I think that it"s not a new ideea, but i could't find a topic with this subject.
Just wonder if this is possible and what modifications are needed.
By achieving this, the CPC would be at the same Mhz level to Atari ST and Commodore Amiga :)
Some of us already tried and there are more about this ;)
6 MHz CPC - CPCWiki (http://cpcwiki.eu/index.php/6_MHz_CPC)
is there a detailed DIY with pictures about this mod ?
On the cpcwiki is only a short description...
The problem is that the 6MHz was done by changing the 16MHz clock with a 24MHz one... That mean all the system is overclocked.
So it is no more compatible with the existing CPC (sound, floppy, display...)
Another way may be to replace the Z80A with a Z80B (no CMOS 6MHz version) and generate a dedicated 8MHz signal to overclock it only.
From the 4MHz clock (Z80 pin6) or the 16MHz main clock (GA pin24) if the signal timing stay OK between the both IC...
But I suggest that some programs (like demo) counting NOP, will no more work properly. No?
Quote from: TotO on 10:02, 25 December 12Another way may be to replace the Z80A with a Z80B (no CMOS 6MHz version) and generate a dedicated 8MHz signal to overclock it only.
From the 4MHz clock (Z80 pin6) or the 16MHz main clock (GA pin24) if the signal timing stay OK between the both IC...
Sadly, that does not work on the CPC.Weirdly, That works with 16Mhz/3 (but it seems to be slower maybe because GA arbitration on memory) but not more, maybe some other components do not support faster CPU (maybe some electronician can say more about this)
On a +, the machine boots with some cartridges but crashes later.
Measuring performance is not easy, i used interrupts because they rely on Vsync and Hsync and, in theory, they remain identical with the new CRTC R0 config (still 312 lines and 64us/line)
Problem with NOP counting is dependant of the CPU clock.
BTW, more complex testing program is not easy as it is not possible to read standart discs.I think i'll experiment more with ROM board if it wants to work with a 24Mhz CPC.
Quote from: radu14m on 08:27, 25 December 12
is there a detailed DIY with pictures about this mod ?
On the cpcwiki is only a short description...
This is just because there is nothing that need a longer description.Just unsolder the 16mhz crystal and replace it with a 24mhz crystal.I can post a picture of the PCB if you want , or TFM if he added a selector.
Yes, I forget to speak about the RAM.
Most CPC get 150ms RAM chip, but some have 120ms embedded. May be that allow to work...
But sure, like we spoke 1 year ago, the overall performances are not proportional to the overclock. (thank you GA)
The interesting thing about the 24MHz way is to get a better overscan mode with 256x256 / 288x224 screens for games.
It"s clear that not all components are able to run at higer frecvency than 4.77Mhz.
Those components who are not suitable, can be replaced with better ones ?
I know, sound easy but it"s not.
Maybe the plus CPC's are more flexible for this mod, due to the fact that they were redesign significantly and enhanced too.
Some pictures with the mod would be nice too.
I imagine something like this:
All the software changes that need to be done in order to run and disply properly the application on higer frecvency can be writen into a special rom/roms.
The frecvency should be selectable at boot time.
With this overclocking possibility, we can have in our CPC the entire PCW familly too :)
A one in all solution, from 4Mhz to 16 Mhz......later some additional memory, an HxC, HDD ...ok i start already to dream 8)
8) 8)
Would it be possible to make a small pcb that plugs into the cpu socket, the pcb would have a faster z80 with clock circuit, cache ram and wait state generator, I remember something similar was made for the amiga 500 but with a 68000 obviously.
In my opinion, you might as well take an FPGA CPC implementation and tweak it to get a dream CPC (that is still clearly a CPC, not something new). A key feature would be supporting the standard CPC mode, with a way to kick into the advanced mode (much like the C64 SuperCPU board starts up in 1MHz 6502 compatability mode).
On a standard CPC you would probably need a CPU+RAM solution that plugs into the Z80 socket, with some logic to allow the rest of the system to access the on-board RAM (rather than the original CPC RAM) as if nothing had changed. It's probably possible, for a hardware guru. This would probably take the form of an FPGA implementing a Z80 core, with the on-board SRAM being used for system memory, and more of the FPGA providing a compatible interface with the rest of the system. However in the CPC IIRC the ULA interfaces to the RAM, not the CPU, as the ULA provides support for memory banks - this probably makes things horribly complicated.
Quote from: radu14m on 08:27, 25 December 12
is there a detailed DIY with pictures about this mod ?
On the cpcwiki is only a short description...
When I did the 6 MHz Mod two decades ago it was just easy, so no detailed description is needed:
- Replace the Z80A with a Z80B (or higher)
- Put a 24 MHz crystal on the 16 MHz crystal and add a switch that allows to switch between them.
Problems: It doesn't work with all motherboards
Quote from: TotO on 10:53, 25 December 12
The interesting thing about the 24MHz way is to get a better overscan mode with 256x256 / 288x224 screens for games.
That's probably the most important point :)
Data transfer: I used the Inicron accumulator buffered RAM-disc to transfer files and play Starfox, Nebulus and much more like never before ;-)
And yes: You have to adjust you monitor depending on the program you use with 6 MHz.
