EEVblog 1376 - Tandy 102 Vintage Computer REPAIR

hi check it out this is the tandy 102 portable computer one of the world's first and most popular uh portable computers of all time this is the upgrade uh to the classic tandy 100 from 1983. this is the 102 data about this model came out in 1986 and it's identical to the model 100 it's just a little bit thinner and lighter and anyway this was uh basically the most popular notebook computer in the 1980s had a modem building every single like uh you know reporter journalist had one of these things because they could type up their stories connect it to their acoustic coupler modem and then dial back in and send their story back to the news desk and stuff like that let alone countless other uh uses but i've done a teardown video of the original 100 but i also have a tandy 102 we can see some classic uh yellowing of the uh bromine in the plastics here it's a fire retardant it causes like them to like go yellow over time and that's fairly typical but anyway i did a tear down of the 100 this is the 102 and it's just it used to work but it's now got a rather unusual fault so let's take a look at it and see if we can fix it oh and by the way the tandy 100 is famous for being one of the last computers actually containing significant code written by bill gates himself so he actually worked on the rom for this thing so the original uh tandy 100 anyway let's switch it on and uh you can see it appears to actually work copyright microsoft text telecom address that had like a spreadsheet a word processor built in and basic building and everything but you can see up the top here um there's something weird going on it's like there's you know look you can see the clock up there it's here 54 55 56 but it's all sort of weedy something's going on this doesn't look correct and well let's go into the basic here and let's have a look uh yeah model 100 software once again this top corner up here something weird is going on we've got two flashing cursors this should say you know something bites free right um i don't know how i can't remember how much memory is in this one but yeah something has gone wrong and microsoft there what's all this going on something weird right let's clear this okay so everything looks to be working just fine and dandy but let's have a look what happens here okay what i'll do is i'll just uh like one two three four so everything seems to be working five six seven but watch this we go to eight oops once we went past eight the cursor has wrapped back here and it's starting to overwrite so nine zero one two three four five six seven eight and it just keeps overriding like that and it's no coincidence that it's doing it after the eighth character so because you know there'll be an eight-bit bass that's shifting whatever it might store like you know one column or something as a byte or you know whatever it's no coincidence that the number is eight and uh if we go down here like this you can see once we get actually down to the four uh the fifth line down here the cursor wraps up here and it duplicates anything on the fifth line and then it duplicates anything on the fifth line on the first line and then once again continues to wrap and if we go down another line it'll duplicate it so the uh the eighth line down here will duplicate it on the fourth line isn't that fascinating so there's obviously something to do with the how the dart is getting into the display drive the display drive is obviously working because the correct characters are showing up it's like it's basically garbage in garbage out pretty much so the actual uh column and line drivers and everything else of the lcd seem to be just fine so we can pretty much uh rule out that as an issue it's it's the data being fed into them and this lcd is going to have multiple uh driver chips for it uh probably like there's probably going to be like one chip controlling like the first four lines here and the first eight characters and they've probably got another chip controlling that one so it's probably that chip up there that's causing some sort of issue anyway luckily we do have the schematics schematics i need schematics for those playing along at home this is 1986 uh 10th month i presume um it's like i presume that's a lowish serial number if anyone knows please let us know product of japan all the best stuff's made in japan oh and if you don't know it has a memory protection uh switch because this didn't have any of that non-volatile rubbish it was volatile sram it kept your uh programming when you switched it off uh so good old sram and um yeah you could protect that or not with your batteries and oh there we go you can plug in expansion ram there so that looks like it has an additional 62256 in there oh okay other stuff if you haven't seen it rs-232 serial a system bus connector a printer connector their phone that was your modem because it does actually have and you know a proper isolation uh transformer in there and your cassette tape uh storage it had a light pen uh you know good for like inventory control for warehouses and stuff like that uh that's the direction answer originate uh for those who are used to your um modem sex you'll know all about that all right i think i've got this mostly off that should lift off we're in like flynn there we go and the keyboard on this feels beautiful by the way it's just ah it's fantastic so there's our lcd up there whether or not we can get oh yeah yeah we can flippity-doo-dah