I got this rare 486 EISA/VLB mainboard

[@blakecasimir]

Another excellent repair. EISA is a rare treat, something I never experienced back in the day. Looking forward to the follow up VGA testing!

[@user-qf6yt3id3w]

Yup. I read about EISA but by the time I got to upgrade it had been clobbered by PCI.

[@Sabz-ul2rc]

EISA was more prominent in multiprocessor systems of the time. They also required a special piece of software to configure properly.

[@JoeBurnett]

These videos always bring back memories of looking in in Computer Shopper at all the numerous motherboards for sale. Great channel!

[@Xaltar_]

Awesome video as always, what a great bit of rare kit! Can't wait to see how it handles the EISA cards and benchmarks.

[@PROSTO4Tabal]

Finally good old Necro. Welcome to New 2025 retro year soon

[@Stratotank3r]

So glad to see that my donated ATI Mach32 is still around and will put to good use in your next vid! This giant board is very special and should perform very good with a DX2 or DX4 Overdrive.

[@necro_ware]

Absolutely! Thank you very much once again for the card, it has a special place in my collection.

[@the_kombinator]

I would have sold it on eBay :P

[@necro_ware]

Well, I'm grateful for the donations, but whenever I know in advance what people want to donate I always honestly warn them how much that parts cost. I got things offered like Voodoo5 and GUS in the past, but I always refused. I only accept donations when people exactly know what they are donating and why they do this. In addition to that, sometimes it is an arrangement or exchange and not a donation.

[@the_kombinator]

@@necro_wareI have a leftover crate of heavily damaged 386/486 motherboards that even Adrian Black would find hard to rebuild. I'm also flying to Europe in a couple months, let me know if you want my *smieci*

[@ronny332]

You have a steady hand (renewing the traces)!. I get tasks of this kind done, but my iron is shaking a lot more.

[@necro_ware]

@@ronny332 so far, yes, but we all get older... 😔

[@matthewday7565]

If you're getting 5.1V after the diodes, it will drain the battery even with the power on - you really need a 3x 1.5V cell pack, anything from 3 stacked button cells to 3x AAA or 3x AA (overkill or what!) - or a 3.6V NiMH with the charge circuit - or a protected LI-ion pack with the charge circuit. Or throw an extra diode in series with the 6V pack

[@nickwallette6201]

Bear in mind, the "0.xx V" drop of a diode is _at a given forward current._ The very low current drain of the RTC circuit may not be enough to get all the way to the rated knee voltage. I confirmed this once when replacing the battery of a Compaq Pentium system that wasn't happy enough at 3V with a single coin cell. Two cells with two standard rectifier diodes in series was still something like 5.4V, which was high enough that the RTC chip was throwing errors on boot. Ultimately, I ended up ordering the lowest quiescent current buck regulator IC I could find and designing an adapter PCB to regulate down to 4.5V. I later found one that claimed even lower quiescent current boosting from 3V to 6V for a different computer. When you're talking about micro-amps, it's hard to say what'll happen until you build them both and see how they behave. Might've been a fool's errand and both are already completely drained. I haven't checked in a while.

[@wskinnyodden]

Hey, enable fast gate A20, that being off is a performance killer of major proportions!

[@tomschmidt381]

Great blast from the past. I was a fan of EISA back then. It was impressive being able to maintain backward compatibility and the industry's first attempt of plug-n-play. I had totally forgotten about the short lived VESA bus.

[@osgrov]

This is really cool. I've never had an EISA board, so this is exciting! Looking forwards to part 2. :)

[@sdavifcom]

Extremely interesting. I'm hooked. Can't wait for the follow up. Thanks for a great video.

