Update: An eagle eyed viewer (thanks @SwitchingPower!) noticed something I missed: the component on the backside of the board underneath the CPU is toast. You can see that around 21:20, directly in the middle of the socket on the back. I'd noticed the dirt on it resembled the spatter from a blown semiconductor during disassembly, but I wiped it off and it seemed like it was just the same deposits I saw elsewhere in the thing, and I didn't see a cracked case, so I thought it was just dirty. Sure enough though, I desoldered it - it looks like the back of the case blew out, even stripped the top layer off the PCB. So yeah, this was never gonna work. If this video feels a bit... unenthusiastic? Like I didn't put in as much effort as it deserved? It's because I got about two hours into shooting it and realized that it felt like this thing SHOULD work, and that I was just not skilled enough to figure out what was wrong, so I lost a lot of drive and just tried to salvage something out of it. Turns out it wasn't my fault; wish I'd realized that at the time. Regarding the little six-pin device on the daughterboard plug: I had tried during shooting to get a good look at the label and couldn't make it out, but while looking at the board again today I suddenly found an angle where I could see the lettering. Turns out it says AAPN, which is a MAX6629 (or similar) temperature sensor! That makes a lot of sense! If you see the early version of this video: There's a quip in there about 240V being used in Canada, which is based on something I saw *years* ago about Canadian houses being wired with split-phase specifically so they could bring out 240 to ordinary wall outlets. True or not, it doesn't seem like it's very common, so I removed that line with the youtube editor and it'll take effect sometime in the next 24-48 hours. Yes I pronounced Ontario wrong. Sorry :( I quipped about FCPGA being basically "the" socket for CPUs with pins - that's true as far as most chips go, but what preceded that was plain PGA, without the FC; "flip-chip" referred to the die being upside down on top of the carrier instead of vice versa. Likewise, apparently ZIF came in with the first Pentium sockets - I suspected it was a bit older than I remembered, that's why I hedged on that point. It's actually the 486 I was thinking of, but those weren't any more pleasant to press-fit into a socket than the 370 chips; I remember breaking pins off of one or two trying to get them out, in fact.
Canadian here : sure, we have 240V and it COULD be connected to a regular outlet, I think I've seen that done in some specific case, especially with specific plugs in kitchens (like for microwaves and kettles) and more recently, electric cars, but way more commonly, it's only used with the oven and the dryer with their own plugs. Our electrical system is basically identical to the US, so really bad, I think our laws are just a little tighter when it comes to safety.
@@MrCed122 yeah, I swear I saw a video from Matthias wandel back in like 2016 where he demonstrated that you can get 240 at any outlet by just joining the phases together, and explained why this was common, but I can't find that now so I can't be sure whether I just dreamed it into existence or what.
our electricity is basically the same but instead of nema 14-50 for split phase on stoves and dryers we usually use nema l6-30R or L14-30R and their higher amp cousins. also "ontahrio" is funny as hell so u get a pass on that. also most timmy hos have the coffee makers right up front and centre which puts them under the TVs, so that rust is probably coffee steam.
I've heard that some American houses are also occasionally wired like that as well, which is extremely cursed. Don't stuff 240V over NEMA 5-15. Ideally, no country would use NEMA 5-15 because it's truly awful, but we appear to be stuck with it.
the same as for the 44-pin connectors smaller pitch, the jumpers on those are smaller too, so maybe that only works with those... guess 3.5" drive jumpers wouldnt work diagonally.. I only ever remember 2.5" drives with 2 jumper pairs, not 3+ like most 3.5 so maybe the diagonal is a solution to limited space/specs
My first thought when you said it was rusted was "Tim Hortons? Did someone decide to put a coffee machine right under it and let it slowly get steamed?"
@@eveypea ~sees dieing computer, pours coffee on it~ "Wake up it's time to work" ~computer is just barely there, and has amnesia, just puts something on screen~ "me work not so good"
I worked at one of the pioneering companies in the digital signage space during the era when this device would have been new. We once got a unit back from a restaurant as "defective" that still smelled of the sanitizing floor cleaner they had mopped into it. Yeah the warranty claim was denied and we let customers know that they should put the unit somewhere where it will not be sloshed with cleaning fluids.
@@timballam3675 sure, i know a guy who designed and built one of those arrays of literal light bulbs with a whole raft of triacs and a 6502. Maybe there's a better term for when it's a flatscreen displaying a video + static images and maybe a chiron at the same time, as delivered over the internet or via satellite
@JohnnyBGoode9 I'm positive it was a genuine error. Gravis is pretty explicit when he's taking a shot at someone/something, and he generally does so to make some kind of point. I'm just busting his chops because I find the mistake funny. I was born in Toronto, I know when someone is really taking a swing at Onterrible, and I can take a joke.
Oh thank god! I'm just about losing my sanity with my two year old today, my partner finally came home and could take over and I open up TH-cam to find a fresh CRD upload. Seriously, you just made my evening. Thank you so much for your videos from a very, very tired mum
This is so random but I know the Violet that sent this in! She's a good family friend and is a wonderful, kind, smart young woman. Absolutely got surprised by her being the one to send this in.. What a small world! Gravis, you are so lucky to have this connection to her! :D Just wanted to share for anyone reading that her game Ooga really is amazing. I've played its demo and it really is wonderful! If there's anyone who deserves a wish list it's her! Life has tried to kick her down a bit lately but she's still so kind hearted and spirited and that totally shows in the game. Give it a look :)
Those 44-pin mini IDE cables were pretty common in the chunky laptops of that era. They were usually pretty short (for obvious reasons), but longer ones did exist. Also in that era, VIA CPUs were *really* cheap. If performance was of no real concern, there was probably no cheaper option. They showed up in all sorts of low-end devices, especially near the end of the Socket 7 era.
The Pentium / Pentium MMX, AMD K5 / K6 were ZIF sockets. 486s were the last (x86) to be non-zif, iirc, and some socket 3 486s were lever style zero force. Now I feel old.
Huh! When I worked in e-waste I handled a ton of P1s and I could have sworn they were often just press-in, but it seems I was thinking of the 486s. For a few years it all blurred together.
@@CathodeRayDude The earlier Pentiums were press in, like the P60, P75, and P90. I have a board with a P-166 on it and it uses a ZIF socket. The board appears to accept a P-120 and higher
You are 100% correct. ZIF was introduced with socket 3, wasn’t common though until socket 5 with the original Pentium. Were there Pentium’s that didn’t have ZIF?, sure, but were not that common. If you bought a socket 5 “generic” motherboard it came with a ZIF.
Yes, all off this. I had to Google to make sure my aging brain hadn't skipped a groove, but yes, my Pentium 66, and the AMD K6 that replaced it were Ziff back in the 90's.
AFAIK digital signage is the more common term for what little guys get used for, narrowcasting is more of a marketing concept you can apply to digital signage. Oh and if you google socket 1, which was used for all 486 variants after the 80486, you see that most of the images are ZIF. So they started using zif quite early, which makes sense as LIF is a pain for cpu's (and zif existed before lif anyway)
sdb4 is the extended partition that's hosting the logical partitions of sdb5-10. You can't mount sdb4 directly. Why the error message when you tried complained about memory, no idea. (I was briefly confused about why the sizes seemed to add up to more than the size of the disk until I recalled how DOS partition tables worked)
For the benefit of others, I'll mention that DOS partition tables are limited to four partitions on the disk. If you need 5 or more, then you use an extended partition to wrap the excess partitions in a trenchcoat to note in the partition table as one partition.
