What did you try out? On a Sony I was putting back together which had a 2.4v nimh pack, I've used a CR2032 with two diodes in series going from battery+ > diode > diode > laptopCMOS+.
That should stop any charge from the motherboard getting to the cell or at least being a really tiny amount. When running the laptop outputs 3.3v to the CMOS battery pack, which is okay for a 2.4v nimh but not good for a CR2032.

I think a tiny bit of voltage / current does get through even with the diodes, so I don't know how safe it is long term but it makes the laptop much more usable for occasional usage.

This should be okay for the 2.4v nimh Toshiba CMOS batteries, but it probably wouldn't be enough for the one with the big standby battery too. I think it was okay on a Satellite 470CDT where the battery was 3-cell 3.6v nimh and it charged at 5v, where hopefully the two diodes would stop the cell being charged.
Do check out some RTC chip documents, because they're intended for battery operation some do actually have quite high voltage tolerances. This one does say 2.5 to 4 volts though so not DS12885 at least: https://www.analog.com/media/en/technical-doc … 5-DS12C887A.pdf
Personally I haven't bothered looking to an RTC battery for the dells that had NiMH packs, because they don't get upset or need any settings changed to boot up even with no RTC battery - although of course that means the time is wrong but that's not really a big deal.

Well, I haven't tried anything out yet, but after a bit of searching I found a hopefully plausible replacement for the Dell (which I'm putting on hold; read below) and Toshiba suspend batt: LIR2032's. These cells are ~3.7v, come in the form of a CR2032 and are rechargeable. Any thoughts?

Completely unrelated: today I found a floppy drive module for the Dell, and since I was having some issues in suspending/resuming in Windows 98 I decided this would be the perfect opportunity to upgrade the BIOS to the newest version. Anyway, the erasing process took forever and a half, but the system fully froze at 12% of writing. After leaving it for a couple hours, just to see it was still the same, I force-powered it off, and of course now it doesn't boot.

This shouldn't be that big of a deal, since flashers and the required adapter aren't too expensive and I figure I'd need one later on as well. The issue is that Dell doesn't release BIOS updates as .bin files, but as a DOS executable meant to be run on the target machine. So now I'm stuck with a bit of a paperweight, unless of course someone is willing to take a dump (heh) of the chip's contents. I'll be making a new thread for this right now, since this one definitely isn't the right place for it.

Sadly, putting a lithium rechargeable into a circuit designed for a NiMH rechargeable really isn't safe - the reason being that NiMH batteries are fine with trickle charging and really don't need any battery / charge management circuitry to keep an RTC battery charged.

Lithium cells though, are quite hazardous if the voltage goes too high, then you can get thermal runaway and fire etc.
Modern computers do now often use Lithium rechargeables like the LIR2032, but that's because they've got circuitry to properly handle charging it and cutoff charge above a set voltage threshold. Since no such circuit exists on an old laptop with a NiMH battery, putting a lithium cell in its place could keep charging up the cell over continuous operation - maybe fine for a few hours turned on here and there, but still a real risk if left charging for a long time.

Maybe there's a plugin solution like those little modules you get for lithium cells that do manage the cutoff, but I haven't seen one that's meant for tiny cells like an LIR2032.
I put a VL1220 into my Libretto 100CT and that should be okay because it sticks at a low voltage, if it was attempting to charge at 5v that wouldn't have been possible.

Something to consider also is energy density - a rechargeable cell can't store as much as a CR2032 can at around 200mah, the LIR2032 is around 40mAh by comparison. Where I put batteries in, I'm probably going to keep doing the one or two diodes (depends whether RTC complains at ~2.4 to 2.2v) with CR2032 cells.

That's unfortunate about your laptop BIOS, hopefully it has a boot block and you can recover it that way. That might have been a bad floppy or bad media, if it doesn't check the file enough first before flashing.

Okay, thanks. I guess I'll be getting a couple of diodes tomorrow! Oh and by the way, I found out the executable has a switch to make a .ROM file out of it, so I think I'll be able to recover this.