phipscube wrote on 2020-08-11, 20:43:

I'm on the case with the battery. I have someone going to my aunts this week to get the base and undo it for me to see... I have […]
Show full quote

I'm on the case with the battery. I have someone going to my aunts this week to get the base and undo it for me to see... I have a vague memory of infact replacing the battery as it didn't hold settings etc, maybe one of the reasons why it was going for £5.. I remember opening it and finding the original batt with the startings of leakage, fluff etc. But of course even the "new" batt is old now so it could have leaked too. The Alt12887 looks good, will that work in a 386 board too?

Thats really interesting about the Mega PC plus.. I thought they had used a 7486 board, but it just seems like Amstrad made a real rush job!... does the Cyrix program basically make it possible in software to use the L1 cache on upgraded Mega PC 386sx boards?.. i'm now wondering if changing the BIOSes to the plus versions would work, and be more stable (should I decide to upgrade the 386sx to a 486slc)

I'm now thinking if its possible to make a Mega PC plus out of my standard board (should it be leak free). Its not as good as the 7486slc.. yes like you say probably not much gains, but just a thought.. Hopefully the 486slc board turns up.. damn I hope I didn't throw it away!

Do you have a photo of the Mega PC plus 7386 board?

The ti chips.. yes I read how some of the 3.3v chips were run on the same process as the 5v so the difference was just expected tolerance.. so a bit risky, but some have done it and have no problems (yet). I saw the bridgeboard mod on amibay, it looks a bit untidy seeing the CPU like that. I prefer it to be neat really though so yes if I could find a 5V 486slc that would be nicer. Its just that right now there is a plethora of those G50 chips on ebay..

Looking forward to seeing your Mega PC running with the 5428!

Thanks for all the info, you guys rock! 😀

Hi Phipscube, (and apologies in advance to everyone for the long post !)

Unfortunately the 7386 uses the barrel battery whilst the 7486 uses the Dallas Chip, so you unfortunately can't use the Alt12887 on the 386 board 🙁

So the MegaPC Plus is an interesting story - this is what I've gathered from a old Amstrad Dealer, it may or may not be accurate but it fits the data I have so far.....
From what I gather the 386 board design (7386) was slightly ahead of the 486 (7486) board design which was its replacement - both were designed to have a "MegaPC" version. However, when the 386 MegaPC hit the shops, it didn't sell fantastically well, Amstrad had a surplus of 386 boards. Given the dealer feedback they got about the MegaPC (the 386 was underpowered) and not wanting to waste the surplus 386 boards, Amstrad added the 486SLC and clock and coupled with a new BIOS were able to call them 486's and thus sell them as MegaPC Plus's - although that didn't succeed either. When the "purpose built" 7486 board arrived, there was still a surplus of the 386 board (these found their way into another PC - see below), and as Amstrad needed the new 7486 boards more for desktop PC's so all of them ended up there.

A good photo of a damaged MegaPC Plus board can be found here https://mcretro.net/wp-content/uploads/2012/06/pcb.jpg and is the same as the working one I've seen.
I think you'll be very lucky to find BIOS images - the collector who owns the MegaPC Plus values it being very rare and the question of making BIOS images when I saw it was a non starter - a bit selfish I know, but sometimes that's why folks like that get into a hobby - and if we had the BIOS images then 7386 boards could be turned into MegaPC Plus boards with very little to tell the difference, something that the collector didn't want happening.

That said, it's possible we may get BIOS images from another MegaPC Plus out there, or another possibility which may be easier to get hold of is the BIOS from the Amstrad PC7486SLC80 (where 80 is the HDD size, not speed). If anyone has photo's of the motherboard of that PC, please let me know as I'd love to see them. this board has a Ti486SL 25Mhz MAB processor on the 386 board, indeed the first post on this link suggests it https://assemblergames.org/viewtopic.php?t=2243&start=230. From what I've heard this PC was made up from reworking the surplus 386sx boards, and using a 25Mhz processor so they could save money on replacing the clock chip (!), but DO have a Quadtel 486SLC bios. Either way these PC7486's have the barrel battery too, so are probably just as rare in working condition as the MegaPC, but if a working one could be found its BIOS would probably do for a MegaPC Plus clone on the 7386 board.

If your board is working ok then I would advise just using a Cyrix "clip on" upgrade chip (486 SRX2 50Mhz) - as good, undamaged working original MegaPC boards are getting rare and it would be a shame damaging a board if its good. The clip on chip has the advantage that you can remove it at any time, returning the board to stock - and the cache etc.. is enabled through a simple program.

