your attached bios file is 961020S.BIN or Release 10/20/1996S (see https://theretroweb.com/motherboards/s/pcchips-m530#bios) Is that your boards current bios ?
The bios update file you use must match the bios chip type, yours being an S type (PcChip/Amptron naming) not an L type. Back then they used a few diff make/models of the bios chips that the bios files were not interchangeable.
refer to : http://web.archive.org/web/19980425122820fw_/ … _mainboard_bios about S and L bios types
I suggest you update to the Release 10/01/1997 S version first from theretroweb.com and see how it goes..
Use the included Award flash v5.2b, if it not work try Uniflash. edited for clarity. i think...

Thanks for the reply. I downloaded and flashed:

theretroweb.com/motherboard/bios/891001s-63e3d9a3ca34a965851111.bin

This is a big improvement, because now the motherboard can boot on the K6-3+ CPU. It doesn't detect it properly, but at least I can now boot and enable the cache and features from the command line with setmul or k6init.

Also, on my particular board the jumper setting for 55MHz is actually 75MHz, so now I have a 450MHz K6-3+! It is a huge speed boost over the original CPU.

The main issue at this point is that the "CMOS checksum error" is always present. The bios defaults allow me to boot, but I can't figure this out. I was getting a battery error, but I opened up the dallas module and attached a new battery. This cleared the battery error and I can see the seconds counting in the setup screen so I think the time crystal is working, but the "CMOS checksum error" persists. Is this definitely a bad dallas module, or is there something else to check before buying a replacement?

Thanks again!

Hmm were you getting the low battery - cmos checksum error before the bios update ? Have you tried clearing the cmos using the j1 jumper ?
Make sure power is off and jump it for a few seconds. Also running at 75mhz bus clock is an overclock, it could be part of the issue but unlikely....
What type battery are you using for the dallas ? Not all coin cells produce enough V's to supply properly, it needs be new and at least 3.0+ volts by multimeter when connected.
Also if you did not cut the one lead when dremeling the RTC the coin cell will drain thru the internal dropping the overall volts. .
added: If you load defaults and save, then turn power off for a bit then back on does it hold the defaults or go back to an error ?

The motherboard I got has essentially no labels, but I've identified it as an M530 because it closely matches the pictures here:

theretroweb.com/motherboards/s/pcchips-m530

In some pictures there is a battery holder and a different real time clock chip. Mine looks the same as V1.6b but came with no battery holder and a dallas module labeled:

dallas ds1287
real time
8813

If the "8813" means the module is from 1988, then it is clearly not the original because other parts on the board date to 1996. Perhaps the dallas module is incompatible? I peeled it open, removed the dead battery, and added a good 3.2v lithium cell. With the pins exposed now I can confirm continuous 3.2v on the battery pin.

Before any modification the post screen reported both the checksum and failed battery errors. With a new battery I only get the checksum error. With the new bios and new battery I still get the checksum error. On the new bios, with the battery removed I again get both the battery and checksum errors. With the dallas clock chip removed entirely there is no post. It will not save time/date or any settings, even with continuous power from the wall. It is nice that the default settings auto detect IDE and allow the board to boot.

The JP1 jumper does nothing. I can see the trace for the JP1 jumper goes straight to pin 21 on the clock chip socket, and the dallas module has that pin folded up and would be completely NC until I opened the module. The datasheet I found for the ds1287 shows pin 21 as the RCLR pin. It measures around 2.7v and grounding it doesn't have any effect.

I've started checking for bad connections between the module and elsewhere on the motherboard, but I don't know what I'm doing and there are no obvious problems. I'm thinking either the dallas module is bad, or it is not original and not fully compatible.

I have tried all the bus frequency settings, and they all work as expected except that the setting for 55MHz as identified by the silkscreen label is actually 75MHz (hapilly 😀 ). This agrees with the datasheet for the CMA8860 clock generator when there is a -75 in the part number.

I am enjoying upgrading and tinkering with this board. It's nice to find others with similar interests. Cheers!

I have two updates to report with this board. First, I believe the Dallas DS1287 is not fully compatible. The other pictures I've found show a Houston Tech HT12888A being used. I saw a video from vswitchzero about a similar problem on a Gigabyte GA-586ATV Socket 7 motherboard, where the machine could boot with a Dallas RTC module but not save any settings. So I installed a necroware BQ3285 module and that fixed it. I can now save BIOS settings, date/time, etc.

