That's a custom LPX motherboard. For finding any info at all you'd be better searching for the make & model of the PC itself instead of just the motherboard. What PC is it?

Beautiful, clear layout with easily accessible jumpers. Just a shame none are labeled...

I checked UH19, but also no LPX form-factor boards with the OPTi 486SXWB chipset known there, or any unknown OPTi chipset board that match this one.

So, unless someone recognizes it. we'll have to consider it undocumented. So let's reverse-engineer 😀

There are lots of jumpers as on most 486 boards, but they are logically grouped around various functions. The interesting ones are the ones near crystal oscillators. Your board has five crystals: 14.31818MHz and 50MHz in the display circuitry by the ET4000AX chip, 12MHz above the riser, 14.31818MHz near RAM and chipset, and 40MHz between CPU and riser. Those last two are the interesting ones. Boards with a PLL usually use a 14.31818MHz input to generate the various clocks. Boards without a PLL derive CPU clock directly from a crystal like that 40MHz unit. I don't actually see a PLL here, but it may be integrated into the chipset (not been able to find a live link to a datasheet for this one). Regardless, even if the CPU is getting its clock from the 40MHz crystal there's *something* going on in between as it's running at half that, which isn't the way 486 usually work. So I find jumpers JP24 and JP25 to be very interesting. My first guess would be that JP27 and JP28 are involved with the Overdrive socket, but if JP24 and JP25 don't do anything it's worth trying them too.

First boot the system as-is to create a baseline. Run a simple benchmark so you're not dependent on what BIOS is telling you the system is doing. Then turn it off, change the first jumper and boot again. See what has changed. Turn off, put jumper back, then change second jumper. Etc. If one of those four is involved with clock you'll find out soon enough. Early 486 boards ran at 16, 20, 25, 33, 40 and 50MHz, so to get all of those speeds would take three jumpers. Most slightly later ones dropped 16 and 20MHz, so could make do with two jumpers. With a bit of luck JP24 and JP25 will select between 25 and 50MHz. Bear in mind that that DX2 will be running out of spec over 33MHz bus speed, so the DX-50 is a better test CPU here.

Other possibility is that the board derives its CPU clock from the 40MHz crystal with minimal logic in between, only allowing 1/2 and 1x clock frequency. If so, one jumper will control it, and you get 20/40MHz depending on setting. In that case replacing the crystal with 33MHz or 50MHz would give you faster options.

PTherapist wrote on 2021-08-10, 12:19:

That's a custom LPX motherboard. For finding any info at all you'd be better searching for the make & model of the PC itself instead of just the motherboard. What PC is it?

thats the challange. no idication on the PC and Box (ovp) all i know is .. it's from Taiwan. the Box is plain brown. there is a number on it but it did not lead to any hint.
Seems like an OEM product imho.

ty dionb,

even i am kind of a PC pro i need to reread your hints. i did not get it all 😁
and yes. clear and easy to access. also .. no dust.
just someone glued the Fan onto the DX-50 and i did not find a way to get the cooler from it yet.

So i booted a couple of times in the past. i actually found a jumper (i need to retry) were if changed the perfomance is cut to half. (acc to sysinfo/benchmarking) but it did not show another MHz.

let me try somethings

so i tested JP25/26
but whatever i set.
like both off. both 1-2 or 2-3 or jp25 1-2, jp26 2-3
the results are not changing.
i added some pics as an impression how it looks like.

i put in the DX-50 CPU btw for that test

some further testing.
i changed JP24/25 and even completely removed them. no effect at all.
altering HP26/27 there is some reaction. but basically just resulting in black screen and non-booting.
no changes on the MHz side.

internaut wrote on 2021-08-10, 13:33:

some further testing. i changed JP24/25 and even completely removed them. no effect at all. altering HP26/27 there is some reac […]
Show full quote

some further testing.
i changed JP24/25 and even completely removed them. no effect at all.
altering HP26/27 there is some reaction. but basically just resulting in black screen and non-booting.
no changes on the MHz side.

Drat. Those would have been the easy choices. It's entirely possible that the board has no settings, just a fixed speed based on that crystal. Unfortunately rather common in early 486 days. I had an IBM PS/1 board that only worked on 33MHz and I have a monster of a Micronics full-AT EISA board fixed at 50MHz (had to get a 486DX-50 just to test that one...). But again, those have 33 resp 50MHz crystals. You're getting 20MHz from a 40MHz crystal (assuming it's that crystal and not the 13.1818MHz or 24MHz crystals delivering the CPU clock here), so I wouldn't give up just yet.