Quote from: steve on 22:06, 25 December 12
Would it be possible to make a small pcb that plugs into the cpu socket, the pcb would have a faster z80 with clock circuit, cache ram and wait state generator, I remember something similar was made for the amiga 500 but with a 68000 obviously.
You talk about the Sprinter card for the Joyce PCW ;)
If I could find a sprinter card, would it work in a cpc?
Not without some mods IIRC. But the principle itself is of interrest.
I didn't kow about the sprinter card...yes, something like this i imagine ;D
Wasn't there a turbo mode on the TREX hardware implementation of the CPC?
How did that do the turbo?
The Amstrad PCW16 is basically a Z80 Amstrad machine at 16mhz with all the video modes of a proper Amstrad CPC...
ok, i'm out... (still couldn't it emulate the Amstrad CPC ?)
Sadly the screen memory layout is sane, and also has the PCW per-scanline pointer system ("roller ram"), and in the colour modes the bit layout within a byte is also sane (AAAABBBB, not ABABABAB; AABBCCDD, not ABCDABCD). It really is an enhanced PCW, rather than a CPC++ refactored into a word processor.
Anyway ... easy vertical scrolling (and stretchy demo effects) because of the roller ram. I don't know about horizontal scrolling, even if the roller ram system let you specify the exact starting byte for a scanline, it would still be byte oriented scrolling rather than pixel oriented scrolling.
From the screenshots of the PCW16 attached to a VGA monitor, it appears that the system also has an extensive colour palette to pick from?
Quote from: Briggsy on 22:15, 26 December 12
Sadly the screen memory layout is sane, and also has the PCW per-scanline pointer system ("roller ram"), and in the colour modes the bit layout within a byte is also sane (AAAABBBB, not ABABABAB; AABBCCDD, not ABCDABCD). It really is an enhanced PCW, rather than a CPC++ refactored into a word processor.
Anyway ... easy vertical scrolling (and stretchy demo effects) because of the roller ram. I don't know about horizontal scrolling, even if the roller ram system let you specify the exact starting byte for a scanline, it would still be byte oriented scrolling rather than pixel oriented scrolling.
From the screenshots of the PCW16 attached to a VGA monitor, it appears that the system also has an extensive colour palette to pick from?
colour palette is the same as the cpc's.
modes are the same, although arranged as nibbles rather than interleaved bits.
You guys are great in getting off topic :P
Quote from: arnoldemu on 16:14, 26 December 12
Wasn't there a turbo mode on the TREX hardware implementation of the CPC?
How did that do the turbo?
That's not a CPC, that's another Chip simulating a Z80, so you can choose speed freely. Check out its HP.
Maybe my questions are naive because I'm not an electronics engineer. Maybe just stick the heatsink on the processor, overclocked by 100% and use VGA monitor to catch the synchronization. Although I do not know whether or not it would be necessary change memory into faster .
Or is it possible to overclock the only processor by 100% and leave the rest of the main board normal? On the Amiga 500, such methods works.
Or by an adapter insert Zilog eZ80. Apparently, despite the 24bit registers is compatible with Z80, and performs commands in 4 times less cycles. So in theory, the CPC would be 4 times faster, even without changing the frequency.
What you think about it? Do you know someone who already tried something like this?
On A500, this method (doubling the clock) allow a non sinificative gain. (around 15%)
Here, it's exactly the same... +50% speed allow 7% gain... Because, in both case, the RAM is not enough faster to take advantage on this hack.
You may have problem with the PPI too... Because, it will not work faster to drive properly the peripherals.
For many reasons, this unfortunately wouldn't work. At least not az 8MHz. If the Z80 was running at 8MHz, the Gate Array would need to be running at 32MHz and I doubt it can handle that. Several other parts that run at 4MHz would also have problems at 8MHz.
Bryce.
The only means by which the Z80 can/may be sped up is with replacing the RAM with much faster SRAM ones, feeding the Z80 with 8Mhz directly (rather than from the VGA) and replacing the ROMs with faster ones as well.
The RAM would not need any refresh by the VGA (or DRAM access controls), so the only issue might be with the RAM configuration since the VGA does control it somewhat.
If the above were possible, the increased speed would only be of limited use anway, since sound would different and disc access might prove an issue.
However, for general program execution like games, there maybe a gain in performance like playability. Also gains might allow for Video playback with quicker/smoother screen updates or even MP3 playback.
rpalmer
QuoteThe only means by which the Z80 can/may be sped up is with replacing the RAM with much faster SRAM ones, feeding the Z80 with 8Mhz directly (rather than from the VGA) and replacing the ROMs with faster ones as well.
You can't replace the DRAM with faster SRAM. You need DRAM too... And don't forget that the GatteArray is the master in board, not the Z80.
QuoteThe RAM would not need any refresh by the VGA (or DRAM access controls), so the only issue might be with the RAM configuration since the VGA does control it somewhat.
You are speaking about an ideal computer, not about the CPC...
The GateArray get control over the CPU for reading the RAM content. You can't avoid that.
