Image Map Image Map
Page 1 of 4 1234 LastLast
Results 1 to 10 of 31

Thread: PT 8KRA board

  1. #1
    Join Date
    Jan 2011
    Location
    Vancouver, BC
    Posts
    4,243
    Blog Entries
    2

    Default PT 8KRA board

    Can bad RAM chips cause other RAM chips to change their contents incorrectly?

    Like on a board with 60+ chips, can one or two bad apples wreck it for the others?

    I think I have found at least one bad 8T98. Initially when I powered the board up, if I dumped RAM contents for the address range it was set for, it was all 7Fs, and you could not change. Now what is happening is you power up, and to DU 2000 2010 let's say, and it looks sort of like this:

    7F 00 00 00 00 49 7F 00 00 00 7F

    And then if you do the same dump command again, you can see the 00s and other numbers slowly changing to 7Fs, until eventually they all are.

    This board came from Israel and looked like it spent time in the bilge of Noah's Ark. I've done a fair bit of cleaning and checking sockets, etc.. and it seems to be alright.. but who knows.

  2. #2

    Default

    I have worked extensively on the 16KRA board, not the 8KRA, but it wouldn't surprise me if there are similarities.

    I have restored 6 boards, all had similar problems. Most used the type of TI brand IC socket which grips the IC pin from side to side, rather than the usual across the thin axis of the pin.

    All of the IC pins, without exception, had problems. Without magnification, a grey or dark line is visible on both sides of each IC pin where the IC pin had been in contact with the socket's receiving claw. Magnification showed this to be a layer of corrosion/oxides which are insulators. Different brands of IC's were affected differently, some IC pins are silver plated, others tin plated etc. Spraying with cleaner alone doesn't help this at all.

    To clean the IC pins of this glassy oxide it required a rectangular metal edge run down them to scrape off the layer, I used the back edge (not sharp edge) of a number 11 scalpel bade. Once the dark line was scraped off and metal visible again I then cleaned the pin further with 2000 grade paper. Then when an IC had all the pin faces cleaned on each pin, I washed it in contact cleaner and coated the pins in WD-40. That was only 50 % of the battle, then came the sockets.

    I found each socket claw needed to be tested for tension with a test IC pin(removed from a worthless IC) held in a pair or forceps or small needle nose pliers. Don't use anything other than a standard IC pin or it will stretch the claw. With lubrication the pin is introduced into each socket hole a few times and the tension can be felt. The occasional one was damaged and needed repair. On these sockets (if the were the TI type) you can lift the plastic off and get access to all the pins. Often these claws get damaged, even from new, when the IC's were fitted when the IC pins were intially spaced too far apart than the standard DIL footprint and they were pushed into the socket with the pins at an angled trajectory, where they would either catch half the claw and bend it, or travel down beside the claw pair.

    It is a very long job to clean all the IC pins and check each socket pin on the card (especially for the 16kRA card). Even then in one case I found a fault remained where there was a resistive contact on an IC socket claw and it required attention twice. This shows up on an IC's input pin with the scope as a slow rising or falling wave without a rectangular edge as effectively a resistance is created between the output pin of one IC and the input of another.

    Before even contemplating any fault finding for defective IC's this cleaning process must be perfect or you can end up chasing your tail with faults when there may in fact not be a single defective part on the pcb. What you have described could still be bad connections. Check some of the IC pins under magnification.

    Once every connection is perfect, you can write a pattern to the memory and inspect the returned bytes. Usually from that its possible to deduce which ram chips are faulty.

  3. #3
    Join Date
    Jan 2011
    Location
    Vancouver, BC
    Posts
    4,243
    Blog Entries
    2

    Default

    Thanks Hugo. I have been bedeviled by problems with the sockets and chips on my Sol. Been a real process cleaning and re-cleaning.

    I've tested all the ICs to no effect, except for the 2102s, which I have done a couple of banks of in my digital group TVC video card to look for defects.. so far they all come up good. So that may turn my attention back to cleaning. However, I did do a very close inspection and discovered that both resistors between IC74 and IC75 have had one leg removed or broken off. I'm not sure if it was deliberate but given they are soldered in place on the working card i have, I'm going to guess they are probably necessary. So I will correct that tonight, test the rest of the 2102s and if no faults/success really get at those awful sockets.

  4. #4
    Join Date
    Jan 2010
    Location
    Central VA
    Posts
    4,380

    Default

    Man, that's a lot of patience for TI sockets! I've ended up stripping entire boards because of them, and resocketing.

  5. #5
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    30,992
    Blog Entries
    20

    Default

    They were the cheapest--and also the worst.

  6. #6
    Join Date
    Jan 2010
    Location
    Central VA
    Posts
    4,380

    Default

    Quote Originally Posted by Chuck(G) View Post
    They were the cheapest--and also the worst.
    I don't know about worst, there are some single-wipe no name sockets that are truly garbage.

  7. #7

    Default

    I'm not trying to start a major debate on this, but in my experience WD-40 eventually dries out and creates a varnish-like residue.

    As an alternative, Tweek is still being produced under the name of Stabilant 22A. Definitely not cheap but only a VERY little is required and I've experienced no problems using it.

  8. #8
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    30,992
    Blog Entries
    20

    Default

    Quote Originally Posted by glitch View Post
    I don't know about worst, there are some single-wipe no name sockets that are truly garbage.
    I'm speaking only from my own experience. Got tons of the TI low-profiles, bubble-wrapped on cards back in the day. Discovered that if you pull on the plastic part, it simply comes off the contacts. Given their performance, maybe that was for a reason--pull the shell off, clean the contacts, put the shell back on. A couple of years later, I discovered Augat machine-pin gold-flashed sockets--shipped in trays. My life changed forever.

  9. #9
    Join Date
    Jan 2010
    Location
    Central VA
    Posts
    4,380

    Default

    Quote Originally Posted by Chuck(G) View Post
    I'm speaking only from my own experience. Got tons of the TI low-profiles, bubble-wrapped on cards back in the day. Discovered that if you pull on the plastic part, it simply comes off the contacts. Given their performance, maybe that was for a reason--pull the shell off, clean the contacts, put the shell back on.
    The nice thing about that is, you can pop the plastic part off, and then remove each pin individually with tweezers and a soldering iron, when you end up having to resocket the whole board

    Quote Originally Posted by Chuck(G) View Post
    A couple of years later, I discovered Augat machine-pin gold-flashed sockets--shipped in trays. My life changed forever.
    Good quality machine pin sockets are all I will order nowadays. Augat probably makes the best you can get, but Mill-Max, R&N, Garry, etc. all make a good product. I prefer not to bother socketing an IC rather than using a cheap, junky socket.

  10. #10

    Default

    Quote Originally Posted by Chuck(G) View Post
    They were the cheapest--and also the worst.

    I'll put Robertson-Nugent sockets against those for total crappiness anyday.

    True story: the first IBM-PC I had was built from a bare board. I used sockets (65 of them, as I recall). It didn't work. I worked up the nerve to ask an extremely good, but nasty and gruff hardware engineer at our place for advice. He hated my boss and by association didn't like me either. He looked at it and said "I won't even look at the #%$#%$ thing until you remove those #@$@%#% R-N sockets!".

    So, one night after work, I sat at the solder station and in a several hour extravaganza, removed all of the sockets except for the UVEPROMS and installed the chips directly. The SOB started up no problem and worked until I tossed it in the mid-90's.

    After that, he was much nicer to me...

    To this day, I will only use Augat-type pin type sockets for anything, period.

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •