arti9m wrote on 2025-01-09, 04:24:

I bet you also wouldn't allow sharing modified BIOS copies you actually sell. Thankfully, those are not needed anymore.

Cheers to that

arti9m wrote on 2025-01-09, 12:18:

Sorry, missed that initially. You're wrong here again, I already booted both max1024 BIOS . it means your CPUs are also not synced properly (just like with my adapters) and they will NOT work in a multiprocessor OS. Possibly due to FSB being too high.

My CPUs are right )) they have right BIOS by mentat-vvo. But the Asus P2B-D has many tweaks to do with soldering to operate properly. My board does not install WinXP in ACPI mode with dual Piii-coppers. MPS mode, however, worked for me. I’d try kmeow’s new bios btw if he had corrected those issues I’ve addressed. It seems he did not.

Ps. Freeway mobo is on its way to receive patched bios.

I have a few questions for the community regarding upcoming revision 2 of the adapter.

1. Do you need jumpers for selection between 66/100/133 MHz bus frequency? Do you need physical jumpers or SMD 0-ohm resistors would be fine? Or dipswitches, perhaps?
2. Do you need jumpers for voltage selection (upto 3.5V, but PPro itself runs at 3.3V anyways)? Can those be SMD 0-ohm parts?
3. Do you need physical jumpers for the BR1/BR3 hack to be able to choose between Dual 440BX board and any other dual board, or SMD jumpers would suffice?
4. If you have Pentium Pro 1M and/or Pentium 2 Overdrive, could you please test if certain pins are shorted with a multimeter? I'd provide a photo of which pins to test.
5. Do any of you plan to use Dual PPro setup on a motherboard that can do Data Bus ECC? Does anyone even have a Slot-1 mobo capable of that?
6. Do any of you plan to do any kind of JTAG debugging?

I'm trying to omit as much unnecessary stuff as possible to make the design simpler.

I have 1M Pentium Pro CPU's, can do some measuring.

Regarding jumpers, always good to have, definitely preferred over 0 ohm resistors.

maxtherabbit wrote on 2025-01-09, 12:50:

arti9m wrote on 2025-01-09, 04:24:

I bet you also wouldn't allow sharing modified BIOS copies you actually sell. Thankfully, those are not needed anymore.

Cheers to that

Try asking Artyom for P2B-D bios from max1024, you will be surprised.

Got successful start of a pair PPro SY048 /200 /512 in Freeway FW-6280BXDR/155 .
Red goes faster! Up to 155 MHz of system bus.
Need to find a proper solution for cooling now. There’s only 35mm gap between CPUs.
Upd: Detects two CPUs up to 112 MHz of system bus at default voltages Vcpu and Vio @ 3.3v.

Back in 1997 I used a Single Pentium Pro CPU computer at work.
It looked something like this:

HP Vectra Tower computer
WinNT 3.5.1
32 or 16mb of RAM, I don’t remember but not more than 32mb RAM.
3com Ethernet NIC
Adaptec SCSI
Sound Blaster
HP CDROM
HP 8mm DAT tape backup
HP Laserjet printer
HP Monitor, HP Clickity keyboard, HP Mouse

Software included:
Office 97
Lotus office
Netscape
Nortons Anti virus

There are some results of overclocking of a pair of Pros / 512k at default voltages. Got a stable run @2x112 MHz in Asus P2B-D. I’m trying to use SETFSB utility to go higher , but I need to overvolt the CPUs to 3.6v. I believe I will have to go as high as 3.8v of Vcore to reach 2x133 . I've noticed that my mobo unluckily has poor clock generator ICS 9150 instead of 9250 which does not allow using 124 MHz FSB. I will find a donor to swap a proped clock gen later.
p.s. the #BR mod has to be applied only to the one of Asus C-P6S1 slotkets to run in a dual iBX mobo. To install a mod you need a 0603 0R resistor, a soldering iron and something sharp to cut traces.

I was able to run Blacktop / 1024kb PPro in red unmodded slotket and and the second CPU was 512kb SY040 in green #BR1-modded slocket.
Both FSB=112MHz overcklocked

Should try 1.5x 133 to confirm the fsb works

At 3.8V with power consumption *that* high I wonder if P2B-D VRMs are even capable of delivering enough power. Especially with those 25 years old capacitors. I'd swap them for solid-states. One more thing that can be done for stability is solder additional SMD ceramics directly to Socket pins, between Vcc and GND.

Also on C-P6S1 there's a tiny ceramic capacitor that goes from FSB clock trace to GND for some reason. I have no idea what impact it has on the clock but I didn't find such cap on ALR adapter and on native Socket 8 boards. Maybe removing it would allow for a few more MHz.

Mentat-vvo wrote on 2025-01-16, 03:00:

p.s. the #BR mod has to be applied only to the one of Asus C-P6S1 slotkets to run in a dual iBX mobo. To install a mod you need a 0603 0R resistor, a soldering iron and something sharp to cut traces.

