 2004 01:43:46 GMT, "Fred J. Scipione"

Post by Fred J. Scipione

As I recall programming the 1702, each programming pulse had to be
short because the chip couldn't take much power into each address, so
you had to pulse each address multiple times to get it fully
programmed. You had to let each address cool down between each
programming pulse. To save time, automating the process so that each
address was programmed in sequence made sense, you could do all the
addresses, then come back and do your second, third etc pulse to each
in sequence.

It wasn't till the 27xx series that you could program each address
with one pulse.

There was also a timing issue with asserting the programming pulse,
but I can't seem to remember what it was right now. I have a few 1702
chips somewhere that never got used since it was much easier to design
in a 2716.
Thanks, John.

Barry Watzman

2004-10-08 02:34:34 UTC

RE: "It wasn't till the 27xx series that you could program each address
with one pulse."

Your description of how the prom is programmed applies to the 2708 also.
It's programmed by repeatedly writing the data, it's not done with a
single pulse.

Frankly, I believe that the 2716 was that way also, based on the code
from the (Cromemco) Bytesaver software.

I have Byte Magazine from Vol. 1 number 1 to about 1983.

R. Steve Walz

2004-10-08 07:20:00 UTC

Post by Barry Watzman
RE: "It wasn't till the 27xx series that you could program each address
with one pulse."
Your description of how the prom is programmed applies to the 2708 also.
It's programmed by repeatedly writing the data, it's not done with a
single pulse.

-----------------------------
Nope. Single pulse. 50ms.

Post by Barry Watzman
Frankly, I believe that the 2716 was that way also, based on the code
from the (Cromemco) Bytesaver software.

-----------------------------
The old algorithm tried twice, but the second was a drop-dead effort
or failure. It merely tested whether the EPROM was busted.

Post by Barry Watzman
I have Byte Magazine from Vol. 1 number 1 to about 1983.

---------------------
Here's a manual programmer from which you could adapt any old 27xxx
programmer to do 2708's:
Loading Image...

-Steve

--
-Steve Walz ***@armory.com ftp://ftp.armory.com/pub/user/rstevew
Electronics Site!! 1000's of Files and Dirs!! With Schematics Galore!!
http://www.armory.com/~rstevew or http://www.armory.com/~rstevew/Public

CBFalconer

2004-10-08 08:58:00 UTC

Post by Barry Watzman
RE: "It wasn't till the 27xx series that you could program each
address with one pulse."
Your description of how the prom is programmed applies to the
2708 also. It's programmed by repeatedly writing the data, it's
not done with a single pulse.

-----------------------------
Nope. Single pulse. 50ms.

My 2708/2716 programmers used shorter pulses (1 mS IIRC), and
applied a series until the location was programmed. It then
overprogrammed by a factor of 2 or 3 (i.e. if the location 'took'
in n pulses, it applied a total of 3n pulses). This adaptive
writing cut programming time considerably, with no loss in
reliability. The sequence was something like:

FOR i := first TO last DO BEGIN
setaddress(i);
n := 0;
REPEAT
setdata(buff[i]); pulse;
n := n + 1;
IF n > maxn THEN GOTO failure;
UNTIL readdata = buff[i];
WHILE n > 0 DO BEGIN
setdata[buff[i]); pulse;
setdata[buff[i]); pulse;
END;
END;

--
Chuck F (***@yahoo.com) (***@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

Barry Watzman

2004-10-08 13:33:17 UTC

I believe that's how the Cromemco Bytesaver worked also.

Post by Barry Watzman
RE: "It wasn't till the 27xx series that you could program each
address with one pulse."
Your description of how the prom is programmed applies to the
2708 also. It's programmed by repeatedly writing the data, it's
not done with a single pulse.

-----------------------------
Nope. Single pulse. 50ms.

