Commodore 486SX25 Restoration. Fixing leaky VARTA battery damage
ฝัง
- เผยแพร่เมื่อ 25 พ.ย. 2022
- Who made this Commodore 486 PC?
Part 2 here: • Commodore 486 Restorat...
Part 3: This video
Part 4: Old-school overclocking: • Commodore 486 overcloc...
Join me on Twitter: / epictronics1
Music by Karl Casey @ White Bat Audio - วิทยาศาสตร์และเทคโนโลยี
I've bought ISA slots off eBay before. I got the most results searching for "edge connector 98", since it's a 98 pin connector (2x49 pin), although there was a quite a few unrelated listings (for some reason a lot of sportsball cards, you can filter them out with "-1998"). The cheapest were obviously from China and Taiwan, but there were a few listings from the US America and UK.
EDIT: Forgot to mention, if you just want the 8-bit slot, it's 2x31/62 pin.
Awesome, I found some right away :) Aliexpress seems to have the best price, thanks!
Haha a friend of mine designed 98 pin daughterboards for a project as luck would have it, and when he ordered random 98 pin connectors, that looked continuous from the description at a glance (forgot to look at detailed datasheet) he turned out to receive ISA sockets when they finally turned up! This was quite a roadblock because the boards were already made and they were continuous 98. Luckily one supplier was found who had real continuous 98 sockets and could express them in. The rest all turned out to be ISA slots.
Here in Germany these computers appeared after Escom had bought up what was left of Comodore. They then provided generic PCs with the comodore label and sold them more expensively, or you had to buy the operating system.
Yes, however, this Commodore was made about one year before the ESCOM deal.
the pc boom era , Escom bought anything they hold off, half of those system came dead on arrival, students like me reseating and swaping boards to get things working.
@@arjanvanraaij8440 Sounds like good times :)
@@arjanvanraaij8440 i loved the store. There were always parts, with the remark that they were defective, for very little money. 90% of the time it wasn't defective, but the staff wasn't properly trained
When lead -tin solders corrode, tin oxide is produced, which removes tin from the solder. This changes the alloy from having a melting point at 183C to having a melting point range where part of the solder melts at that temperature but part melts at a higher temperature-think of it as you making a lead-tin slushie when you try to melt it instead of a liquid. As you found out, grinding the surface away removes the tin-depleted material and allows you to melt the alloy underneath that has the normal solder melting point.
cool, I guessed that some chemical reaction must be messing with the melting point. Thanks
So adding a little modern 99% tin solder to the surface might be more effective than 60/40 or 63/37? Or do I better just use 63/37?
I don't know how viable it is, but citric acid has a pH of 3.2 where white vinegar is around 2.5, so it's not a massive difference. Citric acid powder isn't terribly expensive and you can mix the concentration you want with distilled water. The upside is it will smell like oranges instead of vinegar's nasty smell.
Sounds good to me : ) I'll try with lemon juice next time and we'll take it from there
pH is logarithmic, 2.5 is five times more acidic than 3.2.
@@Squossifrage Oh, dang it. Well, we'll try it anyways. But that is a huge difference. Maybe we'll even check what the actual pH is too. shouldn't be a too expensive test
@@Epictronics1 Nix the lemon juice, what you're after is citric acid powder in the canning section of your local supermarket, next to the mason jars, canning wax, fruit pectin, and pickle brine seasoning.
Stuff also does an absolutely incredible job dissolving lime scale from coffee makers.
@@mfree80286 cool, I'll try it
For badly corroded solder joints I recommend using acid paste flux, it cuts through the corrosion pretty well in my opinion.
Interesting. what brand have tried? Thanks
@@Epictronics1 the brand I use is Kester but I can't seem to find it for sale anywhere, though I think any good acid paste flux would work. They're usually used for plumbing I believe but I had it on hand and tried it on some horrible corrosion on an old 286 Compaq laptop that had the wonderful 1980s era surface mount capacitors. Since then I've used it on battery and capacitor corrosion numerous times as a little bit seems to go a long way.
@@lightmagick I'll see if I can find some. thanks
I've heard of that before, but beware that it's nastier than normal flux: you should spend more attention on washing it away than normal. That's the reason they say not to use plumbing flux on electronics.
@@absalomdraconis oh yeah the cleanup is annoying as heck since it's a sticky paste but the trade-off is how well it works at cutting through the corrosion. Might just be my successful experience with it making me biased towards using it but it's cheap enough that I don't think anyone trying it off of my recommendation would be too upset.
