Cool thing!
I guess this could be very helpful for many vogons!

Maybe it would be a good idea to put as many boards onto a Gerber file as possible on a 10x16 (or similar standard size) board and order a board at a PCB manufacturer, if enough vogons preorder boards.

There was another member here who made a Dallas 12887 replacement in which the battery mounted onto the module itself. I spent 5 minutes looking for the link, but I can't seem to find it. I must be missing the right keyword. If I can remember which computer I have my unit it, I'll take a photo.

Here it is: Re: Possible DS12887 drop-in replacement

From OSH Park: https://oshpark.com/shared_projects/NwcVatSK

This is pretty nice.
A repalcement for the DS1387 would be nice too, but I doubt there is a replacement chip for that one.

I did a mod once in which I tacked on an external 2xAA battery pack, but this is next level.

Oh, hey, that integrated one is probably better since it's a single-board solution. 😀 I'll fix the holes on mine and publish it anyway, for those that might prefer an external battery, such as AAs. The chip claims to "guarantee" protection against reverse-charging, so any source of ~3V would work fine. Do not use a protection diode.

While I'm thinking of it, the BOM is thusly (with Digikey part numbers, and qty 1 price at the moment):

Maxim DS12885T+ -- TQFP-32 RTC (DPN# DS12885T+-ND, $8.93)
ECS ECS-.327-6-17X-C-TR -- 32.768kHz crystal, 6pF load capacitance, 50k ESR (DPN# XC2129CT-ND, $0.70)
MPD BS-7 -- CR2032 coin cell holder (DPN# BS-7-ND, $0.98)
0.1" male header pins of your choosing, e.g. Sullins PRPC040SAAN-RC, 40-pin breakaway (DPN# S1011EC-40-ND, $0.66)

retardware wrote:

Maybe it would be a good idea to put as many boards onto a Gerber file as possible on a 10x16 (or similar standard size) board and order a board at a PCB manufacturer, if enough vogons preorder boards.

I'm not sure how it goes for those outside of the US, but OSH Park is really the ideal solution for things like this. Someone can publish their project, and you can order as many as you want (min. qty. is 3), at any time you want. No need for a middle-man or group buy. For a tiny board like this, it only costs a couple bucks all-in. I can't imagine it would be much (if any) cheaper to do a small-batch run and distribute them individually. Maybe it would be cheaper for someone in Europe to order a few copies and mail them to other Europeans, I dunno.

A gentleman in France is already selling single-PCB DS12887-replacements with similar concept on a popular auction site. I bought one to test and am pretty happy with it.

dionb wrote:

A gentleman in France is already selling single-PCB DS12887-replacements with similar concept on a popular auction site. I bought one to test and am pretty happy with it.

Care to upload a photo?

Finally revisited the RTC replacement PCB. I decided to take a queue from some of the comments here and integrate the coin cell holder. I still have a copy of this project with the pins for external batteries, but in reality, it's just a lot more convenient to use CR2032s everywhere.

So here's the new PCB. First step is to solder in the RTC chip, 32kHz crystal, and the pins.

Then, add the coin cell holder on top.

And presto, RTC sammich.

I replaced the previous version in my 386DX/40, since the pins were only tacked onto the bottom surface of the PCB. This one has large enough holes so the pins can actually go through the PCB. 😉

And now, let's try to get that PS/2 Model 30 - 286 running. I probably should've used different (round?) pins, because this socket is TIGHT. Maybe I just need to replace the socket. For now, it's not fully seated, but it's not going anywhere.

SUCCESS!! After locating the proper startup disk image online, it finally boots without errors 161 and 163!

Since I was ordering from OSH Park anyway, I put together a replacement front panel PCB for my Antec VSK case.

I really like this case. It was about $35 locally, without a PSU. It's definitely not built like a tank, but it seems the corners cut were reasonable compromises. Everything fits nicely and it looks pretty smart.

However, new cases all ship with USB 3.0 now, which obviously uses a different header connector than the USB 1.x/2.0 headers on retro motherboards. I have an Asus TUSL2-M installed in this case, with a Pentium III Tualatin 1.0GHz processor -- definitely not USB 3.0. I don't really like cluttering up the insides with adapters, so I ordered some USB jacks and designed a replacement board.

I bought some USB header extensions from Ebay and chopped off the male end, then soldered it directly to the PCB.

Fits like a glove!

Another completed build. 😀 Well, it's still waiting on a Plextor DVD/CD drive to show up... So almost complete.

Is there an OSH Park link? Are you going to sell kits?

Well, I have no interest in getting in the kits business. 😀 But, I'm happy to make it available for anyone who's interested in making their own. If anyone happens to have an online store and wants to carry them (or whatever), you have my blessing.

OSH Park: https://oshpark.com/shared_projects/EfNaK3A6

The BOM was posted above. I'll attach a ZIP file with the Gerbers and the original project file. I used Robot Room's Copper Connection to design the PCB. It has since been consumed by Express PCB, but there may be an old link out there somewhere.

Another update in the saga of vintage time-keeping. This time, I'm working on two 486 motherboards that had rechargeable barrel batteries. I wanted to convert these to use CR2032 coin cells as well. This isn't a drop-in replacement -- the coin cells aren't rechargeable and thus need some minor circuit hacks to avoid damaging anything.

First victim uh, subject is a Chaintech 4ULD.1:

The second is an Aquarius MB-4DUVC mk 1. This is the board I'm using in my current 486 DX/2 - 66MHz build:

Both of these had a similar circuit, where the battery would run through a small-value resistor, a protection diode, and then appear on one of the pins on a CMOS configuration jumper. Placing the jumper on pins 2-3 would connect the internal battery to the CMOS, while placing the jumper on pins 3-4 would clear the CMOS settings.

The problem is that the original circuit would also connect the battery to the +5V rail via (something like) a 150-ohm resistor and another protection diode. This needed to go. However, simply disconnecting this part of the circuit would mean the CMOS and RTC were always running on battery, potentially draining the coin cell more quickly. So, I wanted to modify the circuit to allow the system to run from the PSU when powered on, and from the battery otherwise.

The modified circuit looks like this, where gray is a removed component and red is an added component:

Finally, a close-up of the rework:

Luckily both boards had multiple footprints where the barrel battery goes. Each had a 20mm spacing that is compatible with standard coin cell holders. Unfortunately, the Aquarius board did not provide ample room for the holder itself, so I had to find a slightly smaller one. Naturally, I forgot to order the required cap that actually holds the coin cell in place, so I'm hoping everything still fits when I finally get around to putting it together.

But there it is. Both boards are now ready to go for a couple decades free of amnesia. I think that concludes all the outstanding issues I have with RTCs.