My 2708/2716 programmers used shorter pulses (1 mS IIRC), and
applied a series until the location was programmed. It then
overprogrammed by a factor of 2 or 3 (i.e. if the location 'took'
in n pulses, it applied a total of 3n pulses). This adaptive
writing cut programming time considerably, with no loss in
FOR i := first TO last DO BEGIN
setaddress(i);
n := 0;
REPEAT
setdata(buff[i]); pulse;
n := n + 1;
IF n > maxn THEN GOTO failure;
UNTIL readdata = buff[i];
WHILE n > 0 DO BEGIN
setdata[buff[i]); pulse;
setdata[buff[i]); pulse;
END;
END;

CBFalconer

2004-10-09 05:16:38 UTC

Post by Barry Watzman

Post by Barry Watzman
RE: "It wasn't till the 27xx series that you could program each
address with one pulse."
Your description of how the prom is programmed applies to the
2708 also. It's programmed by repeatedly writing the data, it's
not done with a single pulse.

-----------------------------
Nope. Single pulse. 50ms.

My 2708/2716 programmers used shorter pulses (1 mS IIRC), and
applied a series until the location was programmed. It then
overprogrammed by a factor of 2 or 3 (i.e. if the location 'took'
in n pulses, it applied a total of 3n pulses). This adaptive
writing cut programming time considerably, with no loss in
FOR i := first TO last DO BEGIN
setaddress(i);
n := 0;
REPEAT
setdata(buff[i]); pulse;
n := n + 1;
IF n > maxn THEN GOTO failure;
UNTIL readdata = buff[i];
WHILE n > 0 DO BEGIN
setdata[buff[i]); pulse;
setdata[buff[i]); pulse;
END;
END;

I believe that's how the Cromemco Bytesaver worked also.

Please don't toppost - fixed.

It doesn't take a raving genius to try such an algorithm. I kept
all the timing in the software, because Intel specifically
recommended against it at the time. This was done about '75 or
so. However the hardware limited the pulse length, so that
software foulups wouldn't burn holes in everything. It seemed to
work reliably, and saved a lot of irritating time.

--
Chuck F (***@yahoo.com) (***@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

R. Steve Walz

2004-10-09 04:07:27 UTC

Post by Barry Watzman
RE: "It wasn't till the 27xx series that you could program each
address with one pulse."
Your description of how the prom is programmed applies to the
2708 also. It's programmed by repeatedly writing the data, it's
not done with a single pulse.

-----------------------------
Nope. Single pulse. 50ms.

My 2708/2716 programmers used shorter pulses (1 mS IIRC), and
applied a series until the location was programmed. It then
overprogrammed by a factor of 2 or 3 (i.e. if the location 'took'
in n pulses, it applied a total of 3n pulses). This adaptive
writing cut programming time considerably, with no loss in
FOR i := first TO last DO BEGIN
setaddress(i);
n := 0;
REPEAT
setdata(buff[i]); pulse;
n := n + 1;
IF n > maxn THEN GOTO failure;
UNTIL readdata = buff[i];
WHILE n > 0 DO BEGIN
setdata[buff[i]); pulse;
setdata[buff[i]); pulse;
END;
END;
--
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

-----------------------
Sure, that's just the Intel "Intelligent Algorithm for EPROMs", but
it wasn't promoted till the 2732 and 2732A's. Before that most
programmers used just the single pulse algorithm.

-Steve

Herb Johnson

2004-10-05 20:14:29 UTC

I did a Google search on "1702 eprom s-100". The FIRST HIT returned
was to my S-100 documents site as below. I searched my own site for
the other references. I think these manuals may be informative, either
for a S-100 programmer or for a stand-alone programmer. But I would
not want to toggle eight switches 256 times without error.

Contact me for prices and terms for these PHOTOCOPIED manuals, part of
my sometimes-called "obsolete" and "self-serving" S-100 manuals
service. ;)

Herb Johnson

http://njcc.com/~hjohnson/d_ssm.html

Solid State Music (SSM) S-100 documentation

MB3 2K/4K EPROM board (1702) 8 pgs

(I believe this is a reader only.)

http://njcc.com/~hjohnson/d_altair.html

Altair 88-PMC PROM memory card, 30 pgs
1702 PROMs. Theory of operation, schematics, assembly.

