Ariakos wrote:

Three new laptops arrived: […]
Show full quote

Three new laptops arrived:

Toshiba 440CDT (Pentium 133 MMX)
Compaq Armada V300 (Celeron 466)
IBM Thinkpad 365XD (Pentium 133)

Hah! I saw the ThinkPad and immediately recognized it as a 365XD as I bought (and still own) the same laptop for university waaay back in the mid to late 90s. I really gotta find my external floppy for it. PITA is find one as you can find them online relatively cheap, but trying to find the _right_ one is what's got me nervous. Seems that even mid-run, IBM changed the connector, so it's really hard to narrow it down. If I cannot find mine (it used to be in the laptop bag, but no more) I might have to bite the bullet and buy a couple different ones. 🙁

The Serpent Rider wrote:

Virgins valued quite high, huh?

Meh, let those who wish fight over them.

Give me the cheap experienced ones with a story 😉

Talking of which, proof that being nice pays off: a vague acquaintance of mine off another forum posted a while back that he was desperately seeking the case his first childhood computer was in. Nothing special, a late 1990s low-end Highscreen minitower - but it meant something to him. A few weeks back I spotted one in a lot of dubious-looking partially complete cases. I went for it and discovered that the Highscreen case was actually in excellent state. The rest was utter scrap, but I got something not mentioned in the advert - the components that had been in them. They were in better condition than I expected based on the state of the cases, and most actually worked too. Not a bad surprise.
Fast forward to last weekend. I at last found time to drop off the case at this bloke's place. He lives quite a distance away and doesn't have any transport of his own. As well as the case I chucked in some colour-matched CD and floppy drives and a CT3670 (Soundblaster 32) that was in the lot with the case. In return, he gave me this:

Two almost identical So7 boards, an MSI MS-5118 and 5122. Both with the underrated SiS5511 chipset. The only difference is the cache - async on the 5118 and PLB on the 5122. Above that is a Jetway/J-Mark/whatever J-446A (with BIOS identifying as Rectron RT-4S3). Very late 486, so I can at last test my 3.45V DX4 CPUs and that IBM/Cyrix 5x86 😀 - it currently has an Am486DX4-100 on it. In the top right corner a Velocity 100 that seems to have been flashed to a Voodoo3 1000, which is exactly what I need to compare performance of my MS-6168 to other contemporary boards without the onboard Voodoo.

And to cap it off I sold one of the motherboards from the lot with beat up cases for an amount that pretty much covered all fuel costs involved. That completes the all-round good feeling 😀

dionb wrote:

[...]Both with the underrated SiS5511 chipset.[...]

In what regard "underrated"? I have an AOpen AP5S mainboard with that chipset (and a layout nearly identical to the MSI 5122) that I actually got with a used Highscreen Sky mini tower pc some time around 1999. While the case is long gone, the board is still there and I am planning to build a system around it. So in what (if any) regard would that chipset be better than, let's say, an i430TX?

gex85 wrote:

dionb wrote:

[...]Both with the underrated SiS5511 chipset.[...]

In what regard "underrated"? I have an AOpen AP5S mainboard with that chipset (and a layout nearly identical to the MSI 5122) that I actually got with a used Highscreen midi tower pc some time around 1999. While the case is long gone, the board is still there and I am planning to build a system around it. So in what (if any) regard would that chipset be better than, let's say, an i430TX?

Some time round 1999? It was ancient by then. Compared to the i430TX the only real advantage the 5511 has is that you can set the cache to WT-mode and cache 128MB of RAM as opposed to the max 64 of the i430TX. In everything else the i430TX blows it away - but that's no surprise given that the 1997 i430TX is a full two years younger than the 5511. If you're comparing it to Intel chipsets, the i430FX is the one to compare it to, as both are 1995 vintage chipsets.

In terms of raw memory performance the i430FX had a slight edge, but the 5511 did a few neat tricks:
- letting you cache 128MB of RAM (not relevant in 1995, but nice now)
- support for 1MB of cache - theoretically letting you cache 256MB.
- the memory controller will work in 32b mode with a single 72p SIMM. Performance is lousy, of course, but makes troubleshooting bad memory a lot easier.
- 5511 can run the PCI bus asynchronously, so it will stay at 33MHz regardless of what you're doing with the FSB (an advantage both for overclockers and for people stuck on old 50/60MHz FSB CPUs)
- the 5511 was the first chipset to support UMA, so letting a VGA core (in this case, a discrete onboard SiS6202 or 6205) share access to system memory instead of requiring its own. This is of course a low-end feature that didn't get used much in the day (note no VGA chips on either of these boards) due to 1995/1996 era systems really not having the memory bandwidth to spare to allow this to work, but is the great-granddaddy of every integrated VGA solution today.

