[Lan Drivers AMPTRON Motherboard
Gildo Santiago
Now I have gone back in time a little further than usual. This computer has many similarities to the Win95 system I fixed up; in fact, I think the same people built them because they have the same motherboard (Amptron PM-8600). That is just a hunch, since the computer case sticker says Kouwell-286, which makes me think someone upgraded the system and one point, replacing almost everything. It is built like a tank and has a solid steel frame.
Another interesting thing is the floppy is connected to ONLY +5v. The ad says that is how it works. Interestingly, it also says there is a version with power provided by the interface ribbon cable. I have two other drives like that- now I can hack together a special cable and get them running.
This really threw me. I found that the video card has a Bus Mouse port on it. And I have a bus mouse I have been using with a PC9821 from Japan. Works great. The PC also has a serial port so I installed Logitech Mouseware so I can use a serial Trackman mouse as well.
The graphics card is neat. It is a full-length ISA card. It has a VGA, and 2 serial data ports- for the mouse! One is Ps/2, and the other is 9-pin serial. It is made by ATI (bought by AMD many years ago). Time for some specs!
I have been trying to get the new ATI card running properly under Windows 98SE. The card is detected as XPERT @Play (some ATI marketing). It took some doing but I got the Rage Pro Turbo drivers installed. The drivers are on archive.org: -524-cd
Changing the multiple for the CPU has no effect on speed on the Pentium 166. Changing the FBS is all I can do, and the max on this motherboard is 75Mhz (up from 66). That raises the MHz to roughly 188 on the CPU. I tried all the clock multiplier settings. 2 had no effect, and one appeared to lower the speed to 15oMhz, but I found it was just that my jumper was damaged, so I replaced it.
I started testing some other drives, and I found that when putting the IDE drives into a Windows 7 machine with SATA and IDE, I had to have the IDE drives set as a master for them to show up in the system. Weird. In any case, the drive shows up as 8MB again. I have to call it dead.
I had some problems initially. I knew I had to install from a command prompt, but I had booted into Windows 3.11 first, then exited to DOS. That does not work, and you will get an error like below about EMM386 being installed. I grabbed a minimal DOS boot floppy, change the BIOS to boot from floppy, and started over.
The second attempt resulted in a new error- the flasher could not read the binary ROM file I was trying to flash. Turns out the old floppy was no good. I had only done a quick format on it. I should have done a complete format. The third time the flashing worked after putting the new BIOS code on a good floppy.
The original hard drive is 730MB. I have newer and bigger ones, but they are too big for the BIOS, which has a 32GB limit. The diagram below (for Wester Digital drives) shows the jumper settings for establishing a 32GB limit on the drive. In my case, I have it set as a single driver at 32GB (top left settings). This allows me to use a 120GB drive that has little usage and will ultimately last longer. And you can install a ridiculous number of old games on it.
I switched to the newest SoundBlaster Live driver I had (2001), despite windows complaining it was not written for my hardware, but that got all my audio back except I still could not play a CD. I also updated the CD ROM driver but that changed nothing.
I almost hesitate to write this, since I have not rebooted since the last patching of the networking OS files. The Intel software (Pro98 10.1) does not kick off a full installation of all OS components required for networking. You still have to add those. Maybe I am a bozo, but I added them by first extracting Net7-10 cab files from the installer (they are zip files) and writing them to a new CD. This is because I could not get the driver update software to read the files off the Windows 98SE installation disk. 3 other OS files disappeared so I manually patched them back into C:\Windows. Another windows system file (msnp32.dll) also kept disappearing. I think it was bad (came from my extracted cabs), so I re-extracted it from the installation disk and place it in C:\Windows\System and it works.
Networking works and it only took 2 hours. This is a record for me on Win 9X systems. I use networking to transfer files between systems, in case you are wondering. Not a lot of web surfing, but I do recommend (works on 98!)
PC Chips (Hsing Tech Enterprise Co.) was a low-end motherboard manufacturer for the IBM PC and its compatibles during the DOS era. Founded in 1984, they later became one of the major shareholders in Elitegroup (ECS) in 1998 and finally merged with ECS in 2005.
The were somewhat famous in the mid 90s to early 2000s for building motherboards with fake cache chips on 486 boards, including modified BIOS to show the presence of a cache when there wasn't one, and relabelled chipsets (socket 5, 7 and slot 1) suggesting they were better than Intel's.
The blue socket in the corner is for a math coprocessor. Another revision of this board has a socket for the main CPU instead of the PQFP (Plastic Quad Flat Package) you see in the rev 3.1 image above. Socketed versions support the Cyrix Cx486DLC or Texas Instruments TX486DLC chips.