(This may be a reader/writer card.)

http://njcc.com/~hjohnson/d_vector.html

Vector Graphics cards:
Prom/Ram schematic (1702s), 2 pgs

(Probably a reader.)

http://njcc.com/~hjohnson/d_misc.html

A. Szerlip's "PROMSetter" (TPS100) card. 144 pgs
includes schematics, operational details, even how to solder!
supports 1702's, 2704, 2708, 2716, some other PROMS.
(Our thanks to Mr. Szerlip for providing this manual.)

-----------end-----------

Herbert R. Johnson, voice 609-771-1503, New Jersey USA
<a href="http://retrotechnology.com/herbs_stuff"> web site</a>
<a href="http://retrotechnology.net/herbs_stuff"> domain mirror</a>
<a href="http://njcc.com/~hjohnson"> my oldweb site expires 2005</a>
my email address: herbjohnson ATT comcast DOTT net
wait a day and if that fails, try: hjohnson AAT retrotechnology DOTT
com
good used Mac, SGI, 8-inch floppy drives
S-100 IMSAI Altair computers, docs, by "Dr. S-100"

Herb Johnson

2004-10-14 16:28:49 UTC

http://retrotechnology.com/herbs_stuff/d_misc.html
A. Szerlip's "PROMSetter" (TPS100) card. 144 pgs
includes schematics, operational details, even how to solder!
supports 1702's, 2704, 2708, 2716, some other PROMS.
(Our thanks to Mr. Szerlip for providing this manual.)

I had a chance to review the above manual. It's large because there is
a lot of information. Mr. Szerlip describes the circuits, the chips
and programming, construction information, even how to solder! The
board is primarily set up to support both 2708's and 1702's, using
external boards for each programming socket. He provides additional
information programming some other odd PROMs. The source code is
simple, it's hand assembly, and was intended to run from the
IMSAI/Altair front panel. So it would not be difficult to take this
design and put it on some other computer, or a stand alone processor
of your choice.

The odd thing about the design is the use of dozens of transistors and
diodes. This may have been the designer's way to deal with the odd
voltages and requirements of the 1702 versus the 2708. But use of
modern parts (this design was from 1976 or thereabouts) and dedication
to only the 1702 may simplify construction.

One reason to preserve old designs is that they tend to be of simpler
components and so easier to interpret. Another reason, obvious in this
case, is to support the use of parts and equipment that have no
current support or simple alternatives. There is always a judgement
call when restoring old equipment: do I use a modern part, or do I use
parts consistent with the times of the design? If you are still
interested in programming 1702's, this document may be informative and
useful in building a programmer.

Herb Johnson

Herbert R. Johnson, voice 609-771-1503, New Jersey USA
<a href="http://retrotechnology.com/herbs_stuff"> web site</a>
<a href="http://retrotechnology.net/herbs_stuff"> domain mirror</a>
<a href="http://njcc.com/~hjohnson"> my oldweb site expires 2005</a>
my email address: herbjohnson ATT comcast DOTT net
wait a day and if that fails, try: hjohnson AAT retrotechnology DOTT
com
good used Mac, SGI, 8-inch floppy drives
S-100 IMSAI Altair computers, docs, by "Dr. S-100"

dwight elvey

2004-10-15 18:46:46 UTC

Post by Herb Johnson

I had a chance to review the above manual. It's large because there is
a lot of information. Mr. Szerlip describes the circuits, the chips
and programming, construction information, even how to solder! The
board is primarily set up to support both 2708's and 1702's, using
external boards for each programming socket. He provides additional
information programming some other odd PROMs. The source code is
simple, it's hand assembly, and was intended to run from the
IMSAI/Altair front panel. So it would not be difficult to take this
design and put it on some other computer, or a stand alone processor
of your choice.
The odd thing about the design is the use of dozens of transistors and
diodes. This may have been the designer's way to deal with the odd
voltages and requirements of the 1702 versus the 2708. But use of
modern parts (this design was from 1976 or thereabouts) and dedication
to only the 1702 may simplify construction.
One reason to preserve old designs is that they tend to be of simpler
components and so easier to interpret. Another reason, obvious in this
case, is to support the use of parts and equipment that have no
current support or simple alternatives. There is always a judgement
call when restoring old equipment: do I use a modern part, or do I use
parts consistent with the times of the design? If you are still
interested in programming 1702's, this document may be informative and
useful in building a programmer.
Herb Johnson

