byte_76 wrote on 2025-04-18, 07:50:

I have this Mecer B870-I2 motherboard which is in a working condition.

https://theretroweb.com/motherboards/s/mecer- … poratio-b870-ii

The Intel DX4 75 which is installed works perfectly at 100MHz in this board, however the voltage regulator gets very hot. Same for a true 100MHz Intel DX4 WT.

Linear voltage regulators tend to get very hot - that's how they work: dissipate the excess voltage as heat.

I can't quite make out the voltage regulator type on the Retroweb pic, but a similar board I have here has a Chinese clone of a TI1084 3.3 regulator, which specifies a junction temperature of max 125C for the control part (and 150C for the output). If concerned about heat, add a heatsink; the datasheet provides extensive notes on how to calculate the required heatsink. These will apply to any similar voltage regulator as well. TLDR: screw a bit of aluminium into it with a bit of thermal compound if concerned.

I'd like to use a few other processors in the board but they're giving me issues. […]
Show full quote

I'd like to use a few other processors in the board but they're giving me issues.

I have installed a Cyrix DX2-66 3.45V and it keeps hanging at the boot screen.
I also tested an ST DX2-66 5V and it behaves in the same way.

An AMD DX4 100 NV8T works fine and the voltage regulator is quite a bit cooler than the Intel processors. Why is that?

I also tried my Cyrix 5x86 100GP and AM5x86 P75 processors but they don't detect properly. So the board POSTS but the processor names aren't displayed properly and L1 cache isn't working.

There are some jumper settings printed on the silkscreen but not all jumpers are listed.

Are you just swapping out CPUs or are you also changing jumpers? If just swapping out this is to be expected as different CPUs need different settings.

The display name of CPUs is dependent on BIOS - apparently your BIOS is from before the release of Cx and Am5x86 CPUs. In itself this isn't an issue, but non-working cache is of course more problematic.

I wanted to backup the BIOS but Uniflash 1.40 reports it can't find PCI bus

Looking at that board neither can I 😜

and AMI flash 3.42 and 8.95 report unsupported chipset. Which utility should I use to just make a BIOS backup?

If you have it, an (E)EPROM reader/writer would be ideal. Otherwise you could consider reverse hot flashing, getting a motherboard of similar age that does have Uniflash support (maybe a PCI 486 board), booting it, hot-swapping the EPROMs and then trying to read it.

When I add identical cache (tested) modules to those empty sockets and configure the jumpers for 256K, the board won't boot past the POST screen. There seems to be some pads for SMD caps next to those open sockets, so not sure if that's the cause. (The pads of the unpopulated SMD caps are a bit corroded but there aren't caps broken off, although it kind of looks that way in the pic. I'm quite sure it was like that from factory)

Hang on...

Assuming you have 128kB in there now, with four 256kb chips (note capitalization B vs b), you need at least 8kB of tag, so a 64kb chip in the 9th position. For 256kB (8x 256kb chips) you would need 16kB of cache, so a 128kb chip.

If the board only had a 64kb tag and you didn't upgrade it when adding cache chips, that would fail POST.

Exactly which chips are we talking about here, in particular which chip is used for tag?

I am looking for any assistance with these: […]
Show full quote

I am looking for any assistance with these:

- Documentation / Jumper settings
- Updated BIOS images
- Correct Utility for BIOS image backup
- Advice on configuring the board correctly for the Cyrix processors.
- Any other information about this board

Challenging, this is an obscure board & brand even by 486 standards. What I do in a case like this is look for boards that are functionally similar (same chipset, same ports/slots) and see if any seem to match in terms of number of jumpers.

To do that, a better pic of the jumper area and the settings on the board than the one on theretroweb would help.

Do bear in mind this looks like an old, not well-supported board and it's entirely possible full support for WB cache and Cyrix CPUs may not be available. If I look at a typical 486 board that supports all CPU types, it has over 15 jumpers for the CPU. This board only seems to have 9 or 10.

The picture in my post is my own board and is a fairly high quality picture. I can post more if that would help.

From the little info that is printed on the board, the current jumper config seems to be correct for both 66MHz and 100MHz CPU’s with 33MHz bus. So I did not change jumpers, except for the voltage when I tested the 5V ST.
That said, I don’t have enough info about the other jumpers and it’s certainly likely that more jumper changes are required for the different brands of processors.

I have an eprom reader but to identify the chip type, I need to remove the sticker and expose the window.
How much light does it take to corrupt the bios, is it generally safe to remove the sticker?

The tag that I installed was a 16kb module. Confirmed working with 128KB cache.

Yes, I assumed the BIOS is too old to support the 5x86 CPU’s.At least they POST.
Can the existing BIOS be modified to add support?

