VOGONS


Amstrad 1512 - CGA(ish?) to VGA

Topic actions

Reply 20 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
sirlemonhead wrote:

Do you know anything about how the 'black' pin on the video is supposed to be handled?

I never noticed that before - how mysterious! My guess (slightly educated by the schematics I was able to find online) is that it's equivalent to an old-style CGA's composite blanking signal, and goes low for a while around the sync signals. This is probably useful for getting the levels right on a mono monitor. I think it's ignorable if you're connecting a CGA monitor or converting the signal to VGA, but it might be useful for connecting a composite monitor.

Reply 21 of 53, by kanecvr

User metadata
Rank Oldbie
Rank
Oldbie
reenigne wrote:
sirlemonhead wrote:

From what I've read, it isn't possible to disable the CGA

My understanding is that it's possible, but doing it involves cutting some tracks on the board.

Hmmm - do you know exactly what traces need to be cut? I've done some googling and came up with nothing.. some pics or a schematic would help greatly.

Optimally I'd like to add dip switches that allow disconnecting the on-board CGA card / and reconnecting it when needed. I don't suppose cutting power to the chip would be enough?

@Jo22 - thanks for the image!

Reply 23 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
kanecvr wrote:

Hmmm - do you know exactly what traces need to be cut? I've done some googling and came up with nothing.. some pics or a schematic would help greatly.

There's a schematic at http://www.retroisle.com/amstrad/pcs/Original … _Agujereado.pdf . Page 31 shows the VDU chip, IC126 labelled 40041.

kanecvr wrote:

I don't suppose cutting power to the chip would be enough?

I wouldn't recommend doing it that way - for one thing, an unpowered chip might give spurious outputs and/or cause excessive load on inputs. For another, it looks like the 40041 also decodes the addresses for the RTC/NVR, so that would also get broken doing it that way.

I think the best way would be to disconnect the BA9 and BA19 address lines (pins 4 and 75 respectively) on this chip. These pins should instead be grounded (connected to pin 64, 29 or 8, which should all be connected together already). Don't ground the other side of the cut - the one that goes off into the rest of the machine, or nothing will work and magic smoke might even escape.

With zeroes on BA9 and BA19, the VDU should not recognize accesses to ports 0x3dX and memory 0xbXXXX, so your VGA card won't conflict. It will have the side effect of mapping the RTC at ports 0x70 and 0x71 also to ports 0x270 and 0x271. However, that only thing that Ralf Brown's Interrupt List mentions for those ports is the Gravis Ultrasound, and that can be configured to use different ports.

Reply 24 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
reenigne wrote:

I think the best way would be to disconnect the BA9 and BA19 address lines (pins 4 and 75 respectively) on this chip.

Thinking about it some more, I'm not sure that would work because the power-on self-test routine at the start of the BIOS is going to be looking for the PC1512 CGA. A VGA card might not pass, and if it doesn't I'm not sure if it's possible to bypass the failed test. So it may also be necessary to replace the BIOS ROM with a modified version that doesn't do this test.

A less surgical option may be to look for a Hercules InColor card, which I think can drive an EGA display but uses the mono addresses. It probably won't be compatible with most of the games you want to play, though.

Reply 25 of 53, by kanecvr

User metadata
Rank Oldbie
Rank
Oldbie
reenigne wrote:
There's a schematic at http://www.retroisle.com/amstrad/pcs/Original … _Agujereado.pdf . Page 31 shows the VDU chip, IC126 label […]
Show full quote
kanecvr wrote:

Hmmm - do you know exactly what traces need to be cut? I've done some googling and came up with nothing.. some pics or a schematic would help greatly.

There's a schematic at http://www.retroisle.com/amstrad/pcs/Original … _Agujereado.pdf . Page 31 shows the VDU chip, IC126 labelled 40041.

kanecvr wrote:

I don't suppose cutting power to the chip would be enough?

I wouldn't recommend doing it that way - for one thing, an unpowered chip might give spurious outputs and/or cause excessive load on inputs. For another, it looks like the 40041 also decodes the addresses for the RTC/NVR, so that would also get broken doing it that way.