Hi Herb
What "modern part" do you suggest for the ~50 volt swings needed
for the 1702A's address and data lines during programming?
Dwight

Post by Herb Johnson

Lee Hart

2004-10-20 15:38:55 UTC

Post by Herb Johnson

I had a chance to review the above manual. It's large because there is
a lot of information. Mr. Szerlip describes the circuits, the chips
and programming, construction information, even how to solder! The
board is primarily set up to support both 2708's and 1702's, using
external boards for each programming socket. He provides additional
information programming some other odd PROMs. The source code is
simple, it's hand assembly, and was intended to run from the
IMSAI/Altair front panel. So it would not be difficult to take this
design and put it on some other computer, or a stand alone processor
of your choice.
The odd thing about the design is the use of dozens of transistors and
diodes. This may have been the designer's way to deal with the odd
voltages and requirements of the 1702 versus the 2708. But use of
modern parts (this design was from 1976 or thereabouts) and dedication
to only the 1702 may simplify construction.
One reason to preserve old designs is that they tend to be of simpler
components and so easier to interpret. Another reason, obvious in this
case, is to support the use of parts and equipment that have no
current support or simple alternatives. There is always a judgement
call when restoring old equipment: do I use a modern part, or do I use
parts consistent with the times of the design? If you are still
interested in programming 1702's, this document may be informative and
useful in building a programmer.
Herb Johnson

Hi Herb
What "modern part" do you suggest for the ~50 volt swings needed
for the 1702A's address and data lines during programming?

Very few modern ICs can deal with this high a voltage. The trend is to
make smaller geometries, which naturally *lower* breakdown voltages. So,
the answer will probably lie in older ICs.

My first thought would be to try the venerable ULN2003. This is an IC
that contains seven 50v 500ma darlington transistors, with input base
resistors so they can be driven from TTL, and output protection diodes
so they can drive inductive loads. This is a 25 cent IC that has many
second sources and has been built for 25 years.

--
"Never doubt that the work of a small group of thoughtful, committed
citizens can change the world. Indeed, it's the only thing that ever
has!" -- Margaret Mead
--
Lee A. Hart 814 8th Ave N Sartell MN 56377 leeahart_at_earthlink.net

dwight elvey

2004-10-08 01:11:11 UTC

Hi Kevin
You can see an example of a programming card that Intel made
to program 1702A's ( as well as 1702's ) at:

http://www.bitsavers.org/pdf/intel/

8008usersManualRev2_Nov72.pdf

One should note that the programming algorithm for 1702's and
1702A's were some different. Of course, there are vary few 1702's
around ( I have one ). 1702A's are more common. I have a data
book with the programming specification for the 1702A's as well
but I'd have to look for it.
If you look at the pages from the pdf, it shows a board called
a MP7-03. These will be on pdf pages 34-37 or manual pages 32-35.
The schematic shows two transformers that are rated at 1 amp out
to create the programming voltage. The board doesn't require
anywhere near that much. Still, although the schematic shows
the transformers being 25.2V, they can be replaced with ordinary
24V transformers that you can get anywhere. One can, of course,
simply bypass these with a DC power supply producing 75-80VDC
at the pins for the external filter cap.
Some of the oneshots are a little hard to find as are most
of the transistors. These can be replaced by more current transistors
of comparible voltage, current and gain.
Someplace, I have the information on what to set the pots to
and how to toggle the control lines to the board. One could run
it from some switches but you need to know that the 1702A's
require about 50 pulses for each location. 50 X 256 is about
12800 pulses. It would make more sense to connect things to
some type of micro controller or computer.
One could use some 555's to control the pulsing and number
of pulses ( a few more or less is not a big deal ).
The last option is that I can program some PROMs for you.
I have a setup to program 1702A's. Let me know.
Dwight

Kevin

2004-10-08 15:45:41 UTC

Thanks to all who replied. I now have several schematics for building
a 1702A programmer. I will probably pick one to build, but given the
wiring and programming sequence, I am first going to take a good look
at making a 2716 adapter. I will have to rewrite the BIOS code
somewhat to allow for the ram stack address change, but that is not
much of a problem. I plan to get cp/m working on this eventually- I
have a versafloppy card now.

