Fixing and improving the IBM PC/AT 5170 BIOS (bonus shorted tantalum fix)

The IBM PC/AT 5170 has some issues with the stock BIOS that causing boot time errors and also prevent you from using a IDE hard drive/CF card/SD card with the machine. Normally people replace the BIOS with a more modern AMI or Quadtel BIOS.
In this video I talk about how to fix the issues with IBM's BIOS so you can run original IBM code (albeit patched) to maintain some of the machine's original purity.
--- Video Links
Download the patched BIOS and write up:
archive.org/details/ibm-5170-...
IBM PC AT 5170 Series:
Part 1: • IBM PC/AT Model 5170: ...
Part 2: • IBM PC/AT Model 5170: ...
Part 3: • Fixing and improving t...
Virtual Box 6.1 running MS-DOS 6.22 and IBM MASM 2.0. Used this to compile the code and generate the BIN (without correct checksum)
www.virtualbox.org/
Managed disk images with WinImage 10.00
www.winimage.com/
Used HxD Hex Editor 2.4 to do binary file editing, and perform the 8-bit checksums correcting the last byte
mh-nexus.de/en/hxd/
Used RomWak to split/merge the binary files into odd/even version
github.com/freem/romwak
RomWak was complied on the PC using TCC 0.9.27
bellard.org/tcc/
To run a virtual 5170 I used Mame 0.207 (version is important)
www.mamedev.org/oldrel.html
Adrian's Digital Basement ][ (Second Channel)
/ @adriansdigitalbasement2
Support the channel on Patreon:
/ adriansdigitalbasement
-- Tools
Deoxit D5:
amzn.to/2VvOKy1
store.caig.com/s.nl/it.A/id.16...
O-Ring Pick Set: (I use these to lift chips off boards)
amzn.to/3a9x54J
Elenco Electronics LP-560 Logic Probe:
amzn.to/2VrT5lW
Hakko FR301 Desoldering Iron:
amzn.to/2ye6xC0
Rigol DS1054Z Four Channel Oscilloscope:
www.rigolna.com/products/digi...
Head Worn Magnifying Goggles / Dual Lens Flip-In Head Magnifier:
amzn.to/3adRbuy
TL866II Plus Chip Tester and EPROM programmer: (The MiniPro)
amzn.to/2wG4tlP
www.aliexpress.com/item/33000...
TS100 Soldering Iron:
amzn.to/2K36dJ5
www.ebay.com/itm/TS100-65W-MI...
EEVBlog 121GW Multimeter:
www.eevblog.com/product/121gw/
DSLogic Basic Logic Analyzer:
amzn.to/2RDSDQw
www.ebay.com/itm/USB-Logic-DS...
Magnetic Screw Holder:
amzn.to/3b8LOhG
www.harborfreight.com/4-inch-...
Universal ZIP sockets: (clones, used on my ZIF-64 test machine)
www.ebay.com/itm/14-16-18-20-...
RetroTink 2X Upconverter: (to hook up something like a C64 to HDMI)
www.retrotink.com/
Plato (Clone) Side Cutters: (order five)
www.ebay.com/itm/1-2-5-10PCS-...
Heat Sinks:
www.aliexpress.com/item/32537...
Little squeezy bottles: (available elsewhere too)
amzn.to/3b8LOOI
--- Links
My GitHub repository:
github.com/misterblack1?tab=r...
Commodore Computer Club / Vancouver, WA - Portland, OR - PDX Commodore Users Group
www.commodorecomputerclub.com/
--- Instructional videos
My video on damage-free chip removal:
• How to remove chips wi...
--- Music
Intro music and other tracks by:
Nathan Divino
@itsnathandivino

[@desmur36]

This video was amazing! Being a programmer, this combined everything I love about retro computing- the detective work (and learning) is where the fun is!