I think the best way would be to disconnect the BA9 and BA19 address lines (pins 4 and 75 respectively) on this chip. These pins should instead be grounded (connected to pin 64, 29 or 8, which should all be connected together already). Don't ground the other side of the cut - the one that goes off into the rest of the machine, or nothing will work and magic smoke might even escape.

With zeroes on BA9 and BA19, the VDU should not recognize accesses to ports 0x3dX and memory 0xbXXXX, so your VGA card won't conflict. It will have the side effect of mapping the RTC at ports 0x70 and 0x71 also to ports 0x270 and 0x271. However, that only thing that Ralf Brown's Interrupt List mentions for those ports is the Gravis Ultrasound, and that can be configured to use different ports.

Good read. I'm not sure BIOS will be looking for the CGA card specifically - it might just be looking for a graphics adapter - we'll see when I get to modding the board. I ordered some tiny SMD dip switches and some 0 ohm 0603 resistors = I'll solder the switches where I plan to interrupt trace conductivity as close to the chip as room allows. The 0 ohm resistors will act as conductive bridges to reastablish trace conductivity if I decide to undo the mod permanently (in case it doesn't work).

reenigne wrote:
reenigne wrote:

I think the best way would be to disconnect the BA9 and BA19 address lines (pins 4 and 75 respectively) on this chip.

Thinking about it some more, I'm not sure that would work because the power-on self-test routine at the start of the BIOS is going to be looking for the PC1512 CGA. A VGA card might not pass, and if it doesn't I'm not sure if it's possible to bypass the failed test. So it may also be necessary to replace the BIOS ROM with a modified version that doesn't do this test.

A less surgical option may be to look for a Hercules InColor card, which I think can drive an EGA display but uses the mono addresses. It probably won't be compatible with most of the games you want to play, though.

I'd really prefer to use a common VGA display adapter for this job - InColor cards are almost imposible to source locally and I don't want to spend a fortune on ebay for one.

Reply 26 of 53, by Jo22

User metadata
Rank l33t++
Rank
l33t++

Hello again,
I've just found a page about PC utilities for the PC1512/1640/2086.

Among these tools are:

Mouse driver
NVR program
Clock fix for PC1512 (for DOS above 3.x)
1512GFX (C code for 640x200 4bit mode)

And several GEM drivers..

Manuals are listed, too.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 27 of 53, by Jo22

User metadata
Rank l33t++
Rank
l33t++

Hello everyone,
I have found some more things related to the PC1512..

Article about Schneider PC (foreign language)

Amstrad PC1512 Review & Retrospective (cool page with background information!)

Also once more some stuff related to DOS Plus. I hope you don't mind..
- It's just that the Amstrads/Schneiders were among the most popular machines running this strange, little OS,
so I thought it would fit in this thread. The only other british computer I know of was the Acorn BBC Master 512
(see PC compatibility cards thread). It also got DOS Plus bundled with its Master 512 module..

A page about DOS+ v1.2 and v2.01
http://www.seasip.info/Cpm/software/cpm86.html#dosplus
Very interesting! Has EXEFIX TSR to make the OS compatible with Pkunzip and other Pacific C programs.
Alternatively, it provides instructions on how to fix the EXE Loader in DOS Plus v1.2.

Also has a patch for the XIOS to support disk up to 512MB and 1.44MB floppies.

A DOS Plus mini-howto can be found, too.

- Technical stuff -
The FAT filesystem in 16-bit CP/M-86
http://www.seasip.info/Cpm/dosplus_fat.html

DOS Plus XIOS
http://www.seasip.info/Cpm/dosplus_xios.html

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 28 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
reenigne wrote:
Jo22 wrote:

And I once read that the internal CGA does scan-doubling, making normal CGA games to appear less striped.

I don't think it does any scan-doubling - the timings are just the standard CGA ones from what I understand. Do you remember where you read that? I may check tomorrow, since I was thinking of getting mine out anyway to check something else.