@@UnlimitedRun for Early software compatibility. 4 primary partitions. The primary partition table shared the boot sector so there was no room to expand, and secondary partitions were actually defined recursively.. an extended partition contains the partition table in the first sector, which may have one further extended partition defined (recursively), and one other partition... Not that the exact details matterred to anyone. It could have been done with four partitions in the "extended" partition tables, and that would have worked just as well. but they did it with 2.
But they were also Low ESR, so not really shocked the replacements he used didn't help. I also would have attacked the whole thing with alcohol, especially the CPU as it looked like it may have grown mould which could have been shorting the probe points or capacitors.
It's a very safe bet that heatsink fan failed first, then the heat from the CPU resulted in the failure of those caps (and the crusting of the fan cable lol)
My random guess is that the little board with the tiny chip is maybe an EDID rom so that the GPU always "sees" a valid monitor and outputs the right resolution even when the external display doesn't have any EDID or is connected with RGB BNC cables
The trip through the innards of that little guy was an entertaining slow-motion train wreck. The mystery of the damage endures. Guess we’ll never know if it was finally done in by fire, flood, lightning, or just a good squirt of cap juice.
i remember these when i worked at timmies way back in the day, they used these for the drive thru's they had them in the menu boxes, i remeber opening the drive thru menu box and that little guy sittting in there, if you lookup the tim hortons dive-thru's from that era you can clearly see the screens, so the rustyness makes sense accounting how many ontairo winters that little guy survived in a metal box
38:30 They went with Via because they were cheaper and lower power consumption. If you didn't need the performance and wanted to save every last possible dime, Via was your choice.
Roku started with boxes like this, though not x86 based. I build a kiosk that played videos in a museum (not advertising, phew) in 2004 with a Roku PhotoBridge media player.
MPEG2 is an extension of MPEG1 for actual industrial usage. For one thing, they had initially forgotten that everything in the video world was interlaced, and the codec had to work that way, too. MPEG2 decoder can handle MPEG1 streams which use a subset of its functionality. However, unless you need to play a postage stamp sized video on some pre-MMX Pentium, or deliver 20 second porn clips to dial-up users, you won't encounter MPEG1. Trying to figure out file properties in VLC is pretty useless, as it works with many kinds of ephemeral sources, not just files, and only shows most generic metadata. You need to look at its module graph and/or the complete logs to learn what data it really tries to munch. MediaInfo is a handy tool that is so common people using it forget that such problems exist .
The words "they were bought by a company called Cineplex" said so casually without a hint of irony threw me for a loop, if I go past a movie theater 9/10 times its a Cineplex. I had just assumed they were a large multinational that was probably US based since I've never seen them advertise themselves as Canadian. It's like if someone said "bought by a company called McDonalds" or "cooked with a vegetable called an onion".
Yeah, I don't have one and I decided it wasn't really worth it. I went at it with a toothbrush off-camera for a bit but I realized there was almost nothing sitting on the surface other than some dust so I lost any enthusiasm for that. I talked about that during shooting, but during editing I went "wow this is just me talking in circles about nothing for 20 minutes" and deleted it all. There's like 2 hours of diagnostics missing from this but trust me, they were completely unwatchable. I decided to just skip to "it's dead, let's see what we can learn from the hard drive." Had the machine been a little more unique I'd have maybe put more effort into it, but I felt like all I'd accomplish is making a boring video even more boring.
Having now worked in two industries putting computers in places they shouldn't be, I am not at all surprised at the rust. There was one location I was providing support for, always wet and occasionally containing some fun chemicals, in which a computer would corrode to death in a matter of 4-6 months on average. I had never seen CPU pads turn black before, and I'd dare say that environment was suited to do it in record time. The closest I came to digital signage was sticking NUCs on the back of TVs in breakrooms and hallways, usually for security camera remote views and employee bulletins. One of those came back once because it would not boot and after examination I determined that a single drop of mouse urine had killed the Intel Optane Memory board in the NVME slot (this was when we were still using Intel branded NUCs) and had rendered the machine unable to boot. Naturally that NUC went into the recycle bin, then my car, then into my homelab with a new SSD.
The reason for the "funky washer" for the hard drive is probably an assembly consideration. Consider that the hard drive has to be added after the main board is mounted, and adding a small washer in that confined space would be fiddly and time consuming, especially if it fell off and had to be fished out of the case. Having an easy to handle arrangement like this could be worth it to simplify assembly during production.
As far as lilo on that thing, it was still the default bootloader on a few distros in 2006, so it's not terribly surprising to see it on there. After all, it's not like you'd be dual-booting the machine, and if the goal is just to boot linux, lilo's much simpler, and harder to screw up.
FCPGA just means Flip Chip PGA, where the die faces the cooler. For the longest time the die faced the socket, and it led to some cooling problems when we hit the P3 era. ZIF sockets were also starting to become common in systems by the early Pentium era. Socket 4/5/7 are all ZIF, and even apple started using it by the time the G3 was coming out.
I was going to say, it makes total sense that the manufacturer would specify FC-PGA since the P-III came in both PGA and Slot 1 packages. So if they just "P-III" it'd be a problem.
Aw man, I was really hoping for some Tim Hortons files on there. I even convinced myself the rust was from it being mounted above a coffee brewer or something.
It still could've been. Campus restaurants tend to have a smaller footprint so there's a non-zero chance that they just winged this front facing sign right over a coffee brewer.
If it is the Tim's in the main food Court at the front of main campus that location was actually a fairly decent sized location. Tim Hortons layout has the menu tvs right above the coffee machines
I was wondering if it referred to a cinema chain. They would have enough locations and demand for centrally-managed digital signage that it could be worth acquiring a (smaller) company to handle it all for them.
Yeah, it makes sense that they'd acquire a digital signage company, they're big enough, and it seems like this London company was a small enough company. To relate to Americans, I'd describe them as like our AMC
Wow! We finally found our first Linux-powered Little Guy! And it was a really early digital signage system! That tiny mainboard is honestly adorable! Even if we didn't get to see it run, I was happy to see it anyway.
Re constant adverts in public. They might not have been TVs with animated ads, but there was those advertising boards with the adverts on a roll, that would switch a new one in every 30 seconds so you were distracted by them. East to blend the two together..
Tip: LILO boot to shell. Instead of "linux runlevel 1", "linux init=/bin/sh" should get you in. Then remount the root filesystem read/write, and reset/clear the root password.
ty, though I'll probably forget it by the next time I need it, hah. my next step was going to be to just google the right command, the only reason I included that clip is because I wanted to include my reaction to the bootloader password.
An extended partition is used to contain other partitions, to get around the MBR limit of four glorious partitions. Never tried mounting it, but intuition tells me that cannot be done. And now you do too.
This video was certainly not boring/dull at all. Watching your efforts towards troubleshooting and research is fantastic content, and is on par with the subject item itself.
