More or less good, but with a few minor problems

IBM 386SLC has 8kb cache, but the IBM 486SLC, SLC2 and SLC3 all have 16kb cache. You are correct they are 386SX bus compatible, but not pin compatible (as far as I've been told at least).

I don't know of any Tx486SLC2s. Do they exist? Can you show me a photo of one? The only one I know of from TI is the 486SXLC2, which is available as either 3.3V or 5V part. These are clock doubled, and have 8kb internal cache, both of which can be enabled or disabled at your pleasure.

Thanks for the feedback!

Interesting points - I thought only the SLC3 had a different pin-out, but you may be right. It doesn't matter though - either way, the IBM parts are not valid replacements for a 386SX. As for the 486SLC2, there are a number of Chinese sellers apparently having them? Yet suspiciously no images anywhere. I'd guess the 486SLC2 is really a 486SxLC2, anyway - and, again, I suspect strongly is 3.3V not 5V (making it incompatible). You mentioned there may be 5V versions, which may work - do you have a part number for a confirmed 5V part?

I am quite certain there are 5V versions, because I own two of them, and I remember checking them in the datasheet to confirm. I bought them as direct replacements for 386sx chips some time ago but never got around to soldering them in.

The other one is a 486SXLC 40MHz version (not clock doubled). I don't have a photo on hand though. It's also 5V.

Last time I checked chinese sellers had a lot of these...some of which were the 3.3v parts. I've only seen 50MHz parts available as 5V. All the 66MHz parts I've seen are 3.3V.

Cx486SLC2-50MP = good! 50MHz power! This is the fastest I've found yet.

That's the one I installed on a Biostar 386SX-25 board here recently. It's a later board that has native BIOS support for the internal cache... older boards will require a software utility to enable it. It does indeed make a very big difference in performance, and it's a direct 'drop-in' swap (assuming you're good at SMD soldering, anyway).

SMD soldering is a piece of cake. Just make sure you're using a low power iron.

That SXLC2/50 looks interesting... I wonder what you can find them on...?
I have a TI 486SLC/25 which I tried on my A2386SX but it didn't work. I don't know if the chip is good, though.

For changing SMT chips I find an SMT rework station is by far the easiest. 😀 The low-end Aoyue beasties are great - not expensive but do the job brilliantly for simple SMT stuff. Certainly a lot safer than a soldering iron, though if you have to use an iron definitely a low heat controlled one is the way to go!

http://www.ebay.com/itm/RARE-486-50MHz-TI-NEW … =item2a0f336ae6

That looks like it may be ideal!

However, I can't find a datasheet anywhere - does anyone know if it's 5V or 3.3V? I guess if there's no specs available the only way to be sure is to try it!

I used to have this information, because at one time I needed to track down the information on programming the CPU init registers which are slightly different than the cyrix chips. As far as I know if the part is 3.3V it will be printed into the top of the chip.

2.31.7 Texas Instruments TI486SXLC-PAF CPU (Potomac)

Intel i80486 CPU instruction compatible, no FPU (Floating Point Unit).
Intel i80386SX CPU bus interface.

8 kbyte cache: write-through, 2-way set-associative, 1024 sets,
4 bytes per line.

33 MHz: february 1994.

Package: QFP (Quad Flat Package).

Technology: CMOS.

ID: step level A: DH = 0x04 (family ID), DL = 0x10 (model ID, revision),
step level B: DH = 0x04 (family ID), DL = 0x11 (model ID, revision).

DIR0 register: 0xFE.

2.31.8 Texas Instruments TI486SXLC-V-PAF CPU (Potomac)

Low power (3.3 V) version of the Texas Instruments TI486SXLC-PAF CPU.

25 MHz: february 1994.
33 MHz: february 1994.

Technology: CMOS.

ID: step level A: DH = 0x04 (family ID), DL = 0x10 (model ID, revision),
step level B: DH = 0x04 (family ID), DL = 0x11 (model ID, revision).

DIR0 register: 0xFE.

2.31.9 Texas Instruments TI486SXLC2-PAF CPU (Potomac)

Clock doubled version of the Texas Instruments TI486SXLC-PAF CPU.