Instead of adding a diode and having that bodge with the vertically soldered holder you can disable the charging circuit on the board and solder the holder normally in place of the original battery without the extra diode.
You most likely need to lift one of those diodes or resistors lined up next to the battery area where the red pin headers are.
To find which one you need to check continuity from the positive of the battery to a diode then to a resistor and then to 5V (diode and resistor might be inverted) in other words find the charging path from 5V and break that circuit lifting a leg of a component.
Then double check you get 0V at the battery terminal and voilà
The card might just need a bios reflash, as for the diode they infact act like an ultra high resistance in the reverse direction and ultra low in the forward but importantly not infinity and zero, since the internal resistance of the meter is so high it can detect the small diode leakage
Oh, I see. I didn't know that, thanks!
Difference between ideal components and real-world components. :-)
@MrTGuru I DO mean the VGA card but not as an upgrade.
EPROM stores data by trapping electric charge within the gate of a FET, However over time his charge can leak away (Oxide layer imperfections, Cosmic rays .etc) causing corrupted data.
Writing a known good (Idendical) copy of the data back to the EPROM should fix this without requiring erasure, since any bits that have become discharged will be recharged and any bits that were never charged will remain uncharged.
@Mr Guru We're in luck. The VGA card turned out to have the jumpers set incorrectly! It's working again : )
Search term '98 pin ISA edge card connector' brings up some good results, first one has gold-plated card fingers.
Thanks, Aliexpress seems to have a good price
I hate the smell of vinegar as well. I buy anhydrous citric acid (Heitmann reine Citronensäure from just about any household goods store around here) and just prepare citric acid by dissolving it in water, you only need a small amount of powder. It works just as well against battery juice.
Digi-Key claims to have ISA connectors. Sullins Connector Solutions EBC49DCWN-S420 seems to be a match. I'd have to dig further into the data sheet to be sure, though.
The picture looks right. but they are silly expensive :/
Great video, what a rabbit hole that machine is!
Thanks Ted : ) We'll get it up and running for some DOS gaming if we persist
@@Epictronics1 I wonder if you've a bit of "bit rot" in the bios of that old gfx card? I'm seeing a lot of that lately. I'm wondering if to flash all my old cards.
@@ted-b I realized when uploading this vid. I may have made a silly mistake. I think I might have had the jumpers on the card set for use with an 8-bit slot :/ I'll check
If you hate the smell of vinegar ... citric acid should work fine, too! It's typically used for decalcifying coffee makers.
I'll get some and I'll try it out, thanks
Need a better soldering station. My Xytronic LF-3200 has no problem with large thermal masses such as PCB ground planes, even when using a small tip. And I always keep it at 325 Celsius.
There are coin cell holders made for upright mounting which may have worked here.
That is pretty good compared to my pincil and TS-100. I usually go with 380 and have to use my large tip occasionally for large ground planes. I'll order some upright holders for sure
With the ISA slots, I would be tempted to assume a poor solder joint or a damaged track on the motherboard, especially with all the flexing when inserting or removing cards. The memory issue seemed fairly obvious, as the problem appeared as soon as you had removed it to be examined.
I have inspected and cleaned the SIMM sockets with deoxit and it still gives occasional RAM errors. I'm considering removing some SIMM sockets and at least one of the ISA slots to inspect underneath
@Mr Guru The thought has crossed my mind. I'll remove at least one or two to check
You can find standard clock oscillators at mouser. Just search for oscillator and filter on dip-14 package.
I checked, thanks. Unfortunately non-stocked
@@Epictronics1 Maybe jameco? www.jameco.com/shop/keyword=full-can-oscillator
@@Farull Yes, they seem to have a few in stock : ) thanks
They do have vertical holders for coin cells, but your solution seems ok, I'd add some hot glue for structural support.
yes, I'll definitely add some hot snort, thanks
The vertical ones don’t usually have a footprint that is compatible with the old rechargeable cell. For this very reason, i made a PCB for a low profile surface mount CR2032 holder that uses 0.1” right angle jumpers, one on each side, and pads for a SMD diode with bypass solder points. Usually I just modify the motherboard to add a diode somewhere (there’s almost always an opportune footprint somewhere) but it’s nice having a drop-in replacement too!