[@Eireman_on_Twitch]

How funny, I had that same motherboard! I bought it at a PC swap meet for $10 with an included EISA IDE/Floppy/Parallel/Serial card. The motherboard was non-functional, but for a high school hobbyist (me) it was worth fixing. I was able to diagnose 3 bad compound resistor packs next to the power connector, and found replacements cheaply at Radio Shack. It came to life immediately on my Cyrix 33Mhz 486SX and a 2MB DIMM. It ended up with an Intel DX4 100Mhz, 4*16MB DIMMs, a Tseng Labs VLB VGA card, a no-name compound controller VLB card, a Pro Audio Spectrum 16, and was my multimedia DOS and Windows 95 machine for quite a while. I put the board (without battery) into storage, but sadly it was carelessly thrown away when my mother sold her house. My mistake for leaving it for 30 years! I think it was a motherboard made for Gateway 2000 machines, hence the lack of silk screen branding. I only say that because the seller was a Gateway 2000 remote warranty service.

Thanks

[@necro_ware]

Thank you very much too. Just bought me a coffee ;)

[@timballam3675]

I remember building/upgrading systems with that board back in the day.

[@vswitchzero]

Great video! I've always wanted to explore EISA and recently picked up a similar VLB+EISA board, but uses the HINT chipset. I believe its called the "NICE mini EISA". Interestingly, the HINT Caesar chipset uses a trimmed-down EISA implementation so there are some compatibility issues. I got it for only $59 CAD, but it has some pretty extensive battery corrosion, so hopefully I'll be able to repair it.

[@necro_ware]

Nice! Than looking forward for a nice video. Just as always ;)

[@stephm5587]

Very interesting video ! Thank you for such a great work !

[@ShadowRune]

Really cool board, it's always neat seeing these unusual ones

[@DotChaotix]

The book Open by Rod Canion has a few chapters about developing EISA. Its a very good read about the history of compaq and gang of nine. Great video as always Necroware!

[@necro_ware]

Thanks for sharing!

[@mcosta3810]

Wow. I have seen and used to have more ISA cards and also some VLB cards than i can remember, but today i learned about EISA that I've never even seen in person before. So this was a good and very informative video 🙂

[@stamasd8500]

I am very interested in the EISA utility that you used. My understanding was (I don't have much direct experience with EISA but would like to experiment at some point) that these utilities were very motherboard-specific, and that if you don't have the exact configuration program for the particular motherboard you're screwed. It's one of the things that prevented me so far in delving more into EISA. If a moderately universal config utility exists... that may change soon. :)

[@awd42]

I believe the Intel configuration utility will work on any system.

[@mrgatogrande4284]

I don't think they were motherboard specific. I tried with an utility that came with a lan card and worked perfectly with my mobo.

[@tonypitsacota2513]

Wow. I literally threw away 200+ of those boards into a dumpster using a wheelbarrow years ago back in the late 1990s at a University I was working at. Out with the old, in with the new.

[@Jivemaster2005]

Now that's an odd motherboard. My main 486 rig has a high end Everex 486 EISA motherboard with VLB slots housed in a Colani full tower case. I am using both an EISA NIC and SCSI controller in it. IMHO most of the EISA motherboards are built with better quality since they were used in servers and workstations. The ISA slots can often be mistaken for the rarer OPTI Localbus slots which look identical.

[@rra022001]

Yep ..had a few everex boards. Picked them up in Irving TX back in 1998.

[@vmisev]

Thanks for the video! Back in ~'93-'94. I worked with PC builder that specialized in higher-end workstations and servers. We did a ton of VLB (mostly with .de SPEA cards), a few of EISA for servers, but I newer saw VLB/EISA combo board before! It will be interesting to see reliability test with some EISA SCSI RAID and VLB graphics, with both cards simultaneously under heavy load 🤔

[@protox07]

Have a happy new year Necroware

[@necro_ware]

Same to you ❤️

[@rs-qq3os]

I look forward for the tests, the results should be very interesting.

[@ocsrc]

There were boards like this in the 90s I remember the most common one, it had about 50 different names silk screen on them, and they were sold by so many places Some brand name companies made machines with them They had the AT style DIN keyboard And they needed an ISA video card A serial printer card A ISA sound card A ISA modem An ISA floppy and IDE controller card I remember building computers using that board It was sold with a 386 40 MHz CPU and a 486 CPU 25 MHz

[@krizator]

Awesome! I like videos with rare pc stuff, like this mobo.