you betcha we can flippity-do that ah that's just beautiful so we can still use this let me switch it on and we should find that yep yep and picked up where we left off oh look oh that's it i hadn't tried that it didn't actually look it didn't actually refresh the contents up there that's interesting ah like it kept everything else but it didn't refresh that but when it's getting new data like it is now i mean if we go down a lot there we go no it's updated but yeah it cleared it that's fascinating all right i know you want to see the rest of it um i can't remember how this differs from the well yet no this is i do believe this is very different from the 100 yep um this is this is the revised model thinner and uh lighter weight double-sided load look at that wow surface mount the other one was all through-hole wasn't it anyway we've got some bodge wires and some bodge caps um the the bodge wires look like big ground stuff it's an inductor or a diode not sure you can see the red glue underneath the chips there that's to hold them on ah yes i remember this from the previous teardown this is absolutely fantastic this is some pcb routing perfection here you'll notice that um there's no markings on this chip that's because it's mounted upside down it's mounted in a cutout in the pcb why if they put one row of them like this and the other like this this is for pcb routing reasons just the way that the pin outs worked if you had these chips up the other way like these ones um it'd all be higgledy piggery and you need extra layers on your board to get around at all it'd be an absolute mess so from a pcb layout elegance point of view they've mounted these chips backwards with a out in the board it's just it it's simply brilliant it really is hats off luckily we do have the service menu complete with all the schematics the theory of operation the whole works they don't make them like this anymore unfortunately tandy 102 custom manufacturer for radio shack and division of tandy corporation look at this look at this ah just a wet dream maintenance disassembly instructions theory of operation troubleshooting oh beautiful i doubt though i actually haven't looked i doubt they'll have a troubleshooting procedure for like this sort of fault in the lcd but yeah you never know look at this so comprehensive look look fantastic bobby dazzler anyway there's the back of our lcd board and look at this ta-da here's the lcd driver board look at this we have two rows here because this is a 40 column by eight line lcd and obviously each row of these chips here will be handling four uh lines and we've got five for the columns so 40 divided by five eight tada is it just simply this chip well if i'm flipping it the right whichever way it's you know actually flipped yeah it could be m1 m1 up there that could be it so i yeah right off the bat you would suspect uh this driver here it's a hitachi because they did all the they still dominate lcd drivers don't they i don't know the standard hitachi chipset anyway the hd 44 102 for those playing along at home so automatically you would either suspect this chip has failed or it's getting bad drive signals coming into it there could be a chip select there could be an address line you know whatever could be some data corruption but you know it's obviously clocking no it's obviously getting the data in there correct so it's got to be some sort of you know like some sort of chip select thing but it may not be that chip it may be off on the main processor board which actually drives it this specs for those playing along at home look at this four double a batteries last like 20 hours you can use this bad boy for unbelievable anyway fantastic ada c85 processor so we've got theory of operation here lcd lcd common drive lcd segment driver lcd waveform uh the block diagram yep okay so an 81c55 drives into the lcd controller and then the lcd but it also like it comes via the cpu bus as well so it doesn't all come via the pio so the pio is probably only like uh driving the chip select lines and things like that they don't make theory of operations anymore do they ah we'll get to the there's the modem interface i do want to do a video trying to use the modem on this thing i think that'd be really cool there's the modem connector interface circuit um i've done some experiments before in the past uh with i can't remember this 102 or the original 100 but uh yeah anyway lcd here we are the lcd using attendee 102 is composed of electrodes in a matrix arrangement 64 common signals 480 segment signals this part is subdivided in three segments the common driver the segment driver and the waveform the common driver here it is that's the hd44103 uses two common driver ics okay m11 and m12 where was that thing we had before so yeah there's your common driver there and down there i don't think it's going to be a common driver because well it's common it's more likely to be the segment uh drivers now i don't think we're going to have to go into like the internal logic of the 44103 and the other uh common driver i don't you know it's just it's just getting i think it's just getting the wrong stuff clocked into it i think that is the problem it could be via any number of reasons there could be a dry joint somewhere one of the driver chips could be failed or something could be wrong with something could else could be loading down the bus which is uh going to be common who knows but uh yeah i don't think we're going to have to get into