[@chrish9482]

Watched both your videos on the 5170 and DEFINITELY can relate to your frustrations. It was my first EVER PC bought for me in 1985 by my Dad, who worked for IBM, when I was 19. With Proprinter, EGA Adapter and Monitor, 10MB HD, and a bunch of software all for $5500 at the 50% employee price :). I loved the thing and it was "State of the Art" in 85. Because of all the annoying glitches and limitations tho I learned a TON tearing it apart, upgrading it, especially breaking it, and putting it back together. Much more than ever actually doing anything productive with it. As a result, I've been a PC tech all of my life. In 2015 I realized it was coming up on the 30th anniversary of getting the AT and was feeling nostalgic for it, so I decided to try to rebuild it EXACTLY to spec with all ORIGINAL parts. I was suprised how easy all the parts were to find 30 years on and more surprised to be be reminded how frustrating the glitches and limitationd in the BIOS were. I've been working in countless ( both long and short) spurts since then to tackle them. So in over a 35 year span I've spent literally a TON of hours working on the AT, and may be able to offer insight as to why those glitches and limitations ( not all of them technical ) are there. Also I HAVE been able to accomplish what you were trying to do ( 1.44mb floppy drive.. Compact Flash and SD bootable readers etc...without error codes ) WITH the original IBM BIOS.. ( no alterations ). TRUST ME.. not claiming to be any smarter... just more stubborn and wasted far more hours. If you'd like to hear how and my other insights on the AT reply back and I'll share them. For now I've been long winded enough. Either way Bravo to you and Stuart for so quickly identifying the sources of some of the EXACT same glitches and problems I've run into, and your very smart solution to them . My bottom line on the 5170?... It is a stately... majestic... beautiful BEAST with a more stubborn... quirky personality than almost any other entity I've ever encounterd, either silicone or carbon based... Just my 3 cents... :) Regards

[@BertGrink]

Stuart is a true Legend! Not only is he incredibly generous, but he´s a Code Wizard to boot (pun intended!)

[@thedungeondelver]

BTW, 601 is half of 1202, the error that came up on the DSKY when Apollo-11 was about to land (basically a data overflow error). Also it's the error that causes the central computer in the original film _The Andromeda Strain_ to crash when it tries to calculate the number of possible mutations that "Andromeda" could take.

[@TranceTrousers]

Everytime the error codes were mentioned I kept thinking of the Apollo 11 '1202' alarms that kept flashing up on the descent to the moon's surface.

[@TheSulross]

which, of course, implies we're in a simulation and everything is scripted

[@adampope5107]

@@TheSulross you want to check out something even trippier than the simulation idea? Search for the Boltzmann brain by pbs spacetime

[@pipschannel1222]

Some excellent BIOS fixes there! So, no overclocking to a blistering 8MHz on this one Adrian? It has a socketed crystal and the original i286 CPU can easily handle 8MHz, so it's easily accomplished and just about everyone with a 6MHz 5170 did it back in the day so it will remain period correct, but it'll give you a nice 33% speed boost 👍 Do it Adrian! Join us at the overclocking darkside!

[@jriccis]

And now that the BIOS is on the bright side of the force... You can try to modify it to remove the code that prevents overclocking the 5170 beyond IBM's 8MHz limit... I bet you can do 10MHz. Maybe more? That would be awesome!

[@sweintz]

he has a type one motherboard here (the ones with the 128K piggyback ram) type 2 and type 3 motherboards (with 2 banks of 256k chips) had a socketed crystal, but most type one boards did not. simple enough to desolder it though. Fun fact, the type 2 and type 3 motherboards are the exact same board with a different crystal, cpu, and bios. WHy different bios? well, the type 2 board bios actually had a timing loop in it to prevent users from swapping the crystal and overclocking it themselves. (type 2 was 6mhz, type 3 was 8mhz). Both 6 and 8mhz versions were sold during the same time, and the cost difference was significant. If yyou tried to overclock a 6mhz type 2 motherboard without swapping out the bios you would get a bad bios beep code if i recall correctly.

[@studmeister1977]

To think, 40 years ago, IBM fought to prevent everyone from editing their BIOS code. This is what Open Source is truly about, mustering the community to fix problems with an original design. WTG Adrian. :)

[@nickwallette6201]

Did they? They weren’t too keen on having clone manufacturers copy the BIOS, which is why it had to be re-implemented with a clean-room approach. But as far as editing the BIOS yourself ... they _gave you the source code_ ... *with comments* ... in the manual. :-) I think IBM’s stance at that point in time is that you should know exactly what it is doing, and to make changes if you need to and are so inclined. That goes for the code and the hardware, since schematics were in there too. Heck, the “game port” card, which they called the analog input card, had a big old wire board area built in for prototyping, so you could read arbitrary analog inputs and do whatever you wanted to do with them.

