VOGONS


Reply 20 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

I was under the impression that the multiplier was locked on this chip, but I have the other settings set correctly.

My board only offers 1.5x and 2x, but I believe 1.5x works as 3x with some chips? I had left it on 2x assuming it didn’t matter if it was locked though.

Though the settings appear to be the same? I assume it’s a mistake on this page… https://stason.org/TULARC/pc/motherboards/D/D … TARION-910.html

EDIT: Added screenshot from the service manual.

Attachments

Reply 21 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

Though if this thing is based on a WinChip C6 and not a WinChip 2 like I had thought, I’m not even sure it’s worth the hassle to get working.

I do now have two 133MHz Pentium Socket 5 machines though, so that’s why I was hoping to upgrade this one a bit. I guess I can toss this on eBay and try to source something else…

Reply 22 of 56, by Meatball

User metadata
Rank Oldbie
Rank
Oldbie

Yes, the Pentiums are normally multiplier locked. A few 166s(?) got escaped unlocked, and I don't know for sure with your IDT chip. I mentioned presuming it was locked (and you'd be aware of course) and to compare the capabilities of your motherboard vs the chip. I have a different motherboard, but a somewhat similar situation. It won't function with a Pentium 233 MMX. Max is 200MMX because the board doesn't support the P233's multiplier. Won't even boot, and no higher FSB is available; again, same as yours.

Anyway, unless someone with more expertise arrives with more ideas, I agree it looks like this particular upgrade path is not available for this machine, unfortunately.

Reply 23 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

Isn’t the voltage wrong for the MMX chips? Unless I’m mistaken they’re all 2.8 volts, while my system is 3.3 volts. Do they just not care about the overvolting? I suppose they’re cheap enough, if you think it should work still.

However, given the service manual claims to only have 1x and 2x multipliers, I’m not sure how I’d get the 3x multiplier needed. Does it reuse the 1x, like 3.5x reuses the 1.5x?

Reply 24 of 56, by Jo22

User metadata
Rank l33t++
Rank
l33t++
Sunoo wrote on 2022-03-12, 19:16:
Jo22 wrote on 2022-03-12, 19:12:
There was an issue with trademarks/brands, I vaguely remember. The proccesor wasn't allowed to be sold under its "WinChip" name […]
Show full quote

There was an issue with trademarks/brands, I vaguely remember.
The proccesor wasn't allowed to be sold under its "WinChip" name in some country (Germany?) - maybe that's why it has the sticker on it.

Windows 3.1 and 98/98SE should be ok with it, also.
These both systems do run well on niche processors in general.

When I did get into Windows, I noticed that it showed up as "CentaurHauls".

But are you saying that this is an issue with Windows 95 and I need to either upgrade to Win98 or downgrade to Win3.1?

I just meant to say that both Windows 3.1 and 98 are useful for troubleshooting. 😅

Windows 95 played an important role in pop culture, but wasn't really finished when it got released.
It had gotten many updates and releases in the years past its original release.

The original version (Win95 RTM), for example, was pretty crashy on Pentium systems, but worked like a charm on Windows 3.x era hardware: 386/486 PCs.

I can only imagine, that this is (was) because Windows 95 was built atop of Windows 3.1x.
Things like VXD components for ACPI/APIC, AGP, PCI etc weren't fully evolved yet or based on drafts for these standards.

Windows 98SE, on the other hand, has a more complete awareness of the new technologies that were released before it.
It can also work around things like incomplete ACPI implementations found in early motherboard BIOSes.

And Windows 3.x.. It pretty much treated every PC like a classic 386 PC with an ISA bus.
Due to its simplicity, it only used the core functions provided by the processor (i286/i386 instruction set) and the AT BIOS.
Because of this, clone processors didn't turn out being incomplete.
The basic 386 instructions were all handled well.

That being said, I'm not saying that Windows 95 isn't worth using - it's often quicker due to being lightweighter. 😺
But it may need quite a few updates to work properly.
Things like new MFC42.DLL, COMCTL32.DLL or OLEAUT32.DLL files. Just to name a few.

These files can be found here: https://www.mdgx.com/web.htm

Good luck! 🙂🤞

Edit: Minor edit.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 25 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

Ah, gotcha, that makes more sense.

For what it’s worth, this machine is running 95B, but I have been considering switching it to 98SE at some point (my other machine is definitely sticking to 95 though), though I’d probably 98lite my way back to the Win95 shell.

However, it really kinda seems like this chip just doesn’t play nice with this computer in a way that’s unrelated to the OS.

Reply 26 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Meatball wrote on 2022-03-12, 22:52:

Yes, the Pentiums are normally multiplier locked. A few 166s(?) got escaped unlocked, and I don't know for sure with your IDT chip. I mentioned presuming it was locked (and you'd be aware of course) and to compare the capabilities of your motherboard vs the chip. I have a different motherboard, but a somewhat similar situation. It won't function with a Pentium 233 MMX. Max is 200MMX because the board doesn't support the P233's multiplier. Won't even boot, and no higher FSB is available; again, same as yours.