@@nickwallette6201 YT isn't great at letting me know about all new comments. I just read this now! Where can I find the project? : ) Thanks Nick
@Mr Guru I actually ordered some of those "external" holders with wires a week ago. I'll try them out in the next project
@@Epictronics1 It is always my intention to give away anything useful that I've done, but I am absolutely terrible at actually setting up any kind of repo for the handful of things I've managed to build to completion... haha
I'll try to get off my duff this weekend and maybe even use the blog hosting site I've been paying for since some number of years ago. I'll keep this tab open to remind me to post a follow-up when I have a link.
I actually have a small collection of RTC / coin cell adapters that I've made, trying to get everything from clone motherboards to PS/2s and Mac Classics using the same CR2032 cells. I even designed a few that adapt one or two coin cells to in-between voltages, like the 4.5V required by a Compaq motherboard that I have. That one in particular is utterly unhappy with >5V, so 2x 3V cells is not an option, and 1x cell isn't enough to maintain the settings. I tried a couple series diodes, but they aren't passing adequate forward current to drop significant voltage, so I ended up with two very low power regulators -- one is a buck (6V -> 4.5V), and the other's a boost (3V -> 6V, I think.) I figured I would let both of them run for a while and see how they do. IIRC, the boost reg actually ended up with a lower quiescent current (oddly enough...), down into uA at no load. But that's still enough to drain the cells more quickly than the idle current of an RTC, so I'm curious to see how long they last. Should be a couple years still, but definitely not 10-20+ anymore.
Could have maybe mounted the coin cell on the case on some velcro and then run wires to it.
Yes, I will order a different type of CR2032 holder for this project
Could you have soldered the battery holder facing the other way towards the ram and you may have been able to get the battery in and out without removing the mb?
Those SIMM modules are a bit higher than the cell. But not a big issue. I'll just replace it with a more suitable holder when it's time to replace the cell. They last a very long time : )
Why did you not used the battery header near the keyboard connector? That way you could have used a battery holder with 2 AA batteries or with a 3V coin battery mounted on the case.
To remove that green corrosion that is a base, actually you neutralize that base with an acid. We usually use vinegar (acetic acid) because is strong enough to do the job before evaporates, is easy to get and the salt formed is easily soluble so easy to wash. But you may use any acid that you have access like lemon juice (citric acid) or lime juice that you can get concentrated at bottle in grocery store. Once, I had no vinegar at hand and to remove batteries form a corroded battery holder I used the acid used to clean toilet; it is hydrogen chloride, an acid quite strong (it will burn you) that I diluted in water until I get it to react with the crusty corrosion (I put drops of acid in water until it was concentrated enough to get a reasonably quick reaction).
I used to clean the ISA slots with a reasonably thick paper/cardboard on wich I put a cloth (linen, cotton) soaked in alcohol or contact cleaner. The cloth being abrasive due to its fibers helped cleaning and the paper was there just to make the "cleaning device" thick enough to bend a bit the contacts; it was almost thick like a pcb.
About the crystal form factor, sometimes there are traces for rectangular (like yours) AND square crystals on the same board, so the manufacturer can use either in the same spot; there may be the same on your board, just check for continuity between the pins in the socket were the crystal is seated.
Thanks, I'll try with some lemon juice. I'll remove the crystal in the next project video and we will have a look at the socket. I will probably order some coin cell holders on wires to use on the pin headers.
@@Epictronics1 Yes, take a look at the crystal, I found on many P1 motherboards and on some newer 486 that under the AT keyboard connector are the pads and sometimes the silkscreen for a PS/2 keyboard connector. On many 386 motherboards I found the same with the main crystal oscillator, the pads existed and were connected for both types of oscillators square and rectangular one.
@@sebastian19745 I had a look and there are only the four obvious pads. However, there is plenty of room to add a modded socket. I have actually ordered one of the square crystals to test. I wonder if we could use the even smaller ones that only have two legs? They are much easier to find
@@Epictronics1 If I remember correctly, those 4 pins crystals are in fact a full oscillator with one pin being ground, one being Vcc (+5V), one being output and one usually NC (but on some models the 4th pin is for fine adjustment). Check the datasheet if you plan to make an adapter. However, despite the different footprint, the square cans and the rectangular ones are identical.
So no, you cannot use a two pin quartz instead a 4 pin oscillator. However, you can use an clock IC (found on more modern motherboards 486+) with its 14MHz crystal to make many clock frequencies. Usually that IC is configured with jumpers (remember the FSB clock selection between 25, 33, 40, 59 or 66MHz on some 486 ?).