QuoteIf the above were possible, the increased speed would only be of limited use anway, since sound would different and disc access might prove an issue.
But, it is not... And you have to change the PIO too... Finaly you change all the computer, only to get more CPU power for doing... What? Nothing... Better, to use an other computer! ;)
QuoteHowever, for general program execution like games, there maybe a gain in performance like playability. Also gains might allow for Video playback with quicker/smoother screen updates or even MP3 playback.
Just in dream...
Theoretically, you could put a faster Z80 and some SRAM on a daughterboard that fits into the old Z80 socket.
There are numerous technical problems to overcome though. The clocks would need to be very accurately synchronised, so you'd need a PLL not a crystal. You'd need to override the WAIT' signal generated by the gate array so as to not be stalled all the time, and you'd want to generate your own WAIT' signal anyway to access any of the RAM/ROM/IO devices on the mainboard, because otherwise the CPU would try to access data before the gate array presents the data to it.
As soon as you start considering that any memory that's being written to in the base 64K (which is where almost everything needs to run from) could conceivably be being using as screen memory and so would need to be slowed down to a 1us access pattern anyway.
In short, it's a very technically difficult job and even if you got it to work, you'd only speed up access to the local RAM that could only be used as a banked RAM expansion. It's really not worth even thinking about doing.
That said, you could build a CPC-like system from scratch but it'd be a complete redesign.
You are speaking about heavy hacking just to enjoy a faster CPU. Nobody will sacrify his CPC for that and nobody will program for that.
If you need to change the RAM, better to use dual-ported chip... So, you don't get more conflict betwin the GA and the Z80 and can unconnect the WAIT pin.
Quote from: ralferoo on 10:46, 06 March 13
That said, you could build a CPC-like system from scratch but it'd be a complete redesign.
Sure. And not a all-in-one FPGA...
Yeah, a 16MHz Z80 core in an FPGA accessing SRAM that can handle 16MHz accesses sounds like the best plan.
Even better, if the clock speed is controllable via a port, so you can have RSXs like |16 and |4, with the default being 4MHz for compatibility.
But the limited number of these CPC++'s would limit the software created for the configuration.
Quote from: TotO on 09:16, 06 March 13
On A500, this method (doubling the clock) allow a non sinificative gain. (around 15%)
Here, it's exactly the same... +50% speed allow 7% gain... Because, in both case, the RAM is not enough faster to take advantage on this hack.
You may have problem with the PPI too... Because, it will not work faster to drive properly the peripherals.
I'm not goint to change that in the Wiki page, because I don't want to have another war here. However my 6 MHz CPC runs definitivelz 50% more quick. But I did not use a BASIC program for it (which uses the software clock), but I did check out a couple of games, f.e. Starfox, Nebulus etc... it's a 50% gain!
Okay, you convinced me that overclocking creates so many problems that it is too difficult and not worth the effort. But what about the processor eZ80?
http://en.wikipedia.org/wiki/Zilog_eZ80 (http://en.wikipedia.org/wiki/Zilog_eZ80) If it's compatible enough, it could accelerate the programs even without overclocking. Perhaps this is only hope for a significant acceleration of the classics. Maybe someone smart and brave to consider? Or try out in redundant 464 or Spectrum. If this works, you can gain fame and start promoting it. :) RSX |16 and |4 sounds great. :)
I've heard about Turbo CPC FPGA such as C-One and T-Rex, or would they fit in a standard case 6128 or CPC+?
With the right 6128 keyboard (quick RAMs!) overclocking is as easy as adding a second crystal and some wires 8)
The eZ80 brings on problem, it's its IO, better use a Z280 (Actually there are cards replacing a Z80 with a Z280 and RAM).
But if you want a solution like the Super-CPU for the C64 then just make a new computer ;-)
You may also like to check out the CPCnG project.
Quote from: TFM/FS on 17:54, 07 March 13
With the right 6128 keyboard (quick RAMs!) overclocking is as easy as adding a second crystal and some wires 8)
Yes, some CPC get faster RAM chips.
Yes, I guess they were made around 1988, but I'm not sure. However it's written on them.
Also some hardware expansions run at 6 MHz and some don't.
RRB does not work, Eprom-card does ;-)
I guess you talk about LC version with small ASIC, palette 4096 without sprites (1988-89). I just looking for this version. :)
No, actually not the cost down model, but the cost down model has a good chance for quick RAM - on the other hand the GA may be not quick enough.
The MegaFlash would also run quite happily on a 6 or 8MHz CPC.
Bryce.
Quote from: ZbyniuR on 01:48, 08 March 13
I guess you talk about LC version with small ASIC, palette 4096 without sprites (1988-89). I just looking for this version. :)
I don't think the cost down has a 4096 colour palette??
if it does, how is this accessed?
Quote from: Bryce on 08:58, 08 March 13
The MegaFlash would also run quite happily on a 6 or 8MHz CPC.
Yeah! And now I have something I
REALLY can look forward to! Will enjoy it during my holidays in May :) :) :)
Quote from: arnoldemu on 14:47, 08 March 13
I don't think the cost down has a 4096 colour palette??
That's correct. Of course you're right.