Before you consider the G50 - have a look at this https://mcretro.net/amstrad-sega-mega-pc-486- … pgrade-failure/ where a similar CPU swap failed on the MegaPC board (apparently even caught fire if the previous post is to be believed !). If I had a damaged board then I would probably go down the route Amstrad did and consider the Ti486SL 33Mhz MAB, you can still get this chip on ebay and would run at 25Mhz if you didn't want to risk changing the 50Mhz clock chip to 66Mhz. The cache could still be enabled using the cyrix program for the upgrade chip, and disabling it's line in autoexec.bat would still be an option if you encounter instabilities - so its a bit safer option.

Just dropped the board in for its 5428 mod - looks like it'll be the end of the week before I get it back as there is a backlog in jobs at the moment -but when I have a result, I'll let you all know !.

Zalog wrote on 2020-08-12, 08:16:

Before you consider the G50 - have a look at this https://mcretro.net/amstrad-sega-mega-pc-486- … pgrade-failure/ where a similar CPU swap failed on the MegaPC board (apparently even caught fire if the previous post is to be believed !).

Without any Mega PC, but with a tiny bit of QFP soldering in the book, I just watched him do the replacement and almost couldn't help talking and pointing at my screen... 😁
To my eyes, a few pins might not even be properly soldered on and some others look VERY much like sitting beside the pad and maybe even touching.
I wouldn't be surprised if that's at least one of the reasons why it stayed silent 😁 Not saying I'm having any ad-hoc idea if such a transplant could or should work or not (didn't look into it at all , just saw the soldering work 😀

As for the fire, I'm thinking that was an exaggeration in the post before. When he turned it on, there didn't even seem to be any magic smoke. 😀

MJay99 wrote on 2020-08-12, 16:52:

I just watched him do the replacement and almost couldn't help talking and pointing at my screen... 😁

As for the fire, I'm thinking that was an exaggeration in the post before. When he turned it on, there didn't even seem to be any magic smoke. 😀

Agreed, my thinking is with these rare boards its definitely best left to the professionals if your personal skill isn't at a high level and when things aren't easy to replace it's best to be very honest with yourself about that.
Indeed, that's why my 5428 upgrade is being done by folk who do chip replacement every day for a living, as frankly me doing the replacement would be a very good way to trash a good board very quickly !!
That said even with the best skills sometimes things do go wrong, so always best to go in with your eyes wide open on these things.

Yes though that was probably the case with the magic smoke comment 😉 but it does highlight the need for caution when doing these things !

very interesting info Zalog about the 386sx board with a 486slc cpu... too bad we can't locate the correct bios - most likely the only change is the activation of the 1kb of cache 🤣

keropi wrote on 2020-08-12, 18:30:

very interesting info Zalog about the 386sx board with a 486slc cpu... too bad we can't locate the correct bios - most likely the only change is the activation of the 1kb of cache 🤣

Looking into the whole issue has been a really interesting bit of Amstrad PC history, Indeed the whole of the later (6000 series to 9000 series) set of Amstrad PC is definitely worthy of a bit more study as they do tend to get overlooked (MegaPC being a notable exception !).

I'm sure we'll locate a copy of the 7386 (486)SLC Quadtel BIOS eventually - I'm keeping my eye's peeled for a 7486SLC80 at the moment, but these things do take some time to turn up....... but if anyone out there on Vogons has a 7486SLC80 and wants to share it's BIOS here, please do 😀 - totally agree though that the only difference we'll probably see is a tiny bit of code for L1 cache activation !! still would be nice to know 😀

Zalog, thanks for the interesting information!

Its crazy how Amstrad failed with their desicions.. I think if they had just waited a bit and released the Mega PC with the PC7486slc board it may have been a different story. I understand the collector not wanting to let go of their BIOS so hopefully an image of the PC7486SLC80 BIOS.. and yes it seems very much like its just activating L1 cache!

I think if my board isn't damaged I will keep it as it is, I didn't realise how rare they are TBH... I saw in your list of BIOS and boards that you know of 12 units, most of them damaged... thats not many good ones left

The MCretro stuff made me cringe to be honest! I think I would have experimented on a more common board before attempting to transplant the CPU. I have the ability to perform that level of soldering nicely so if I ever try it, it should be ok.

Zalog wrote on 2020-08-12, 08:16:

That said, it's possible we may get BIOS images from another MegaPC Plus out there, or another possibility which may be easier to get hold of is the BIOS from the Amstrad PC7486SLC80 (where 80 is the HDD size, not speed). If anyone has photo's of the motherboard of that PC, please let me know as I'd love to see them. this board has a Ti486SL 25Mhz MAB processor on the 386 board, indeed the first post on this link suggests it https://assemblergames.org/viewtopic.php?t=2243&start=230. From what I've heard this PC was made up from reworking the surplus 386sx boards, and using a 25Mhz processor so they could save money on replacing the clock chip (!), but DO have a Quadtel 486SLC bios. Either way these PC7486's have the barrel battery too, so are probably just as rare in working condition as the MegaPC, but if a working one could be found its BIOS would probably do for a MegaPC Plus clone on the 7386 board.

