Hi, it's repair time. The Classic: IBM PC Junior Keyboard: Now this comes from my older IBM PC Jr. teardown video and where I got that partially up and running. I'll link in both of those videos main channel in the second channel video.

If you haven't seen it, it's a rather interesting and in-depth teardown of probably the most successful failure in IBM's history. The IBM PC Junior Anyway, um, this keyboard didn't work it. Actually, it's an infrared wireless sight keyboard. I've already.

Yes, I've already taken it apart a bit infrared wireless keyboard, but it's also got a wired connection as well. Now the wired connection was causing the computer not to boot up. So yeah, it's a problem. and I've tried to put the batteries in there I've tried it.

the computer works, but the wireless keyboard doesn't either wired or wireless. so I thought we'd take it apart, have a look. Of course all the yellow eggs happened and from people using it all you know the smooth edge down here and all the rest of it. So let's take a look inside.

here. it's got pot. I Mentioned this in my previous video. Probably the worst design battery holder I've ever seen is with the rails in there and you can slide the batteries under.

It's just really, really bad. Anyway, let's take this thing apart and let's have a squiz inside. And today that's why people hated it is because it's a rubber membrane keyboard. It's not a proper IBM keyboard look at that.

but I think it feels ok. There's nothing hugely wrong with it I won't take all that out anyway. I Think it's okay. What we're looking at is the mainboard.

There's something wrong with this Now the first thing. you may not notice it. but I Can I can just bend that board like that? That is not a 1.6 millimeter. PC Because you wouldn't be able to bend to 1.6 millimeter PC be that big like that? That's point eight millimeter piece of why they've used a point Eight millimeter.

PCB Why? Why to give it a little bit of give when you and push down on a key? I didn't Unbelievable. Not a huge amount of stuff on here, just of my 4000 series see much you want it to be at low power. Of course the main processor 8 OC 48 which was a common mask programmable number flash or a squared prom. Rubbish.

It was a one-time programmable and very common in the IBM Kiba. In fact, keyboards today still have the 8 O 48 OC 48 don't they I think so? Anyway, I'm just some, you know for Oh 66 Si Oh, I've got a 7 for Hco3 in there and there's not much else. We got the L infrared LEDs there and yes, I've actually checked it because you can actually hold up any infrared remote control up to a video camera like I'm using at the moment and it'll be out of the sensor in the video camera. We'll be able to see the lids flash and they don't flash.

So Houston, we have a problem. So first thing first. like visuals. of course it looks ok.

the caps aren't leaked in. there's nothing else in there really. So I'll pair it up for the lab power supply. We'll set that a 6 volts hundred milliamp current limit because well, you know it shouldn't draw more than a hundred and milliamps.

I'm pretty sure I've got that around the right way. You can see the negative pointed towards there so that looks correct. It looks like we've got some diode protection there, so let's let's power that up and see what's what that's good. It's only drawing 4 milliamps.

That's what you'd expect. Kay: I'm shorting some of those pads out and I don't see the current changing I would have expected the current to go up because it would have been driving the lids. So yeah, and of course first thing first, thou shall measure voltages so ground and power on our micro five point, eight volts. Well, that's alright.

I'm this because it's a CMOS R version. It's an it can go up to six volts so that's alright and check to see that we've got a clock. Yep, Six megahertz? No worries. Alright, so what I did is look for a schematic for this thing.

Unfortunately, I can't find it. They do have an awesome technical reference manual for this and they also have a hardware repair service manual as well. Unfortunately, that doesn't actually contain the schematics. Um, but I Did found find something useful in the manual that said, when you plug in the keyboard, it actually disables the infrared transmitter.

So obviously our process is running We because this is not an oscillator so it's just a crystal. so the oscillator needs to be running in the processor so it's running and it's doing something. So when you plug it in, there's a signal apparently that disables the infrared, which would explain it. So I'm just gonna give this a nice visual.

I'm not gonna do it on the screen here I Go look at it under the microscope just to see if there's any funny business. see if there's any. there are any pins short it out inside there. I'm just probing the pins here.