This is one of the PC Chips boards that have the fake cache chips, as can be seen in the top-right corner of the image above - this means that unless you have a COASt module installed in the slot, the motherboard has no L2 cache. Either a 256 KB or 512 KB module of asynchronous SRAM or a 256 KB module of pipeline-burst SRAM can be installed. The board can support a cacheable area up to 64 KB of cache.
All three PCI slots support bus mastering, with support for up to 4 IDE drives in PIO Mode 0 up to Mode 4 (max. transfer rate of 16.67 MB/second) and master DMA Mode 2 for a maximum of 22 MB/second. Hard disks of between 528 MB and 8.4 GB are supported without device drivers.
This motherboard often had its chipset relabelled to "VX Pro II", but is in fact an Ultron (aka Hintcorp) UT801X. The chipset does NOT support UDMA or ACPI. The "1437" in the AMI BIOS string indicates the board is by Hsing Tech, PCChips' actual company name.
Board revision 4.1 added support for 50 and 55 MHz FSB speeds. These were missing on the 3.1 and 3.1A revisions. 3.1A uses a replaceable Lithium-Ion CR2032 battery coin cell. 3.1 used a RTC chip. 3.1A also fixed a small hardware issue related to the HDD LEDs.
Four BIOS versions are known. The 1126S.rom removes the "TX Pro" text from the BIOS string. 0214s.rom added support for Cyrix 6x86MX (MII) 266 MHz, resolved an issue with formatting SCSI hard disks and a hang-up on COM port 2. Version 0410s.rom added support for hard disks over 8.4 GB and also added Trend's Chipaway ROM. Finally version 0624s.rom fixed a problem when using Cyrix 6x86MX (MII) and booting from a ZIP drive on the ATAPI interface.
On the non-LMR version, the 72-pin SIMM slots support memory from 8 MB up to 512 MB of either FPM or EDO 60ns DRAM. The DIMM slots (fitted to both non-LMR and LMR versions) can be fitted with either 16Mx4, 8Mx8 or 4Mx16 of PC-66 or PC-100 SDRAM modules.
Reports on the web also indicate that this motherboard has serious stability problems when trying to run FSB at 100 MHz, e.g. with an AMD K6-2 350 at 100 x 3.5 multiplier. Instead running it at 83.5 x 4.5 multiplier (375 MHz). Apparently the SiS chipset on this board is not rated for above 90 Mhz, and is very unstable even at that speed. When you set FSB to 95 MHz it actually runs it at 90 Mhz, which causes the stability problems, such as registry errors.
This board supports a FSB speed of 90 MHz, comes with onboard sound which uses 3D Sound Pro, internal AGP 2x (pins on the board connect to a ribbon cable ending in a 15-pin SVGA output) using a VIA 6326 (as used on the Diamond SpeedStar A50), and 1 MB of L2 cache.
Despite the fact this board has the highly publicised fake cache chips (see top-right corner), it's still an interesting one in that it has a combination of ISA, VESA Local Bus, and PCI slots. There is an argument that combining VLB and PCI on the same board is a bad design choice due to the differences in frequency that are supported between these two standards. Some have noticed the presence of VLB slots on boards like this slows down the PCI slots to avoid incompatibility.
The M919 is also a flexible board memory-wise, with both 30-pin and 72-pin SIMM slots. The latter accepts either Fast Page Mode (FPM) or Extended Data Out (EDO) SIMMs. Note that only boards with the UM8881E or UM8881F can use EDO memory - the older UM8881B cannot. If configuring for EDO RAM, you must specify the DRAM Type as 'EDO' in the BIOS. It can take a total of 256 MB of memory.
It also has a COAST slot to 'expand' the level 2 cache, populated in the pictured board on the middle-right above the CPU socket (Actually it's not a standard COAST slot! Be warned this slot only support PC Chips proprietary cache modules - apparently the pinouts are different so you may damage a standard COAST SRAM card or your motherboard if you use it in this board!!). The user manual indicates modules were in sizes of 128 KB, 256 KB, 512 KB and 1 MB, and that initially only asynchronous SRAM versions were available but they would be releasing pipeline burst SRAM modules at a later date. Furthermore, these cache modules came in both 3.3V processor and 5V processor versions (configured via JP4 and JP5 A-D - see manual for settings) - modules with model M919-01S3 are the 3.3V ones and model M919-02S5 are the 5V ones. They will only work with processors of the same voltage as the cache module. My cache card is the asynchronous type with nine Alliance Semiconductor AS7C256-15JC ICs, meaning it's a 256 KB module.
Because the actual level 2 cache chips are simply black squares of plastic, it's a slightly redeeming fact they included this so you can at least have some level 2 cache if you put an SRAM cache card in this slot.
795a8134c1