I remember those Via Chips, what i could decipher: 1) Its a Nehemiah Package. The first Row is not readable, the second row: (133x 7.5) or (133x 9.0) followed by 1.25V The instructions to setup the Chip, making this a Via C3 Nehemiah+ 1ghz or 1.2ghz with a maximum CPU TDP of 12-15 Watts. A Pentium III would start at 13 and would go up to 37 Watts. Even a Coppermine Celeron or a Tualatin-256 would use 30Watt+ at higher clocks. But most remarkable is the release Date, which unvails why this CPU is chosen: The Nehemiah+ was released in 2003, Coppermine und Tualatin Production stopped in 2002. So the only new Processors to source for those S370 Boards were the Via C3 Nehemiah+ The Nehemiah Clockspeed gives about halv the Speed in PIII Power so a Nehemiah 1200 would be roughly as fast as a Pentium III 600. Which draws around 17 Watts of power. But there were no alternatives in 2006: Netburst Pentium/Celeron were drawing much more power, same goes with the Athlon (64) and Duron line. Intel Atom started in 2008, so there were literally no other low Power Processors cheap between 2003 and 2008 except the Via C3.
I love these great long videos about products I would never think about and learning things I only just understand. Turns out having someone who just likes talking about stuff they're passionate about is a real cool vibe.
Worked in a digital signage outfit for a few years and saw lots of stuff like this. The "media engine" branding was a trend at the time. We largely used AOpen "Media Engine" mini-PCs based on C2D laptop procs and an external power supply, but otherwise basically the same thing.
the wow stick can actually be quite easily used as a ratchet since it has metal gears, I have put one through hell and back and only after drilling actual long ass wood self tappers into beams (20+ times) did I manage to break it. you just hold down the button and go ham on it. 100% would buy again
Hi, yes there was an interesting part in this video, it's the one from 00:00 to 49:04. Thanks for trying to repair this little guy, and for showing off this creepy jumper connection.
Wow a london shout out! Sitting in my living room 5 minutes from Fanshawe and about the same from ek3. Good ole Fanshawe Love the cathode Ray dude her in london on tar eee oh
I fell asleep to this video earlier today and had a lucid dream where I was hearing the video in my dream while living out wild scenarios of my own making. Thanks, CRD!
The Via C3 did make a lot of sense from a TCO perspective.. while less performant than its Pentium3 counterparts at the same speed, they came with a TDP of around 6 Watts, compared to 30-35 Watts for the Intel chips.
Old hardware brings back dim memories. AST multifunction boards (RAM expansion, Parallel and Serial ports) Hercules mono graphics boards, actual separate RAM chips, Hard drives on an expansion card and separately socketed Floating Point chips. My first PC compatible was a no-name 286. Overlocking a Celeron 300A to 450 MgHz. Those were the daze.
I absolutely love this series.😚 I would not be surprised if Diebold Canada was involved in this project as they had a presence in London at the time in the tech industry in Ontario.
I feel like the lil guys line of videos doesnt need one big epic thing in them its just neat looking at all the stuff that was avaivlable over all these years :3
I am quite familiar with that power supply from equipment I work with and a very common failure from them, they will feed AC voltage through to the output. If caught early (or lucky) doesn't tend to do any damage, just a failed system, but every so often when one of those power supplies fail, they can send a pretty high spike of AC current to whatever they have been powering. When I see a fluctuating DC reading, also check to see if it is reading an AC current.
ZIF sockets for processors started becoming common on the 486 and became the norm by the original Pentium. I may have seen some early Pentium (Socket 4) boards that didn't use a ZIF socket but I can't remember a single Socket 5/7 Pentium that wasn't.
Regarding the electric screwdriver. We use a Panasonic EY7410 cordless screwdriver at work. I think that it fits perfectly in what you are looking for, between a small electric screwdriversand a drill. It might be a bit pricier because it's mainly intended for industrial but they have really solid torque with 2speed gearbox with adjustable clutch. I'm starting to sound like a salesman :D Anyway really interesting little guy once again. We used to say that computers with a VIA CPU is ''Viallinen'' and that can be interpred as broken or a thing with a via, and in this case both :D They weren't really good desktop chips compared to P3 or a Celeron. They are probably good enough for an embedded system like this.
The "Flip Chip" in FCPGA means the wafer of silicon that "is" the chip, is mounted upside down instead of right side up. When it's upside down, the bonding wires would be on the bottom side, pushing the silicon toward the top, where heat can be extracted more efficiently. Sometimes there are other reasons, but this one is pretty common.
God. The amount of digikey bags I save because "I'm going to use that for something" but end up just storing them all in a bigger digikey component bag....
Thank you; I honestly had never encountered that pronunciation of Ont-air-io. On-tar-io; makes complete sense that someone could find that sound but I had never thought of it. 🏴
i honestly love that you explain electronics and engineering concepts even if you think most of your viewerbase already knows it. i find myself learning something new in every video and names for things i only tangentially knew about by taking apart shit out of boredom.
My Sweet Summer Child, you don't know what an Extended Partition is? I envy you It is a container for extra partitions, just take a look at the start and end clusters.
What they were dconfused by wasn't extended partitions as much as the error message not really indicating what it was about and instead saying something about memory I think?
I do wanna say, even if the video is "boring" honestly I just enjoy being here for the ride. I love both styles of vids, your more "I'm presenting a topic" and "I'm at the workbench", both are just good to vibe to and maybe learn some things along the way Like cursed pin jumper configurations
@@CathodeRayDude Probably so, but Linux is Linux and I bet mount actually _attempted_ to mount that instead of seeing it and "oh, this isn't a *real* partition, let me tell the user that"
It tells you where to look for meaningful info. ;-) What happens when you blindly try to mount something is both well documented, and quite unpredictable. mount tool knows little about filesystems that potentially exist on the third planet from the Sun, or are supported by a given system at a given time, it relies on kernel and filesystem drivers. If type is not defined by the user, it tries each available filesystem driver in a certain order (see option -t in the fine manual). After ext family and other filesystems with well defined metadata formats decline to work with that device, basic ones try their best (e. g. raw floppy, maybe some raw flash fs with hard-coded sizes and offsets for embedded use). It is totally possible that one of them finds some arbitrary bytes, interprets them as filesystem size of 555 petabytes, honestly tries to allocate enough memory to work with such a giant, then fails. mount is traditionally for administrators who know what they mount. It may not succeed with a perfectly fine partition just because of some missed critical option (even though ext family should be generally backward compatible). A full-featured graphical disk/partition management utility from any general distribution should show you both partition metadata and filesystem metadata before mounting anything. There is also an option to ask testdisk to perform an analysis if you're unsure. This was mostly for curious readers, I'm sure our host is no stranger to manual calculation of disk offsets.
@@DarkKnight32768 oh right, of course, I was thinking at the wrong abstraction level. Mount is unaware of the partition table because you give it the block device representing the partition after it's been carved up by the OS. Presumably extended partitions do not contain a magic number that would make sense for mount to look for.
the Amiga 600 and 1200 are suited with 44pin headers on the mobo for 2,5" drives via a ribbon cable. back in the days we often modded those from 2,5" formfactor to 3,5" by splitting the cable or via an adapter to fit a bigger and mostly cheaper 3,5" drive, cuz 2,5" drives where pretty expensive back then.