It appears that if you can get hold of a C-One, then you can run a CPC core on its FPGA at up to 80MHz - from 2009: C-ONE Reconfigurable Computer (http://c64upgra.de/c-one/)
Looks like there is source code on the download page too. Might be useful for relferoo's project?
Quote from: Sykobee (Briggsy) on 10:15, 13 March 13
Looks like there is source code on the download page too. Might be useful for relferoo's project?
Well, I've already implemented everything myself from scratch apart from the Z80 core, so I don't really need the C-one core. I'm sure it could be fairly easily ported to my board though. I have some recollection of looking at it a couple of years ago, but I can't remember why I didn't like it now...
I'm currently using the opencores T80 softcore which claims to be able to be clocked at 100MHz, although I've never tried it that far. Certainly, when I synthesised it on the Atmel chip I was prototyping with before I built my own board it wouldn't synthesize above 16MHz but the Xilinx tools are happy letting me run it at 48MHz.
That said, there's a couple of bugs in the T80 implementation that I have mostly fixed in my source, but I still don't really like the way they've implemented the Z80 soft core, so I'm planning to rewrite my own Z80 core at some point too. There are certain things they've done that don't accurately reflect how the Z80 actually works, e.g. sampling WAIT on the clock rising edge not the falling edge that means I have to implement the gate array logic to work to a broken Z80 implementation. But mostly, I want the satisfaction of having independently recreated every aspect of the CPC... :)
The C-One has a Z80 core in which a lot of bugs are killed, which you usually find in the normal Z80 core.
What about overclocking a PLUS and its Asic?
(enjoy it explode...)
Haven't tried it, because the crystal it tightly linked to the PCB and I don't want to screw it up. However if may be worth trying to increase the frequency of the whole system by 25% (5 instead of 4 MHz Z80). That should stay inside the reality tolerance levels. But nobody knows how the ASIC will react...
Did anybody try to overclock the cpc @5Mhz ? ( if 6Mhz creates problems )
It would be good if the wait states could be removed for extra ram and ROM, these are not in contention with video hardware. If Tue z80 was clocked at 8mhz we could have a bit more throughput? 2 z80 clocks instead of 1. No change to system clock but separate clock for z80. Has anyone tried giving the z80 a different clock and keeping system one at normal?
Quote from: arnoldemu on 13:27, 24 August 14
It would be good if the wait states could be removed for extra ram and ROM, these are not in contention with video hardware. If Tue z80 was clocked at 8mhz we could have a bit more throughput? 2 z80 clocks instead of 1.
With a dedicated memory bus (for ram/rom), that should give us twice the throughput, but we end up with amiga like slow/fast mem issue when accessing the on board HW and video RAM.
Quote from: arnoldemu on 13:27, 24 August 14
No change to system clock but separate clock for z80. Has anyone tried giving the z80 a different clock and keeping system one at normal?
This cannot work because of the GA that deals with the DRAM address muxing. Only one RAM access is possible per µs. 8MHz Z80 could try 2.
The only way is to either overclock the GA, with all the associated problems (video), or a dedicated bus with an interface to access the slow HW.
@Gerald: And what do you think about using Dual Ported RAM for disabling completely the GA contention? Because GA and Z80 could access simultaneously to the RAM (GA = Read and Z80 = Read/Write).
Quote from: SyX on 18:08, 24 August 14
@Gerald: And what do you think about using Dual Ported RAM for disabling completely the GA contention? Because GA and Z80 could access simultaneously to the RAM (GA = Read and Z80 = Read/Write).
That's a possibility. However, you still need to heavily modify the GA controlled datapath. Also dual ported RAM are not that common (at least for CPC hacking :D ).
Then you hope that other HW can cop with the 8 MHz Z80 (PPI/FDC, extension)
But we are no more talking CPC, are we ? ;)
Once my friend had the Amiga 500, with system clock frequency doubled but only for the CPU, motherboard and the rest work normally. With glued cooler on processor, and it worked well. This idea was described in Amiga Magazine.
As I know in CPC - 16MHz goes to GA and this generates a 4MHz for Z80 and 1MHz for AY, maybe something else I don't know. It may be possible to speed up the CPU only in the same way as in Amiga, just doubling frequency somewhere between GA and Z80? And if necessary, change the RAM and ROM into faster one. And I hope disc drive and picture on screen still will be fine. :)
Quote from: gerald on 18:52, 24 August 14
But we are no more talking CPC, are we ? ;)
Well, those are too many hacks for an old CPC motherboard, but not for a new one, jejeje.
Quote from: ZbyniuR on 20:36, 24 August 14
Once my friend had the Amiga 500, with system clock frequency doubled but only for the CPU, motherboard and the rest work normally. With glued cooler on processor, and it worked well. This idea was described in Amiga Magazine.