[@harleyarmstrong5947]

@@nickwallette6201 If I recall correctly, my original IBM PC had the BIOS on a ROM chip without a socket. I suspect that would be done more for cost than anything else, I suppose we could have copied some of the bios routines and modified them into our own programs that ran on PC DOS, but not sure how modifiable the BIOS and its POST logic would have been without reworking in a new chip.

[@nickwallette6201]

@@harleyarmstrong5947 Yeah, would likely have to hook into interrupts and bypass the ROM. But, I would lean toward that being a cost-cutting measure, or perhaps even something silly and pedestrian, like they had a supply chain shortage of sockets. Afterall, they did sometimes make available BIOS upgrades -- so they were aware of, and planned for, the need to replace the ROM ICs regardless.

[@kaitlyn__L]

@@nickwallette6201 so that ISA card with breadboard on that was unpacked a while ago now wouldn’t have been necessary there.. pretty sweet. I’m reminded of people hooking up thermistors to a C64 or a Sinclair’s port and writing software to calibrate and making their own homemade digital thermometer. I wonder how many of IBM’s customers actually did that. Not that many kids buying Usborne books AND original IBM PCs… still neat to know though.

[@marlenestockton2269]

'Open source' anything is a fail from the 'Get Go'. Linux (a big open Source fail) began in 1991 and really is a failure at 30 plus years later. If there is no incentive to innovate (money/profit) things fall to the side of the road. Microsoft, IBM, Honeywell and all others would have failed if 'Open Source' were their business plan. Nothing would have been done or completed and that's the truth. Ever look at the latest Linux code and its documentation? Almost shockingly bad. See what 'free' gets you!

[@VincentGroenewold]

Ok, this is a very cool one, not going to settle for the easy route but tackling the original problem. Very nice Adrian and especially the Assembler wizard!

[@8o86]

Note the BIOS was not disassembled and commented. It's just scanned and OCR'd from the IBM manuals (with some manual tweaking afterwards).

[@daghtus]

Stuart is a true retro geek king. Good job both of you.

[@gbotti82]

I love the startup of those old machines. Thank you for getting in so much detail and your great videos.

[@Jody_VE5SAR]

This type of code editing takes me back to the days of doing the exact same thing to remove copy protection from C64 games. :-)

[@BertGrink]

I have done similar hacking on the Spectrum +3 (the one with the disk drive) in order to be able to load games that were originally on tape, from the disk drive.

[@patrickbetts5504]

Amazing work Adrian and Stuart! Really love this kind of low level debugging!

[@DaveGagliardi]

What an amazing follow up! Thank you for keeping these vintage machines running! They are a fascinating look at the history of coding and computing in general.

[@saifal-badri]

This is such an amazing job, thanks for both of you for sharing the result of these fixes!

[@warpedmetalhead]

How frickin cool. Looks like I now have a fun mission for today! Thanks so much for your contributions to the community!!

[@exidy-yt]

I couldn't believe it when that controller card shorted out! wtf was wrong with the tantalum capacitors of the mid-80s that IBM used? Fascinating video as always, Adrian!

[@janpedersen9120]

i love this sense of fixing, old computers and the habbits we tried to workaround for years ;) thanks to Adrian and his followers for doing all this fine work ;)

[@megaglowz8540]

Thank you guys for allowing me to fetch my old 5170 from the closet lol. Gonna play with this stuff all weekend!!

[@stefanocrespi5424]

I saved all the parts and saw them one after another in a 5 hours marathon. What a show.

[@renatoamaral8259]

You are a genius! I admire the way you fixed the bios problem! A+ work. Congrats!

[@OldmanGamerYT]

Love the t-shirt! I had a Tandy TRS-80 back in the day and later a 1000 TL/2. Brings back memories.

[@VenomStryker]

Stuart seems like a really smart dude. I would have no idea how to do something like this unless it was explained to me like in this video.

[@patbreen3859]

Excellent episode. I think getting IBM BIOS going is really the key to whether this system passes the purity test or not. Using the a 3rd party BIOS (like an AMI or Phoenix) really would have this system basically like a clone. Nice work Adrian (and Stuart)!

[@RetroJack]

Great stuff - another great watch, as usual!

[@wildcat189]

Wow that series was a journey!