[@jasmijndekkers]

Hello Necroware. Happy new year 2025! All the best wishes for you. Greetings from Steven from the Netherlands

[@necro_ware]

Same to you ❤️

[@JaimeOrtega-i5r]

Had many of them back in the day. From what I recall not many 32-bit slots board were ever created for them, so the board and the standard went away.

[@ironhead2008]

That's a sick board. It'd be nice for benchmarking and comparing EISA and VLB cards if you can find them that is. EISA cards are hard as hell to get hold of.

[@glenncaughey5044]

I know!

[@matthuegera5096]

The rechargable battery removed was probably 3.6V (they usually are) and the button battery you tried was only 3.0V. The RAM was needing the extra 0.6V to retain backup power. You were just lucky both the RTC chip and RAM could handle a wider/higher voltage input, otherwise you could have fried them before checking the datasheets! Always check the datasheets FIRST before applying/testing a higher supply voltage!

[@helmargesel3972]

Thanks for sharing your knowledge

[@awd42]

I'd like to see you test EISA ethernet cards. While 100Mb ISA cards did exist, they were barely any faster than 10Mb cards -- nowhere near 100 -- but EISA should be able to achieve that. 3COM 3C597 was the model, I think. Fun fact: if you put the common 3C509 ISA card into an EISA system, you can set it to EISA mode with the 3c5x9cfg utility. No performance change, but I/O ports and IRQ will be assigned by EISA, so kinda plug-and-play before ISAPnP was a thing.

[@selami32]

This BIOS setup screen i wanna cry how nostalgic :') 17:37

[@VShuricK]

30:28 AFAIK 62xxx SRAM and 68xxx RTC drop in low power mode when power supply lower than 4V. So in my taste I will try 3*1.5V batteries, without charging circuit. I saw 3*AA external batteries modules in some old PCs.

[@mal2ksc]

I think a single 3.7V lithium AA would be enough, although swapping out the diode for one with a lower forward voltage drop might be in order if working that close to the ratings.

[@VShuricK]

@@mal2ksc good point!

[@FlipidusX]

Was für ein interessantes Board. Gefällt mir sehr gut!

[@necro_ware]

Freut mich, danke!

[@MrQuist125]

Super video sir

[@danielleblanc5923]

If I am not mistake the "barrel" batteries are 3,6v and not 3v nominal which could cause the voltage after the diodes to become too low (below 3v). I would try to increase the battery voltage with an external pack or change the diodes to models with a lower drop than 0,6v.

[@SHSPVR]

A truly rare motherboard is the Hauppauge 386 AT/16 and the Hauppauge 4860

[@erickvond6825]

I'd heard of E-ISA before but never seen a board or card that had it. Oddly the cards remind me of AGP which used a similar architecture for the card edge connector. Even PCIE uses a similar connector style. I know the spacing is different but the contact geometry is similar. It makes me wonder how much of AGP and PCIE owes it's existence to this antiquated technology.

[@tspawn35]

I am sure we will find out later. The cpu has a copyright of 89 that was the original run of 486s by intel.

[@lemon3rd800]

A friend of mine had gifted me a couple of his left-over computer boards where an EISA compatible board was included, with a Pentium CPU and some RAM on it. I even have two IDE cache controller driver disks (Novell, OS/2, Windows 3.10, UNIX 3.2) which seem to have belonged to it. The only part I'm missing, however, is the mainboard you showed in your video's thumbnail.

[@rra022001]

Had a few of these boards. Fun overclocking dx4s.

[@user-qf6yt3id3w]

Subscribed and clicked the Bell icon for the benchmark video.

[@josephnealeUKscratchcards]

nice repair and see you soon

[@madfox2000]

WOW!!! I Had one of these in the 90's!! I used an overclocked AMD 5x86/133 processor on this board and had excellent results.

[@JoCrt]

A silicone rubber abrasive bur may be gentler to the traces than that diamond bur which can remove too much copper in one pass.