He can update the CPU for a 16MHz version to not have to cool it... But, in both cases the fact to increase the CPU clock and not provide it the adequate fast RAM, is just useless... Probably, he had increased it by 5% or 10% ... Not 50%... Wrong? ;)
Quote from: radu14m on 12:56, 24 August 14
Did anybody try to overclock the cpc @5Mhz ? ( if 6Mhz creates problems )
I have one CPC6128 in which I replaced the Z80A with a Z80H (had no Z80B at hand), and I overclock the whole system witha 24 MHz crystal instead of a 16 MHz crystal. In my hands it works like a charm. But some expansion hardware is to slow to work with it.
it"s possible to use a 20Mhz crystal, right ? to get 5Mhz :)
As it was already explain on the first page, changing the main clock as a side effect on the sound and the fdc.
The result is a CPC that no more work properly.
Quote from: radu14m on 18:31, 25 August 14
it"s possible to use a 20Mhz crystal, right ? to get 5Mhz :)
Yes.
Quote from: TotO on 18:42, 25 August 14As it was already explain on the first page, changing the main clock as a side effect on the sound and the fdc.The result is a CPC that no more work properly.
Well, it works properly for me. The picture on the monitor has a higher resolution (using 32 KB fits on screen w/o a problem). The soundchip is moved to a higher region, more like Atari ST (half the way there). And the FDC can write more on discs (50%). All that side effects are very welcome to me. :)
Quote from: TFM on 19:01, 25 August 14Well, it works properly for me. The picture on the monitor has a higher resolution (using 32 KB fits on screen w/o a problem). The soundchip is moved to a higher region, more like Atari ST (half the way there). And the FDC can write more on discs (50%). All that side effects are very welcome to me. :)
Sure, I know that. But, in reality, that made the computer no really running faster. (something like 5% instead of the 50% expected)
The PIO should fail because it is not intended to run 50% faster.
The Audio sound "halftone" and original floppies can't more be read. (and 50% more disc need a custom format)
Well, if your PIO fails then the whole system does run not much more quick.
I did a benchmark and in my case everything worked well, it is 50% faster, tested with some simple BASIC programs and stopped time by hand.
It probably depends which keyboard you are using. Some seem to work at 6 MHz, others don't.
@TFM (http://www.cpcwiki.eu/forum/index.php?action=profile;u=179):the used cpc6128 was one of the latest model (Cost Down Version) ?
What memory chips & speed ( ns ) are build in ?
It was a 1987 CPC with CRTC1. The RAMs are usually fine (regarding speed). IMHO the GA is the one that makes problems first. But if you try 5 MHz everything is fine.
So.
would need heavier mods to get proper display resolutions and this would then need more RAM used as VRAM then... so the 64K or main RAM would be quite fastly used as VRAM.
on the other hand a properly X2 overclocked CPC would basically become an 8bit Atari ST. ;D
Minus the megaRAM management and the 16/32 CPU and the MIDI port and the 512 palette... also what about mode2 ? can it be turned into 640x400x2 ?
Depend if you want a full overclock on every thing, or jsut the CPU running at 8mhz (or 6) but the Video staying at same speed...
if gates array handles both from the original clocking, you can't separate unless you redesign the GA.
Also as some video stuffs are to be synched with the CRTC... (50hz cathodic stuff) there are signal that must not really be overclocked.
BTW... 6mhz and 192x256x16 ? is that a SAM coupé ? ::)
Anyway wouldn't be the same machine, and wouldn't be really compatible unless you want all display divided by 2 on screen...
Not even sure "fullscreen" could be achieved... if the video if also overclocked (=384x256x16 per exemple)
would need to be able to run 48k of RAM for one screen.
hell a simple normal 320x200x16 creen would be actual "fullscreen" in 32K...
Will not really work on a 464 lol.
in order to do it properly you'll have to replace :
= crystals
=RAMs
=CPU
=GateArray
=Also not sure the AY used can be run at 2mhz (think they could)
Not sure about the CRTC though...those were also used with EGA or VGA cards (may not be the same models though) if I remember well, so may perhaps handle some 320x200x16, 640x200x4 or even 160x200x256... :laugh:
as a result you would have to design a pin compatible Gate array/Asic that would be somewhat 100% compatible yet with a few tweaks : some clockings would be doubled, others wouldn't... and a few stuff would be redone in order to handle some extra monstruosity.
I mean we all agree a 1280x200x2 mode wouldn't work... :laugh:
otherwise, such souped up CPC on cocaïne would still have the same CPU to Video ratio... wouldn't be faster, jsut display more colours.
If you keep the existing classic CPC video modes, then you may have a thing quite faster but it would still be slowed down by the video interruption when "VRAM" is read.
As a result such device should be able to handle both overclocked video modes and classic video modes.
The aleste 520ex has two speeds. One where 16Mhz is master clock. This is the one with the CPC's modes.
But you can switch to 13Mhz master clock, in this mode the pixels are wider I think, but this is the mode with the MSX modes.
I don't know how the CPU suffers, it should run slower in 13Mhz mode, but it may be delayed less. I have not looked into the possibilities.
What I am saying here is that the aleste can give a resolution with more colours I think (e.g. 320x200 in 16 colours). So faster speed may not always mean higher resolution.
CPC and MSX (or spoeccy) wouldn't produce pixels on the same size nor same display on the same monitor.
I like to say that the CPC can manage to put 320x200 where Speccy can only put 256x192... (need to verify betterly)
So the "video" clock difference and pixel ratio difference on the Aleste may come from the need to stick to original machines display.