43:10 it seems to be a mostly sensible partition table for this kind of device, from what i can see: sdb2 -> / (system root directory) sdb3 -> /var sdb5 -> /usr While sdb10 i _think_ would be /root (the root users "home"), but it's possible it could be /opt (though that's more likely the 75gb volume), either way the rest should nest below it like so: sdb6 -> /sdb10/incoming sdb7 -> /sdb10/bin sdb8 -> /sdb10/lib sdb9 -> /sdb10/etc With these partitions containing the bare minimum utilities necessary to remote in and configure/diagnose the device, seperated to be less prone to breakage from changes and updates
The 75 GB partition is the "Extended partition table" which contains all of the remaining positions. The MBR extended partition scheme is another one of those historically significant horrible kludges that it's best not to dive into right now; but suffice to say it is a horribly fragile hack and best avoided for all but legacy applications.
ZIF sockets were around since at least socket 1, at least for PCs. So... much earlier than the Pentium 3. Socket 1 was the same 169-pin socket that 486 overdrive CPUs used but enhanced with a ZIF+lever locking. EDIT: I see this is in your pinned comment. Nevermind!
28:42 - years of working for NCR support (COMPRIS systems) confirms this. We had backup systems, and when the restaurant was closed, we'd remotely log in and troubleshoot whatever broke, never during opening hours (unless it was a main KVM or the host system)
There were ZIF sockets starting with the 80486. Before that there were many 80286 CPUs in the LCC (Leadless Chip Carrier) package that had a single row of flat contacts around four edges. The CPU was held onto the socket, which had spring contacts, with a metal clip and a small finned heatsink. Not that the heat sink was actually needed, except possibly with the rare 16 Mhz version. Most of these 286 systems I encountered were 12 Mhz. I'd call that 286 style the great grandpappy of the Land Grid Array CPUs and sockets.
Not surprised those caps gave up the ghost and took other power regulation components with them. That board was manufactured right in the middle of the capacitor plague era. I seen dozens of mainboards from the later Pentium 2 boards all the way through the Core 2 days die from bad caps that often times took power regulation components with them. So there was no way that board was going to work even with the caps replaced.
With caps you can go with a higher voltage and you could probably get away with a slightly lower capacitance but the correct value will always be best.
The most important characteristic in this case is that they _must_ be low ESR capacitors, else they will not work at the frequency they are being operated at. Power supply circuits, as a general rule, have the capacitors specified at twice the design value, to account for deceased capacitor performance due to ageing; so as long as you don't go too much higher or lower than what's there, the voltage regulator ought to work just fine.
Like you said, the Wowstick works great for driving lots of small screws that aren't stuck or threadlocked. But reef on it like it's a boring unelectrified screwdriver and the gearbox will break. And that's why I bought it twice. Also, that power supply looks like it came from an external SCSI drive enclosure.
@43:00 sdb4 didn't mount because that partition is an MBR extended partition than includes all the partitions after it. Take note of the size of sdb4, which roughly is the sum of the following partitions. This exists because MBR data structure had a limit of partitions of 4 originally, but it was later extended by producing a fake, unmountable partition containing the extra partition's metadata. No ideas on the weird linux mount error though
Take a look at an MBR extended partition table in Norton Disk Editor sometime. What you find will utterly horrify you; especially when there are this many partitions involved.
Power screwdriver recommendation: Vessel Ball Grip. They're about the size of that Whia insulated driver you used to remove the HDD screws and have a reasonably ergonomic slide switch. They come with two different gear reductions, the gray handle is higher speed and works well for general computer screws into metal, if they're tight you break them loose by turning the driver, then speed them out with the motor, but screws into plastic often have more drag for the entire length of the thread, the low speed black handled driver works better for those. Bonus: They're made in Japan and come with a rather nice #2 JIS bit.
![Little Guys: Episode 1 [Logic AU912]](http://i.ytimg.com/vi/GQmj1P-DSoY/mqdefault.jpg)
![Little Guys: Episode 1 [Logic AU912]](/img/tr.png)
Update: An eagle eyed viewer (thanks @SwitchingPower!) noticed something I missed: the component on the backside of the board underneath the CPU is toast. You can see that around 21:20, directly in the middle of the socket on the back. I'd noticed the dirt on it resembled the spatter from a blown semiconductor during disassembly, but I wiped it off and it seemed like it was just the same deposits I saw elsewhere in the thing, and I didn't see a cracked case, so I thought it was just dirty. Sure enough though, I desoldered it - it looks like the back of the case blew out, even stripped the top layer off the PCB. So yeah, this was never gonna work.
If this video feels a bit... unenthusiastic? Like I didn't put in as much effort as it deserved? It's because I got about two hours into shooting it and realized that it felt like this thing SHOULD work, and that I was just not skilled enough to figure out what was wrong, so I lost a lot of drive and just tried to salvage something out of it. Turns out it wasn't my fault; wish I'd realized that at the time.
Regarding the little six-pin device on the daughterboard plug: I had tried during shooting to get a good look at the label and couldn't make it out, but while looking at the board again today I suddenly found an angle where I could see the lettering. Turns out it says AAPN, which is a MAX6629 (or similar) temperature sensor! That makes a lot of sense!
If you see the early version of this video: There's a quip in there about 240V being used in Canada, which is based on something I saw *years* ago about Canadian houses being wired with split-phase specifically so they could bring out 240 to ordinary wall outlets. True or not, it doesn't seem like it's very common, so I removed that line with the youtube editor and it'll take effect sometime in the next 24-48 hours.
Yes I pronounced Ontario wrong. Sorry :(
I quipped about FCPGA being basically "the" socket for CPUs with pins - that's true as far as most chips go, but what preceded that was plain PGA, without the FC; "flip-chip" referred to the die being upside down on top of the carrier instead of vice versa. Likewise, apparently ZIF came in with the first Pentium sockets - I suspected it was a bit older than I remembered, that's why I hedged on that point. It's actually the 486 I was thinking of, but those weren't any more pleasant to press-fit into a socket than the 370 chips; I remember breaking pins off of one or two trying to get them out, in fact.
Canadian here : sure, we have 240V and it COULD be connected to a regular outlet, I think I've seen that done in some specific case, especially with specific plugs in kitchens (like for microwaves and kettles) and more recently, electric cars, but way more commonly, it's only used with the oven and the dryer with their own plugs. Our electrical system is basically identical to the US, so really bad, I think our laws are just a little tighter when it comes to safety.
@@MrCed122 yeah, I swear I saw a video from Matthias wandel back in like 2016 where he demonstrated that you can get 240 at any outlet by just joining the phases together, and explained why this was common, but I can't find that now so I can't be sure whether I just dreamed it into existence or what.
our electricity is basically the same but instead of nema 14-50 for split phase on stoves and dryers we usually use nema l6-30R or L14-30R and their higher amp cousins. also "ontahrio" is funny as hell so u get a pass on that. also most timmy hos have the coffee makers right up front and centre which puts them under the TVs, so that rust is probably coffee steam.
I've heard that some American houses are also occasionally wired like that as well, which is extremely cursed. Don't stuff 240V over NEMA 5-15. Ideally, no country would use NEMA 5-15 because it's truly awful, but we appear to be stuck with it.
can't believe you would do this. smh my head.