Found a 400 kilohertz signal there. so obviously the chip is doing something so so it's not like it's dead. I Can get some pins to vary when I touch some of the pads, but apart from that getting zippity-doo-dah if I short it out, we can get it's all the switchy bounce. No matter what I do I can't get it to actually trigger the infrared LED output.

so I don't know what the deal is and I'm getting no signal on the keyboard either like any of the keyboard lines look, that was just a static impulse and we get this little wiggly signal here on the panel rail. but if you see where one microsecond per division, count the rough number of cycles in there and there's six six Megahertz. We've got a six megahertz crystal that's just the AB like lack of bypassing on the rail, so nothing doing there. Okay, apart from the 400 kilohertz on one of these pins.

there is absolutely no activity on any other pin on that microwave. Checked every pin when I'm like activating a button like this and like there's no scanning, there's no nothing. So I've just got the clock, six megahertz clock and a 400 kilohertz signal on one of those pins and that's it. I'm gonna, how did I trace that 400 kilohertz? see where it goes And of course it went to the very last chip that scan because you go around and you scan all the pins like this with power off of course and I finally get across here and bingo got it! Last one bloody.

Murphy And here's where your studio lights come in handy. You can see right through the board cuz some of these like I can't see I couldn't actually trace that 400 kilohertz our signal out of there. so I went up here somewhere, couldn't see it and I can't see where this pin goes cuz it's on the top side. A buggers off under here somewhere.

Well that's interesting I Hooked the board back up having a polka Doak and it's now drawing. Nothing sails George For millions Before now it's drawing bugger all. Wow! Wow. this is interesting.

Some sort of intermittent fault. we got any five volts on our chippy, but we're getting nothing I Now oscillator anymore. All that 400 kilohertz signal? course. I'm not gonna get the 400 if theif my own oscillators not working, the oscillators jet dead, the micro is getting 5 volts.

Interesting. I Just replaced the crystal with another six megahertz one and we're back up to four milliamps. and it's we're back to our oscillator. So, a dodgy intermittent crystal, but that wouldn't explain why the keyboards not working though.

Okay, so we're back to it. Let's follow the money. This is this 400 kilohertz signal and we don't get any variation in that when we do the keys. So if we've followed money, it actually goes down to including.

the chip up here goes down to this resistor pack. Here sits a 104 and on the other side nothing. Hmm, let's trace where the other side goes to. I Think it goes to this simple Hco3 here, but why? you get a signal on one side of a 10k resistor and not on the other? Maybe you need to pull up? Uh-huh That's I Thought this was a through resistor like that like individual.

once it's not, 104 is a hundred K And sure enough, if we measure I've powered it off, measure across set 200k 200k like that. Which means one of them's comments probably this one up here. Yep, 100k 100k you? 100 K A comment. Okay, so that's just a pull-up so it's not going through.

They resisted to there. That's just to pull up on that 400 kilohertz line. That's okay. Okay, there was nothing wrong with that crystal cuz I Turn this thing off and on and I'm still it's gone back.

like it's failed again. So it's gone back. So what I'm gonna do is just I'm gonna look at my current sleigh and just give that a bendy. Bend Like that, not my current just stays on zero.

So we've got an intermittent startup. Jeez! I Hate these things when you get intermittent problems like this. just makes troubleshooting a real pain. Okay, so next thing I'm going to do is trace the reset line of the micro which is pin 4 and see if it's a getting a reset I got lucky this time I was just dragging along there and it's connected to pin 7.

there of this for 11 and then a trace that over goes over to this V R which goes up to pin 1 of this 4 o 13 up here again. Jeez, it's going everywhere man and it goes off to another pin this 401 3. So 401 3 is pretty important then you might think oh okay, this is some sort of RC thing happening, but there's nothing like you. You scan over all the passives.

Good thing about through-hole like this and just scan across and like there's nothing. So there's no like RC like reset thing. at least not directly right. So at this point it starts to get a little bit ugly.

Without the UH schematic. you know there's so much there as if like you're starting to reverse-engineer this. Okay, the the micro is not starting up so you look at the reset pin. where's the reset pin going? Oh, it's going to over to a couple of chips and pins over here of this gate and so without a schematic you go well.

what do I do you know like suck out the 401 3 and test that. get out an old-school IC tester. Geez, that'll be fun. and when you pair it on, the reset line is permanently low like that.

so there's something forcing that reset line note load and they're destroying. Now for because it's in static condition, they're all CMOS and maybe I should have traced this, but just for kicks I sucked it out and all the pins are bent over so you got it with your soda. Suck a gun. You've got a sort of like bend them up and then you've got to get in there and you got to just give them a little bit of wiggle wiggle wiggle yeah and should pop out.