When you get a 256x192 on a CPC, you feel like having only a post-stamp displayed despite the exact same resolution.
A but like the 256x192x16 on MSX2 or 192x256x16 on CPC in mode0 ful screen... same resolution but not the same border, not the same pixel ratio, not the same size on the screen, and so on..
I haven't proved it yet but the differenc in pixel clock says the spectrum has almost square pixels. Cpcs mode 1 has thinner pixels, so yes bigger border on CPC. Aleste Msx resolution probably has wider pixels than spectrum.
I swapp yesterday the 16Mhz with an 24Mhz oscilator on my cpc.
Unfortunately i did not get a stable image >:( [size=78%] ( even tested on 2 TV's )[/size]
did i miss something ?
Quote from: radu14m on 08:32, 30 August 14
did i miss something ?
Yes ;) .
You now need to update the ROM so the CRTC is programmed according to its new frequency (1.5MHz istead of 1MHz) and generate a 15kHz/50Hz syncs.
As a side note, your floppies can't be read anymore (unless you make them spin at 450rpm :-\ )
But TFM did not update the ROM....it worked directly.
Quote from: radu14m on 09:39, 30 August 14
But TFM did not update the ROM....it worked directly.
He may not have updated the ROM, but you still need to adjust the CRTC timing.
Check there : 6 MHz CPC - CPCWiki (http://www.cpcwiki.eu/index.php/6_MHz_CPC) . You will have to adjust the value to your clock (20MHz instead of 24MHz)
is the updated ROM available here somwhere ?
:)
Quote from: gerald on 09:29, 30 August 14
Yes ;) .
You now need to update the ROM so the CRTC is programmed according to its new frequency (1.5MHz istead of 1MHz) and generate a 15kHz/50Hz syncs.
As a side note, your floppies can't be read anymore (unless you make them spin at 450rpm :-\ )
No, the floppy works fine. Of course you can't read old discs. But where is the problem. Data transfer can be done using a RAM disc. The big gain is that you can use a format which gives you nearly 0.3 MB on a 3" disc. It worked in my hands and the floppy is a regular 3" floppy.
I didn't alter the ROM. Just typed in a couple of OUT instructions under BASIC :)
Can you post here the BASIC OUT instructions that you type in ?
Thanks ! 8)
Don't remember the value I used back the day. But it's enough to increase the number of total characters per line by 50% :)
understand but how ? :)
Check out the CRTC page in the Wiki. Then...
CRTC - CPCWiki (http://www.cpcwiki.eu/index.php/CRTC)
Try this in BASIC
OUT &BC00,0 selects CRTC Register 0 (horizontal total)
OUT &BD00,95 sets number of total CRTC characters to 96 (95+1)
That should work IIRC. You can further adjust the horizontal positioning.
A brand new addon with overclocking cabilities and speedy ram, very nice :o :laugh:
it's working TFM ! :)
I manage to overclock the cpc to 6Mhz, but unfortunately the hxc is not working anymore :(
the same situation on 5Mhz. ( OUT &BD00,77 )
do i have to change something else too ?
Well, first! Awesome that it works!!!
Second, the FDC is (at 6 MHz) working 50% more quick, so this is out of tolerance for the HxC probably. On the other hand you could try to use a HD format (just an idea, I don't expect it to work).
You should be ok with 3" or 3.5" disc. Now how to get data to them? I would suggest to use a RAM disc. Back the day I used the !COPY command of the X-DDOS ROM and the RDOS which provides a RAM disc C:
The X-DDOS also allows to format discs. And you need to format discs for 6 MHz.
You can try this: Try to format a .HFE file on the HxC, with some luck this could work.
looks that i need o wait for my X-MEM to test this ;)
did you try to overclock the cpc to 8Mhz ? i struggle to find a 32Mhz crystal....
Yes, I tried, but in this case the Gate Array is definitely not quick enough. I think that most CPC's can make it with 6 MHz though.
I don't know if the X-MEM is quick enough to run with 6 MHz. But let me know it you do some tests.
Slowly I think again to get my 6128 up to 6 MHz (switchable back to 4 MHz of course) and make some stuff for it, because the screen resolution is far better. Also if the HxC would work - this would be awesome. If it doesn't work with 6 MHz, then a Firmware-Update could help IMHO.
well, i swap also the z80 with a z80b, but it was not necesary. The original Z80 from zilog run also @6Mhz without a problem. ( and did not get hot )
Memories are with 15ns not a problem for the new speed.
Some pictures:
(http://i60.tinypic.com/2poruqu.jpg)
The ideea with the switch is good, i use now a crystal socket and took him ouside the case - in that way i can swap the crystals easy. Now we need also a ROM for 6Mhz :)
Looks like this :
(http://i60.tinypic.com/2w5nfh4.jpg)
some z80 used for tests :) :) :)
(http://i57.tinypic.com/2e21e2t.jpg)
About the X-Mem maybe TotO can tell us more.....i modded my schneider cpc in order to test the Hxc and the X-mem @ 6 Mhz :)
Setting the CRTC every time is boring....