Dwight, thanks for the offer to program, but I first need to get the
original setup working, then work on cp/m support. That would be a lot
of mailing eproms back and forth. If I can't do it any other way, I'll
contact you.

I'm trying to get a 8080 S-100 system I built in the 70's working again,
but the cpu card is missing one of the 1702 eproms that hold the bios. I
have the code and several blank 1702s, but no programmer that can handle
a 1702 (multiple voltages, relatively large programming voltage). I have
been unable to find any information on making a 1702 programmer and
don't want to spend several hundred dollars to buy one that can. Does
anybody have a schematic or link to a site that has information on
building a 1702 programmer? I'm only going to program a few of them, so
I am willing to set the address/data manually with switches.
Kevin

dwight elvey

2004-10-11 15:59:44 UTC

Hi Kevin
You shouldn't need to move any addresses around. Just
don't select all the address space of the 2716. Just use
part of the EPROM and tie the unneeded address lines to ground.
This would be the minimum work.
Dwight

Post by Kevin
Thanks to all who replied. I now have several schematics for building
a 1702A programmer. I will probably pick one to build, but given the
wiring and programming sequence, I am first going to take a good look
at making a 2716 adapter. I will have to rewrite the BIOS code
somewhat to allow for the ram stack address change, but that is not
much of a problem. I plan to get cp/m working on this eventually- I
have a versafloppy card now.
Dwight, thanks for the offer to program, but I first need to get the
original setup working, then work on cp/m support. That would be a lot
of mailing eproms back and forth. If I can't do it any other way, I'll
contact you.

I'm trying to get a 8080 S-100 system I built in the 70's working again,
but the cpu card is missing one of the 1702 eproms that hold the bios. I
have the code and several blank 1702s, but no programmer that can handle
a 1702 (multiple voltages, relatively large programming voltage). I have
been unable to find any information on making a 1702 programmer and
don't want to spend several hundred dollars to buy one that can. Does
anybody have a schematic or link to a site that has information on
building a 1702 programmer? I'm only going to program a few of them, so
I am willing to set the address/data manually with switches.
Kevin

Kevin

2004-10-11 19:31:43 UTC

Dwight,

This board was custom made at the Univ. of Iowa Physics dept. It was
set up so that you could have either 1 K of ROM (4 1702s) or 768 bytes
of ROM and 256 bytes of RAM (used mainly for the stack). That's the
way I wrote my BIOS. I could use only 256 bytes of three different
2716s, or 512 bytes of one and 256 of a second, or 1K of the 2716, but
not 768 bytes. The ram addresses overlap the 4th 1702 address (or
maybe the 1st, I don't remember now), so if I use 1K I have to remove
the ram, and put the stack in main memory. However, if I do it this
way, I could have two different BIOS code, 1K each in the upper and
lower halves of the 2716. By flipping an address switch and rebooting,
I could go from the cassette based version to the cp/m floppy based
version. I don't think I can fit both into 1K.

Kevin

Post by dwight elvey
Hi Kevin
You shouldn't need to move any addresses around. Just
don't select all the address space of the 2716. Just use
part of the EPROM and tie the unneeded address lines to ground.
This would be the minimum work.
Dwight

R. Steve Walz

2004-10-12 02:07:33 UTC

Post by Kevin
Dwight,
This board was custom made at the Univ. of Iowa Physics dept. It was
set up so that you could have either 1 K of ROM (4 1702s) or 768 bytes
of ROM and 256 bytes of RAM (used mainly for the stack). That's the
way I wrote my BIOS. I could use only 256 bytes of three different
2716s, or 512 bytes of one and 256 of a second, or 1K of the 2716, but
not 768 bytes. The ram addresses overlap the 4th 1702 address (or
maybe the 1st, I don't remember now), so if I use 1K I have to remove
the ram, and put the stack in main memory. However, if I do it this
way, I could have two different BIOS code, 1K each in the upper and
lower halves of the 2716. By flipping an address switch and rebooting,
I could go from the cassette based version to the cp/m floppy based
version. I don't think I can fit both into 1K.
Kevin

-----------------
No, you're thinking about this wrong. Just make an adapter for each
2716 to socket into that then plugs into the 1702 socket, and use only
the address lines the 1702 needs, ground the rest in that adapter.
The board cannot tell the difference.

