Maker Pro
Maker Pro

Please help advise me on building a RAM tester

J

Joe Magiera

I posted this to a few different newsgroups, sorry for the triple postings,
I wasn't sure of the most appropriate place to post.

Since I can't find anything on the 'net that will test them, I am
considering building a RAM tester, specifically a 4116 DRAM tester. It is a
very old form of RAM. To those of you that may not remember it, the pinouts
are

___ ____
-5v -|1 \_/ 16|- Gnd
Din -|2 15|- /CAS
/WE -|3 4 14|- Dout
/RAS -|4 1 13|- A6
A0 -|5 1 12|- A3
A2 -|6 6 11|- A4
A1 -|7 10|- A5
+12v -|8 9|- +5v
-----------


I admit up front, I have no background in this type of stuff, but I do have
a willingness to try. My long range general idea so far is to have a fairly
large breadboard with 25 ZIF sockets on it (to test 25 at once). From there,
wire the appropriate voltage pins to a common point so I can supply power
via a switching power supply. That in itself is probably enough of a
challenge for me, but assuming I can do that (and that's a pretty big
assumption), then what? I'd like to have it interface with a PC to run a
program to test the RAMs at various speeds. I know pretty vague, but that's
where I'm looking for your input.

Thoughts? Ideas? Advice? Any comments appreciated. Thanks,

Joe

[email protected]
 
I

Ian Stirling

In sci.electronics.design Joe Magiera said:
I posted this to a few different newsgroups, sorry for the triple postings,
I wasn't sure of the most appropriate place to post.

Since I can't find anything on the 'net that will test them, I am
considering building a RAM tester, specifically a 4116 DRAM tester. It is a
very old form of RAM. To those of you that may not remember it, the pinouts
are

___ ____
-5v -|1 \_/ 16|- Gnd
Din -|2 15|- /CAS
/WE -|3 4 14|- Dout
/RAS -|4 1 13|- A6
A0 -|5 1 12|- A3
A2 -|6 6 11|- A4
A1 -|7 10|- A5
+12v -|8 9|- +5v
Hmm.
Inputs: 7 address + 2 strobes + 1 enable + 1 data = 11 pins.
Outputs: 1 pin.

There are 12 outputs and 4 (?) inputs from a standard parallel port.
At 100Khz rep rate, that's 320ms or so to read/write the entire chip.
Write 1, read, write 0, read, write 'random pattern' (to check 'stuck
at'), read.

Maybe 2 seconds.
Oh, connect 4 chip outputs to the PP inputs, and you're testing them
way, way faster than you can plug them in.
This will catch 99.99% of errors.
 
R

Rich Grise

I posted this to a few different newsgroups, sorry for the triple postings,
I wasn't sure of the most appropriate place to post.

In a case like that, crosspost. Whoever answers should be smart enough to
set followups.
Since I can't find anything on the 'net that will test them, I am
considering building a RAM tester, specifically a 4116 DRAM tester. It is a
very old form of RAM. To those of you that may not remember it, the pinouts
are ....
Thoughts? Ideas? Advice? Any comments appreciated. Thanks,

Just one thought - why are you making a tester for chips that haven't
even been made in about ten years?

Just build a general purpose Z80 SBC and use them for a bank of RAM. If
they work, they work. If not, trash them.

You use the Z80 because it has a refresh cycle built into every fetch
cycle. :)

But why in the world try to resurrect such obsolete chips? Even if they
_do_ work, they're a pain in the ass to try to use, and 4K of storage is
practically nothing!

Good Luck!
Rich
 
C

Chris

Rich said:
In a case like that, crosspost. Whoever answers should be smart enough to
set followups.


Just one thought - why are you making a tester for chips that haven't
even been made in about ten years?

Just build a general purpose Z80 SBC and use them for a bank of RAM. If
they work, they work. If not, trash them.

You use the Z80 because it has a refresh cycle built into every fetch
cycle. :)

But why in the world try to resurrect such obsolete chips? Even if they
_do_ work, they're a pain in the ass to try to use, and 4K of storage is
practically nothing!

Good Luck!
Rich

Hi, Rich. Actually 2KB, which is even more worthless.

If all the OP wants is a basic go-nogo test, and isn't concerned about
sorting for timing, most all of the old TTL IC testers had DRAM checks
built in. Googling

"IC Tester" +DRAM

produces any number of TTL/CMOS/DRAM/RAM testers. The first one goes
for the princely sum of $252 USD. It will give a go-nogo on these
fossils. The gently used IC testers are even better priced. I know of
one which works, as far as I know, and could be picked up for a six of
Corona, a lime or two, and a few old UFW songs -- but I'd have to ask
the guy, and haven't talked with him in a few years. I believe he got
out of engineering about the time of the big collapse in 2001, and
retired out west.