20/40 MHz: february 1994.
25/50 MHz: february 1994.

Technology: CMOS.

ID: step level A: DH = 0x04 (family ID), DL = 0x10 (model ID, revision),
step level B: DH = 0x04 (family ID), DL = 0x11 (model ID, revision).

DIR0 register: 0xFE.

2.31.10 Texas Instruments TI486SXLC2-V-PAF CPU (Potomac)

Clock doubled, low power (3.3 V) version of the
Texas Instruments TI486SXLC-PAF CPU.

20/40 MHz: february 1994.

Technology: CMOS.

ID: step level A: DH = 0x04 (family ID), DL = 0x10 (model ID, revision),
step level B: DH = 0x04 (family ID), DL = 0x11 (model ID, revision).

DIR0 register: 0xFE.

Thanks for that info!

I've tried several CPUs on my A2386SX: an AMD AM386SX/SXL-25, an i386SX, a Cyrix Cx486SLC/e-33-MP, a Cyrix 486SLC2/e-50 of some kind and a TI Tx486SLC/E-33PAF.
The only CPU that didn't work was the TI.
Is that a dead chip, or are the TI chips not drop-in 386SX replacements like the Cyrix modules?

Every 386SX upgrade that I've seen that uses a TI chip also has a logic board. The Cyrix 486 chip literally just clips on top of the 386SX - no other ICs at all - but the TIs all use extra ICs. I'm thinking therefore that only the Cyrix will work as a replacement CPU.

Has anyone ever actually managed to use a 486SxLC from TI to replace a 386SX chip?

I've used TI486SXL as a direct replacement for 386DX in several boards. Some work okay and others don't. I suppose it's a bit more of a pain in the ass if you solder in an SXLC chip to your 386SX board and it doesn't work. At the very least all of my boards booted up just fine, some just didn't like certain features turned on. It was years ago and I don't remember specifics.

The Tx486SLC shouldn't have worked any differently than a Cx486SLC. It is an exact copy. Cyrix used TI's fabs to make the chip. The only difference between the two should be the label. TI 486SXL(c) chips are basically the same as the 486DLC except that they have a larger cache and remap one of the registers to turn clock doubling on or off.

I would guess that the reason you see chips on circuit boards is that they are 3.3V parts and need a voltage regulator to work in a regular 386 socket.

One thing I've noticed about the Tx486SXLC upgrade chips is that not only are they on circuit boards, but they're all marked the same. They don't have a "V" marking for 3.3V, nor do they have "3.3V" written on them anywhere.
I think I'm going to stick to Cyrix - at least until I see word from anyone who's managed to put a TI chip in place of a 386SX, or I see a TI upgrade chip which doesn't involve an extra circuit board. A shame, really - those TI chips in the eBay auction are $5 or whatever instead of $45 for a Cyrix!
It just seems like there's loads of remaindered 3.3V chips around and very few 5V ones.

Can you show me a picture of one of these upgrade chips?

There's one here:
http://www.cpucollection.ca/TI486SXLC2-G50-WN.jpg
It's clearly on a little board of its own so I think it must be an upgrade.
I did have a 50MHz upgrade I think at one point (ages ago) and it looked like this. I think it was a Make-It 486 board.

There are two chips on the adapter. One is a triple input NOR gate. Have no idea what that would be for. The other appears to be a VRM, which is strange since the chip doesn't appear to be a 3.3V model.

Spirantho wrote:

There's one here:
http://www.cpucollection.ca/TI486SXLC2-G50-WN.jpg

I think it was a Make-It 486 board.

That's an Evergreen Tech., "REV TO 486" upgrade.

http://web.archive.org/web/19961104092136/htt … om/2sxinfo.html

I have one of these installed in a Tandy 1000 RSX. Works well enough, as long as a heatsink is attached.

What is the name of the utility which enables L1 cache on 386 motherboards that do not have this option in their BIOS ? I suppose it should work with a 486DLC system as well.

I think it might be M5.EXE.

For the most part the DLC and SXL utilities are interchangable...except that you won't be able to turn on clock doubling if you use a DLC utility on a SXL.