[@BrassicGamer]

I really need to get a microscope for this kind of work as I have many boards I want to fix. I have a similar board to this with EISA and VLB, it's awesome. It has a weird problem with the serial port not working though (it was fine, then it just stopped working), so hopefully I can fix it with similar techniques to you.

[@K10driver]

What a unicorn!

[@alexthemorgan]

I had this same mobo, at least one very similar. I had a Mach32 VLB card in it, along with a VLB IO card and it was an excellent overclocker. I had a dx4 running with a 50mhz FSB and eventually an 83mhz Pentium Overdrive that was an excellent gaming machine.

[@Miltiadis_Vouzounaras]

Happy new year!

[@necro_ware]

Same to you ❤️

[@isilder]

The bios information may be wrong, due to similarity of the chipset they may have just replicated the bios from the 407 into this motherboard, even if this is not a Shuttle Holco HOT motherboard.... it might be prototype. Seems to be looking cheap when the expense of having VLB and Eisa would put it in a high price bracket...

[@LFOSyncToo]

"this one is thru" means "this one is broken" for those who do not understand what is said when necroware tests the traces...

[@mikes989]

I have a couple of ALR's EISA motherboards in the testing queue

[@JaimeOrtega-i5r]

WIth the exception of one 32 bit video card that also seemed to go away quickly.

[@Roxor128]

Heh, I see you've modified your multimeter probes to have extra-small tips for the fiddly work you do with them. Nice!

[@uwepolifka4583]

As I remember I used 4 x 1,5V battery packs. That gives ~5V after the diodes. This works if 3V is to less.

[@watchmakerful]

Typically 3, not 4.

[@misterkaye7796]

I seem to remember that they used 3 Nicad cells which would be 3.75VDC. Remember that it was a rechargeable.

[@goUkraine]

Wow. Blast from the past. I remember constructing 386 and 486 machines with OPTi chipsets, plus my preferred C & T chipsets that always seemed to work well. Constructed a couple of VLB machines, never had an issue I noticed. Never got to play with EISA, though. EISA seemed rather specialty at the time. "Server stuff" in my mind at the time, so skipped from VLB to AGP. Oh...TURN ON the Fast Gate A20 option. Unless you WANT to go slow.

[@tigheklory]

I wonder if you put the CR2032 batteries in parallel instead of series it would still work and be safer. That way are just increasing the amperage to handle the greater load.

[@djpirtu2]

Seems that you have exact same Benq TFT like I do. It's special one because it takes 15KHz signal from Amiga through VGA-adapter. 👌

[@necro_ware]

Yeah that's why I have it. I made a series of videos about this topic in the past.

[@djpirtu2]

@necro_ware I have totally missed those 😅

[@Pulverrostmannen]

I had the same voltage issue on my 286 where the RTC stopped running while on CR2032 backup. it had to be a battery above that so I used a 4,2V lithium pouch cell instead and it been working fine. and I can just charge that battery when it gets low too

[@BlackEpyon]

Oh daym! Never even heard of this before! Now what if they put the EISA slots in line with the VLB, and put PCI slots where the EISA slots are!

[@MonochromeWench]

Nickle Hydroxide from NiCd Battery leakage is greenish and can look like copper oxide and would be the main suspect if there isn't much exposed copper around

[@AK-vx4dy]

@17:08 In this moment you should relax some minds - yes you could put ISA to EISA socket :)

[@luispanaderoguardeno3306]

Nice ... I have a unknow 486 MB with VLB bus (and battery removed) that I need to know how are set the pins to use a external battery. That it was stooped me of doing interesting things with that computer.

[@EBIKEScott]

Great vid! Something was draining your new CR2032 battery... every time you checked the voltage it was lower and lower

[@FSK1138]

fantaftic find 486 was my fist intel computer cpu .. fun times watching defrag while playing mystery jumper 🥰🤓

[@BigDaddy_MRI]