[@rpeetz]

Just what i needed for my coffee break

[@SlyEcho]

Patching the BIOS would have seemed such wizardry back in the day but now it’s very understandable.

[@thedungeondelver]

ADB making a terrible day go just a little bit better. Love the vids, Adrian.

[@senilyDeluxe]

I just wondered for a second what Apple Desktop BUS has to do with this... my stupid brain...

[@thedungeondelver]

@@senilyDeluxe Hah! When I typed that I thought, hey, ADB = Apple Desktop Bus, too :)

[@MrVipeg]

Thank you, Adrian! This video is really, really great!

[@Rx7man]

I had modified a JNE->JMP in a binary to bypass a key check way back when.. was kinda rewarding to have figured it out.. it was for personal use only of course :)

[@youshouldjustknowit8583]

Thanx Adrian, another great video from the basement. Before this video I would have considered every Bios error code as definitely a hardware problem only.

[@AndrewTubbiolo]

I've been hoping for a open source PC architecture for decades. Lovely to see more bits and pieces being developed.

[@pdppanelman5889]

Now this is really good. More like this please (Other BIOS problems in other computers)

[@itstheweirdguy]

Man this is crazy! I found a forum that had scripts that I was able to use to create my own BIOS'es for newer AMI APTIO bios systems, which was cool. This is miles ahead of that! On my end I was just updating ME firmware, CPU microcode, adding PCIE NVME support. In your case you are doing things never ever considered with those BIOS versions...

[@elmariachi5133]

Nice you are searching the true retro way, now! :)

[@Vermilicious]

Fixes long overdue. Neat.

[@janman]

Albeit very niche, this is an incredibly useful video. Thanks!

[@dregenius]

I love this type of content!!! Also, it reminds me I've always been curious about how Classic Mac "BIOS" (or I think properly, bootstrapper?) worked and I'd love to see any videos of people doing similar ROM racking and talking about how Classic Mac OS boots... if anyone knows of such videos, let me know!

[@Chriva]

More files added to the seed machine. Hope more people do it. Really useful stuff :)

[@TopVersnelling]

Thin THICK I appreciate your labelling

[@ddognine]

I remember my granddad who was a WWII vet and worked for IBM in the 50s, bought an IBM PCjr in the 80s. He then proceeded to tinker with the BIOS and write his own OS for that machine. He only knew how to program in assembler, and couldn't be bothered using the PCjr for stuff like games. I also remember him ranting against Intel and MS when he couldn't get a hold of the CPU architecture for newer PCs effectively shutting guys like him out. I also remember him snickering at viruses and secretly telling me how he could easily write a virus that made your monitor explode or crashed the head on your drive. He was a certified genius even though he never went to college. RIP Granddad.

[@1990chrism]

Disassembling the IBM bios, risky move xD

[@blockbertus]

Why?

[@ovalteen4404]

It wasn't disassembled. It was already distributed in source form by IBM in their Technical Reference Manuals. The "Reconstruction Project" just scanned it in and wrote it into .ASM and .INC files.

[@blitzwing1]

I would say it's more of a CISCy move to be honest.

[@TheSulross]

if the disassembly assembly code can generate the same binary image when assembled, then its a valid disassembly

[@krz8888888]

Lmao

[@fellipec]

This was crazy fun!

[@artursmihelsons415]

I think, that modifying BIOS is best way to go.. Love the results. Great work and video! Now I'm thinking, why I didn't thought about that earlier.. 😂 Because, I still have IBM main board with 8088 CPU that works (in Basic), but can't boot in DOS after recommended update to MS DOS 6.22.. Older DOS worked and booted fine from hard drive.. I assume, that BIOS is expecting some signature in in MBR, that original older DOS had.. Thanks for idea! 😉👍 It's assemblers time again for me..

[@thesteelrodent1796]

just an FYI: LBA addressing was retired a good 5-10 years ago when hard disks went beyond a terabyte. The replacement is ADF - Advanced Disk Format - which uses 4 KB sectors. These disks can obviously only be used with a modern BIOS that understand ADF, but some BIOSes (probably not any as old as the 5170 though) can be updated to a version that has ADF support, though the change notes will usually just say it supports larger disks. If you use one of these disks on a system that does not support ADF it should fall back to an LBA translation mode which limits the usable disk space to whatever the LBA limit is on that system. I discovered this when I put a 500 GB HDD in an old Toshiba laptop and it told me that I had to update the BIOS

[@JustinEmlay]

Silent touch on that keyboard is hilarious in a day and age where people try to get the loudest keyboards ever.