Have you tried running one BR-modded slocket together with an unmodded slocket on 440LX or Via, but in reverse? Does it work?
I no longer have any adapters on hand to test it 😒

Would be nice if we could just swap adapters to make it work on non-BX boards, that way there would be no need to create BR1/BR3 jumper, which could negatively affect max possible FSB.

myne wrote on 2025-01-16, 05:51:

Should try 1.5x 133 to confirm the fsb works

Pentium pro does not have 1.5x. Do you think 112 and 124 MHz does not confirm the FSB to “work”? I had only one mobo that was able to post (but not boot) with my capable 512kb PPro @ 2x133 at 3.8v. It was VIA-based iWill WD-133. And 3.8v was the result of Vcache shortening.

arti9m wrote on 2025-01-16, 06:35:

Have you tried running one BR-modded slocket together with an unmodded slocket on 440LX or Via, but in reverse? Does it work?
I no longer have any adapters on hand to test it 😒

I don’t have any dual LX or Via mobos since I sold them. The #BR Jumper may affect the performance the same way as any other jumper. You have too many of them 🤣

Mentat-vvo wrote on 2025-01-16, 07:18:

The #BR Jumper may affect the performance the same way as any other jumper. You have too many of them 🤣

No, BR signals are part of the GTL bus, which works at the FSB speed, and all the limitations apply (eg. trace length, stub length, impedance, etc).
BR pins always have activity on them.

BSEL and VID pins are just shorts or n/c and they just tell the motherboard circuitry what to do. They are not part of any signal group.
BSEL pins are not even connected to the CPU at all at any time.
VID pins are either n/c or tied to GND on the CPU package, depending on CPU model.

arti9m wrote on 2025-01-16, 06:35:

At 3.8V with power consumption *that* high I wonder if P2B-D VRMs are even capable of delivering enough power. Especially with those 25 years old capacitors. I'd swap them for solid-states. One more thing that can be done for stability is solder additional SMD ceramics directly to Socket pins, between Vcc and GND.

The best way is to solder them directly to the ceramic CPU package on the footprints intel so kindly left for us. I've done all of my 512/256 chips that way. Of course the 1M doesn't have that as an option...

arti9m wrote on 2024-12-21, 02:48:

So eh.. A certain somebody PMed me asking for gerbers, but for the first time ever I haven't figured out how to reply :-/ Must b […]
Show full quote

So eh.. A certain somebody PMed me asking for gerbers, but for the first time ever I haven't figured out how to reply 😒
Must be getting old 😁

Anyways, here are pre-production gerbers (checked & modified by NextPCB). Use at your own risk.
I still haven't tested it for stability because honestly I'm not too interested in this revision since it doesn't work in dual setup.
You absolutely must use thick inner layers (prefereably 2oz) for adequate power delivery.

Final reason for failing to work in dual CPU setup is thought to be excessive traces length. I didn't think about the traces leading to the intermediate CPU as "stubs", but I should've. Any traces that go off the main path (from chipset to last CPU) shouldn't be longer than 7.5cm in length, and my adapter traces are almost twice as much as that.

Also I am now left with ~30 boards that I don't really need. I'd just sell them for their production price but I most probably won't be able to mail them to you =(

First of all, I apologize for my poor English. I am from Taiwan. I have used the gerber file provided by Arti9m to ask the manufacturer to make a CPU adapter. I don't want dual CPU operation, I only want single CPU operation. I only use 5 passive Components, PCB thickness is 0.062 inches, 2oz thick copper inner layer, four-layer board design, the following is my list, is it possible to work normally? I am very worried that it will burn directly after power on, please help me check if there is any problem
Components List:
1. Chip resistor 0805 56 ohm (front, back)
2. Tantalum capacitor D case 6.3V 470uf (front)
3. Chip resistor 0805 150 ohm 1% x3 (front R1-R3)
4. Ceramic capacitor 0603 1uf (front, back)
5. Ceramic capacitor 1206 0.1uf (one on the back)
I know there will be a big language barrier, but I will try my best to understand

I want to ask arti9m again
In addition, I have made jumper pins at positions jp1, jp2, and jp3.
I would also like to ask, can I leave the positions of these three chip resistors on the front blank?
Or should I solder 56 ohm chip resistors at these three locations?

I choose 0603 1uf ceramic capacitors to solder in these positions. Is this specification OK? I also choose 0603 ceramic capacitors of the same specification on the back. Is there any problem with this?

I find this location very strange, with only one 0805 or 1206 0.1uf ceramic capacitor? What is this capacitor used for, and why only 0.1uf specifications are used?

I used the S8S1_FINAL Gerber files provided by Arti9m to make these CPU adapters. I spent more than $1000 on this. My motherboard is Tekram P6F40-B5, which uses the 440fx chipset. It can use pentium pro without bios modification. I spent more than $300 on the motherboard, and I am still a little worried whether the CPU adapter board produced can work properly, so I would like to ask for the detailed bill of materials (BOM) of S8S1_FINAL? Can it help my manufacturer produce the correct CPU adapters?