On Tue, 15 Dec 2009, Mr Ian Primus wrote: ... and when you get that fixed, there's an excellent chance that the foam pads in the keyboard have disintegrated. My Model II had only 5 working keys when I got it, but I was able to rebuild the keyboard with the help of this article: http://www.solivant.com/sol20kbd/ Mike Loewen mloewen at cpumagic.scol.pa.us Old Technology http://sturgeon.css.psu.edu/~mloewen/Oldtech/

--- On Tue, 12/15/09, Bob Brown <bbrown at harpercollege.edu> wrote: According to the manual, the video board should be in the 4th slot from the power supply. The back of the board has the numbers 8709048. Assuming that the documentation here is correct, the video RAMs should be in sockets, and are U19-22. They are 2114's. Swap the chips around in the sockets and see if you can get the symptom to change at all - you should be able to get the messed up sections of the screen to change. If so, then one is definitely your culprit. As for obtaining replacement parts, I know that Unicorn Electronics has 2114's (www.unicornelectronics.com, look under "Static RAMs") -Ian

How do I identify which board is the video/keyboard interface board? (I see 4 boards in the card cage)? How do I identify which chip has failed? Where do I get replacements? ======== Use the force :) You know that a byte has 8 bit. You know that a space (chr 32) is 00100000 in binary You know that a exclamation point (chr 33) is 00100001 in binary You know that in the place of a space, there is a ! on the screen So, you know that the least significant bit, or D0 line, is stuck in "1" Get the schematic and see which 2114 is connected to the A0 line. Change it, and you're set :o)

The entire screen is having the trouble. (it is odd that when it asks me to mount the disk, the !'s are inverse...as soon as I put the disk in, they become normal-video !'s. Also, the oddly swapped around characters... How easy is it to pull those chips and swap them around? Looking at the web page, do I need just plain 2114's? (I see several variations of them including different speeds and low-power etc)? (I've never swapped chips before..hence the questions). Thanks. -Bob -----Original Message----- From: cctalk-bounces at classiccmp.org [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Mr Ian Primus Sent: Tuesday, December 15, 2009 9:55 AM To: General Discussion: On-Topic and Off-Topic Posts Subject: Re: trs-80 model II trouble --- On Tue, 12/15/09, Alexandre Souza - Listas <pu1bzz.listas at gmail.com> wrote: Don't you mean, D0 line? <grin> According to the schematics, both U21 and U22 are on the D0 line. One thing I just thought of - is the ENTIRE screen affected, or just half of it? If one chip had a stuck line, it should only mess up half the screen. In any event, the chips should be socketed, so doing a couple quick swaps should narrow it down quickly. -Ian

Hi It could still be a single RAM failing. If it is a stuck output line, it could over ride the others output. RAM failure is still the most likely. Swapping the RAMs for the high and low nibbles would tell which it is. If they are socketed, they should be able to be removed by carefully prying them up with a small screw driver. I do this by doing a little bit on each end. If you have two screw drivers it is a little faster. Of course, they do have special tools and I think I even have one, someplace in one of my tool boxes or drawers. Just don't pry the socket off the board or you'll need a soldering iron as well. Dwight _________________________________________________________________ Your E-mail and More On-the-Go. Get Windows Live Hotmail Free. http://clk.atdmt.com/GBL/go/171222985/direct/01/

On Tue, Dec 15, 2009 at 10:03 AM, Bob Brown <bbrown at harpercollege.edu>wrote: The space was swapped for a ! because of the stuck bit. So any character that has a 0 in that bit position will be mapped to the character with a 1 in that position, but all the other bits the same. The end result is that half of the characters will be wrong. All the characters with a 1 in that position will be correct. I guess a stopped clock is right twice a day. :-) If they're dips, it's easy. Just don't bend the pins when you put them back in a socket. There's special tools for pulling them out, but you can use a small flat head screwdriver if you're careful. Socketed dip chips have a habit of slowly walking out of the sockets from years of heat expansion and contraction when the machine is switched on and off. It's possible that reseating the chips (just pushing them back into the sockets) will fix the problem. But if they've walked out a bit, it's easier to remove them since a flat screwdriver will fit between the chip and the socket. If they're all the way in, i've used an exacto knife to get between the chip and socket before. Just don't break off the blade. Try to get the chip to come out straight, that is work both ends of the chip. Otherwise, the pins will bend when you take it out. Oh, and don't put them in backwards or they'll fry. They'll get hot enough to burn you. So note the orientation before taking them out. Also, I've received new chips (back when I routinely installed cache chips on 486 motherboards) where the pins sort of splayed out a bit. This made them not line up with the sockets quite right. You may need to bend them so that they're slightly more upright making them line up with the holes in the socket. This is easier than it sounds. brian

On Tue, 15 Dec 2009, Brian Lanning wrote: When the fancy new "Master/Slave" Simplex clock in the library went dead, it was stuck at 12:00. The administration explicitly ordered me to not fix it. So, in order to avoid confusion, I hung a sign next to it saying, "This is not a clock. This device predicts the exact time that NOON will occur." It was either going to be that, or "This is not a VCR"

On Tue, 15 Dec 2009, Bob Brown wrote: "oddly swapped"?? You are RIGHT!! Every even numbered character has been converted to the next ODD number! Now, make that chart. They should be plugged into sockets. VERY caarefully pry them out, without bending any of the pins.