The diagonal jumper is the worst thing I’ve ever seen in my entire life
Would you even be able to fit a regular jumper between these pins? It would be a greater distance
@@philipc4272 I was wondering the same thing!
the same as for the 44-pin connectors smaller pitch, the jumpers on those are smaller too, so maybe that only works with those... guess 3.5" drive jumpers wouldnt work diagonally.. I only ever remember 2.5" drives with 2 jumper pairs, not 3+ like most 3.5 so maybe the diagonal is a solution to limited space/specs
Yeah, I physically shuddered and kept saying "no"
the worst thing ever? i dunno ive seen some sh** on the internet back in the day lmao
I think I need a “Boy Howdee counter” in the corner of the screen. Makes me smile every time.
I came here to say this
New drinking game?
Pour a drink every time Gravis says "Boy", down a drink every time he says "Howdy"?
My first thought when you said it was rusted was "Tim Hortons? Did someone decide to put a coffee machine right under it and let it slowly get steamed?"
Or did someone pour a long black into the case?
Mmmmm... Steamed little guy. GHHhHhHhH
@@eveypea ~sees dieing computer, pours coffee on it~ "Wake up it's time to work"
~computer is just barely there, and has amnesia, just puts something on screen~ "me work not so good"
@@Kafj302Like something out of The Kids in the Hall.
100% those devices tend to live in the worst of places
I worked at one of the pioneering companies in the digital signage space during the era when this device would have been new.
We once got a unit back from a restaurant as "defective" that still smelled of the sanitizing floor cleaner they had mopped into it. Yeah the warranty claim was denied and we let customers know that they should put the unit somewhere where it will not be sloshed with cleaning fluids.
aaaaaaAAAAAA
I can just imagine it now…
“What? The outside was dirty so I cleaned it! You sell a product for restaurants that can’t handle getting dirty??”
Yeah just slide it under a counter and leave it on the floor, it's fine
Did digital signage at the tail end of the 80s, hardly makes the 2000s a period of pioneering..... 🤣🤣🤣🤣
@@timballam3675 sure, i know a guy who designed and built one of those arrays of literal light bulbs with a whole raft of triacs and a 6502.
Maybe there's a better term for when it's a flatscreen displaying a video + static images and maybe a chiron at the same time, as delivered over the internet or via satellite
I can't get past that pronunciation of "Ontario."
Me too; i couldn't decide if it was on purpose as a joke or just a lack of familiarity.
@Pablonmon I think it was a genuine lack of familiarity. FYI for anyone unaware, we pronounce the "ar" like "air."
@JohnnyBGoode9 I'm positive it was a genuine error. Gravis is pretty explicit when he's taking a shot at someone/something, and he generally does so to make some kind of point.
I'm just busting his chops because I find the mistake funny. I was born in Toronto, I know when someone is really taking a swing at Onterrible, and I can take a joke.
Me too. I wanted him to say it just once more 😅
If you live in California, there's an airport in Ontario California so he should be familiar with that.
Oh thank god! I'm just about losing my sanity with my two year old today, my partner finally came home and could take over and I open up TH-cam to find a fresh CRD upload. Seriously, you just made my evening. Thank you so much for your videos from a very, very tired mum
This is so random but I know the Violet that sent this in! She's a good family friend and is a wonderful, kind, smart young woman. Absolutely got surprised by her being the one to send this in.. What a small world! Gravis, you are so lucky to have this connection to her! :D
Just wanted to share for anyone reading that her game Ooga really is amazing. I've played its demo and it really is wonderful! If there's anyone who deserves a wish list it's her! Life has tried to kick her down a bit lately but she's still so kind hearted and spirited and that totally shows in the game. Give it a look :)
@CCaleighC If you don't mind sharing the steam link for the game?
@@arlandithat gets filtered now
The “Shedcat” one? (There’s a few games and board games with similar names.)
@@kaitlyn__Lthis is the one!!
Thanks so much for the kind words Caleigh! ❤
Those 44-pin mini IDE cables were pretty common in the chunky laptops of that era. They were usually pretty short (for obvious reasons), but longer ones did exist. Also in that era, VIA CPUs were *really* cheap. If performance was of no real concern, there was probably no cheaper option. They showed up in all sorts of low-end devices, especially near the end of the Socket 7 era.
I was always happy to find some weird little laptop or device running a VIA C3 or Transmeta Crusoe
If this was in place of a P3/celeron, would be a Via C3, in an analog of the socket 370.
@@RealJonDoe Oh good catch, that's a whole generation newer than I was thinking. I forgot Intel ever licensed that socket out to third parties.
The Pentium / Pentium MMX, AMD K5 / K6 were ZIF sockets. 486s were the last (x86) to be non-zif, iirc, and some socket 3 486s were lever style zero force. Now I feel old.
Huh! When I worked in e-waste I handled a ton of P1s and I could have sworn they were often just press-in, but it seems I was thinking of the 486s. For a few years it all blurred together.
@@CathodeRayDude The earlier Pentiums were press in, like the P60, P75, and P90. I have a board with a P-166 on it and it uses a ZIF socket. The board appears to accept a P-120 and higher
You are 100% correct. ZIF was introduced with socket 3, wasn’t common though until socket 5 with the original Pentium. Were there Pentium’s that didn’t have ZIF?, sure, but were not that common. If you bought a socket 5 “generic” motherboard it came with a ZIF.
Yeah I had a 486 socket 3, I upgraded it to a DX4 100 overdrive and it could absolutely shred Descent.
Yes, all off this. I had to Google to make sure my aging brain hadn't skipped a groove, but yes, my Pentium 66, and the AMD K6 that replaced it were Ziff back in the 90's.
AFAIK digital signage is the more common term for what little guys get used for, narrowcasting is more of a marketing concept you can apply to digital signage.
Oh and if you google socket 1, which was used for all 486 variants after the 80486, you see that most of the images are ZIF. So they started using zif quite early, which makes sense as LIF is a pain for cpu's (and zif existed before lif anyway)
sdb4 is the extended partition that's hosting the logical partitions of sdb5-10. You can't mount sdb4 directly.
Why the error message when you tried complained about memory, no idea.
(I was briefly confused about why the sizes seemed to add up to more than the size of the disk until I recalled how DOS partition tables worked)
For the benefit of others, I'll mention that DOS partition tables are limited to four partitions on the disk. If you need 5 or more, then you use an extended partition to wrap the excess partitions in a trenchcoat to note in the partition table as one partition.
@@UnlimitedRun for Early software compatibility. 4 primary partitions. The primary partition table shared the boot sector so there was no room to expand, and secondary partitions were actually defined recursively.. an extended partition contains the partition table in the first sector, which may have one further extended partition defined (recursively), and one other partition... Not that the exact details matterred to anyone. It could have been done with four partitions in the "extended" partition tables, and that would have worked just as well. but they did it with 2.
Fellow Ontarian, here. I’d like to commemorate this Little Guy putting in his time at Tim Hortons. He’s braver than any US Marine or Canadian Mountie.
Not just a Tim's, but a Tim's on campus, staffed almost entirely by college students
Those caps look like they've suffered a catastrophic failure and vented. So great call to replace them
But they were also Low ESR, so not really shocked the replacements he used didn't help. I also would have attacked the whole thing with alcohol, especially the CPU as it looked like it may have grown mould which could have been shorting the probe points or capacitors.