<wry plug for an unfortunate business associate>
By the way, if you or anyone knows of anyone who wants to start up
production of something with 4116s, I know a poor soul who got stuck
with a few pallets of 'em when the music stopped (new, 1991 date code),
as well as 4164s, and might be talked into getting rid of them at a
reasonable price. Feel free to email me if you happen to be trapped in
that Series Of Unfortunate Events, and I'll put you in touch with him.
10K minimum quantities, 100K+ preferred, fast delivery. Serious
inquiries only.
</plug>

Hey, Rich! Viva La Causa!
Chris
 
G

Glenn Gundlach

Hey Rich, those are 16K x 1 IIRC. I bought a set of 8 for my Apple][
back in '80 for $50. At least 3 of them failed in the next few years.
Whatever would some one want with those dinosaurs? Slower than mud.
Dang, in November I bought a PC3200 512 Meg DIMM for $60. Its a might
faster, too. GG
 
K

Kryten

This smells like a homework type project to me.

It was even on my university's list of final year projects to choose from.

You can certainly do it with a printer port.

I hacked the top off a ceramic 4164 and experimented getting it to work like
a crude camera!


A more interesting challenge would be to think of a project that found a use
for these old DRAMs. I hoarded loads of 4164 and 41256 myself. 8K or 32K is
quite big compared with many small micro's on-chip RAM.
 
Z

Zak

Chris said:
<wry plug for an unfortunate business associate>
By the way, if you or anyone knows of anyone who wants to start up
production of something with 4116s, I know a poor soul who got stuck
with a few pallets of 'em when the music stopped (new, 1991 date code),

Did he got stuck with these because of speculation on even higher memory
prices?
as well as 4164s, and might be talked into getting rid of them at a
reasonable price.

If it was speculation he'd better aceept his loss.


Thomas
 
Joe said:
I posted this to a few different newsgroups, sorry for the triple postings,
I wasn't sure of the most appropriate place to post.

Since I can't find anything on the 'net that will test them, I am
considering building a RAM tester, specifically a 4116 DRAM tester. It is a
very old form of RAM. To those of you that may not remember it, the pinouts
are

___ ____
-5v -|1 \_/ 16|- Gnd
Din -|2 15|- /CAS
/WE -|3 4 14|- Dout
/RAS -|4 1 13|- A6
A0 -|5 1 12|- A3
A2 -|6 6 11|- A4
A1 -|7 10|- A5
+12v -|8 9|- +5v
-----------


I admit up front, I have no background in this type of stuff, but I do have
a willingness to try. My long range general idea so far is to have a fairly
large breadboard with 25 ZIF sockets on it (to test 25 at once). From there,
wire the appropriate voltage pins to a common point so I can supply power
via a switching power supply. That in itself is probably enough of a
challenge for me, but assuming I can do that (and that's a pretty big
assumption), then what? I'd like to have it interface with a PC to run a
program to test the RAMs at various speeds. I know pretty vague, but that's
where I'm looking for your input.

Thoughts? Ideas? Advice? Any comments appreciated. Thanks,

Joe

[email protected]

With DRAM timing is critical for proper operation, any test that
doesn't test this aspect is useless. The simple way is to find a piece
of equipment that uses them and test by substitution, even that is a
very poor test.
 
R

Robert Lacoste

Joe Magiera said:
Since I can't find anything on the 'net that will test them, I am
considering building a RAM tester, specifically a 4116 DRAM tester. It is
a very old form of RAM.

A very low cost and easy solution : Buy an old Apple II or Apple II+ (but
not an Apple IIe !), which is equipped with 3x8 4116 RAM sockets for a great
48Ko total. Remove the top 16 of these 24 original RAM chips and put your
RAM to be tested in these sockets, and run one of the "RAM test" software
the used to be available for the Apple II... You will have an indication of
the failing chip if any...

Friendly,

--
Robert Lacoste
ALCIOM - The mixed signal experts
www.alciom.com

To contact us, thanks to click on the antispam link below :
http://www.cerbermail.com/?dCSHUxvwpw
 
M

Michael A. Terrell

Joe said:
I admit up front, I have no background in this type of stuff, but I do have
a willingness to try. My long range general idea so far is to have a fairly
large breadboard with 25 ZIF sockets on it (to test 25 at once). From there,
wire the appropriate voltage pins to a common point so I can supply power
via a switching power supply. That in itself is probably enough of a
challenge for me, but assuming I can do that (and that's a pretty big
assumption), then what? I'd like to have it interface with a PC to run a
program to test the RAMs at various speeds. I know pretty vague, but that's
where I'm looking for your input.

Thoughts? Ideas? Advice? Any comments appreciated. Thanks,

Joe

Wasn't the 4116 used in some early video arcade games? If it was the
people who collect and repair the old games might be interested in the
chips. I remember a lot of static damaged memory from careless handling
of the 4116 chips.
 
J

Joe Magiera

Hi,

No it isn't a homework project. One of the later replies did capture it
perfectly though. Many classic video games from that era (1980-1984) used
as many as 24 of the 4116's on a single boardset. I know they can be
replaced with 4164's, but if really for nothing more than a "nothing else to
do" project, and to test few thousand 000 of the 4116's I have lying around,
I thought maybe I'd give it a try. Thanks for the replies so far though.

Joe
 
Top