[@adriansdigitalbasement]

HAHA! Yeah it's a cheap keyboard but I have to say I kind of enjoy the tactile feel + the silent typing. Sadly whatever makes it quiet seems to wear out rather quickly, as I have two of them and they are both getting more noisy as they wear.

[@stamasd8500]

Adrian, one trick I have come up with in dealing with older machines in which I want to make BIOS changes (or add BIOS extensions) is to use the boot ROM socket on ethernet cards. If your NIC has a boot ROM socket, you can put the XTIDE BIOS in there and it will work.

[@FPVphilly]

The IBM 5170 was my father PC when I was a little kid

[@MonochromeWench]

Debugging and fixing a race condition in someone else's assembly code Something most coders would run away crying from. Amazing work by Stewart. I wonder at what clock speed the original code starts to fail at when using an original mfm hard drive. I doubt anyone would want to underclock an AT just to find out

[@krz8888888]

Outstanding stuff

[@TR19]

On this episode of Antoine's analog attic

[@ConexSpot]

So helpful, thanks!

[@phantom2012]

A milli-ohm meter is awesome for finding which cap is bad. Even trace lengths can be figured out with one.

[@wojiaobill]

dude that tshirt is awesome

[@organiccold]

Nice Job Adrian... Now jumping to the 2nd chanel :)

[@c.m.1537]

I'll be honest , I put a BIOS chip in a fancy 80286 AT/Clone many moons ago that would take bigger drives or something. I watched the label melt off and smoke came from it. Gone in a flash. Expensive chips. I was a stupid teen though. This is some serious time warp machine fixing / upgrading. The IBM PS/2s are of more interest to me though. nice work.

[@ropersonline]

To think that almost 37 years later, IBM's original AT BIOS is working correctly _for the very first time!_ EDIT: Second thought, _and having actually watched the 601 error part :D,_ I think "working correctly" might be overstating it, because the unconditional jump is more of a hack than a fix. Not to be anal about it, but that's not the same thing as understanding exactly what they were trying to do and why this was failing even on original or 100% compatible hardware. I still applaud this, but it's a hack not a fix. The same is true for the 162 error. The hard drive NOT_BUSY code is an actual fix, IMHO.

[@kcinplatinumgaming2598]

my first ever pc at .. i miss that heavy chunk of metal ... :D truth is i had the motherboard in a bag in my cupboard for over 20 years and decided to sell it on ebay sold as seen and the buyer said he tested it and it still worked .. omg lol it went to a good place.. just shows the hardware of today does not last as long .. i glad i did take that battery off because of the chance of acid leak ... i glad i made that decision...

[@frdsofficial]

The harder the project the better!

[@brianlhughes]

I worked with a guy that programmed assembly code on mainframe IO routines on a custom version of Honeywell GCOS. I did lots of assembly code on a standard Honeywell back then. I bought an Atari ST because I studied the 68000 before buying and really liked the instruction set and 24 bit addressing. I haven't coded assembly in 30 years! I miss it but I refused to try it on my first PC clone purchased in 1993 because of the segment registers.

[@gato38]

In my IT career and hobbyist side of things , Notepad++ has saved me so much time over the years its amazing. I love it. Nice to see someone else using it.

[@alerey4363]

if the bios check for the original controller card was something exclusively IBM-branding thing then there's no problem in bypassing it at all; but if that check is part of a generic functional test (regardless of a closed-branding thing) then there will be side effects to bypassing it and you will find them in the future while using some feature of the hardware (i.e. in the middle of a game) by Murphy's Law

[@chrish9482]

Have over 35 years experience with the 5170 and there are several possible reasons for the check in the bios ( including both possible reasons you state ) My guess is more on the branding side. In learning a bit of the history of the 5170 development, it seems IBM reluctanly "threw in" support for 1.44 floppy drives into the bios of the Type 2 and Type 3 motherboards in the AT, because of pressure to keep up with the industry but I'm 99% certain they never offered a 1.44mb drive of their own brand for the AT at the time, so they were only covertly supporting it. This is probably the reason why the 1.44mb floppy drive option doesn't appear in even later versions of the AT Diagnostic/Bios Setup Disks. If memory serves, I believe the specific controller card its looking for are IBM original ones that have specific Western Digital Chips on them. IBM may have thought ( possibly through internal testing ) that these controllers were the only ones that either supported 1.44 drives reliably OR at all.. in the AT. I believe this is only half true but that's only through alot of trial and error experience, not any forensic technical analysis.