-Steve
--
-Steve Walz ***@armory.com ftp://ftp.armory.com/pub/user/rstevew
Electronics Site!! 1000's of Files and Dirs!! With Schematics Galore!!
http://www.armory.com/~rstevew or http://www.armory.com/~rstevew/Public

Post by Kevin

CBFalconer

2004-10-12 07:19:53 UTC

Post by Kevin
This board was custom made at the Univ. of Iowa Physics dept. It was
set up so that you could have either 1 K of ROM (4 1702s) or 768 bytes
of ROM and 256 bytes of RAM (used mainly for the stack). That's the
way I wrote my BIOS. I could use only 256 bytes of three different
2716s, or 512 bytes of one and 256 of a second, or 1K of the 2716, but
not 768 bytes. The ram addresses overlap the 4th 1702 address (or
maybe the 1st, I don't remember now), so if I use 1K I have to remove
the ram, and put the stack in main memory. However, if I do it this
way, I could have two different BIOS code, 1K each in the upper and
lower halves of the 2716. By flipping an address switch and rebooting,
I could go from the cassette based version to the cp/m floppy based
version. I don't think I can fit both into 1K.

Don't toppost. Do snip.

Somewhere you are generating a chip select for the 256 byte ram.
Apply that to inhibit the chip select for the 2716, and you will
cut it to the 768 bytes needed.

--
Chuck F (***@yahoo.com) (***@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

Randy McLaughlin

2004-10-12 07:26:22 UTC

Don't toppost. Do snip.
Somewhere you are generating a chip select for the 256 byte ram.
Apply that to inhibit the chip select for the 2716, and you will
cut it to the 768 bytes needed.
--
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

I have a question:

Once you have decided to replace the 1702 with a newer chip why not use a
flash ROM?

They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).

Randy

dwight elvey

2004-10-12 17:43:26 UTC

Once you have decided to replace the 1702 with a newer chip why not use a
flash ROM?
They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).
Randy

Hi Randy
Are you talking about those NVRAM's that Dallas sells. These
would be an option if he didn't worry about accidentally writing
over the bios part. Still, I think Steve's idea is the simplest,
just make an adapter for each socket and use 3 2716's. These
are cheap. HSC has used ones for 99 cents.
When I do this kind of thing, I use two machine pin sockets.
I can then pop out pins or break off the tips of pins and
cross wire what I need. I do this often to read mask ROMs
using a EPROM programmer, since the select signals are usually
different.
Dwight

Randy McLaughlin

2004-10-12 19:54:00 UTC

<snip>

Post by Randy McLaughlin
Once you have decided to replace the 1702 with a newer chip why not use a
flash ROM?
They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).
Randy

No I am refering to Flash devices such as 29f010, etc. Howard uses such
devices on the SuperIO and most PC manufacurers use them for BIOS storage.

They are extremely friendly to use.

I recomend using a programmer like the Willem programmer but it is not
needed, all programming can be done with standard TTL levels.

It is written to basically like RAM, the programming cycle is built into the
chip and is complete when you can read the same byte back that was just
written.

The only "special" needs are erasing. Anyone really interested can lookup
the datasheets and download the PDF's.

Anyone replacing an eprom on a vintage system might consider it.