Nice. No pads lifted. No nothing. Beauty.

Yeah, we can get an old school. Well it's not an old school one because I actually designed an IC tester way back we're talking 25 years ago or something. It doesn't work anymore, but I have one that does. Let's go.

It's the upgraded one to that I originally had it's the Tl8 6 6, 2 + @ z GE ECU Calm. Anyway, it's got an IC tester built in so check this out. It's got logic ICS and devices 401 3 which is exactly what we have. Place it in the socket I have let's press, test and test normal test.

No more tests on our test. Normal wat wat wat? wha? Well ruled that out like I don't know how comprehensively it tests it, but it knows that there's X gates in there and it I assume it exercises all options for all the stuff. So yeah, okay, tell you what. just for kicks I Decided to start this thing back up without the chip in there and yes, it oscillates.

Now we're getting that 400 kilohertz that we'll get in before and I thought and I just probed the lead over here and then just do the shorty short thing. we're actually getting Trent packet transmission. There we go. Look, Wait, we're now sending data we're sending Darla's without the chip Wow Look at that.

something's going on there. It's a signal level. it's just that's the voltage drop of the lead there. the infrared LED from the positive rail Wow I might actually see if the keyboard actually talks to the PC Now I can't remove the chip.

It's a hilarious I like I thought I'd just start up I just just for kicks I thought I'd do that. This is hilarious. Okay I'm just going hockey like this and no, it doesn't seem to be doing anything. so you know, like this, the 401 three could have something to do with the encoding of course.

Alright, don't try this at home kitties. But what I'm gonna do is put my meter in the micro amp range and I'm actually gonna short out the reset pin the chips back in there now solder back in. It doesn't work. We're back to our original wire configuration.

It's drawing 4 milliamps and and ya know, jeez it just it started drawing zero again. Ah, give me a break. Alright, it's working again. I Just kept pairing up and mucking around with it until I drew four milliamps.

Instead, there's our let's drop the zero again. I Just bumped it. What's going on? Okay, I might have sussed it if I hold a key down, not like short one out. when I'm pairing it up, it seems to come good anyway.

I'm not gonna worry about that now. What? I want to do? Okay, so I'm gonna probe my lead here and don't try this at home kitties. but I'm gonna put my meter on my current mode. so just a micro amps here and I'm gonna short out that reset line cuz that reset line pin four here is currently low so it's keeping the processor in reset.

and if you read the datasheet, the oscillator does actually start up in that mode. so the oscillator is are actually running. but it's obviously, but it can't execute its program obviously. Okay, so what I'm gonna do is hold that pin.

the reset pin. I'm gonna actually force it to the positive rail here. it's drawn an extra 6 milliamps on my power supply so you can't see that. But so we're forcing it to operate and look, you saw that we're now getting data on our scope.

So now I forced it out of reset mode and it seems to be doing the business working sending the packets. So obviously this thing is being held in reset mode. So we've got to figure out why it has to do with the 401 3. It's flippity-flop time by the way.

I've never heard of synced semiconductor. ease wise, component ease like what? Anyway, I've got no ideas that like Spanish or something I'm sorry I got no idea. Anyway, never heard of them. it just first one that pulled up when I search for PDF Anyway, all I know is that the Q1 output of the flippity-flop here that goes to our not reset line of our processor.

That's what's staying low so it needs to go high in order for this thing to work. and that's actually tied to pin 9 which is the second data pin of this flippity-flop here. I Haven't traced anything else. This is where the schematic would be.

Yeah, really quite handy. Anyway, we're gonna check out see that our inputs here I'll just try and that single-shot capture some of these like a clock. for example, when we pair it on. So this is our clock signal.

this is when it was powering down like this. so powers on we. There's no extra clock in there, but uh, we don't know. We haven't time correlated that against the power on.