that sort of detail that's sort of like the last rabbit hole we want to jump down show us the schematic dave okay after the bill of materials here it is no isn't this just brilliant absolutely brilliant fantastic takes a second just to render a new sharper version but anyway all the goodness it's all down here right so what we want here's our lcd connector here okay okay we've got some cmos logic doing some stuff here not worried about that yet let's have a look at the address and data lines here right so let's follow this here id so this is the uh this is the data bus it's not actually the address bus these chip selects are the address so if we follow the data here okay this comes over here it's going yep it's coming it's all the work it's coming from here which uh let's not go there but let's follow the money up here up here up here here you go there you go we've got a 245 there and that's hooked into the 8085 processor databus right so i don't expect there really like to be any issue with the processor addresses and data buses because everything hangs off that right so it doesn't everything else works so i don't expect that to be an issue so that's kind of like the last thing that you would look at here so i'm not worried about the data bus there all the data seems to be getting in it's just getting latched wrong so there's uh looks like there's all these well the thing is it says chip select one two three four here these ones have no label so we'll go check the lcd schematic in a sec but these come from pa00 pa1 so these come from yeah and 81c55 up here okay so that's being driven from there so that's an io coming from the processor so it's right into this and then four of those lines are going into driving the lcd and then other stuff you know look the the read write pin so it looks like it does it can potentially read back i don't know where s1 is coming from or si but i wouldn't expect there like to be any issue with the reset because the lcd is just working just fine and i don't expect it you know there's no problem with any of the voltage drivers or any of that sort of stuff you know so all the enable here so you know you wouldn't bother looking at that uh i don't know what this second what this chip select here does but you know we're looking at one specific segment of the lcd so there's nothing to see here move along now to the lcd schematic they've got a nice looking lcd and it's like you know it's very it's laid out very well look you know like m1 up here is you know they've done it overlaid it into the specific thing they've got uh information with driving the commons and the segments and the commons and stuff like this absolutely brilliant so anyway so this is looking face on like we saw and as you saw we were having problems in this quadrant here the first four lines and the first eight characters well as we said that's going to be controlled by m1 here so what's coming into m1 let's have a look tilt your head with me and once again i wouldn't expect any of like the voltage things to be a problem because it's you know the segments aren't faded they're not doing anything like that and they're all just you know common from with all the other chips it's you know it's not a problem so it's you know data going into this sucker so it could be you know i'd be looking at the first thing i'd be probing if i got the scope out would be i'm probably going to have to unless you just want to go well it's probably that one and you know if you had like an lcd swap or something like that to swap over you might swap it over and you know that it's nothing on the driving side of things it's on the lcd module and then you might go well okay it's got to be this chip here and then you know but even if you do that you've still got to like probe it to see like it could be a dry joint it could be a bad shoulder joint and one of these uh control pins or something like that once again i don't think it's the data because the data actually gets in there you saw it as we were able to write into these particular segments it's just that yeah the interesting thing is when we were driving this one here in this bottom corner of the lcd like this so if we were driving like line five here line one would like duplicate this stuff on uh line five here so when we're writing data to this it's also appearing on this and you might think okay you know there could be like some short between maybe there's some i don't know there's there's some tin whiskers or something happening and there's a little short between you know just throwing that one out there uh you keep that in the back of your mind but this is where you've got to like go back to here and look nine i all these other lines all these chip selects are being driven so i don't know why the labels were missing they're certainly going up here it's a bit hard to read but you can see that's a cs28 there for example so that's cs2 one oh they're actually out of order six seven one two three like so that chip for example is there and that makes sense because we need ten of them okay because we've got ten uh pins here ten chip selects and we've got ten chips total so it's almost as if like the chip select for this one is short into the chip select for this one yeah so we've got cs21 here and that'll be this will be cs26 yes so okay it matches the m1 m2 so yeah we're looking at that line there and that one there that's what's happening uh so i you know like it's not like those two lines i can't see