Since this is a "standard" 486 board with an ALI 1429 chipset without any onboard controller or I/O ports, you can try to use a BIOS image from a different board with the same chipset. TRW has a couple of these for download.
UV-erasable EPROMs won't be corrupted that fast if you take off the sticker, just don't let it lie around in bright sunlight and tape the UV window over with some sticky tape.
Cyrix CPUs definitely need special jumper settings in order to work properly. Most of the seetings seem to be silkscreened on the board. There are a couple of jumpers between the 8-bit and first 16-bit ISA slot which are labeled "M6", "M7" and "486". M6 and M7 are the codenames for Cyrix DX33 and DX2-66 cpus, so you probably need to at least change those.

byte_76 wrote on 2025-04-18, 11:17:

The picture in my post is my own board and is a fairly high quality picture. I can post more if that would help.

In that case you have bad battery leakage damage that needs cleaning up...

Probably not the source of your current challenges, but unchecked the corrosive stuff will continue eating into traces until something fails.

From the little info that is printed on the board, the current jumper config seems to be correct for both 66MHz and 100MHz CPU’s with 33MHz bus. So I did not change jumpers, except for the voltage when I tested the 5V ST.
That said, I don’t have enough info about the other jumpers and it’s certainly likely that more jumper changes are required for the different brands of processors.

I'm having trouble making out what is printed on the board, in particular there seem to be some legends around the CPU jumper block that are either obscured by the jumpers themselves or in shadow between jumpers and socket.

I have an eprom reader but to identify the chip type, I need to remove the sticker and expose the window.
How much light does it take to corrupt the bios, is it generally safe to remove the sticker?

UV light erases EPROMs, so if you're indoors and using LED light (or incandescent), there no risk. Sometimes UV EPROMs have the type info on the bottom, might avoid having to mess around with the sticker.

The tag that I installed was a 16kb module. Confirmed working with 128KB cache.

That should work. Sure the new SRAM chips are all OK? If so, maybe play around with JP14 (12 and 13 seem clear enough).

Yes, I assumed the BIOS is too old to support the 5x86 CPU’s.At least they POST.
Can the existing BIOS be modified to add support?

BIOS modding isn't something I have much experience with, but you should at least be able to get correct recognition. WB caching also requires physical support from the board, so even with BIOS support that might prove an issue.

There is no battery damage.

The CPU that I would like to use in the board is the Cyrix DX2 66 at 3.45V. (It keeps hanging. Sometimes at the boot screen, sometimes at the command line)
I’d be grateful for advice on which jumpers to change to which positions)

Here are some more detailed pics:

More pics

If you simply want to dump the BIOS and not reprogram it then use NSSI in DOS

As for the different regulator heat production with different CPU that's a simple case of the CPU using less current. The more current flowing through the regulator at a given voltage will result in more heat.

maxtherabbit wrote on 2025-04-18, 12:38:

As for the different regulator heat production with different CPU that's a simple case of the CPU using less current. The more current flowing through the regulator at a given voltage will result in more heat.

What I meant was, the regulator is significantly hotter with the Intel DX4 100 than it is with the AMD DX4 100.

Since both CPU’s operate at 3.3V and AMD processors were almost clones of the Intel processors, it just seemed strange that the Intel causes the regulator to run so much hotter.
I guess, as you say, it just uses more current for some reason. I do understand that it shouldn’t really be anything to worry about but I suppose there’s always more potential for failure with heat. I’ll look into fitting a heatsink as suggested.

Since the BIOS chip has a UV window, does that mean my TL866 would be unable to reprogram it? (I haven’t tried anything yet, so that’s why I’m asking before messing with anything)

byte_76 wrote on 2025-04-18, 13:59:

Since the BIOS chip has a UV window, does that mean my TL866 would be unable to reprogram it? (I haven’t tried anything yet, so that’s why I’m asking before messing with anything)

It probably can reprogram it, but you need to erase it first with a proper UV light source (EPROM eraser).

byte_76 wrote on 2025-04-18, 12:18:

There is no battery damage.

Hang on, that's not the same board as in the Retroweb link. Must have misunderstood you there.

The CPU that I would like to use in the board is the Cyrix DX2 66 at 3.45V. (It keeps hanging. Sometimes at the boot screen, sometimes at the command line)
I’d be grateful for advice on which jumpers to change to which positions)

Did you already try pyrogx's suggestion for JP2/JP3 for M6/M7?

That Cx486DX/2-66 is probably an M7, so that would mean putting a jumper on JP2.

Also, what is the exact name of the CPU? Cx486DX2-66 is a 5V part. The 3.45V version is Cx486DX2-V66. If yours isn't the V66, it needs 5V, which could explain the crashes.

Here are some more detailed pics:

Tnx. Very clear.

Since the BIOS chip has a UV window, does that mean my TL866 would be unable to reprogram it? (I haven’t tried anything yet, so that’s why I’m asking before messing with anything)

Depends on the chip. Normally UV wipes the EPROM which makes it writable again. Early EPROMs like that frequently do require higher write voltages, which the TL866 may or may not support.