I looked carefully at the resistor between the power supply +5vdc and the diode matrix. I looked very closely, and the resistor appears to have gotten hot. I would lift one side of that resistor and measure it. I'm thinking that if that resistor is open, then the reverse bias steering of the diodes would fail. When the board is powered up the 2 IC's you were looking at should have near to +5vdc on the Vdd pins, because the steering diodes will switch off the battery voltage and allow almost all of the +5vdc from the power supply to the two IC's that require it to function fully. You also may want to lift all the steering diodes on one side and do a diode check (front to back) and see if they are still working properly. Just a suggestion, but that will allow you to better understand the way the steering diodes are biased when no power is applied, and when power is applied. Also, do you know the battery voltage of the battery that was originally installed? Because it might not have been 3.3vdc, but higher, like 6vdc or just above, and then you also will learn more about the charging circuit and how the steering diodes are configured. The higher battery voltage (if the battery was higher than 3.3vdc) will change how the steering diodes are biased (from power off to power on) and how the steering of the diodes actually works. I do know this, the 2 IC's should not work via the battery when the board is powered on; the 2 IC's should run from the +5vdc when the board is powered on. Hope this helps. I really enjoy your channel. Well done!

[@itstheweirdguy]

Of course it needs 2 cmos batteries! It's 2x as awesome!

[@aintaintaword666]

21:50 won't the new battery interfere with the motherboard placement in the case now?

[@sebastian19745]

Why do not use 3 AA or AAA NiMH cells in series instead the onboard NiCd? They have a similar chemistry and are a direct replacement for NiCd. Also, they last a long time, use the same charging circuit, are cheap and can be mounted with the holder away from the motherboard. I think this motherboard deserves this mod instead the 3V Li coin cell that is unstable.

[@michvod]

Yes, that is probably the best option. Modern NiMH cells last 20-25 years without leaking, especially if they are kept charged. Else, I would use a modern lithium 3.6V cell. Or just modify the onboard CMOS circuit to completely lose the diodes, so there would be no voltage drop.

[@the_kombinator]

6:22 - so much hate for the VLB - I never had problems with it in the dozens of 486 and even 386 machines I had over the decades. I recall at 16 I built a DX 50 (yes, the not-clock doubled one) with a 2 Mb VESA video card and a 32 bit I/O card - not a single hiccup. Maybe I got lucky?

[@necro_ware]

No hate at all. In opposite, I like VLB builds a lot, but due to its design it is indeed a matter of luck. This bus was designed with 486 in mind, so the VLB slot is basically hard wired directly to the CPU. No buffers, no drivers in between, just direct connection. Up to 33MHz most systems are fine with that, but if you go over it you need to choose the hardware wisely. Having VGA and I/O in VLB at more than 40MHz is a russian roulette. Then it depends on many aspects, like how far the VLB slot is located from the CPU socket? How many VLB slots there are on the board? How many are populated? How good the VLB is shielded from the other parts? Then things like RAM and Cache timings and again, the physical distance of the memory/cache slots from the CPU, because that has to remain stable and in sync with the VLB bus all the time. A very high quality well designed 486 VLB board could run at 50MHz, but most probably only with one card, usually VGA, and by far not all boards offered that level of quality. Usually 386 and 586 boards with VLB were much more stable, than 486 VLBs. Ironically that was because VLB was not made for 386 and 586. On that boards the bus was bound through an additional controller to the CPU, which was more timing tolerant, than directly bound VLB on a 486. Of course that was at costs of performance. Effective bandwidth of VLB in combination with 386 and 586 was around 25-50% slower, than on directly bound 486. All that was usually not an option on servers and workstations, that's why VLB was mostly used on desktop PCs. And this is why the board in this video is so rare.

[@the_kombinator]

@@necro_ware Apparently I won at Russian roulette every time - Trust me, I know about horror stories from the 90s about having to declock a 486 so that the bus could catch up with the I/O card on VLB systems. As a tech, this was a bit of an issue, but by the time I got there in 1996, there were very few EISA systems coming in for upgrades. I actually just sold an EISA/VLB board for pennies on eBay. Had I known, I would have shipped it to you - it's a good design, but I could not get it going.

[@morantaylor]

I wonder if the board was stacked with another board with leaking battery contacting the underside.