[@marksterling8286]

Had a similar problem with an AT about 12 years ago. Used an adapter aha1542 maybe aha1642 scsi card then had an adapter from scsi to compact flash card.

[@VidaDigital]

In that regard the 1542 was doing the same thing as the XT IDE replacing the hard drive routines in this case for SCSI

[@marksterling8286]

@@VidaDigital it was doing exactly that. So the AT though it was just a scsi disk with a scsi card with its own bios. Then the pcb converted standard scsi to CF. It was quite a find back in 2008/9. Still working running sco xenix 286

[@marksterling8286]

Actually it was easier to install sco on to a scsi disk

[@chris-tal]

Hand soldering tantalums in reverse polarity is very common I think. The fact that most components are marked on their negative side vs tantalums on the positive generates some human error. :)

[@DesertRaven365]

9:02 - please put that one TDK tape the same way up as all the others, my OCD will thank you forever!

[@elbiggus]

Damnit, I hadn't noticed, but now it's all I can look at!

[@jaycool428]

Yes, it's bugging me as well for months.

[@runeodin7237]

Just think of it (the reversed tape) as a piece of art

[@kaitlyn__L]

@@runeodin7237 it’s a deliberate statement about Life .. :)

[@chriswareham]

This sounds like a number of changes I've had to make to code much more recent than this IBM BIOS code. The code reports error conditions that make sense when the it was originally written, but changes to hardware make those conditions no longer valid in all situations. All too often the errors just get ignored, but other errors are then ignored along with them... I guess that's one definition of bit rot.

[@sweintz]

even when set to secondary, i believe the IBM controller still uses IRQ 14. The jumpers only change the IO base address. So you have an interrupt conflict with the IDE card.

[@alexandrecouture2462]

Sometimes sd to ide adapter works well in computers freezing with cf to ide.

[@LeftyPem]

That silver TRS shirt is awesome! Where can one procure such a gorgeous garment??

[@Dukefazon]

29:15 - I was watching in 1.5x speed and this transition was strangely interesting :D

[@awilliams1701]

It's really cool you were willing to do all this, but at the end of the day, I'd rather stick with a non crappy bios. The IBM one seems like junk. lol Honestly I would have stuck with the better machine though.

[@lauram5905]

So for the people in the know here, is the HDD issue a race condition?

[@tw11tube]

Yes, it is. You might want to read how I analyzed and described the problem in kzbin.info/www/bejne/gXnLdpedmZymns0&lc=Ugwf7gnYFwD0lBASv3h4AaABAg

[@Zontar82]

i had that same bios on a machine i got from a warehouse of office pc, and man is it really the worst just to navigate

[@jyvben1520]

V'ger in the making ...

[@NielsHeusinkveld]

Stuart is pretty handy for a guy who used to run a comic book store!

[@mal2ksc]

Worst. BIOS. Ever.

[@enginerd80]

Was it found out why the “too” fast hard drives were a problem with the IBM BIOS, but apparently not with other BIOS’es? Did the others already work the way like the IBM BIOS would after your modification? Or did they do something completely different?

[@absalomdraconis]

I would suspect that the other BIOSes were accepting interrupts before they issued the command to the drive, while the IBM BIOS was probably only starting to listen after the command was issued.

[@cheapasstech]

Adrian, where can I buy the CMS ic’s ? As my 8bit soundblaster does not have those on board .

[@mal2ksc]

If you're going to hack the BIOS at all (which is justified, given the IDE read timeout), you might want to go back to the third-party BIOS but with all the splash screens and messages stripped out and/or altered to look like the equivalent 5170 routine. Then it will still look stock and can quickly be reverted to the original state, but secretly have all the advantages of later 286 machines. I'm not sure what this would do as far as trying to share your work, though.