Not much else for it other than testing remaining jumpers one by one.

yep. kind of on the end of my wisdom. (i tried every single yet but not any combination of them orchestrated)
there is definitly a 486-DX33 running somewhere with the exact same setup and board (i contacted the seller to get a pic of the jumpers but no response yet)

The Chrontel IC is a clock generator, but it might just be to do with the Tseng video chipset, the jumpers near it may be worth trying though

Matth79 wrote on 2021-08-10, 18:23:

The Chrontel IC is a clock generator, but it might just be to do with the Tseng video chipset, the jumpers near it may be worth trying though

i tired but these jumpers seem to be graphic related. if i change them the screen stays black mentioning a underrun of the refresh rate.

pfew. a hard challange. not sooo many jumpers but so far no reaction on any jumper change apart from disabling Ports, running into blank screens or beeps or getting FDD/HDD errors. sure .. cmas clear is also there ...

so to summerize:

JP21/22 next to the 50MHz Crystal: set something on the vga. approx refresh rate
JP13 south of TSENG: Replace the top, you get a long loud beep. remove the lower does not do anything
JP4-10 (next to the AMI Bios) Disable PS/2, serial ports and so one. JP9/10 changes result in beeps w/o booting-
JP11: could not locate a change
Jumper between PDU and Nec Chip: No reaction on changes
Jumper12 (southeast of the Power Connector): Clear CMOS
JP15/16 (next to the IDE / FDD Ports): Disables the Ports / drops errors
JP20/21 (south of the IDE ports): Connect both result in no boot. Connect the JP21 no reaction. connect JP22 no boot
JP29: (south of the CPU Sockets): connect will cause a Cache BAD error
JP27: (between the CPU Sockets): disconnect both will cause a black screen. disconnect one no reaction
JP24/25: no reaction so far. promising location though. (similar to 21/22 next to a crystal)
Jumper at the front-LED/Button connectors: Does enable/disable the internal beeper.
CN10(right of the riser slot): no reaction

Is that QA sticker a clue and it's a Victor Technologies model? Very similar formfactor Victor 433D video on youtube... but that's barely any help because teh interwebs has forgotten Victor Technologies

internaut wrote on 2021-08-10, 13:55:

yep. kind of on the end of my wisdom. (i tried every single yet but not any combination of them orchestrated)
there is definitly a 486-DX33 running somewhere with the exact same setup and board (i contacted the seller to get a pic of the jumpers but no response yet)

Bigger question: which crystal does he have next to JP24/25? If that's 66MHz, there's no point in looking at the jumpers.

BitWrangler wrote on 2021-08-10, 19:18:

Is that QA sticker a clue and it's a Victor Technologies model? Very similar formfactor Victor 433D video on youtube... but that's barely any help because teh interwebs has forgotten Victor Technologies

hmmm besides that sticker there is no evidence.
i thought maybe the QA guy was called victor ^^
All i know is that it originally has been manufactored in Taiwan. But thats also very common for that time i guess

dionb wrote on 2021-08-10, 19:36:

internaut wrote on 2021-08-10, 13:55:

yep. kind of on the end of my wisdom. (i tried every single yet but not any combination of them orchestrated)
there is definitly a 486-DX33 running somewhere with the exact same setup and board (i contacted the seller to get a pic of the jumpers but no response yet)

Bigger question: which crystal does he have next to JP24/25? If that's 66MHz, there's no point in looking at the jumpers.

i hope for an answer (or someone else offering such an machine. i have seen the case at least 3 times in the last couple of weeks)

dionb wrote on 2021-08-10, 19:36:

internaut wrote on 2021-08-10, 13:55:

yep. kind of on the end of my wisdom. (i tried every single yet but not any combination of them orchestrated)
there is definitly a 486-DX33 running somewhere with the exact same setup and board (i contacted the seller to get a pic of the jumpers but no response yet)

Bigger question: which crystal does he have next to JP24/25? If that's 66MHz, there's no point in looking at the jumpers.

abd this is the DX33 pic i have. so unfair you can see the crystal but not the actual Mhz print 🙁

I would say that the 40 MHz crystal by the riser slot (above the CPU in the top pic) sets the 20MHz bus. If you want it to run faster, just replace the crystal with a socket (14 pin DIP with all but the corner pins removed, IIRC), then use crystals of double the FSB you want.

debs3759 wrote on 2021-08-10, 20:13:

I would say that the 40 MHz crystal by the riser slot (above the CPU in the top pic) sets the 20MHz bus. If you want it to run faster, just replace the crystal with a socket (14 pin DIP with all but the corner pins removed, IIRC), then use crystals of double the FSB you want.

kind of my fears come true. i've never done something like that before ^^
is there any good howto/guide how to approach that?

ok. found some hints. (like here: 386 dx 33 running at 40 mhz )

so for the DX-50 i would need an 100Mhz Crystal. for an DX2-66 i would need an 66MHZ (same for an DX33 i guess)
then i would need something like this: https://www.amazon.de/St%C3%BCck-Sockel-Adapt … e/dp/B00Q6ZC0BU
and on top a fitting crystal from .. where ever i get them. right?