[@Jivemaster2005]

Those EISA gfx cards you got there have VRAM so they will be slower in DOS compared to gfx cards with DRAM

[@enilenis]

I have 2 VLB boards, but neither has EISA. For graphics, it's hard to find a good card that takes full advantage of the bus speed. BIOS has interesting settings, that are easy to miss. Interrupt settings control speed of VLB as well as whether 1, 2 or 3 cards can be installed simultaneously. Order matters as well slot-wise. I had instances when swapping 2 cards between slots made a difference, even though it logically shouldn't have mattered. I have on S3 and one Tridaent VLB videocards with 2MB memory each, fancy for the time, but inferior to cheap PCI cards that came later on. One 486 with VLB is model 601 and the other has PC-Chips branding, but model... I don't even remember. Have to pull it out of the box to look at. I mostly run the 601. It supports DX2 and DX5, but not DX4... not without a voltage adapter, that I don't have.

[@SergiuszRoszczyk]

Some slots were bus master capable, others not. You had to match card and slot in that case. Also some cards didn’t like bus master slot while others required it. It was crazy especially with 3 slots, where 2 were BM ones. Sometimes also slot closest to CPU allowed higher FSB while others were too far

[@AncapDude]

I cannot confirm your problems with VLB. I heavily used this bus on multiple systems and never experienced such issues. Just stable and fast as it should be.

[@necro_ware]

Well, those are not my problems in particular. This is a well known limitation of the nature of VLB and how it is bound to the CPU. Most 40MHz CPUs would already run into many issues with VLB, especially if you try to run VGA and I/O through that bus. If you go for 50MHz, most of the VLB cards will not work. There was also more stable buffered VLB bus, which was used on 386 and 586 boards to make those compatible, but that was a lot slower. VLB was designed for 486 CPU and was meant to be used directly, which made it fast, but unstable.

[@AncapDude]

@necro_ware Most Boards have Jumpers to address high Bus speed like VLB Wait state.

[@nickwallette6201]

@@AncapDude That _helps_ -- particularly if the card cannot process data fast enough at 33MHz and above. But it doesn't do anything to address what happens if the capacitance of the bus is too high, from the distance between CPU pins to chipsets on VLB cards and all the traces in between. The CPU was never designed to drive a load like that. That it worked at all was a hack. That it worked well enough for most of us, was a coincidence. That it didn't work as well for some, was inevitable.

[@gkcnalaka]

❤

[@1kreature]

HM6818A does not operate below 3v. It is the absolute minimum.

[@AK-vx4dy]

Those kinds appeared when i strarted to build PC proffesionaly, i personaly don't have experience but my boss told nightmare stories of configuration jumpers on those boards with EISA... 20 years a go shelf of board instructions (wich we stored at least one) existesd but now i'm afraid it is lost :(

[@mophus6461]

For lower voltage drop you can use BAT85. For the batery it is better to check the current draw first. Maybe CR2032 is to weak. 3xAAA Ni-MH would be better.

[@pvc988]

Doesn't RTC chip go into full power operation mode when supply voltage is over 4 V? I think it'll drain the battery in no time. BTW. I wonder if supercap would be suitable replacement for the battery . If it managed to hold settings for a week or two it would be usable.

[@KKarl1337]

Gute Arbeit

[@AK-vx4dy]

in case of 486 3 of 4 postions were "safe", one caused short of some kind but in my case didn't fry CPU completly, but Windows stoped workind, DOS was fine...

[@GL1zdA]

I have 2 EISA/VLB boards. Probably the sweetspot for them is to use a VLB graphics card, an EISA SCSI controller and an EISA NIC. Not sure how it would behave, if busmaster cards were to be used in both a VLB and an EISA slot at the same time.

[@teknologyguy5638]

Great work, I would like to know where you found that dual external CR2032 battery holder. Thanks. What you searched for in English would be sufficient.

[@SuperDavidEF]

If you just type that "dual external CR2032 battery holder" into a search engine, this exact one comes up like he used in the video, from several sources. I've never seen it before this video, but apparently it's common enough.

[@teknologyguy5638]

Thanks, I was watching on mobile and should have just done that lol

[@cocusar]

remove the diodes thag try to charge the battery, and then bypass the remaining diode that comes from the battery.