On Dec 16, 2009, at 9:10 AM, Bob Brown wrote: Excellent. Good job! There's always eBay.. Most any CP/M software. That's the whole point of CP/M. Can't help you there.. -Dave -- Dave McGuire Port Charlotte, FL

Hi I don't recall if Jameco had a minimum purchase. Dwight _________________________________________________________________ Hotmail: Trusted email with Microsoft?s powerful SPAM protection. http://clk.atdmt.com/GBL/go/177141664/direct/01/

On Wed, Dec 16, 2009 at 3:21 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote: Absolutely! I second that. I've replaced many 2114s over the years. One commercial board I worked on in the past (the COMBOARD-I), had a field of 32 chips. Memory failures were one of the most common reasons boards ended up back on the bench. I learned to cleanly unsolder parts from 4-layer boards because of them (clip the leads at the IC body, then extract the pins one-by-one). I got lots of practice. -ethan

What is it with the 2114? I've replaced them in devices from many manufacturers, using ICs from many manufacturers, and of different ages. Some plastic DIP, some ceramic. Why is this chip so darn unreliable? I could accept that one batch from one manufacturer could have problems, but there seems to be nothing in common between the ones I've had to change -tony

On Dec 17, 2009, at 10:54 PM, Geoffrey Reed wrote: No, the 2114 is a 1Kx4 SRAM, while the 4116 is a 16Kx1 DRAM. -Dave -- Dave McGuire Port Charlotte, FL

On 17 Dec 2009 at 19:54, Geoffrey Reed wrote: No, the 8K chips were the Intel 2109--half of a 2117. They had a two- piece part number; the second half was S6xxx. If the xxx was even you set address bit 14 high, otherwise you set it low. I probably still have a bunch of these--they were castoffs from work when we got some "real" 16K DRAMs, I got to keep the '09s. Curiously most of the 09s worked just fine in full 16K mode and passed every diagnostic I could throw at them. So calling them "bad" was perhaps a little misleading. It could well have been that half the chip didn't meet specs. --Chuck

On 17 Dec 2009 at 21:07, Fred Cisin wrote: Let's put this in its historical context. When Intel started making 16K 16-pin DRAM, yields weren't terrific, so they sold a limited number of "half-good" parts--2109s early on, probably to mark their territory. About 1978-79, the US got in a trade spat with Japan over DRAM "dumping" and the supply of DRAM abruptly dried up (pesaging other similar periods yet to come). You couldn't find 16K DRAM for any price in quantity. There was at least one break-in for the sole purpose of stealing DRAM that I'm aware of, nearly putting the firm out of business.. (Later, there was a hijacking of a truck carrying mostly CPU chips, but that's another story). The Mercury-News archives doubtless has the story somewhere. So, given the situation, sifting through 8K DRAM to find decent 16K parts was quite reasonable. I recall that our 2109s were tendered somewhat apologetically by our Intel sales rep when we asked for 2117 DRAM engineering samples. These were the purple ceramic carriers with gold lid version. Plastic was later. When the antidumping order was lifted, we received a bunch of NEC uPD416s. We were surprised at how much better they were than the Intel parts (I still have those too). --Chuck

On Fri, Dec 18, 2009 at 12:57 PM, dwight elvey <dkelvey at hotmail.com> wrote: Intel did this sort of thing with pentium 2s (or was it celerons?) at one point. iirc, the stepping number was SL2W8. It turned out that the "slower" pentium 2s were just relabeled copies of the faster one. It made the overclockers specifically ask for that stepping number. I guess their yields were too good for the marketing department. brian

Been a long time since I've seen either in person :) at least it stimulated some conversation :) On 12/18/09 11:30 AM, "Tony Duell" <ard at p850ug1.demon.co.uk> wrote:

On 12/19/09 11:18 AM, "Tony Duell" <ard at p850ug1.demon.co.uk> wrote: True, but most of my classic systems were in a storage unit that got emptied :(

UGH! :oO My bad, sorry :oO (Repeating to my friends: "I must engage BRAIN before engaging HANDS" :oP)

Last time I got a model 2 or 16 that was "broken" I pulled all the boards, one at a time, put them on an antistatic mat, and pulled and reinserted all socketed chips, I also cleaned the corrosion off the card edge fingers.

On Tue, Dec 15, 2009 at 5:51 PM, Fred Cisin <cisin at xenosoft.com> wrote: Until they sell one. Then 385 drives will be available. -ethan