Anyway, unless someone with more expertise arrives with more ideas, I agree it looks like this particular upgrade path is not available for this machine, unfortunately.

The multiplier situation is rather complicated.

I have a number of mmx chips of all speeds that are unlocked entirely, but most are locked out of higher multis than their rated speed.

All pentiums that I have tried have more than one multiplier option but the locked ones only go lower.

This makes me wonder though what my pentium 75 supports. (I haven’t checked) perhaps it only supports 1.5x?

The 233 uses the 1.5x multiplier setting.

It sets 3.5x when you set 1.5x

But you need a dual plane motherboard.

If OP can post a good front and back pic of the mobo I can offer some upgrade paths. (If the winchip still isn’t working out)

Would need to see the components and numbers for help identifying and seeing what voltages it’s equipped for (for this I really need to see the actual board) some volt meter testing may also be involved

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 28 of 56, by Sunoo

User metadata
Rank Member
Rank
Member
Sphere478 wrote on 2022-03-13, 01:26:

If OP can post a good front and back pic of the mobo I can offer some upgrade paths. (If the winchip still isn’t working out)

Would need to see the components and numbers for help identifying and seeing what voltages it’s equipped for (for this I really need to see the actual board) some volt meter testing may also be involved

I’d appreciate that, it’ll take some effort to strip it down and get the board out, so I may not get to it this weekend though. I’ve got a multimeter if needed as well, but I’m pretty sure the board is just 3.3v only.

Beyond that, in the off chance it helps, the service manual is available on the Wayback Machine: https://web.archive.org/web/20050115202908if_ … pdf/star9xx.pdf

Reply 29 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Sunoo wrote on 2022-03-13, 01:50:
Sphere478 wrote on 2022-03-13, 01:26:

If OP can post a good front and back pic of the mobo I can offer some upgrade paths. (If the winchip still isn’t working out)

Would need to see the components and numbers for help identifying and seeing what voltages it’s equipped for (for this I really need to see the actual board) some volt meter testing may also be involved

I’d appreciate that, it’ll take some effort to strip it down and get the board out, so I may not get to it this weekend though. I’ve got a multimeter if needed as well, but I’m pretty sure the board is just 3.3v only.

Beyond that, in the off chance it helps, the service manual is available on the Wayback Machine: https://web.archive.org/web/20050115202908if_ … pdf/star9xx.pdf

Looks like socket 5 (single plane)

Installation Procedures
Installing a Higher Performance CPU
The Starion main logic board is equipped with a Pentium processor installed in a ZIF socket (Socket 5 type) and a voltage regulator that supports 3.3V dc Pentium Processors.
To install a higher-performance CPU:
1) Turn off the computer, disconnect external devices, ac power and monitor power.
2) Unlock and remove cover.
3) Lift up on lever to release old CPU.
4) Remove old CPU.
5) Install new CPU.
6) Make sure pin 1 on CPU is aligned with pin 1 on ZIF socket (A).
7) Return release lever to its original position and then set all appropriate CPU jumpers. Refer to “Main
Logic Board Jumper Settings”.
8) Replace and lock cover and connect external devices and restore power.

Okay, so this mostly limits you to 3.3v cpus.

Pics could still help though, but sounds pretty conclusive that single plane socket 5

Amd I believe made chips supporting 3.3v core up to the k6 233 (decent chip, cheap alternative to intel overdrives)

Cyrix/ibm/st looks like 3.3v options exist up to 166gp (not 166mhz) it’s a rating

Winchip: these are kinda slow cpus but 3.3v options I believe exist up to 266mhz but are hard to find and rare and don’t perform well. (More of a collector piece these days)

Intel (non mmx) options exist up to 200mhz, and mmx overdrive options up to 200mhz, the regular mmx chips should not be used on socket 5 despite reports of them “working” it’s improper and over volting, but your call.

Rise: not aware of 3.3v options.

————————-
But, there are technically other options:

Options past 3.3v:

You can get interposers to upgrade to socket 7

Powerleap

Evergreen

Kingston

These are devices that you can use to turn socket 5 into socket 7 and use dual plane voltage cpus in the motherboard which basically opens you up to any of the remaining socket 7 cpus and and even super socket 7 cpus.

The one you will most likey want if you do this is the k6-2+ 570 which you can get cheap (3 for $70) on ebay and can be unlocked into the infamous k6-3+ (see my signature)

Problem is these adapters are expensive.

Your max clock will be limited by the fsb that the mobo can do, the k6 offers multipliers up to 6x so if your max fsb is 66 your max core is 400mhz, if 75, then 450mhz, if 83mhz then 500mhz.

Last edited by Sphere478 on 2022-03-13, 02:22. Edited 6 times in total.

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 31 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++

Updates done, here is a upgrade adapter.

If you keep your eye out you can find them for 150 or so.

https://www.ebay.com/itm/194587220915?hash=it … vUAAOSwst1hrtHc

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 32 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

Thanks, I’ll keep an eye out. $150 is probably doable, but $350 is pushing things a bit.