It's a very safe bet that heatsink fan failed first, then the heat from the CPU resulted in the failure of those caps (and the crusting of the fan cable lol)
The bulged caps are the very first thing i noticed ! One should totally replace them all.
My random guess is that the little board with the tiny chip is maybe an EDID rom so that the GPU always "sees" a valid monitor and outputs the right resolution even when the external display doesn't have any EDID or is connected with RGB BNC cables
The trip through the innards of that little guy was an entertaining slow-motion train wreck. The mystery of the damage endures. Guess we’ll never know if it was finally done in by fire, flood, lightning, or just a good squirt of cap juice.
i remember these when i worked at timmies way back in the day, they used these for the drive thru's they had them in the menu boxes, i remeber opening the drive thru menu box and that little guy sittting in there, if you lookup the tim hortons dive-thru's from that era you can clearly see the screens, so the rustyness makes sense accounting how many ontairo winters that little guy survived in a metal box
Shout out to the boys over at Funshawe, London is my home town. always a joy tuning in to the little guys eps.
I grew up near london :P
38:30 They went with Via because they were cheaper and lower power consumption. If you didn't need the performance and wanted to save every last possible dime, Via was your choice.
Roku started with boxes like this, though not x86 based. I build a kiosk that played videos in a museum (not advertising, phew) in 2004 with a Roku PhotoBridge media player.
MPEG2 is an extension of MPEG1 for actual industrial usage. For one thing, they had initially forgotten that everything in the video world was interlaced, and the codec had to work that way, too. MPEG2 decoder can handle MPEG1 streams which use a subset of its functionality. However, unless you need to play a postage stamp sized video on some pre-MMX Pentium, or deliver 20 second porn clips to dial-up users, you won't encounter MPEG1.
Trying to figure out file properties in VLC is pretty useless, as it works with many kinds of ephemeral sources, not just files, and only shows most generic metadata. You need to look at its module graph and/or the complete logs to learn what data it really tries to munch. MediaInfo is a handy tool that is so common people using it forget that such problems exist .
The words "they were bought by a company called Cineplex" said so casually without a hint of irony threw me for a loop, if I go past a movie theater 9/10 times its a Cineplex. I had just assumed they were a large multinational that was probably US based since I've never seen them advertise themselves as Canadian. It's like if someone said "bought by a company called McDonalds" or "cooked with a vegetable called an onion".
runs software made by some company called micro soft?
Your idea of a good cleaning disappoints me and my sonic cleaning tub.
Yeah, I don't have one and I decided it wasn't really worth it. I went at it with a toothbrush off-camera for a bit but I realized there was almost nothing sitting on the surface other than some dust so I lost any enthusiasm for that. I talked about that during shooting, but during editing I went "wow this is just me talking in circles about nothing for 20 minutes" and deleted it all. There's like 2 hours of diagnostics missing from this but trust me, they were completely unwatchable. I decided to just skip to "it's dead, let's see what we can learn from the hard drive." Had the machine been a little more unique I'd have maybe put more effort into it, but I felt like all I'd accomplish is making a boring video even more boring.
@@CathodeRayDude I was only making a snarky comment, my man, I love your videos
@@FeliciaByNature Oh not at all, I just figured I'd explain my reasoning because I had the same thought!
@@CathodeRayDudethis is why having creators who actually interface is so valuable! We cherish you Gravis!
You make sonic the hedgehog clean your computers for you? At least CRD did it by his own hand :P
11:28
CONTINUITY ERROR. You have been fined 100 bison dollars
Same with the "first" opening of the case. The disk is upside down and disconnected. ;)
@@MartinaD That Is exactly where the first opening happens.
Having now worked in two industries putting computers in places they shouldn't be, I am not at all surprised at the rust. There was one location I was providing support for, always wet and occasionally containing some fun chemicals, in which a computer would corrode to death in a matter of 4-6 months on average. I had never seen CPU pads turn black before, and I'd dare say that environment was suited to do it in record time.
The closest I came to digital signage was sticking NUCs on the back of TVs in breakrooms and hallways, usually for security camera remote views and employee bulletins. One of those came back once because it would not boot and after examination I determined that a single drop of mouse urine had killed the Intel Optane Memory board in the NVME slot (this was when we were still using Intel branded NUCs) and had rendered the machine unable to boot. Naturally that NUC went into the recycle bin, then my car, then into my homelab with a new SSD.
The reason for the "funky washer" for the hard drive is probably an assembly consideration. Consider that the hard drive has to be added after the main board is mounted, and adding a small washer in that confined space would be fiddly and time consuming, especially if it fell off and had to be fished out of the case. Having an easy to handle arrangement like this could be worth it to simplify assembly during production.
As far as lilo on that thing, it was still the default bootloader on a few distros in 2006, so it's not terribly surprising to see it on there. After all, it's not like you'd be dual-booting the machine, and if the goal is just to boot linux, lilo's much simpler, and harder to screw up.
Particularly as industrial computers like this wont necessarily be using the latest versions of anything, just whatever works with the hardware.
"This is not unreasonable, but it does suck" I think you nailed it there.
FCPGA just means Flip Chip PGA, where the die faces the cooler. For the longest time the die faced the socket, and it led to some cooling problems when we hit the P3 era. ZIF sockets were also starting to become common in systems by the early Pentium era. Socket 4/5/7 are all ZIF, and even apple started using it by the time the G3 was coming out.
I was going to say, it makes total sense that the manufacturer would specify FC-PGA since the P-III came in both PGA and Slot 1 packages. So if they just "P-III" it'd be a problem.
Aw man, I was really hoping for some Tim Hortons files on there. I even convinced myself the rust was from it being mounted above a coffee brewer or something.
It still could've been. Campus restaurants tend to have a smaller footprint so there's a non-zero chance that they just winged this front facing sign right over a coffee brewer.
If it is the Tim's in the main food Court at the front of main campus that location was actually a fairly decent sized location. Tim Hortons layout has the menu tvs right above the coffee machines
As a Fanshawe grad of 2007, this video was delightful. 👍
Shout out to the little bug on the table at 33:00
Now that's a real 'little guy'.
buge :)
It's a feature.
Cineplex as in Cineplex Odeon? That’s one of the largest movie theatre companies in Canada
I was wondering if it referred to a cinema chain. They would have enough locations and demand for centrally-managed digital signage that it could be worth acquiring a (smaller) company to handle it all for them.
Yeah, it makes sense that they'd acquire a digital signage company, they're big enough, and it seems like this London company was a small enough company. To relate to Americans, I'd describe them as like our AMC
Im still mad they only decided to show Godzilla Minus One in like 3 theaters in the country... Untill the last week
@@forivall They do digital signage too, I saw one of their digital signs advertising itself in a mall recently
Wow! We finally found our first Linux-powered Little Guy!
And it was a really early digital signage system!
That tiny mainboard is honestly adorable!
Even if we didn't get to see it run, I was happy to see it anyway.
New genre on TH-cam: forgotten tech (mostly PC lately) ASMR. I dig it
Re constant adverts in public. They might not have been TVs with animated ads, but there was those advertising boards with the adverts on a roll, that would switch a new one in every 30 seconds so you were distracted by them. East to blend the two together..
Tip: LILO boot to shell. Instead of "linux runlevel 1", "linux init=/bin/sh" should get you in. Then remount the root filesystem read/write, and reset/clear the root password.
ty, though I'll probably forget it by the next time I need it, hah. my next step was going to be to just google the right command, the only reason I included that clip is because I wanted to include my reaction to the bootloader password.