So if you're looking for the fastest So7 chipset, look elsewhere (jury still out between the higher clocking Aladdin V and the bigger cached MVP3), equally if you're looking for the fastest chipset clock-for-clock (a virtual tie between the i430TX and Ali Aladdin IV+, although the Aladdin has a much better feature set). But for a late 1995/early 1996 system it's a very decent choice and a nice bit of history to boot 😀

Wouldn't the HX chipset be the fastest.

Wouldn't the HX chipset be the fastest.

EDO RAM only, so nope. Also no UDMA33.

kixs wrote:

Wouldn't the HX chipset be the fastest.

In mid-1996: certainly, albeit neck-and-neck with the SiS5571. And its EDO performance remained competitive to the end - but EDO got blown away by SDRAM, so both clock-for-clock and in absolute terms SDRAM chipsets were the fastest So7 beasts.

430TX is also compatible with Throttle.

Thallanor wrote:

Ariakos wrote:

Three new laptops arrived: […]
Show full quote

Three new laptops arrived:

Toshiba 440CDT (Pentium 133 MMX)
Compaq Armada V300 (Celeron 466)
IBM Thinkpad 365XD (Pentium 133)

Hah! I saw the ThinkPad and immediately recognized it as a 365XD as I bought (and still own) the same laptop for university waaay back in the mid to late 90s. I really gotta find my external floppy for it. PITA is find one as you can find them online relatively cheap, but trying to find the _right_ one is what's got me nervous. Seems that even mid-run, IBM changed the connector, so it's really hard to narrow it down. If I cannot find mine (it used to be in the laptop bag, but no more) I might have to bite the bullet and buy a couple different ones. 🙁

I have an IBM 755c that I tried to restore but can't find parts. Basically the ribbon connector from the motherboard to LCD connector goes bad. It's a cheap design.
The Ribbon is rolled up solution that eventually wears out from rubbing against the plastic thinkpad housing. Causes the display not to work or show lines of distortion.
You can still find NEW displays but this ribbon cable is hard to find. But you can always use them with an external display if the LCD does not work.

So I stopped trying to restore these really old Thinkpads. I settled on the IBM 380 with CDROM drive. These are really easy to fix and parts are easy to find.
Also they have a built in CDROM.
And IBM has drivers for DOS, Win 3x, WinNT and Win 95 and 98 and OS2.
The 380 works very good and reliable. The Display is not to good however and they use Crystal Audio.
That Toshiba probably has the best specs.

Intel486dx33 wrote:

I have an IBM 755c that I tried to restore but can't find parts. Basically the ribbon connector from the motherboard to LCD conn […]
Show full quote

I have an IBM 755c that I tried to restore but can't find parts. Basically the ribbon connector from the motherboard to LCD connector goes bad. It's a cheap design.
The Ribbon is rolled up solution that eventually wears out from rubbing against the plastic thinkpad housing. Causes the display not to work or show lines of distortion.
You can still find NEW displays but this ribbon cable is hard to find. But you can always use them with an external display if the LCD does not work.

So I stopped trying to restore these really old Thinkpads. I settled on the IBM 380 with CDROM drive. These are really easy to fix and parts are easy to find.
Also they have a built in CDROM.
And IBM has drivers for DOS, Win 3x, WinNT and Win 95 and 98 and OS2.
The 380 works very good and reliable. The Display is not to good however and they use Crystal Audio.
That Toshiba probably has the best specs.

I have fortunately never had any issue with the ribbon cables, despite my 365XD seeing serious use. In my case, it really is just the FDD that I need to find. And the worst part is I know I own it. Somewhere. 😀

Until then, I cannot do much in the way of nuke and start over with my ThinkPad as I believe I looked just a couple months ago and it does not allow you to boot from the optical drive. So I'm kinda boned, shy of pulling the HDD, nuking, repartitioning, and setting up the basics on another computer, and then putting it back in. I'd just rather have the FDD to do it from scratch, nice and easy. 😀

dionb wrote:

Some time round 1999? It was ancient by then. Compared to the i430TX the only real advantage the 5511 has is that you can set th […]
Show full quote

gex85 wrote:

dionb wrote:

[...]Both with the underrated SiS5511 chipset.[...]