Hello there. I'm the one who made the post linked above. I've upgraded to a 486SLC33 motherboard since (the one with the CL5420 chip and AMI BIOS) but I still have the older PC7486SLC80 one. And the processor is indeed a Ti486SL 25Mhz as the following picture I've just taken of the naked motherboard suggests (click to enlarge):

I'm taking the opportunity to also include a download link a to PC7486SLC80 dumped BIOS image. I have no idea if that would work into a regular 386SX25 one but who knows. Here's a mirror of mine where I keep all Mega PC related stuff.

Above all, I'd like to greatly thank keropi for heading me here and providing a link to the original built-in software of the machine! Right when I almost lost hope. This is a goldmine we have here and I'm just realizing all the upgrade possibilities this PC7486SLC-33 motherboard has.

hey Dioxaz! thanks for the BIOS and pics!!!!
do you have any idea what this patch wire does? was the mobo repaired or this is needed for the 486slc to work?

turns out the mega-pc crowd is plentiful 🤣

Dioxaz, thank you so much for sharing the BIOS!! Your site has lots of awesome stuff too, all the software!! Man, this thread is becoming epic!

What is the ow362e16.exe program for?

I'm still waiting for my Mega PC to be picked up and checked.. Kinda nervous on what condition it will be in. No luck as yet finding the 486slc board or the Monitor. My Mums attic might be worth a try but its rammed full of stuff so if they are in there still its gonna be a bit of a job..

Dioxaz wrote on 2020-08-13, 19:42:

Hello there. I'm the one who made the post linked above. I've upgraded to a 486SLC33 motherboard since (the one with the CL5420 chip and AMI BIOS) but I still have the older PC7486SLC80 one. And the processor is indeed a Ti486SL 25Mhz as the following picture I've just taken of the naked motherboard suggests (click to enlarge):

Hi Dioxaz

Welcome and thankyou so much for sharing your BIOS, it is really useful ! From your photo, I can see the BIOS part number which is really handy, it comes in the series after the MegaPC Plus, but before the 7486SLC33Mhz so it confirms things nicely. Would be really interesting to replace the two bios chips (big number one and Quadtel one) some time on a damaged 386SX board to experiment - it should work in theory 😉

Like Keropi, I'm really interested in where that additional wire link goes before as I haven't seen that- it appears to connect to the Ti486SLC near the KEN, FLUSH and A20M pins - not sure which one pin its connected to though (nor where it connects on the back of the board!), but must be there to improve stability and help L1 support, probably based on previous experience with earlier boards ! It is very interesting 😀

Thanks from me as well for the link to your site its a really handy resource .

You're welcome.

@phipscube: That's a 16-bit version of Opera. It turns out to be the most "usable" browser with that machine (most up-to-date, most functional but still prone to crash).

@Zalog: The patch wire goes to pin 19 of the keyboard BIOS socket (labelled "KEY", possibly keyboard controller), underneath the motherboard. The CPU pin where the wire is soldered is cut off. Which means the wire is connected to the matching pad on the board but the CPU pin isn't connected to anything. I can't tell which CPU pin it is exactly due to the insulating thing blocking the view. The cut-off pin is the 5th one from the bottom left (on my photo). I can't tell if the wire is soldered to a single pad or if there is a bridge.

Edit: There are 2 BIOS chips. One is the regular 30-pin one (labelled "486 058859") and the other is a keyboard(?) BIOS chip and is a 40-pin one, labelled "KEY". The patch wire is connected to the 40-pin one. Here's a close-up (click to enlarge):

Thanks for that extra info Dioxaz. So basically #FLUSH pin on cpu is cut, and the pad goes to pin22 of the Keyboard controller.
I wonder why is that... I am tempted now to attempt upgrade on a spare 386sx mobo since we have the missing BIOS thanks to you 😀

Are you sure it's not connected to the A20 pin, that would make sense being connected to the keyboard controller. It looks like the flush pin was just cut to make room for the wire.

Sounds pretty feasible especially if it's actually pin 22 on the keyboard controller (which is the mirrored position of pin 19). That's bit 1 of the keyboard output port which software uses to control the A20 gate so should be routed to the A20M# pin on a 486. The A20M# pin was an addition to the 80486 because it had internal cache and needed to know when a20 was meant to be masked, rather than the "gate" sitting on the actual A20 line between the CPU and RAM, hence the bodge wire if this board wasn't originally designed for a 486.

well I decided to be the annoying guy and asked Dioxaz to double-verify pin19
I will attempt the upgrade soon-ish 😀

Because I might be wrong in identifying the correct pin the keyboard controller, here's the front and the back of it (click to enlarge):

Looks like pin 22. Easy way to check, if the wire isn't connected himem.sys will likely complain about not being able to control the a20 line.