@@sebastian19745 Yes, I think the sensible thing is to get a bunch of ICs. I just ordered a few crystals and with shipping, it ended up at $20. Better to have a few ICs and adapt them to whatever project is at hand
you might want to try some vinegar directly on the solder joints, before you add vast amounts of flux... then it should flow more easily. The bad news is, that it's going to be very stinky.
Alternatively you might want to try a different kind of flux (i.e. for soldering steel).
Do you mean mix vintage and flux? Hmm... That might work. And yeah, that might stink too lol
@@Epictronics1 : I actually mean to apply the vinegar first and let it work its magic. And then apply flux without cleaning the vinegar off. It should drop the pH of the joint and make it easier to de-solder.
Though I would refrain from using the de-soldeing gun, and instead use a cheap hand pump or wick. I don't know how well the tip of the gun takes that mixture.
@@bikkiikun ah, yeah, that might end up an expensive experiment : )
Damn, forgot that this was happening, so didn't watch the Premiere.
bummer.
I'll try to make a new vid next week : )
You can use the smaller crystals with 4 legs
cool, I just received some of those in the mail yesterday. We'll try them out
Remember that there is some reverse leakage volt in a diode from approx. 0.5v to 1.5v some diode types have less than others. zener diode I mean I have less. How much reverse leakage current depends on how much you load the diode
I used a Schottky 1N5817
Since the cathode end isn’t referenced to anything, it’ll float pretty high. There are nA to uA leaking through in the fwd direction that absolutely won’t hurt the battery, but they will dissipate and clamp down to the battery voltage once you install the cell.
@@Epictronics1 That explains it. Schottky diodes typically have much higher leakage than PN junction diodes. By my calculation, you are seeing about 0.5 microampere, assuming that your meter has the typical 10 Mohm input impedance. Probably not enough to worry about--although I'm not really sure how much current a CR2032 can safely sink before you start to get gas buildup.
@@danmenes3143 Someone suggested adding a protective 1k resistor in series. (a recommendation from a CR2032 datasheet) Maybe this would also take the current closer to zero?
I think baking soda (sodium bicarbonate) is used for neutralize both acid and oxidation. The thing is that you need a lot of time. You can use also with vinagre and or oxygenated water. But I haven't tested that much which method is better and faster.
Ok, do you sprinkle sodium bicarbonate on the board when it's time to clean off the vinagre?
@@Epictronics1 since it work for corrosion you can use both, but you need to clean after or remains like a solid paste.
i think they make rechargeable lithium coin cells, might could use them as replacements
I did in another video. I used the LIR2032 but some viewers commented that the onboard charger was too crude and there was a risk of things going really bad. th-cam.com/video/7AIkJctDP8w/w-d-xo.html
@@Epictronics1 Yup. LiOn batteries require some special circuitry to charge/discharge. Honestly the CR2032 + diode is likely the ideal solution.
Hi! What diode are you using? I have a 386 sx board with a varta that I cleaned. I would like to perform the coin cell mod and I have the holder - just don't know what diode I need to prevent the reverse charging. Do you have the model or a reference link to this mod? Thank you! Great video!
Thanks, I use 1N5817. good luck with the project
Just about any small signal or low current rectifier diode will do. Schottky would be nice, since they have low fwd voltage drop and you want as much of that 3v as you can get.
@@Epictronics1 Hey there! One last question.. For the diode, what direction (lead) do you have attached to the coin cell holder and what end (lead) is soldered to the Mb. I just want to make sure I put the diode in the right direction before attaching my coin cell holder.
@@TheRetroRecall Sure, here is a picture. Hope it helps. The resistor is optional. 1.bp.blogspot.com/-2t1V8k063ts/XVDCtN3swMI/AAAAAAAAAic/5ZH0aG0Fdd86i2NaR5PiS2e5StrJeIZXQCEwYBhgL/s1600/20190812-a501-coincell-battery-mod-schematic.jpg
@@Epictronics1 thank you!
samtech might have new isa slots, the at least provide pcie slots as new
Thanks. I did a search for "ISA" but got nothing. Not sure what search word to use
@@Epictronics1 ask them directly, the website is cluttered and not working
Diode on its own is not enough. You need to also add a 1K resistor.
That could be a problem. The CR2032 is 3v and the 1N5817 has a voltage drop of 0.4v. So we are already down to 2.6v instead of the original VARTAs 3.6v
@@Epictronics1 When disabling the charging circuit, I use a SCHOTTKY diode (like the BAT85,113) When the computer operates, the BIOS SRAM chip gets 5V from the PSU. The battery is used only when power is off. Typical SRAM can still hold its data with only 2.2V.