That said, I’m a bit worried that if this PNY/WinChip thing I’ve got doesn’t play nice with this board, do I have any reason to believe this other chip would work… I’d like to avoid just dumping $150 down a hole if possible. 😜

Reply 33 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Sunoo wrote on 2022-03-13, 02:22:

Thanks, I’ll keep an eye out. $150 is probably doable, but $350 is pushing things a bit.

That said, I’m a bit worried that if this PNY/WinChip thing I’ve got doesn’t play nice with this board, do I have any reason to believe this other chip would work… I’d like to avoid just dumping $150 down a hole if possible. 😜

I can’t think of a reason the winchip wouldn’t work. (Have you tried another os or a reinstall?)

Oh btw, many of these cpus, especially the s7 and ss7 options will work best with a bios update.

You’ll want to check out this thread, and talk to chkcpu and check out his website.

http://web.inter.nl.net/hcc/J.Steunebrink/k6plus.htm

Re: Diy modding support for k6+And 120gb hard drives into bios roms

Btw, if that’s a ceramic winchip (1) 200 I think I have one. Dono if it will help though.

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 34 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

I’ll definitely check into that, if I manage to catch one of them at a good price. I doubt anyone will want to bother modding the BIOS to a rare-ish OEM machine though.

There are some switches on the bottom of the PNY chip, not sure if they matter? I haven’t touched them, and they’re taped over, but J1 shows some evidence of possibly having been moved through the tape.

Attachments

  • image.jpg
    Filename
    image.jpg
    File size
    1.6 MiB
    Views
    506 views
    File license
    Public domain

Reply 35 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Sunoo wrote on 2022-03-13, 02:39:

I’ll definitely check into that, if I manage to catch one of them at a good price. I doubt anyone will want to bother modding the BIOS to a rare-ish OEM machine though.

There are some switches on the bottom of the PNY chip, not sure if they matter? I haven’t touched them, and they’re taped over, but J1 shows some evidence of possibly having been moved through the tape.

If you look, there are three missing pins, and three traces.

Those switches program pins bf0, bf1, bf2

These are the pins for setting the multiplier.

Dono if the fourth switch does anything, or what that may be.

Looks like you have one of the nicer adapters.

Some are just a thin pcb and no riser

Jan (chkcpu) has helped many people with their boards. Doesn’t hurt to ask he is very nice and helpful.

You can play around with those three switches that have the traces. Any combo of them shouldn’t hurt anything, just no post or different multiplier.

(Solved) Cyrix 400gp/366gp multiplier settings

More info

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 36 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

I tried switching the one that appeared to have been changed, and the system no longer POSTed or anything, so I put it back.

And I tried changing the multiplier jumper on the main board, with the Pentium it showed as 100MHz as expected, but no change with the PNY (also expected, I guess, if the switches on it control multiplier).

Also, I’ve discovered that I cannot get the Windows 95 setup to launch no water what I try at this point. No matter which CPU is installed, and attempting to install from both hard disk and CD, it gets to “Please wait while Setup initializes.” then hangs forever. I don’t have Win98 (or blank CDs) handy at the moment to test that.

Reply 37 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Sunoo wrote on 2022-03-13, 02:52:

I tried switching the one that appeared to have been changed, and the system no longer POSTed or anything, so I put it back.

And I tried changing the multiplier jumper on the main board, with the Pentium it showed as 100MHz as expected, but no change with the PNY (also expected, I guess, if the switches on it control multiplier).

Also, I’ve discovered that I cannot get the Windows 95 setup to launch no water what I try at this point. No matter which CPU is installed, and attempting to install from both hard disk and CD, it gets to “Please wait while Setup initializes.” then hangs forever. I don’t have Win98 (or blank CDs) handy at the moment to test that.

Yeah, with that adapter your mobo settings shouldn’t do anything.

Try taking cards out, extra ram, go back to the 133, go minimal using known good parts, get back to working and go from there. Nothing you’ve done should have harmed anything.

Try a bios reset. Disable system bios cachable. (Sometimes that gives issues)

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 38 of 56, by Sunoo

User metadata
Rank Member
Rank
Member

I did try resetting the BIOS, but I’ll try pulling out other cards and disabling that cache tomorrow and see if that makes any difference.

That said, it’s now the same build as what I had and was working before I started touching anything with this chip. Also, the old Windows 95 instance started working again while the 133 is installed, not sure why it had stopped for a while there, and not sure why the installer would have stopped working.

Reply 39 of 56, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Sunoo wrote on 2022-03-13, 03:19:

I did try resetting the BIOS, but I’ll try pulling out other cards and disabling that cache tomorrow and see if that makes any difference.

That said, it’s now the same build as what I had and was working before I started touching anything with this chip. Also, the old Windows 95 instance started working again while the 133 is installed, not sure why it had stopped for a while there, and not sure why the installer would have stopped working.

Sometimes windows 9x will crash once, then second time your try it works. Seems like mine does that whenever I change hardware. Maybe try and make sure it’s crashing twice in a row?

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)