So yeah, we need to do that. This is where you need a two channel scope. Lucky you got one. So I traced out pin3 the clock line there.

it goes to Q 4 of the 4060 which is down here. and that's a good old binary ripple counter here. And yeah, so that's a Q4 output is. So obviously this is maybe some power on delay or something for the reset that they're doing.

So I'm gonna now focus on the 40 60. I'm not going to bother checking the rest of this anymore if we're not getting a clock in there. Which is the well. there's two ways to change your output.

here. your reset. you can use your set and your reset pin. But what's the point? Anyway, so it.

but the fact that the clock goes off to this ripple counter tells me that this ripple count is important. Probably some pair on timer or something else that is not working. So mmm-hmm Just a quick check of the cap down in there. It's gonna be that smaller one, not the yellow one.

That'll be the bypass. 10 Nano Farad all right and I measured those two resistors in there as well. There are being honors class being on as they will be in circuit so that's not a problem. 10 Nano Farad sounds about right.

Yeah, Okay, so the 4060 is a binary ripple counter of course, doesn't have a know that Q 0, 2, Q 2. Rubbish. don't need that. and it basically you can hook up an external crystal or an RC 'm typically and in this case an RC oscillator on here about the master reset pin will disable that.

I've checked pin 9 for any oscillation. there is no oscillation. I've checked pin 12. What do you know? it's high, so it's an active high input so this thing ain't working.

No wonder that no oscillation. That means no clock on Q 5 Which means no flippy-floppy action over here to change your reset to toggle your reset pin. Aha, what's happening to the reset here? Okay, I've done a little bit more trace in here and this reset pin. it goes off to this NAND gate here.

the 401 1 that's a baby next to it there and this bug is often gifted the first time. This goes through a cap and a resistor over to the keyboard connector over here. So you remember how we had mentioned before how it had some sort of plug-in detection and disabled the the LEDs on here. Well, I think there might be some plug-in contact in there so you have a look inside there.

Focus your bastard. There's those little lever things in there. Maybe one of those is not making contact when it should. Hmm when it's not plugged in with a shot.

I Think we're getting somewhere because this buggers off to an internal switch. a plug-in switch inside the keyboard connector which goes up to +5 volts. So watch this. This is the connector on the side of the pin there.

So there you go. we're getting our 5 volts. Sorry, it's hard to see that, but that's basically on the output side of the that are see there. If I get in there and I disconnect that like I'm plugging in the connector.

Boom! It goes to zero and then it goes to five. It's got a little lever switch in there that detects the plug-in but that seems to be working. Actually, all right, something's going on here. If I look at pin 1, Let's measure the voltage on Pin 1 here.

2 point 2 volts. What won't won't Wha? That's smack in the undefined region of the CMOS gate. No wonder the outputs going to be doing silly buggers well. The Apple toy which is forcing our reset and what's pin 2 doing? Pin 2 is high as well, so it should be low so they're both.

Yeah, something's wrong there. There you go. Well, that's supposed to be a 2.2 mic. That's just a little while those bugger 10 lumps I'm Bennett measures.

Okay, yeah, confirmed it's okay. something else and hold on to your hats because I've left that capacitor out Ivory powered it up to have a play around with it. and you have to see my power supply for this. I haven't used the scope to look at anything, but it's drawing no current at the moment.

But if I short out a key, bingo, it draws Neff all like microamps. That's probably until you press a key. Maybe that's what you'd expect. So all along for millions, we could have been coming to Gutter.

Anyway I think she's working now. Okay, so what I care about now? let's just turn it on and Bob's your uncle Now we're transmitting. So yeah, I could go into the intricacies of how or why I mean this cut seems to be okay I don't think it's leakage is a an issue. but anyway, oh well, it could be.

Oh no no, it's sort of. light. stays on five well and then it drops back down. but maybe that's it.

could be some powers down, stayed or something. There you go. It's actually drawing 65 micro amps there, which is that's kind of what you'd expect for a keyboard. isn't it before? I should have tweaked to that.

People probably scream it at me here. 4 milliamps. that's a lot. no double A's but still and I would have been sucking the juice and we'll put it in that max mode and 50 milli amp range there.