Also, I flashed M520_J2.BIN from Jan Steunebrink. This modified BIOS for the M520 works on my M530, correctly identifies the K6-2/3+ and enables the features I previously had to do using k6init/setmul. My board has a UMC UM8670F chip on it, and this seems to be BIOS-compatible with the UMC UM8663 used on other versions.

I've maxxed out the RAM and I'm now installing linux. Thanks to everyone in the community for helping to make this possible!

nickles rust wrote on 2025-05-13, 01:02:

I have two updates to report with this board. First, I believe the Dallas DS1287 is not fully compatible. The other pictures I […]
Show full quote

I have two updates to report with this board. First, I believe the Dallas DS1287 is not fully compatible. The other pictures I've found show a Houston Tech HT12888A being used. I saw a video from vswitchzero about a similar problem on a Gigabyte GA-586ATV Socket 7 motherboard, where the machine could boot with a Dallas RTC module but not save any settings. So I installed a necroware BQ3285 module and that fixed it. I can now save BIOS settings, date/time, etc.

Also, I flashed M520_J2.BIN from Jan Steunebrink. This modified BIOS for the M520 works on my M530, correctly identifies the K6-2/3+ and enables the features I previously had to do using k6init/setmul. My board has a UMC UM8670F chip on it, and this seems to be BIOS-compatible with the UMC UM8663 used on other versions.

I've maxxed out the RAM and I'm now installing linux. Thanks to everyone in the community for helping to make this possible!

The attachment M530.png is no longer available

Great news.

M520 and M530 shared the same BIOS, even on the old page on the original site (retroarchive) it's explicitly mentioned.
Only difference to mind is L or S (size).

To clean up the wiring some I added a power tap to the motherboard for the CPU (simply a wire soldered to a jumper) and fitted a passive heat sink. The power supply has a loud fan so I've ordered some that should be much more quiet. Otherwise I think bench testing is done and it can go into the case. I'm pleased with it so far. A 512kB cache module would be nice to find, but they seem to be pretty rare.

Is there a simple way to determine the hard drive size limit for this board? Is it 128GB?

nickles rust wrote on 2025-07-01, 21:39:

Is there a simple way to determine the hard drive size limit for this board? Is it 128GB?

When using the patched M520_J2 BIOS on the M530, the IDE HDD limit is indeed 128GiB/137GB. For all other M530 BIOSes, the limit is either 32GB or 8 GB.

Cheers, Jan

Thanks for the reply! I tried a 120GB SSD and it could detect it but not use it. I've ordered a different disk to try, and I'll report back...

Also, I found a 512kB coast module that is compatible. Even when using a k6-2+ or 3+ with built-in cache, the extra motherboard cache also helps some.

nickles rust wrote on 2025-07-05, 19:12:

Thanks for the reply! I tried a 120GB SSD and it could detect it but not use it. I've ordered a different disk to try, and I'll report back...

I assume that 120GB SSD is SATA and that you use an IDE-SATA adapter to connect it to the M530’s IDE port.

A lot of discussions/experiences about these adaptors can be found here on Vogons.
Member douglar did a nice summary in the Re: The hunt for a good IDE to SATA adapter thingy. thread. But for your M530 you can disregard the discussions about UDMA because the 430VX chipset can’t do UDMA and is limited to MWDMA mode 2.

Jan

The SSD is a 2.5" 44-pin version. I opened it up and there's not much to it. There is a "SM2236" chip that is probably the controller and a couple big flash chips. The SSD does work on a different motherboard.

Those SATA converter boards are inexpensive, so I'll order one of those to try also. Thanks for the info!

Chkcpu wrote on 2025-07-02, 08:08:

nickles rust wrote on 2025-07-01, 21:39:

Is there a simple way to determine the hard drive size limit for this board? Is it 128GB?

When using the patched M520_J2 BIOS on the M530, the IDE HDD limit is indeed 128GiB/137GB. For all other M530 BIOSes, the limit is either 32GB or 8 GB.

Cheers, Jan

Does your bios have the boot from cd-rom option?
Right now it's only floppy and hdd.
If it does I don't have to play patch doctor (and probably fail again 🤣).