An extended partition is used to contain other partitions, to get around the MBR limit of four glorious partitions. Never tried mounting it, but intuition tells me that cannot be done. And now you do too.
This video was certainly not boring/dull at all. Watching your efforts towards troubleshooting and research is fantastic content, and is on par with the subject item itself.
I remember those Via Chips, what i could decipher: 1) Its a Nehemiah Package. The first Row is not readable, the second row: (133x 7.5) or (133x 9.0) followed by 1.25V The instructions to setup the Chip, making this a Via C3 Nehemiah+ 1ghz or 1.2ghz with a maximum CPU TDP of 12-15 Watts.
A Pentium III would start at 13 and would go up to 37 Watts. Even a Coppermine Celeron or a Tualatin-256 would use 30Watt+ at higher clocks. But most remarkable is the release Date, which unvails why this CPU is chosen: The Nehemiah+ was released in 2003, Coppermine und Tualatin Production stopped in 2002. So the only new Processors to source for those S370 Boards were the Via C3 Nehemiah+
The Nehemiah Clockspeed gives about halv the Speed in PIII Power so a Nehemiah 1200 would be roughly as fast as a Pentium III 600. Which draws around 17 Watts of power. But there were no alternatives in 2006: Netburst Pentium/Celeron were drawing much more power, same goes with the Athlon (64) and Duron line. Intel Atom started in 2008, so there were literally no other low Power Processors cheap between 2003 and 2008 except the Via C3.
these videos are never boring, always delightful, whether the little guy shows up for work or not
I love these great long videos about products I would never think about and learning things I only just understand. Turns out having someone who just likes talking about stuff they're passionate about is a real cool vibe.
Worked in a digital signage outfit for a few years and saw lots of stuff like this. The "media engine" branding was a trend at the time. We largely used AOpen "Media Engine" mini-PCs based on C2D laptop procs and an external power supply, but otherwise basically the same thing.
RIP little guy, glad you could look at the hdd and get some of its story tho.
the wow stick can actually be quite easily used as a ratchet since it has metal gears, I have put one through hell and back and only after drilling actual long ass wood self tappers into beams (20+ times) did I manage to break it. you just hold down the button and go ham on it. 100% would buy again
Hi, yes there was an interesting part in this video, it's the one from 00:00 to 49:04.
Thanks for trying to repair this little guy, and for showing off this creepy jumper connection.
Wow a london shout out! Sitting in my living room 5 minutes from Fanshawe and about the same from ek3. Good ole Fanshawe
Love the cathode Ray dude her in london on tar eee oh
Honestly, little guys might just be the best series you've done. I love your take on the stories of these random devices.
I fell asleep to this video earlier today and had a lucid dream where I was hearing the video in my dream while living out wild scenarios of my own making. Thanks, CRD!
The Via C3 did make a lot of sense from a TCO perspective.. while less performant than its Pentium3 counterparts at the same speed, they came with a TDP of around 6 Watts, compared to 30-35 Watts for the Intel chips.
Yeah, if it was fast enough for what you needed then it was cheap as shit and ran very cool esp compared to PIIIs.
Old hardware brings back dim memories. AST multifunction boards (RAM expansion, Parallel and Serial ports) Hercules mono graphics boards, actual separate RAM chips, Hard drives on an expansion card and separately socketed Floating Point chips. My first PC compatible was a no-name 286. Overlocking a Celeron 300A to 450 MgHz. Those were the daze.
I absolutely love this series.😚 I would not be surprised if Diebold Canada was involved in this project as they had a presence in London at the time in the tech industry in Ontario.
I feel like the lil guys line of videos doesnt need one big epic thing in them its just neat looking at all the stuff that was avaivlable over all these years :3
I love Little Guys! One of my favourite series on TH-cam atm. Keep up the good work Gravis!
I am quite familiar with that power supply from equipment I work with and a very common failure from them, they will feed AC voltage through to the output. If caught early (or lucky) doesn't tend to do any damage, just a failed system, but every so often when one of those power supplies fail, they can send a pretty high spike of AC current to whatever they have been powering. When I see a fluctuating DC reading, also check to see if it is reading an AC current.
Nice vid! Cool that the hard drive files were recoverable ^_^
ZIF sockets for processors started becoming common on the 486 and became the norm by the original Pentium. I may have seen some early Pentium (Socket 4) boards that didn't use a ZIF socket but I can't remember a single Socket 5/7 Pentium that wasn't.
It's never boring, dude. I need to find out about all the stuff I didn't know I needed to find out about.
The way you pronounce "Ontario" really brings a smile to my face, lol.
41:05 You missed an opportunity to add the internet jingle. The internet jingle is the sole reason I'm watching this channel.
Also missed the very first one; not say least we got _one_ in the middle.
I love your videos so much. I love this series so much.
You're a gem.
Regarding the electric screwdriver. We use a Panasonic EY7410 cordless screwdriver at work. I think that it fits perfectly in what you are looking for, between a small electric screwdriversand a drill. It might be a bit pricier because it's mainly intended for industrial but they have really solid torque with 2speed gearbox with adjustable clutch.
I'm starting to sound like a salesman :D
Anyway really interesting little guy once again. We used to say that computers with a VIA CPU is ''Viallinen'' and that can be interpred as broken or a thing with a via, and in this case both :D They weren't really good desktop chips compared to P3 or a Celeron. They are probably good enough for an embedded system like this.
The "Flip Chip" in FCPGA means the wafer of silicon that "is" the chip, is mounted upside down instead of right side up. When it's upside down, the bonding wires would be on the bottom side, pushing the silicon toward the top, where heat can be extracted more efficiently. Sometimes there are other reasons, but this one is pretty common.
First, the "God's most unloved jumper configuration", and then my instinctive reply of WhoopWhoop to the ICP reference, Great episode!
Even if the 'Little guys' don't work ultimately, you still make the journey interesting. 👍 Thanks Gravis
God. The amount of digikey bags I save because "I'm going to use that for something" but end up just storing them all in a bigger digikey component bag....
I absolutely love this series
Thank you; I honestly had never encountered that pronunciation of Ont-air-io. On-tar-io; makes complete sense that someone could find that sound but I had never thought of it. 🏴
i honestly love that you explain electronics and engineering concepts even if you think most of your viewerbase already knows it. i find myself learning something new in every video and names for things i only tangentially knew about by taking apart shit out of boredom.
ALSO I NEED THE ICP ELECTRONICS LOGO ON A SHIRT
Wasn't having a great day today, seeing the Packard bell logo redone to represent ICP made it substantially better. ❤
thank you for your professionalism despite the disappointing computer.
My Sweet Summer Child, you don't know what an Extended Partition is? I envy you
It is a container for extra partitions, just take a look at the start and end clusters.
What they were dconfused by wasn't extended partitions as much as the error message not really indicating what it was about and instead saying something about memory I think?
I do wanna say, even if the video is "boring" honestly I just enjoy being here for the ride. I love both styles of vids, your more "I'm presenting a topic" and "I'm at the workbench", both are just good to vibe to and maybe learn some things along the way
Like cursed pin jumper configurations
I had a WTF moment when you read where the little guy was from. I’ve been to that Tim Hortons hahah
really enjoy the Little Guys series good, bad, or ugly there's always something to learn.