Who can fix the Hxc Firmware ? Jeff ?
The FDC gets a direct signal from the main clock, you need to isolate this signal (at IC205 pin 13) and add a seperate 4Mhz clock crystal. The FDC clock doesn't need to be synced to the main clock, so it will work fine and you can then read standard disks. This could also be switched so you have "compatiblity" or "high density" modes.
Bryce.
CAN You explain me why do you want to overclock our cpc ? What Will really be the gain ?
Would this overclock improve things with laggy areas of games? Like the R-Type homebrew on second level, Rick Dangerous 2 with laser beams etc going off. Shame we cant get an extension instead of changing crystals. Like an ACA accelerator :o
Quote from: TFM on 20:14, 03 September 14I don't know if the X-MEM is quick enough to run with 6 MHz.
:laugh:
TotO's answer probably tells: YES! It is!
However, the question is valid. I know for example that the RRB has problems at 6 MHz, while my EPROM board has not. So, yes, IMHO it's needed to check expansion hardware.
The MegaFlash would probably have problems, but it's not a hardware problem, it's a software issue. Delays required in the writing routines would be too short for the Flash. You'd need to write a 6/8Mhz version to correct the delays.
Bryce.
Instead of a fixed waiting time one can check bit 7 to wait for end of cycle.
If the flash code wait the IC chip to be ready instead of making a wait loop, it should be not a problem.
About the X-MEM, I have not made tests but the RAM is 55ns (90ns for ROM) and the CPC RAM is 120/150ns.
Great TotO!
@TFM (http://www.cpcwiki.eu/forum/index.php?action=profile;u=179): it is possible to prepare a ROM that will automaticaly check the bit 7 ( end cycle check)and adjust the CRTC accordly for 4 or 6 mhz?
Quote from: TotO on 22:00, 03 September 14
If the flash code wait the IC chip to be ready instead of making a wait loop, it should be not a problem.
About the X-MEM, I have not made tests but the RAM is 55ns (90ns for ROM) and the CPC RAM is 120/150ns.
On the MegaFlash this isn't possible, because you would have to keep switching the read/write switch to check the ready signal.
Bryce.
@radu14m (http://www.cpcwiki.eu/forum/index.php?action=profile;u=724): I don't see your pictures. Please post them again.
Quote from: Bryce on 21:14, 03 September 14
The FDC gets a direct signal from the main clock, you need to isolate this signal (at IC205 pin 13) and add a seperate 4Mhz clock crystal. The FDC clock doesn't need to be synced to the main clock, so it will work fine and you can then read standard disks. This could also be switched so you have "compatiblity" or "high density" modes.
Bryce.
i like that ideea. Do you have a schematics for this ?
Also would be nice to have an integrate frecvency switch possibilty, without need to restart the system.
like a turbo button :o
yestrday the pictures were fire..damn.
No, but I could do some when I have time. Switching frequencies "on the fly" would be rather difficult as you would need to add circuitry so that only one crystal is being used and then being divided for the lower frequency. Not impossible, but complicated.
Bryce.
Quote from: arnoldemu on 09:07, 04 September 14
@radu14m (http://www.cpcwiki.eu/forum/index.php?action=profile;u=724): I don't see your pictures. Please post them again.
I see them now.
I see lots of Z80's including soviet ones. :)
And I see a 68000, planning to swap the z80 for a 68000?? :laugh:
No, these are from my CPU collection :)
Quote from: radu14m on 07:19, 04 September 14
Great TotO!
@TFM (http://www.cpcwiki.eu/forum/index.php?action=profile;u=179): it is possible to prepare a ROM that will automatically check the bit 7 ( end cycle check)and adjust the CRTC accordingly for 4 or 6 MHz?
That should be doable with the X-MEM and the FlashGordon. ;)
Quote from: Bryce on 12:22, 04 September 14
No, but I could do some when I have time. Switching frequencies "on the fly" would be rather difficult as you would need to add circuitry so that only one crystal is being used and then being divided for the lower frequency. Not impossible, but complicated.
Bryce.
Thats great Bryce !
Someyhing like ACA solution should be done for cpc too. 8)
Pity that i cant use my HxC @6mhz
Unfortunately, the CPCs architecture wasn't made for extrernal accelerators and second processors like the Amiga was.
Bryce.
Quote from: TFM on 20:54, 04 September 14
That should be doable with the X-MEM and the FlashGordon. ;)
And symbiface, presumably? This would be very cool :D
Ok, just messing about here with a concept for a possible "Accelerator" for the CPC. The idea is this. A single 24Mhz Clock with a divide by three circuit to bring it back down to 16Mhz. Solid-state switching to allow you to switch frequencies without resetting and a seperate 4Mhz clock for the FDC (this could also be made switchable if required). The PCB would be installed by removing IC117 and plugging this board into the (newly added) socket. It would also have a single wire that would need to be connected to IC205 Pin 13 (after disconnecting the pin from the mainboard). It looks complicated in the schematic, but actually it's only 4 ICs including the 74HC04 that was on the mainboard.
It's just a first draft/scribble and probably full of mistakes and potential for simplification, but I thought I'd upload it anyway, because (A) radu14m asked for one and (B) To get the discussion rolling.