I have not updated this thread because I've tried a bunch of different stuff with quirky results. The 120GB SSD doesn't really work, but an 80GB spinning IDE disk does. Different PCI disk controller cards have varying degrees of function. I have not been able to discern much of a pattern. It could be the BIOS, or software, or hardware, or some combination. It has been frustrating. I've posted some results in this thread:

Re: Socket 7 NAS aiming for the stars

I tried Jan's BIOS and it has all the BOOT options of the era.
I have yet to test HDDs = or > 80GB

When I first powered on the PC I entered the BIOS and was prompted with a password check.
All 0 (zeros) made it; I was lucky. No idea why, probably there's something with the cmos reset that didn't work or something, but it happened also with another BIOS.

The board seems faster now, but I'll do a W98 reinstall soon.

Edit: now I have artifacts at memory count that weren't before. Also a conflict with IRQ 3 with SCSI card and S3+... I'll have to tune that.

Nexxen wrote on 2025-10-05, 12:53:

I tried Jan's BIOS and it has all the BOOT options of the era. I have yet to test HDDs = or > 80GB […]
Show full quote

I tried Jan's BIOS and it has all the BOOT options of the era.
I have yet to test HDDs = or > 80GB

When I first powered on the PC I entered the BIOS and was prompted with a password check.
All 0 (zeros) made it; I was lucky. No idea why, probably there's something with the cmos reset that didn't work or something, but it happened also with another BIOS.

The board seems faster now, but I'll do a W98 reinstall soon.

Edit: now I have artifacts at memory count that weren't before. Also a conglict with IRQ 3 with SCSI card and S3+... I'll have to tune that.

I have this BIOS on an M520 and it too has those colored rectangles that show up during the memory count. I asked Jan, and he isn't sure of the cause, but the board works fine with that BIOS.

Also, I don't think we can say this too often; thank you Jan for the work that you do, it is appreciated!

Repo Man11 wrote on 2025-10-05, 14:44:

Nexxen wrote on 2025-10-05, 12:53:

Edit: now I have artifacts at memory count that weren't before. Also a conglict with IRQ 3 with SCSI card and S3+... I'll have to tune that.

I have this BIOS on an M520 and it too has those colored rectangles that show up during the memory count. I asked Jan, and he isn't sure of the cause, but the board works fine with that BIOS.

ChkCPU,

Could this be due to the bios code attempting to display the epa logo ?

weedeewee wrote on 2025-10-05, 14:53:

Repo Man11 wrote on 2025-10-05, 14:44:

Nexxen wrote on 2025-10-05, 12:53:

Edit: now I have artifacts at memory count that weren't before. Also a conglict with IRQ 3 with SCSI card and S3+... I'll have to tune that.

I have this BIOS on an M520 and it too has those colored rectangles that show up during the memory count. I asked Jan, and he isn't sure of the cause, but the board works fine with that BIOS.

ChkCPU,

Could this be due to the bios code attempting to display the epa logo ?

Hi weedeewee,

Yes, these colorful artifacts are indeed caused by a scrambled EPA logo.
When I talked with Repo Man11 about this issue back in 2022, we couldn’t pinpoint the source of this cosmetic issue. But we did noticed that all was well right after flashing the M520 patch J.2 BIOS, but that the artifacts appeared after the first reboot.

I did found out that the ESCD block contained a lot of scrambled data, although the Plug&Play data was intact. This BIOS has the EPA logo as a separate compressed module and it seems that it uses a part of the ESCD block to store the uncompress logo data, instead of storing it temporarily in RAM. When the BIOS writes to the ESCD block at the end of the first boot after the flash, I believe it tramples on the logo there.

I was unable to find this bug and I don’t know if only my M520 patch J.2 BIOS has it, or if the original M520 Rel. 1.1A (12/07/98) has this bug as well.
I do know that an earlier M520 Release 10/16/1997 S BIOS doesn’t have this bug.

Cheers, Jan

Repo Man11 wrote on 2025-10-05, 14:44:

I have this BIOS on an M520 and it too has those colored rectangles that show up during the memory count. I asked Jan, and he isn't sure of the cause, but the board works fine with that BIOS.

Hi!
As far as I tested it no issues so far and it performs as expected.
I tested all boot options and LS-120 and Zip (IDE) work as well. It was purposeless but why not 😀

The only strange thing is why a password was present on the bios chip I programmed, coming from another board. Maybe some piece of code didn't get erased? First time ever.
The original M530 had a OTP chip.

Edit: with HDD over 80GB, as Jan mentioned, you have to allow AUTO for the m/b to correctly set it, otherwise with HDD tool in BIOS it'll hang.
It'll fail to work in W98. Not recognized and hdd utils fail to initialize (throwing allignement errors).
I'd stick to max 80GB.