So it's peaking at about 8 odd milliamps there to drive the leads we know. Winner chicken dinner with no cap. It may not actually be the cap, maybe something else, but the cap is making it work. So via that silly plug in a keyboard switch thing, hang on.

The story is not over yet. Take a take a look at this. This is our reset pin is still low, but let me touch a key. it wakes up.

go down. Could I get it to stay on? Let me get it to latch on sometimes I can get it there we go. it's latched on and 5.7 volts so it's high and then what goes and it goes low. Ah brown, black, green.

you'll know what that is. High value job'. And look, those two pins are shorted. This is where I've removed the cap from.

There's nothing. Look over this side here. they're not connected to anywhere yet. If I measure that resistor Wow What? 170 Hey What the well? I found it remarkable and I won mega carbon film resistors can fail and sure enough it hasn't have I missed a trace on the PCB Oh yeah, look at that.

You can easily come a gut. so I couldn't see it under the resistor right angle trace. Look at that. it was hidden by the resistor mongrel.

Okay I just sold it in a ceramic job' over here. it's not 2.2 but it's near enough. Good enough for Australia doesn't have to be that and we're getting our thing again. This seems to be like yeah I think if you do the keys quick, it's sort of like latches on like that I think there's something wrong with that cap in that measures okay.

could be some our voltage our dependency issue on there. and yep, there we have it at. 6 volts. I've got 150 micro amps leakage so combine that with our one mega resistor and you're gonna come a guts ER And that's why we'll get in that mid rail voltage that we'll see in so leak eternal them doesn't have to be much, but when you combine it with that, there's just the design of having that one Meg resistor down in there.

it's coming got sir. There you go. Now it's jumped up to 190. It's all over the shop actually if I change the voltage so a point where it drops to Knuffle not leaky.

bloody tantalum. Unbelievable. Well, it's not. They're notorious for it.

So there you have it. We know we're not chicken dinner. It was a little too an alumni wear those tag tents. one of those axial tantalum zat that a leaky tantalum.

and when you look back on it, even without the complete schematic, a little bit of reverse engineering we did is that it was. Yeah, that would explain it. A leaky tan in there was causing a non valid art logic level in here. This was probably oscillating doing all sorts of funny things and that went to the reset which and the buggered everything up.

So that explains why I was seen those like it sometimes to be drawing for milliamp. so it's probably due to the oscillation or something like that doing something weird and then when it went into zero, maybe when it actually went to zero. it would have worked if I happen to or tested at that exact point. but it's like it's just all over the shop.

So that explained all the intermittent operation and how we chased a red herring down a rabbit hole there a bit you know it seems this is what's this video. like 30 minutes or something. It seemed like a long time doing this and it might be. There's some things in here that I wouldn't have done.

I was just doing it for the sake of the video like sucking out the chip and just wanted to use the IC tester and show you that you know like it's not something I would have ordinarily done and you could have said oh, if you reverse engineered it from the get-go would have been a little bit quicker maybe. but all up like if I wasn't shooting a video for this that it's not and not a really long repair. Yeah, it was a pain in the butt to get to that point. had to go through most of those processes I went through unless you've got, you know, I'm lucky in some other way to track that down to the axial tent on there and some people might have gone I Yeah, I would replace the axial tent from the yet go.

Okay, well you know we were methodically going through tests and everything, but that was an absolutely classic I hope you enjoyed that as much as I did I Love it when I get faults like this that takes some time to track down and they come down to. you know that cap if you test it on your meter or your LCR meter. It measures fine, but it's actually leaky and that might not ordinarily be a problem. This cap could still work in many, many other circuits, but when you combine it with the one Meg resistor in there, you're gonna come a gutter.

So oh look, yeah, it's just it's all over the shop that's hopeless. Anyway, that was fascinating as I hope you really enjoy that. Love it When I get repairs like that, it's not very often if you did enjoy it. Please give it a big thumbs up and as always, discuss down below.

catch you next time. Oh and by the way, this repair came about because I was actually trying another thing which involves this monitor over here so I was in the middle of shooting that video I was most of the way through it and I wanted to get this PC Junior working I thought I'll get that I know I want to get the keyboard working and I so how'd it go at that and you one thing led to another. So this video come our first. Catch you next time.