how they're going to be shorted or whatever but it could be something as simple as like an open line some they're wearing and then it's getting crosstalk uh from the other lines so yeah i look you know you can analyze this until their cows come home you have a bunch of hypotheses and you just go in and check them all out so let's open this sucker up and i'd just start with just probing around here for example and just seeing hopefully we can pro get access to the chips haven't taken apart yet get access to the chips and probe them while we're actually using the thing otherwise uh you know you gotta like solder in mod wires to get test signals out and things like that you know but first of all open up give it a bit of a visual inspection and then yeah we'll get probing anyway there's our suspect up there m1 so what i'm going to do is i'm going to have a get that under the microscope and have a good look at it do the visuals looking for any you know dodgy shoulder joints any hairline cracks any uh bridges uh caused by uh tin whiskers uh for example which you know a growth over time they can grow out and short out pins and stuff like that um yeah just generally have a squeeze and you can just really see it quite uh clearly here how it just duplicates and go to the next line look and shift that up check this out i wonder why they've exposed the solder mask along there just a little slither there there there it's interesting they've done it on every chip like this one at the top here it's gone on all four sides look at that you'll also notice on the board here how they've got all these traces running off the edge of the board like this up here they've got up here like this they're just buggering off over here one of the column drivers over here just goes off the uh even head down these signals down here why are they doing that well that's probably for some sort of uh production panel testing would be uh my guess and then they um shear the things off that route them off uh later all right let's have a look m1 that all looks pretty smick i've seen any tin whiskers or shorts or anything might want to go over this see if there's any bad solder joints like as in physically not soldered see if i can move any of these pins they're all looking pretty good nothing's loosey-goosey there really is no issue there whatsoever i mean it could certainly be the chip like i would not rule out the chip hang on for a second i was just looking for a place to probe and i was looking at the pins i'm interested in down here and i see a little dag hello mr dag i'm gonna i'm actually gonna measure between these two pins and see what's what can't be that easy surely those two pin are not 925 ohms my meter's beeping at 125 ohms i'm going to get rid of that dag now it's still 125 ohms i think that's a nothing burger but geez that was that was coincidental okay it's probing time what i did is i just soldered a little uh 0.1 inch header pin onto the ground pin there which is uh pin 5 and pin 6 is ground so that just allows me there was no other like convenient ground point unfortunately and i'll include the schematic here so what we're looking at we want to probe the chip select pins um that's like definitely the first port of call so we're talking pins 16 15 14 cs21 pin 14. that's the culprit that we actually want right there so pin 12 10 8 so you know there's a whole bunch of uh they'll be on this side of the connector over here so it's working it's powered up so two four six eight wasn't one of them pin eight i think there we go we're getting bus activity five volts and yep yeah if i press the like just pressing the enter key uh tens 10's doing nothing okay oh no no sorry i just wasn't probing it properly you got to get through you might have some flux on the pins or whatever always good to have sharp pointed probes to get through if you ever that's a tip like ever have like if your signal's missing or something like that and you think this should be a signal there what's going on don't chase that red herring down a rabbit hole just check that you're actually piercing the uh like any uh flux or any other contaminant on the uh joint that you've got a nice sharp pin that's number twelve yeah it's all these chip selects they're all working now 14 this is the bad boy hello hello hello number 14. let me yeah i'll push really hard on that yep number 14. let's go to 16. yeah 16's there bingo winner winner chicken dinner we found it exactly as i uh suspected yep pin 14 there cs21 which drives our m1 chip which drives the quadrant up here which is the chip select for this chip up here so obviously um it's now the chip select on this one here it's just flapping around in a breeze it's just floating so it's picking up crosstalk from basically everywhere else so that's why it's like latching in the same data um you know it's because it's a common database among them so they're all getting the same data so there's no wonder that this shift this chip is shifting in the same data as you're typing in over here somewhere for example so it's yeah the chip select is just you know it's been enabled i think the re yeah the read write will be all the same they'll all be the same right so it's right in the data it's only the chip select that makes a difference between these chips and this sucker is floating so that means our fault is actually on this border it could actually be on the ribbon cable it could have like a broken ribbon cable or something like that um yeah it