[@christiankarsten664]

I have a nearly same looking board, but is has a Dallas RTC and a SiS chipset.

[@matthewrichardson2467]

Hi Necroware. I used to work on old Apple computers in the early 2000s, and they used to use a 4.5v alkaline battery with a 4 pin header for their pram (static ram). I wonder if thats what this board wants? If you google "4.5v apple pram battery" you'll see them

[@aliencray7269]

Looks like on this board manufacturer "saves" two transistors that used to switch between battery and power supply 21:06

[@DemianTowers]

won't such battery holder be a problem when assembling the MB into any chassis? as it goes by the keyboard connector, if I'm not wrong that edge leave tiny or no room at all for components out of the perimeter... I saw you remove it at the end, anyways! thanks for another awesome video!

[@necro_ware]

Usually there is a gap of around 5mm between the PCB and the case, since the expansion cards brackets are located there. As long as it doesn't stand off by more than that gap, it fits.

[@thecriss88]

Go get a Shottky diode in series with the battery to limit the voltage drop from ~0.6v to ~0.1V.

[@r4z4m4t4z]

how could it not work? tech E said it works OK!

[@cleitonfelipe2092]

I think some boards of this time used a 5v battery, no?

[@strelatronics]

That's quite a chunky piece of wire at 12:30

[@necro_ware]

That's 0.1mm just to give an idea.

[@strelatronics]

@@necro_ware Damn small traces

[@2dfx]

Proof that replacing original 3.6V batteries with 3V parts doesn't always work. The retro community seems to think that it's okay to substitute these and believe nothing bad will happen (the retro Macintosh community is notorious for this). Replacing like-for-like always provides the least headaches.

[@necro_ware]

Sometimes you don't have good enough like-for-like. In this case a replacement for a NiCd/NiMH battery would be the same one, but I would never ever suggest it. It's just a matter of time until they destroy your hardware and I saw many modern "replacements" which started to leak just after 2-3 years and people noticed it as it was already too late.

[@BlackEpyon]

Rather than a CR2032 mod, I prefer to replace the original NI-Cd battery with a Ni-MH battery. They cost more, but they are rechargable, and has the higher mAh rating, so you don't have to worry about modding it, it's just a drop-in replacement. best part is, Ni-MH doesn't LEAK!

[@necro_ware]

Unfortunately they do leak as well. Not quite as often as NiCd, but they still do.

[@sebastian19745]

Yes, they leak just like NiCd ones. If they discharge too much, they start to leak. I had many hand tools destroyed by the leaky NiCd batteries. In this case, the AA or AAA cells have much, much more capacity than the minuscule Varta and will not discharge too fast. The advantage of replacing the onboard NiCd battery with standard AA or AAA NiMH is that the batery holder can be mounted far away from the motherboard so if they leak do not affect the PCB. And, as you said, they are a drop-in replacement.

[@mcosta3810]

​@@sebastian19745Exactly, while also making the batteries easy and convenient to remove, when the computer inevitably goes back on the shelf for storage until its next use.

[@sebastian19745]

@@mcosta3810 Add that for this specific system (the EISA config SRAM), the high capacity of AA cells can make the removal not necessary, even if powered once a year, they will recharge and retain the charge for a long time.

[@snakezdewiggle6084]

The Shuttle boards I have, use VIA chipset, and 3.6volt battery Ni-Cd. You can upgrade to NiMH 100% compatible. Did you check for a BIOS update.? I hope OPTi are a better chipset. 👉 VIA are CPU Killers.👈 if you have any VIA, thow them in the Ocean...

[@necro_ware]

I prefer not to throw anything into the ocean. It is already dirty enough.

[@snakezdewiggle6084]

Oh, sorry.! Its a metaphor meaning, "stay the F_ away from them". 🚫VIA🚫 Nice job on the repair btw.👍

[@MichaelRusso]

I own two! I got one as a box of parts for sale.

[@bmw518rijder]

External AT style batteries are 3.6 volts or 6 volts.