• Please review our updated Terms and Rules here

Anyone put a Am5x86-133Mhz on a Soyo SY-25N?

njroadfan

Veteran Member
Joined
Jan 21, 2011
Messages
1,629
This board: http://stason.org/TULARC/pc/motherboards/S/SOYO-COMPUTER-CO-LTD-486-SY-25M-N-P.html

I have the original manual here, this board predates the 5x86 so no 4x multiplier, but does support 3.3v 100Mhz DX4 CPUs. The AMD chips seem more readily available then the DX4s. Right now the machine has a DX2-66 in it. From what I have read, apparently the board can be set to 2x multiplier and the AMD will automatically clock up to 4x for 133Mhz operation.
 
From what I have read, apparently the board can be set to 2x multiplier and the AMD will automatically clock up to 4x for 133Mhz operation.

That was standard for all Socket 3 boards... there was never a dedicated 4x setting, it's just that all the chips that had a 4x multiplier used the 2x setting on the board.

I see no reason why a 5x86 wouldn't work, all the main issues seem to be covered... voltage, multiplier, and even write-back cache support.
 
as long as the voltage is correct, its not going to hurt anything is you give it a shot. provided you dont put it in backwards or something. I have a pc chips board that I use for testing because the jumpers are quite literally FSB and multi. It doesnt care if its an intel or amd or cyrix. all the same jumpers.
 
Well scratch that, the "N" version of the board is 5v only, the jumpers and voltage regulator for 3.3V are missing. :mad:
 
Does the motherboard have an open section for the missing voltage regulator? Could just be missing that and a linking fuse/resistor. If it was my board, I next figure out what it was missing, and duplicate it. Sounds like it be a fun easy project! :)
 
Does the motherboard have an open section for the missing voltage regulator? Could just be missing that and a linking fuse/resistor. If it was my board, I next figure out what it was missing, and duplicate it. Sounds like it be a fun easy project! :)

Sure does. The silver pads are there for the voltage regulator although I have no clue what part to source. As for the jumper block, its missing and there are two small metal pieces serving as a permanent connection in the 5v position, easy to de-solder and put the proper jumper block in. I don't see any other missing components, but I haven't pulled the board out of the case to check.

Otherwise.... anyone got a spare Pentium P24 or DX4 100Mhz 5v overdrive chip?... or one of those 3.3v socket adapters?
 
Well first step then would be to identify the existing voltage regulator. It may be possible to tweak the output voltage to 3.3-3.6v with a variable resistor. Can you please post the details found on the VRM, or post a clear enough picture they are readable? I may be able to help more with this information, provided you feel comfortable using a soldering iron. :)
 
You should be able to get this to work iwth a simple 3.3V LDO, such as the LT1085CT-3. (there are similar SMT versions available, such as the LM2675M-3. The datasheet says that the operating voltage tolderance is +/- 0.3V and less than an amp. Make sure that the voltage regulators have an adequate heat sink; include a decoupling cap on the output and you should be fine.
 
Managed to find a picture of the "P" version of the board that has the 3.3V VRM. Looks like the only parts missing is the VRM and the jumper block. I might try this mod assuming I can source the parts cheaply locally (most places online have minimum order requirements). Which pin in the socket does one test for CPU voltage? Obviously I don't want to fry a CPU, although I have read people running AMD 5x86 chips at 5v without blowing them up!

http://imagestore.ugbox.net/image/IMG_2448_ac1cdec2f380a83c56bb39e.jpg
 
Apparently, imagestore doesn't allow linking from other websites. I get a "Your domain isn't allowed to access this image" type of message.

Maybe a different link would help.
 
http://www.hardwaresecrets.com/article/373

scroll to socket 3 > click pinout

Only trick I know of to perhaps help is, I recall someone telling me when checking voltages are correct on a 486 motherboard, always use a 5v cpu to test. 5v on a 3.3v could fry it, but 3.3v on a 5v cpu would do no damage other then not boot.

Hope someone can help more, as I will most likely use the same method to check 3.3v on my Dell 466v. It seems my machine was also sold as a dx/4 100, and supposedly has socket 3. But dell makes no references to 3.3v settings in any their manuals. Only 1 jumper labeled pent overdrive/dx4. I assume this is for 3.3v but I would love to know for sure. So if you can figure out which pins to check to 3.3v, it would definably help me out also! I just scored a 5x86 133 on fleabay for $6.61 shipped. =)
 
Last edited:
Just ran into this searching for what pins to bridge and which to monitor for vcc, and came onto this gem!

See http://web.inter.nl.net/hcc/J.Steunebrink/amd5x86.htm for the whole page this came from, it has the pinout chart needed


... if you have a multimeter, you can use the following procedure to check this out.

Unplug and then open the system.
Lift the ZIF socket handle and take the CPU out. Store it in an anti-static bag or box or place it with it pins on a piece of conductive foam.
The VOLDET pin is pin S-4. It is tied to ground (Vss) internally in the CPU. Simulate this by inserting a wire from pin S-4 to Vss pin R-4.
Now insert two wires for the multimeter. The ground wire goes in Vss pin S-10 and the live wire in Vcc pin R-10.
Close the ZIF socket handle to connect the wires.
Switch the system on and check for 3.45V + 0.15V DC on the multimeter. If you see 5V, then your motherboard is not autodetecting.
Restore everything as it was.