Randy

dwight elvey

2004-10-13 17:26:23 UTC

Post by Randy McLaughlin
Once you have decided to replace the 1702 with a newer chip why not use

Post by Randy McLaughlin
flash ROM?
They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).
Randy

No I am refering to Flash devices such as 29f010, etc. Howard uses such
devices on the SuperIO and most PC manufacurers use them for BIOS storage.
They are extremely friendly to use.
I recomend using a programmer like the Willem programmer but it is not
needed, all programming can be done with standard TTL levels.
It is written to basically like RAM, the programming cycle is built into the
chip and is complete when you can read the same byte back that was just
written.
The only "special" needs are erasing. Anyone really interested can lookup
the datasheets and download the PDF's.
Anyone replacing an eprom on a vintage system might consider it.
Randy

Hi Randy
These work fine for EPROM replacements but he has a problem
with overlapping a 256 byte chunk with RAM. The flash can't be
used like a RAM but the NVRAM can.
I often use flash ROM's for various things. They are generally
great for most things. There are also some flashes that have
specialized chunks to be boot flashes. These are intended to
have some minimal boot code. I find it interesting that there
is often enough room allowed to include an entire Forth kernel.
Dwight

Randy McLaughlin

2004-10-13 20:55:30 UTC

Post by Randy McLaughlin
Once you have decided to replace the 1702 with a newer chip why not use

Post by Randy McLaughlin
flash ROM?
They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).
Randy

<snip>

Post by dwight elvey
Hi Randy
These work fine for EPROM replacements but he has a problem
with overlapping a 256 byte chunk with RAM. The flash can't be
used like a RAM but the NVRAM can.
I often use flash ROM's for various things. They are generally
great for most things. There are also some flashes that have
specialized chunks to be boot flashes. These are intended to
have some minimal boot code. I find it interesting that there
is often enough room allowed to include an entire Forth kernel.
Dwight

I was only suggesting using a flash ROM in place of an EPROM, not use it as
RAM. Replacing a 27XX EPROM with a 29FXX is easy. With a little effort it
can even be made to write.

I saw the reference to RAM on the board that he is working on but ignored
that reference since the talk of wiring in a 27XX EPROM was being discussed.

Randy

dwight elvey

2004-10-15 18:48:01 UTC

Post by Randy McLaughlin
Once you have decided to replace the 1702 with a newer chip why not

use
a

Post by Randy McLaughlin
flash ROM?
They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).
Randy

<snip>

I was only suggesting using a flash ROM in place of an EPROM, not use it as
RAM. Replacing a 27XX EPROM with a 29FXX is easy. With a little effort it
can even be made to write.
I saw the reference to RAM on the board that he is working on but ignored
that reference since the talk of wiring in a 27XX EPROM was being discussed.
Randy

Hi Randy
This makes sense.
Dwight

Jeff Jonas

2004-11-09 12:01:24 UTC

Once you have decided to replace the 1702 with a newer chip why not use a flash ROM?
They are cheap.
They are easy to program.
They are fast.
They can usually be removed from and old/dead PC MB (free).

No I am refering to Flash devices such as 29f010, etc. Howard uses such
devices on the SuperIO and most PC manufacurers use them for BIOS storage.
They are extremely friendly to use.

I was going to reply to this but all I can say is "I agree!"
- I salvaged many flash roms (happily, many are still DIPS instead of leadless)
- I recently found a diagnostic ROM in a correct-a-dip / piggyback socket
with a switch for what part of the ROM to use!
(older Promise disk controllers with the separate ROM
had PCB traces that selected what half to use).

As to retrofitting 3 chips where there was one:
are you sure the board can handle the additional load?

Kevin

2004-10-13 16:14:03 UTC

As others have suggested, I will probably use three different 2716s,
at least for now. I happen to have several handy, and a UV eraser and
programmer that will handle the 2716. I'm going to make a 'long
strip', consisting of 2 rows of header pins to plug into the card
socket, and a socket for the 2716, using wire wrap to make the adapter
connections. The pinouts of the 1702 and 2716 are quite different;
stacking one on top of the other would be a more difficult wiring job.
I'm going to replace just the missing 1702 first and see if that
works. If it does, I'll replace the other two. The problem with the
ram is that I don't have the schematic for the card. I tracked down a
friend who has the same system, with schematics, but he moved a year
ago and hasn't yet been able to find the schematics.

Kevin