(Obviously I'm not proposing you disconnect it to find out, but for keropi's info the board should still be able to at least boot without it.)

Hey folks, I'm very excited to come across this thread with so many actively working on enhancing/repairing Mega PCs! After years of wanting one (having played on one back around ~1994 while a guest at someone's house), to supplement my Sega collection, I took a chance on acquiring a stock Mega PC with monitor and manuals a couple of months back and have been slowly repairing it and renovating it; cutting out the battery and cleaning up the leak (thankfully no serious damage), upgrade to 16MB RAM and 512KB VRAM, CF card adapter, building an extended floppy cable to fit around the CF card, rigging something to keep the monitor button pressed in as it no longer stays stuck in. I was delighted to find the original hard drive still working absolutely fine with Counterpoint still set up as it had been left, though I won't rely on that drive going forward.

My electronics skills are very rusty; I'm hoping to attach an external battery pack with 3xAAA batteries to the Ext Bat pins which I understand should work ok despite the original battery only being 3.6V, does anyone advise against that with the voltage difference?

I'm strugging to work out the +/- pins for the Ext Bat on the motherboard, using my aging multimeter, the resistance appears to be approx zero on pin 4 (furthest from the edge of the MoBo) which I think means that's negative? This is based on what I read at http://pc-restorer.com/replacing-cmos-batteries-in-old-pcs/ and how the pins for the original battery also behave but I would value a second opinion.

One other thing that's troubling me is that the PC display on the monitor is bowing in a little at the vertical edges as per screenshot attached. The megadrive display however appears to be absolutely fine. Can anyone advise on whether/how this is fixable? I've just ordered a Benq BL-702A on the back of @Zalog's advice as love the idea of a flat screen that supports 15KHz.

Also very excited to hear about @keropi's Orpheus sound card (with other utilities), put me down for one please!

I'd also love to do the 60Hz megadrive mod at some point but while I'm happy enough cutting that cable, my soldering is very rusty and was terrible to begin with so removing and replacing R75 leaves me feeling a bit nervous.

Dave

smsdave wrote on 2020-08-15, 22:42:

I'm strugging to work out the +/- pins for the Ext Bat on the motherboard, using my aging multimeter, the resistance appears to be approx zero on pin 4 (furthest from the edge of the MoBo) which I think means that's negative? This is based on what I read at http://pc-restorer.com/replacing-cmos-batteries-in-old-pcs/ and how the pins for the original battery also behave but I would value a second opinion.

One other thing that's troubling me is that the PC display on the monitor is bowing in a little at the vertical edges as per screenshot attached. The megadrive display however appears to be absolutely fine. Can anyone advise on whether/how this is fixable? I've just ordered a Benq BL-702A on the back of @Zalog's advice as love the idea of a flat screen that supports 15KHz.

Dave

Hi Dave,

Welcome - seems like we're gathering a fair number of the megapc owners here on vogons 😉 In answer to your questions - Don't know what type of external battery you'll be using, but if it's a lithum or other non-rechargeable, you'll definitely need to have a diode in the link to the motherboard otherwise the megapc will try and recharge the battery which is a no-no. If it's a rechargable then don't worry about this. There is a bit on adding an external battery on this video https://www.youtube.com/watch?v=6dWb-uw27Yg - can't remember which of the 4 video's its on, but you may find it helpful.

Glad the monitor idea is helpful - its certainly an improvement over the old monitor - don't worry my old one is permanently switched on too (!)

Zalog wrote on 2020-08-11, 08:05:

Will have a bit more to report at the end of the week as I've found a local PCB maintence firm who's going to do the graphics chip replacement for me - and I have a few upgrades in progress at the moment - until then I'll tantalise you all with the image of the NOS 5428 chip I managed to find for the upgrade !!

So a progress update - the chip is in...

All the Video RAM is socketed !

And new BIOS is installed - a BIG thanks to Keropi for making this !

All boots fine, but when I run the test - its this result ?!?

So what's going on here....??? We'll after some careful analysis it appears that I have picked up a geninue "fraud" chip from the 90's (!) - someone relabelled a batch of 5422 core and passed it off as a 5428's. Fortunately the seller is a reputable one, who once notified of this has withdrawn the chips and has refunded my cash (not that it was expensive) - but folks trying this in the future should perhaps use a chip from a known good graphics card in th future like Keropi did, to avoid this happening. Still, I'm not annoyed the chip change has fixed the memory issue (as in the picture now accessing all 512KB of the ram in there) so I can expand the RAM, plus the 5422 is faster, has 32bit memory access and more powerful than the 5420 so I really haven't lost out, and its only cost me the cost of the chip install which wasn't too bad. All in all fairly happy and will follow the mantra if it 'aint broke, don't fix it - so will live with this for now 😀