[@SparksNZeros]

Take a shot every time Adrian says 'Actually' ;D

[@pkhoury1212]

Now we just need a fix for the rarer XT 286 (5162). I prefer that over my 5170 any day of the week.

[@EsotericArctos]

Trying to keep it pure would also be not using hard drives that were not available at the time the PC was made. IF you want to use modern hardware the machine was not designed to utilise, then you need to use more modern ROM's. Any modifications you make to stock ROM, even if it is just changing conditional to permanent, is NOT keeping it pure.

[@kompzec]

Which eprom chips did you use to program the new bios to?

[@scharkalvin]

Didn't IBM publish the source code for the PC and AT bios in the USER manuals? If running MSDos version of the assembler (not windows), then can use DOSBox utility to run it.

[@fortmax8370]

Technical Reference Manuals, but yes. Says so on the site shows: they scanned and used OCR on the printed BIOS code in the manuals.

[@jazbell7]

I had one of those on my desk in the mid 90s. It's only purpose was to manage the company telephone system. My job was to work on industrial control hardware and software and manage the telephone system. The PC guys hated this machine. I was glad to walk away from it after a few years.

[@retropcdurham]

I wonder if those error codes were put in place to call out if you didn't have the approved IBM options installled. They knew they couldn't prevent you from using some else, but they could make it look like it wouldn't work

[@michaeljuneau4085]

Splitting this ROM file into even and odd required a 3rd party utility? I managed to do it with the xgecu tl866 pro software in its features for Mac Plus ROMs, no external utility needed.

[@Angelgreat]

28:03 It can run Planet X3!

[@TheSulross]

which is now sufficient reason to not send it to an e-waste center

[@stephendouglas684]

Maybe add a POST message regarding bios mod and mention Adrian and Stuart?

[@technattic6409]

I thought that MFM and IDE harddisk controllers used the DMA controller, but it looks like the harddisk routines of the 5170 isn't using it at all?

[@ESDI80]

That probably explains why the IBM HDD/FDD controller could only handle a maximum interleave of 2:1 and the Advanced Diags would format the hard drive with a 3:1 interleave.

[@AntonyTCurtis]

Hmmm ... could be worth thinking about modifying the 286xt bios so that it would tolerate running at 8mhz.

[@knghtbrd]

Better solution to the checksum problem: Write a little Python tool that calculates the checksum and generates a ROM with the checksum before using minipro to write that ROM. I think you can get minipro working on Windows, and it should run alongside the Windows minipro software (maybe not while the windows software is running), preserving the BIOS checksum feature while also giving you the convenience of not having to remember to fix the checksum over and over again.

[@rickyvanass7192]

Wasn't there room in the bios memory for your desired floppy disk swap?

[@RockyMountainWest13]

Is this the model John Titor was after?

[@tech34756]

How do you decide which capacitors to buy? Do you just go to somewhere like Digikey and buy a load of them or do you just buy a kit somewhere? If it’s a kit do you have any recommendations? I know they’re available on ebay but you never know with these things in regards to quality.

[@adriansdigitalbasement]

Sadly no kits -- I just buy from Digikey and yeah, I just looked at several boards and ordered an assortment. For regular caps -- 1uf, 10uf, 100uf, 1000uf, 4.7uf, 47uf, 470uf, 2.2uf, 22uf, 220uf and 2200uf -- those seem to cover a lot of things. For tantalums -- yeah kind of a range. Honestly the exact value usually isn't too important. Just the voltage must be at least the same or higher.

[@tech34756]

@@adriansdigitalbasement Thanks, wasn’t sure what values to have to hand.

[@Scoopta]

13:14 IO ports are not memory and are an entirely independent and separate address space from memory

[@elmariachi5133]

Why can't you connect the sd adapter to the big original IDE/Floppy-controller?

[@garybalanesi610]

I wish I was smart enough to resolve issues, regarding bios, machine language and the like. And I'm not challenged...Retired com tech...

[@adamreavis8214]

I have an old Super Socket 7 motherboard (FIC VA-503+) I have recapped. It has a cracked inductor that has no markings. Is there a way to identify a replacement?

[@The1RandomFool]

Didn't you have an error on the PC Jr. as well similar to this? I think the BIOS had to be modified on that to get rid of a printer error since it wasn't using one of the stock cards. It's like IBM didn't anticipate people changing out the hardware.