@@leonkiriliuk Ok, have done any longer tests with a CR2032 and 1k resistor? Will it last for a while?
@@Epictronics1 I've had an X68000 computer run for 5+ years now on a CR2032 this way. This is how NES, SNES, etc carts all save SRAM and those lasted for 25+ years now.
@@leonkiriliuk Ok, cool. I'll replace the diod lead with a 1k resistor in the next project video. Thanks
Lemon juice smells better and neutralizes battery and capacitor juices.
Sound good to me : ) Do you use regular Lemon juice?
@@Epictronics1 I’ve been using the Sicilia which is sold in a small, lemon-shaped bottle. Or you can just squeeze the juice of a lemon. I also have citric acid grains from an agricultural store that you can mix into water.
@@terosaarela4555 cool, I'll try the stuff they have at my local supermarket. Thanks
@@terosaarela4555 yeah, citric acid is a good not smelly replacement for vinegar.
Could try to throw the motherboard in the dishwasher? Or ultrasonic cleaner? And then do tests again....
I wish I had an ultrasonic cleaner. Especially the size of a MOBO : ) Not sure if I'm comfortable throwing it in the dishwasher though... Maybe I'll try with something lesser first
@@Epictronics1
I understand for the ultrasonic cleaner. But you've got such a nice youtube studio setup that an ultrasonic cleaner wouldn't be out of place for the serious repairman? ;-)
Regarding using a dishwasher: There are absolutely 0 risk to damage electronics doing this.
You would have to remove the newly installed coin-cell and let dry thouroughly after the cleaning.
I think it would bring some consistency and more reliability to your repair process, which is always a plus.
BTW you're probably aware that the 8bitguy has done this in one of his video (or was it retro-receipe)?
@@LFOSyncToo yeah, I've seen a bunch of TH-camrs cleaning mobo's in dishwashers. I'm sure it's fine. but it feels really weird. I guess I should try it out
i have something that looks just like this on my desk right now
cool, have had a look inside?
@@Epictronics1 its just a motherboard and power supply i am trying to get to boot. i have a few 386 and 486 motherboards and a small pile of ram and cards. i'm tring to get a operational ISA bus system running.
@@borgheses Oh, ok, I thought you meant you had a Commodore 486 on the bench : ) Good luck with the project
@@Epictronics1 thanks. i cant find the motherboard anywhere online. its got a cyrix 486 dx 40 and math co-processor and a slot for anther processor.
@@borgheses You could ask the Vogons. They are very helpful in such things
code 31 mean booting errors
other code u said whiles agos was ram died or crash just tips help you
cool, thanks
After doing a little digging, I'm not sure you make the right choice to use a Schottky diode rather than a PN junction diode. Yes, the forward voltage drop is nominally lower, but you are seeing something like 0.5 uA of leakage current, which seems a bit much to be pushing into a CR2032 batter. I might lean towards something like a 1N4148, which lists a max leakage current of 25 nA at 25 C and 20 V. Typical leakage at 1.5 volts reverse bias will be even lower.
According to a Panasonic design guide (AAA4000COL15.pdf, if you want to Google it), in a backup design like this, the total allowed leakage through the protection diode is 3% of the battery capacity, cumulatively, over the lifetime of the product. A CR2032 has a capacity of about 220 mAh, so the cumulative allowable charging current is 6.6 mAh, which at 0.5 uA, gives you 13,200 hours before you have pumped enough electrons into the battery to risk popping it. That's probably fine, given how infrequently a vintage computer is likely to be used, but for a computer that was on all the time, it would be less than two years.
If you are worried about the forward voltage drop of the 4148, that's probably less of a concern than you think. The rule of thumb is, "silicon rectifiers drop 0.7 volts," but that's not going to be true under the extremely low currents that the backup battery has to supply. I would be surprised if you see more than two or three tenths of a volt in practice--although confirming my prediction would require a meter with an extremely high input impedance.
Cool, I'll switch to using 1N4148. I have some leftovers from another project. What do you think about adding a protective 1k resistor as suggested below?
@@Epictronics1 The 1K resistor certainly won't hurt anything. The additional voltage drop will only be a few mV. I think the risk that it is intended to prevent is pretty low to begin with, but if you like the belt-and-suspenders approach, go for it.
To be clear, I'm not an EE or anything, just a hobbyist. So take my advice for what it's worth.
@@danmenes3143Same here. But data sheets and math apply to us all. I did a quick search and 1N4148 doesn't seem to be an uncommon choice