may not be an electrical fold at all it may be a mechanical job or we could have a bad contact on the ribbon cable the first thing you do is just reseat the ribbon cable on both ends okay those contacts look pretty shiny no wuckers there let's plug it back in nope same problem it's not that it looks like the drive side or the other side of that connector so unfortunately this ribbon cable goes through the board here so i have to take out the entire board well not unfortunately means we get to look at the other side there you have it there's the rest of the board yeah it's still all uh through-hole jobs mostly um although we do have some service mount i don't think the previous version had any surface mount apart from the lcd did it but i could be wrong anyway yeah there's all our modem stuff out there there's our isolation transformer for our modem and the 8085 made by oki in japan all the best stuff made in japan 81c55 is down here and there that chip contains bill gates himself copyright 1983 microsoft um so yeah that's the the rom is soldered in hmm let's measure our backup battery there sorry you can't see it 3.42 volts oh that's still pretty good and again there doesn't appear to be any corrosion on the contacts at all nothing doing there plug that back in and i've reseeded that and uh no same old same old so you know good thing about this is you can actually get this right out and work on the bench i mean look we can actually it's actually better to work on it now on the front of the board like this because we can just have our lcd it's just facing up fantastic oh that is probe in heaven beautiful beautiful good thing is is that we can actually probe the top of this connector because the even numbered pins actually come out over here so let's go pin 14 again 14 tada there it is same problem all right so it's not the cable at all it's not the the ribbon cable or the connectors or anything to do with the lcd it is definitely coming from drive on the circuitry although i haven't inspected the solder joint now hang on check this out pin two four 6 8 10 12 14. that trace going through there what is that gauge take there is no way that is a coincidence is that some sort of like gouge chunk taken out of there that's that's way too much of a coincidence that it's there i'm going to get that under the microscope there's no way that's a coincidence is that being reworked i don't know but look come on that's not a coincidence is it that looks black that looks like it's rotted away or something am i wrong that looks dodgy ass let me buzz it hang on i'll see if i can get yet yeah yeah there we go that side right but it won't go over to here yep there you go we don't win a chicken dinner found it it's a track wow wouldn't a bet on that a track on the main board that somehow like corroded oh away i don't know and these chips you tell me has that been i'm looking it's like the other chips on here they don't have the same flux residue on them i reckon this has probably had a repair done to it just look at some of the other chips out you know these don't have any of that because this is all um wave uh soldered that's why it's got the glue although that one still does have the glue under it look so i'm not sure what the deal is yeah it's actually corroded away let me clean that up not entirely sure how i i have to lift my microscope up here because um it's designed for like because the board's too high so and add extreme zooms like this but look at that yeah yeah there's some sort of corrosion that's got into that and that's eaten through that poor little track that was it i found it by pretty good deduction there that was one of our hypothesis that uh the chip select was flapping around in the breeze maybe it was just getting crosstalk from something else and that turned out to be the case so yeah all i gotta do is fix that should be as good as new let's just scrape away some of the solder mask on the pad there and this is post editing dave here yeah my capture froze i'm gonna like uh probably replace that with a um hardware capture solution soon like dedicated hardware like a uh black magic some atem probably anyway little jump link yes we have a winner winner winner chicken dinner here it is no worries basic january 1st 1900 uh yeah y2k bug um yeah it didn't end the world though trust me that was a thing those who remember the y2k bug oh jeez i can yeah i was the y2k engineer at uh was it telly's or so seller no it was probably tallies australia at the time and yeah i had to go around and certify that everything i'm talking every little piece ant project that contained a microcontroller or anything that had no real-time clock in it i would still have to write a report certifying that it was y2k compatible anyway basic ta-da trs-80 model 100 software copyright 99 because you want to save a few uh characters every character was precious in 1983 microsoft bill gates himself might have written that um 29 382 bytes free done fantastic winner winner chicken dinner hope you liked that repair video if you did give it a big thumbs up and as always you can discuss uh down below and catch me on odyssey and all the other platforms you know what to do catch you next time [Music] you

Channel: EEVblog

Views: 49,983

Rating: 4.9728417 out of 5

Keywords: eevblog, video, tandy 100, tandy 102, tandy, tandy portable computer, repair, computer repair, notebook repair

Id: uUXxY6gA-7g

Length: 29min 57sec (1797 seconds)

Published: Thu Feb 18 2021