I have now installed a jumper onto JP2.

The system still hangs at the info screen.

I have these two processors and both of them hang on this screen.

Cx DX2v66 - Motherboard set on 3.3V
ST DX2-66 - Motherboard set on 5V

It’s possible that the board has a fault but it does work without hanging with the Intel and AMD DX4 100 processors.

Okay, hang on.
Should have done the obvious first.

I loaded failsafe options in the BIOS and so far, it’s booting and stable…..but slow.

Enabled internal and external cache without changing any other settings and still booting and stable.

I found that if I set the BIOS to optimal and just disable the power mode options, the system boots and performs well with the Cyrix DX2 v66.

I've been messing with this board trying to get my Am5x86 133 working on it.

I've had some success in that I was able to get the CPU working at 133MHz and 150MHz.

In both cases, the boot screen shows 100MHz but I can see in CHKCPU that the processor is clocked at the speeds that I have configured.

There are two issues:

1. The BIOS code does not recognise the CPU which causes the incorrect CPU to be displayed.
2. The L1 cache is running in Write-Through mode and I can't figure out how to get it into Write-Back mode. (I'm fairly sure the board does support WB mode because I have seen it displayed in CHKCPU with other processors)
Could this cache issue be related to lack of BIOS support as well?

The board uses an AMI BIOS. (One of those that has the mouse support)

Would it be possible to add support for the Am5x86 P75 in the BIOS?

byte_76 wrote on 2025-04-26, 18:32:

The board uses an AMI BIOS. (One of those that has the mouse support)

Would it be possible to add support for the Am5x86 P75 in the BIOS?

Unfortunately, modding an AMI WinBIOS is not (yet) possible. But modding an Award BIOS for Am5x86 support is no problem.

To help you find a replacement Award BIOS for your board, please upload a dump of your present AMI BIOS here. You can use your TL866 to make a copy of your BIOS. The TL866 should have no problem reading the 64KB UV-EPROM when selecting a 27C512 or equivalent chip, even if the ID doesn’t match.

Jan

Chkcpu wrote on 2025-04-27, 17:57:

Unfortunately, modding an AMI WinBIOS is not (yet) possible. But modding an Award BIOS for Am5x86 support is no problem. […]
Show full quote

byte_76 wrote on 2025-04-26, 18:32:

The board uses an AMI BIOS. (One of those that has the mouse support)

Would it be possible to add support for the Am5x86 P75 in the BIOS?

Unfortunately, modding an AMI WinBIOS is not (yet) possible. But modding an Award BIOS for Am5x86 support is no problem.

To help you find a replacement Award BIOS for your board, please upload a dump of your present AMI BIOS here. You can use your TL866 to make a copy of your BIOS. The TL866 should have no problem reading the 64KB UV-EPROM when selecting a 27C512 or equivalent chip, even if the ID doesn’t match.

Jan

Thank you for the feedback Jan.
I’ve only just seen your message now but I’ll get the file uploaded asap.

I dumped the BIOS yesterday and will upload it after work.

The eprom chip type is attached.
I’d like to replace the chip with an eeprom because I’m not sure I can replace the BIOS image on the current chip with my TL866.

I’ve ordered a couple of SST27SF512 chips and I’m hoping they will be compatible.

Don’t over clock you might over heat and ruin your motherboard or CPU.
It’s NOT worth it because the performance improvement is minimal and might make your computer unstable.
If you want the Best performance of a 486 use a DX4-100
Or an AMD 5x86-133 or for the Very Best Performance use an Intel Pentium 83mhz over drive.
64kb of Cache is good enough
8mb to 16mb of RAM is Enough
Windows 95 was Designed to Support the 486dx-33 CPU

In the End. Its Still a 586 cPU
It can’t play MP3’s or Video very well.
You Need a Pentium Class CPU for that.

Chkcpu wrote on 2025-04-27, 17:57:

Unfortunately, modding an AMI WinBIOS is not (yet) possible. But modding an Award BIOS for Am5x86 support is no problem. […]
Show full quote

byte_76 wrote on 2025-04-26, 18:32:

The board uses an AMI BIOS. (One of those that has the mouse support)

Would it be possible to add support for the Am5x86 P75 in the BIOS?

Unfortunately, modding an AMI WinBIOS is not (yet) possible. But modding an Award BIOS for Am5x86 support is no problem.

To help you find a replacement Award BIOS for your board, please upload a dump of your present AMI BIOS here. You can use your TL866 to make a copy of your BIOS. The TL866 should have no problem reading the 64KB UV-EPROM when selecting a 27C512 or equivalent chip, even if the ID doesn’t match.

Jan

Here is the BIOS. (zip file attached)

I'm not sure if I saved it correctly.
Please let me know if this file is complete so that I can send it to The Retro Web

https://theretroweb.com/motherboards/s/mecer- … poratio-b870-ii