Not boring at all. We love you, Gravis.
Can you imagine being the guy watching this being a 2010 Fanshawe College Alumni?
43:06 MBR partition table is limited to four partitions; the type "Extended" is just a hack to add more partitions.
@@doq yeah, but you'd think mount would understand that and give a meaningful error
@@doq yeah, but you'd think mount would understand that and give a meaningful error
@@CathodeRayDude Probably so, but Linux is Linux and I bet mount actually _attempted_ to mount that instead of seeing it and "oh, this isn't a *real* partition, let me tell the user that"
It tells you where to look for meaningful info. ;-)
What happens when you blindly try to mount something is both well documented, and quite unpredictable. mount tool knows little about filesystems that potentially exist on the third planet from the Sun, or are supported by a given system at a given time, it relies on kernel and filesystem drivers. If type is not defined by the user, it tries each available filesystem driver in a certain order (see option -t in the fine manual). After ext family and other filesystems with well defined metadata formats decline to work with that device, basic ones try their best (e. g. raw floppy, maybe some raw flash fs with hard-coded sizes and offsets for embedded use). It is totally possible that one of them finds some arbitrary bytes, interprets them as filesystem size of 555 petabytes, honestly tries to allocate enough memory to work with such a giant, then fails.
mount is traditionally for administrators who know what they mount. It may not succeed with a perfectly fine partition just because of some missed critical option (even though ext family should be generally backward compatible). A full-featured graphical disk/partition management utility from any general distribution should show you both partition metadata and filesystem metadata before mounting anything. There is also an option to ask testdisk to perform an analysis if you're unsure.
This was mostly for curious readers, I'm sure our host is no stranger to manual calculation of disk offsets.
@@DarkKnight32768 oh right, of course, I was thinking at the wrong abstraction level. Mount is unaware of the partition table because you give it the block device representing the partition after it's been carved up by the OS. Presumably extended partitions do not contain a magic number that would make sense for mount to look for.
Was just thinking about not having seen a Little Guy in a while, and here we are!
Did the filesystem say Splashtop Business Linux on it? We have a crossover episode!
the Amiga 600 and 1200 are suited with 44pin headers on the mobo for 2,5" drives via a ribbon cable.
back in the days we often modded those from 2,5" formfactor to 3,5" by splitting the cable or via an adapter to fit a bigger and mostly cheaper 3,5" drive, cuz 2,5" drives where pretty expensive back then.
43:10 it seems to be a mostly sensible partition table for this kind of device, from what i can see:
sdb2 -> / (system root directory)
sdb3 -> /var
sdb5 -> /usr
While sdb10 i _think_ would be /root (the root users "home"), but it's possible it could be /opt (though that's more likely the 75gb volume), either way the rest should nest below it like so:
sdb6 -> /sdb10/incoming
sdb7 -> /sdb10/bin
sdb8 -> /sdb10/lib
sdb9 -> /sdb10/etc
With these partitions containing the bare minimum utilities necessary to remote in and configure/diagnose the device, seperated to be less prone to breakage from changes and updates
The 75 GB partition is the "Extended partition table" which contains all of the remaining positions. The MBR extended partition scheme is another one of those historically significant horrible kludges that it's best not to dive into right now; but suffice to say it is a horribly fragile hack and best avoided for all but legacy applications.
@@BrendonGreenNZL ah, right, been so long since I formatted a drive with an MBR scheme I forgot about that 🤦
"Little is a philosophy not a physical fact." Excellent :3
ZIF sockets were around since at least socket 1, at least for PCs. So... much earlier than the Pentium 3. Socket 1 was the same 169-pin socket that 486 overdrive CPUs used but enhanced with a ZIF+lever locking.
EDIT: I see this is in your pinned comment. Nevermind!
"Gods least used jumper configuration" is absolutely entering my dictionary
You're absolutely right in erring towards uploading a video. I enjoyed this video.
Apparently Via C3 had integrated MPEG decoder which is why they might have chosen it over Intel CPUs.
Ontario' s second syllable is pronounced like tear, as in tearing a sheet of paper.
i thought you meant tear, as in crying, at first, and i was SO confused
28:42 - years of working for NCR support (COMPRIS systems) confirms this. We had backup systems, and when the restaurant was closed, we'd remotely log in and troubleshoot whatever broke, never during opening hours (unless it was a main KVM or the host system)
The hue of that Dell laptop is astoundingly beautiful LOL
There were ZIF sockets starting with the 80486. Before that there were many 80286 CPUs in the LCC (Leadless Chip Carrier) package that had a single row of flat contacts around four edges. The CPU was held onto the socket, which had spring contacts, with a metal clip and a small finned heatsink. Not that the heat sink was actually needed, except possibly with the rare 16 Mhz version. Most of these 286 systems I encountered were 12 Mhz.
I'd call that 286 style the great grandpappy of the Land Grid Array CPUs and sockets.
The ZIF socket was actually invented way back in the 386(!!!) era, though only commonly used starting at the end of the 486 era with Socket 3.
Not surprised those caps gave up the ghost and took other power regulation components with them. That board was manufactured right in the middle of the capacitor plague era. I seen dozens of mainboards from the later Pentium 2 boards all the way through the Core 2 days die from bad caps that often times took power regulation components with them. So there was no way that board was going to work even with the caps replaced.
That jumper is gonna give me nightmares.
With caps you can go with a higher voltage and you could probably get away with a slightly lower capacitance but the correct value will always be best.
The most important characteristic in this case is that they _must_ be low ESR capacitors, else they will not work at the frequency they are being operated at. Power supply circuits, as a general rule, have the capacitors specified at twice the design value, to account for deceased capacitor performance due to ageing; so as long as you don't go too much higher or lower than what's there, the voltage regulator ought to work just fine.
Like you said, the Wowstick works great for driving lots of small screws that aren't stuck or threadlocked. But reef on it like it's a boring unelectrified screwdriver and the gearbox will break. And that's why I bought it twice.
Also, that power supply looks like it came from an external SCSI drive enclosure.
@43:00 sdb4 didn't mount because that partition is an MBR extended partition than includes all the partitions after it. Take note of the size of sdb4, which roughly is the sum of the following partitions. This exists because MBR data structure had a limit of partitions of 4 originally, but it was later extended by producing a fake, unmountable partition containing the extra partition's metadata. No ideas on the weird linux mount error though
Take a look at an MBR extended partition table in Norton Disk Editor sometime. What you find will utterly horrify you; especially when there are this many partitions involved.
18:18 Now that you showed us the cursed jumper config you totally have to try it and show us Reduced Power Spinup Mode! Gaiden! Gaiden!
Power screwdriver recommendation: Vessel Ball Grip.
They're about the size of that Whia insulated driver you used to remove the HDD screws and have a reasonably ergonomic slide switch. They come with two different gear reductions, the gray handle is higher speed and works well for general computer screws into metal, if they're tight you break them loose by turning the driver, then speed them out with the motor, but screws into plastic often have more drag for the entire length of the thread, the low speed black handled driver works better for those. Bonus: They're made in Japan and come with a rather nice #2 JIS bit.
I need to know the story behind the nail polish on the DELL keyboard