In what regard "underrated"? I have an AOpen AP5S mainboard with that chipset (and a layout nearly identical to the MSI 5122) that I actually got with a used Highscreen midi tower pc some time around 1999. While the case is long gone, the board is still there and I am planning to build a system around it. So in what (if any) regard would that chipset be better than, let's say, an i430TX?

Some time round 1999? It was ancient by then. Compared to the i430TX the only real advantage the 5511 has is that you can set the cache to WT-mode and cache 128MB of RAM as opposed to the max 64 of the i430TX. In everything else the i430TX blows it away - but that's no surprise given that the 1997 i430TX is a full two years younger than the 5511. If you're comparing it to Intel chipsets, the i430FX is the one to compare it to, as both are 1995 vintage chipsets.

In terms of raw memory performance the i430FX had a slight edge, but the 5511 did a few neat tricks:
- letting you cache 128MB of RAM (not relevant in 1995, but nice now)
- support for 1MB of cache - theoretically letting you cache 256MB.
- the memory controller will work in 32b mode with a single 72p SIMM. Performance is lousy, of course, but makes troubleshooting bad memory a lot easier.
- 5511 can run the PCI bus asynchronously, so it will stay at 33MHz regardless of what you're doing with the FSB (an advantage both for overclockers and for people stuck on old 50/60MHz FSB CPUs)
- the 5511 was the first chipset to support UMA, so letting a VGA core (in this case, a discrete onboard SiS6202 or 6205) share access to system memory instead of requiring its own. This is of course a low-end feature that didn't get used much in the day (note no VGA chips on either of these boards) due to 1995/1996 era systems really not having the memory bandwidth to spare to allow this to work, but is the great-granddaddy of every integrated VGA solution today.

So if you're looking for the fastest So7 chipset, look elsewhere (jury still out between the higher clocking Aladdin V and the bigger cached MVP3), equally if you're looking for the fastest chipset clock-for-clock (a virtual tie between the i430TX and Ali Aladdin IV+, although the Aladdin has a much better feature set). But for a late 1995/early 1996 system it's a very decent choice and a nice bit of history to boot 😀

Thank you very much for this detailed write-up! I have a complete 430TX based system with a Pentium 233 MMX which does everything I need it to do pretty well, so I am not on the hunt for the fastest chipset. Just have this board sitting in a box for years now, booted it up a few weeks ago to see if it still works, and it does. So I thought I should do something with it, and since I didn't do much investigating on the chipset yet, your statement made me curious. And yes, when I got that system around 1999 it was already completely outdated.

dionb wrote:

- the 5511 was the first chipset to support UMA, so letting a VGA core (in this case, a discrete onboard SiS6202 or 6205) share access to system memory instead of requiring its own. ...

That would certainly explain these oddball socket 7 boards with four SiS chips on, one of them being the 5511 and another being the 6205 without any discrete video memory chips. I originally assumed it was just the typical "let's put the video card on the MB" stuff without looking closer and had no idea it was actually more-or-less a proto-IGP implementation...

SW-SSG wrote:

dionb wrote:

- the 5511 was the first chipset to support UMA, so letting a VGA core (in this case, a discrete onboard SiS6202 or 6205) share access to system memory instead of requiring its own. ...

That would certainly explain these oddball socket 7 boards with four SiS chips on, one of them being the 5511 and another being the 6205 without any discrete video memory chips. I originally assumed it was just the typical "let's put the video card on the MB" stuff without looking closer and had no idea it was actually more-or-less a proto-IGP implementation...

Yep, that's the idea. And it's a great one if you want your Pentium performing a bit worse than a similarly-clocked 486... 😵

The next step was the SiS5596 chipset, which was basically a 6205 integrated into a 5571 (different package though, the 5571 was single-chip, the 5596 had a separate southbridge). Performance was still as crap, but with a K6-2 or 6x86 M2 you had more wiggle-room - at least it would perform like a slower Pentium. Still gave the later Cyrix chips they were inevitably paired with a bad name though, which helped kill Cyrix 😢

On to happier things. Today is keyboard day:

The disappoinment is the white Alps board. Bought based on a fuzzy picture, though mercifully I didn't pay much for it. Before today I didn't know Alps made (bad) rubber dome keyboards. Now I do...