Caution: Anytime before you touch any component inside your system, place one hand on the chassis, such as the power supply case, to prevent electrostatic discharge dam
 
Well scratch that, the "N" version of the board is 5v only, the jumpers and voltage regulator for 3.3V are missing. :mad:
I have a similar board as yours ( SOYO 486 VESA 4.2: http://stason.org/TULARC/pc/motherboards/S/SOYO-COMPUTER-CO-LTD-486-486VESA-VER-4-2.html ). Mine has the voltage fixed at 3.3V. The area where the voltage selecting jumpers were supposed to be, does not have jumpers, but it is soldered with soldering iron in the same way as it would be if there were jumpers set for 3.3V operations (I might post some pictures). It can easily be desoldered and resoldered the other way around if you know about soldering

This board: http://stason.org/TULARC/pc/motherboards/S/SOYO-COMPUTER-CO-LTD-486-SY-25M-N-P.html

I have the original manual here, this board predates the 5x86 so no 4x multiplier, but does support 3.3v 100Mhz DX4 CPUs. The AMD chips seem more readily available then the DX4s. Right now the machine has a DX2-66 in it. From what I have read, apparently the board can be set to 2x multiplier and the AMD will automatically clock up to 4x for 133Mhz operation.

It is true that setting the multiplier to 2 sets an Am5x86 to x4, but with my board, there is no jumper for multiplier selection. Unfortunately this must be done under BIOS automatically. I tried all possible combinations and it's a no-go. And I managed to get 2x and 3x multiplier settings with the exact same jumpers settings and different CPUs.
But this didn't stop me from forcing an Am5x86 to 4x. I bent the CPU-pin that is responsible for the multiplier and grounded it myself (manual 2x multiplier force) and it worked.
 
Last edited:
You can run a 3.3v chip on a 5v board just fine in most/all cases. I grant that this probably isn't good for the chip, but it does function. Maybe my Am5x86 was especially hardy, but before I realized that there was a voltage disparity I'd put it in a 5v Socket 0/1 system, and had been running it for months that way. I figured "what the hell" and kept it in there until I could get an appropriate upgrade chip (and eventually a POD83, sometimes I swap between the upgrade Am5x86 and that.. I love advanced 486s.. speaking of which, I gotta get myself a Cyrix 5x86 100 and 133mhz to add to my experience with the various late 486 chips. :3).

Also, though this has been said in various ways in the thread, I'm not sure if it was gone over like this:

A 486 board with multiplier selections is *not* the norm - in fact, I've never seen one. 486 chips with multipliers are all internally multiplied AFAIK, and some of the more advanced motherboards that understand multipliers/etc. send things automatically (as you put it) from the BIOS/CMOS, which the chip can then interpret. However, most boards don't even have that, nor the jumpers, they just rely on the chip's internal workings, because that was the norm for the vast majority of the earlier 486 boards, and these were upgrade chips (yay run-on sentence!).

As such, ignore the multipliers. Use chkcpu (http://web.inter.nl.net/hcc/J.Steunebrink/chkcpu.htm) or a similar tool to check the speed and specs for the chip, once installed, and you'll know whether the multiplier is working right, as you know what the FSB is I'm sure (and usually that either runs at a fixed speed or *does* have a jumper - most often 25/33 selectable, with better boards having 25/33/40/50/60 or 66 or combinations thereof, and the earliest 486 boards running at 16).

Excuse my rambling, but I love 486 machines, hah. I hope some of the information I spouted here was useful, and if you'd like clarification don't hesitate to ask :3
 
Last edited:
Also, though this has been said in various ways in the thread, I'm not sure if it was gone over like this:

A 486 board with multiplier selections is *not* the norm - in fact, I've never seen one. 486 chips with multipliers are all internally multiplied AFAIK, and some of the more advanced motherboards that understand multipliers/etc. send things automatically (as you put it) from the BIOS/CMOS, which the chip can then interpret. However, most boards don't even have that, nor the jumpers, they just rely on the chip's internal workings, because that was the norm for the vast majority of the earlier 486 boards, and these were upgrade chips (yay run-on sentence!).
Well the mainboard does have a setting for the CPU multiplier. It is not a direct multiplier selection (couldn't actually be, as multipliers were selected by the CPU eitherway). This is a certain pin on the 486 socket and the mainboard can either:

1) Ground it
2) Leave it with no connection (or connected to 3.3V)

Early 486 CPUs that were designed to work with either 1x or 2x multiplier have no effect from this, but when CPUs with 3x multiplier arrived, they were operated in either 2x (ground) or 3x (no connection or +3.3V) speed according to input they receive from this mainboard pin. When the Am5x86 arrived, it still used that same pin but worked at 4x (ground) or 3x (no connection or +3.3V).

Anything other than a very early 486 board, has a way to switch between these 2 options, and in a way, set the multiplier for some CPUs. If this is done via a jumper, then it is good, as you know what you are doing. If it is done via automatic BIOS detection mode thought, then more likely an Am5x86 will be set at 3x multiplier.
 
One of the advantages to that setup is that upgrade chips with VRMs - and probably some stand-alone VRMs have this too - can have jumpers on-board to select that pin setting, among other settings; such is the case with my Am5x86-133 upgrade chip by Evergreen.

The pin setting had slipped my mind - I thought you meant a proper multiplier jumper or some such, which became common on Pentium boards (particularly late ones, inc. SS7). I remember before I was aware of that, I had a machine that I couldn't get my 133Mhz chips to run any higher than 100Mhz in - a PS/Valuepoint - and that's almost certainly why.
 
Last edited:
Back
Top