I finally got around to doing this today. It definitely does not look double-scanned - the visible scanlines correspond to the pixel lines. I checked with my oscilloscope and found that the composite sync signal consists of a 64us horizontal (~15.6kHz) and a 16.7ms vertical (~60Hz), consistent with a standard CGA signal.

http://www.reenigne.org/misc/IMG_6471.JPG
http://www.reenigne.org/misc/IMG_6472.JPG

Reply 29 of 53, by Jo22

User metadata
Rank l33t++
Rank
l33t++

Thanks, reenigne. I had no possibility to check this myself, sorry for the trouble.
I only have vague memories of the article I read. It showed two screenshots (real screen photos actually) with a green-ish landscape and
a games character (a goblin ?). One picture was decribed as normal CGA(?) and the other one as Amstrad graphics (chip or mode ?).
And one was striped, while the other was not. That's pretty much all that I do remember. Maybe I do find that magazine again..
I think it is more than ten or fifteen years ago that I read it. Double that for the release date of the issue of the magazine.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 31 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
sirlemonhead wrote:

Does anyone know what and where to get the chips to upgrade from 512 to 640k memory for these machines?

They're just standard 4164s (16 of them). According to http://www.z80.eu/blog/index.php?entry=entry141221-233000 the 120ns version will work (not sure if the 150ns version would be fast enough or not). Looks like AliExpress is the way to go: two lots of https://www.aliexpress.com/item/10PCS-4164-M5 … a2-217e4c904704 and setting the jumper as described on that z80.eu page should do the job.

Reply 33 of 53, by sirlemonhead

User metadata
Rank Member
Rank
Member

Well, shit.

I finally got those memory chips and just went to put them in.. uh on. My 1512 is a later model (marked as 1988 on the motherboard) with a different memory config. Argh!

There are empty sockets, and the 512 to 640k jumper is present.

Here's what I have:

ram_slots.jpg

I thiiink the portable PPC512 has the same config? Could someone with more know how than me try confirm this? Here's some detail on that machine and how to upgrade it's memory: http://oldcomputer.info/portables/ppc512/index.htm

Reply 35 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
sirlemonhead wrote:

I finally got those memory chips and just went to put them in.. uh on. My 1512 is a later model (marked as 1988 on the motherboard) with a different memory config. Argh!

Ah, sorry about that! I wasn't aware there was a variant of this machine with a different RAM configuration.

Going by the pin counts, this does appear to be the same configuration (41464 has 18 pins and 32kB and there are four of them giving 128kB total). I guess the two 4164s (and the 41256s that are already installed) are for parity checking.

Reply 36 of 53, by sirlemonhead

User metadata
Rank Member
Rank
Member

No worries, not your fault - I had no idea either and didn't check. Sneaky Amstrad 😁

I can use two of the chips I bought at least! If anyone needs any for a '86 model 1512 let me know, I'll have 18 left.

Reply 37 of 53, by kanecvr

User metadata
Rank Oldbie
Rank
Oldbie

Well I gave up on trying to get a VGA card to work on my 1512. Frankly the only reason I did that in the first place is because my machine came with the PC-MM monocrome monitor and I really wanted color.... If anyone has a working or dead (psu) PC-CM for trade / sale I'm all ears.

acb.jpg
Filename
acb.jpg
File size
26.99 KiB
Views
1753 views
File license
Fair use/fair dealing exception

Reply 38 of 53, by sirlemonhead

User metadata
Rank Member
Rank
Member

When I have something working for converting to VGA, I'll let you know.

reenigne, will 4 of these in addition to 2 of the chips I bought previously work? http://www.ebay.co.uk/itm/5PCS-D41464C-D41464 … yEAAOSwqfNXmcuH

Reply 39 of 53, by reenigne

User metadata
Rank Oldbie
Rank
Oldbie
sirlemonhead wrote:

reenigne, will 4 of these in addition to 2 of the chips I bought previously work? http://www.ebay.co.uk/itm/5PCS-D41464C-D41464 … yEAAOSwqfNXmcuH

Yes, I think so.