The one behind it is nicer, a Unicomp buckling-spring board, i.e. supposedly the IBM Model M reincarnated. I wouldn't go that far, the build quality and sheer weight and stability doesn't come close - but it still types gloriously and has USB to boot. Needs a bit of a clean though.

And finally I at last got my hands on a keyboard for my Olivetti M24 (pre-XT PC-compatible with some AT-class features - and a totally non-standard keyboard interface). Just one 'problem': it came with another M24. This one's much less pristine than my original one, and is supposedly dead or dying - booting maybe one time in ten. But at best I can fix that, at worst I have a pile of spare parts for my other one. Also, this one has a beautiful and absolutely huge Microscience MFM controller, with matching 20MB drive. The other one also has a 20MB drive, but stepper, not voice coil like this one. I'll probably be moving this drive over for starters 😉

I even have a purpose for these dinosaurs: my partner wants to start our son playing Sokoban. She wanted to do it in Dosbox on a tablet. I have a better idea 😉

I have been wanting to put together a 386 machine for a while now and my first piece of it arrived today.

It's actually a hybrid board - Opti 495SLC. It is pre populated with a AMD DX40 and a 40mhz ULSI Maths Co Processor and 4mb ram. Battery hasn't leaked but I'll be replacing this with a new one. Very happy with this pick up.

Secondly a Creative TNT2 Ultra

That board is pretty darn cool. I didn't know they used the same style battery the Macs used. Does that also take a 486 processor as well, or am I wrong about that?

It's a combo 386 or 486 CPU board. It's a shame the 386 is soldered on and 132-pin socket is missing.

kixs wrote:

It's a combo 386 or 486 CPU board. It's a shame the 386 is soldered on and 132-pin socket is missing.

Yes but in saying that, I’ve found that chipset to be a little fussy with VLB cards so sticking to 386 and ISA should keep things simple in that regard.

In saying all THAT tho it is a good 486 board because the turbo function actually under-clocks the CPU, which I’ve found to be a nicer slowdown resulwould than adding wait states, etc like other boards I’ve tried.

Great pick up!

I have two of these boards but with 386 socket. It can be clocked to 16/20/25/33/40/50MHz with a 486 cpu (not sure for 386 any more). VLB works fine with both cpus.

Intel486dx33 wrote:

I have an IBM 755c that I tried to restore but can't find parts. Basically the ribbon connector from the motherboard to LCD conn […]
Show full quote

I have an IBM 755c that I tried to restore but can't find parts. Basically the ribbon connector from the motherboard to LCD connector goes bad. It's a cheap design.
The Ribbon is rolled up solution that eventually wears out from rubbing against the plastic thinkpad housing. Causes the display not to work or show lines of distortion.
You can still find NEW displays but this ribbon cable is hard to find. But you can always use them with an external display if the LCD does not work.

So I stopped trying to restore these really old Thinkpads. I settled on the IBM 380 with CDROM drive. These are really easy to fix and parts are easy to find.
Also they have a built in CDROM.
And IBM has drivers for DOS, Win 3x, WinNT and Win 95 and 98 and OS2.
The 380 works very good and reliable. The Display is not to good however and they use Crystal Audio.
That Toshiba probably has the best specs.

I have 2 Thinkpad 755CX (one is mine from back in the day, one I bought about a year ago), they both have issues but not display-related. The LCD and all connections to it are good. My old one has keyboard issues, and I tried 2 different replacement keyboards with the same result: the left half of the new keyboard I plug in doesn't work. I don't know if it's the keyboard's problem (though it seems suspicious that 2 keyboards obtained from different sources have the same defect) the motherboard's or something in-between. The other one that I bought more recently works fine, but the original owner never upgraded the BIOS and because it has an ancient one, the Mwave modem+sound doesn't work in Win98, only in 95. And I cannot flash a newer BIOS because the geniuses at IBM decided that you cannot flash a BIOS unless the laptop is both plugged in and has a battery that is 100% charged. And all batteries for this vintage of laptop don't hold a charge, I tried about 8 or 10 of them with no luck. I even tried rebuilding one but failed miserably.

dionb wrote:

pre-XT PC-compatible with some AT-class features

So it's a proprietary system? 😉 By the way I don't think that this Olivetti predates the 1983 IBM PC XT, it's probably from '87 or so.