All electronics members (and non-electronics members), feel free to criticise and pick holes in the design / schematic / concept :)
Bryce.
It's amazing Bryce !!!! :D :D :D
thanks !
Quote from: Bryce on 14:12, 05 September 14
A single 24Mhz Clock with a divide by three circuit to bring it back down to 16Mhz.
Did you mean this? Because 24/3 = 8. You mean 2/3 (divide by 1.5) right?
@Bryce (http://www.cpcwiki.eu/forum/index.php?action=profile;u=225): What would the circuit look like if the main 16Mhz clock was kept for gate-array, crtc, fdc and sound, but a seperate 6Mhz/8Mhz clock was used for z80?
This would keep the sound, graphics and fdc the same but give more power to z80.
Just wondering...
Quote from: Bryce on 14:12, 05 September 14
Ok, just messing about here with a concept for a possible "Accelerator" for the CPC. The idea is this. A single 24Mhz Clock with a divide by three circuit to bring it back down to 16Mhz. Solid-state switching to allow you to switch frequencies without resetting and a seperate 4Mhz clock for the FDC (this could also be made switchable if required). The PCB would be installed by removing IC117 and plugging this board into the (newly added) socket. It would also have a single wire that would need to be connected to IC205 Pin 13 (after disconnecting the pin from the mainboard). It looks complicated in the schematic, but actually it's only 4 ICs including the 74HC04 that was on the mainboard.
It's just a first draft/scribble and probably full of mistakes and potential for simplification, but I thought I'd upload it anyway, because (A) radu14m asked for one and (B) To get the discussion rolling.
All electronics members (and non-electronics members), feel free to criticise and pick holes in the design / schematic / concept :)
Bryce.
With this circuit the Gate-Array is clocked higher so the resolution is higher? And the sound would be closer to 2Mhz and similar to ST?
It would be interesting to see a circuit for dma transfer... but finish the accelerator one first ;D
Quote from: arnoldemu on 17:47, 05 September 14
@Bryce (http://www.cpcwiki.eu/forum/index.php?action=profile;u=225): What would the circuit look like if the main 16Mhz clock was kept for gate-array, crtc, fdc and sound, but a seperate 6Mhz/8Mhz clock was used for z80?
This would keep the sound, graphics and fdc the same but give more power to z80.
Just wondering...
Quote from: arnoldemu on 17:48, 05 September 14
With this circuit the Gate-Array is clocked higher so the resolution is higher? And the sound would be closer to 2Mhz and similar to ST?
I think the z80 and gate-array, crtc and sound have to be synchronised, only the fdc can have a separate clock (I could be wrong). And so this would make the resolution higher (not sure about sound...)
Quote from: Munchausen on 18:48, 05 September 14
I think the z80 and gate-array, crtc and sound have to be synchronised, only the fdc can have a separate clock (I could be wrong). And so this would make the resolution higher (not sure about sound...)
I don't agree. Only croc and gate array needs to be synchronised. Fdc,ay and cpu can be different speeds. Ga is master and will hold off cpu when it needs ram. Cpu will try to use available slots and instruction timings will not be so predictable.
Quote from: arnoldemu on 19:52, 05 September 14
I don't agree. Only croc and gate array needs to be synchronised. Fdc,ay and cpu can be different speeds. Ga is master and will hold off cpu when it needs ram. Cpu will try to use available slots and instruction timings will not be so predictable.
Ah ok, I haven't checked , I was just going with my impression. You would know better :)
Quote from: Munchausen on 16:32, 05 September 14
Did you mean this? Because 24/3 = 8. You mean 2/3 (divide by 1.5) right?
Yes, I meant divide by 2/3 (quite a difficult division to do in electronics), not 3.
I'm not sure how much the clocks can be split otherwise. I think the AY could have its own clock, but every split means cutting tracks on the mainboard which I don't like to do. I'm pretty sure that if the CRTC, GA or CPU weren't synchronised that things would start getting pretty random and/or instable.
Either way, it is just a concept to get ideas out.
Bryce.
Bryce, can wait to see the first prototype !!! ;D ;D ;D
i try today the X-MEM @6Mhz and it"s working ! ;D
And if you increase the BIT RATE of the HxC FLOPPY EMULATOR by 50% (do that in the menu of the HxC itself for testing), then it will work with 6 MHz too....[nb]Well, I guess ;-)[/nb] :)
great ! i will test :)
the only problem is to set every time the CRTC :(
can you prepare a ROM file ? with an rsx command would be more easy !
Quote from: radu14m on 04:18, 29 September 14
great ! i will test :)
the only problem is to set every time the CRTC :(
can you prepare a ROM file ? with an rsx command would be more easy !
Since you got the X-MEM you could just patch the firmware. Adjust the CRTC values of the ROM in a way that it works. The ROM listing is somewhere in the wiki / internet (writing from PC at work, so got nothing with me).
Quote from: TFM on 15:11, 29 September 14
Since you got the X-MEM you could just patch the firmware. Adjust the CRTC values of the ROM in a way that it works. The ROM listing is somewhere in the wiki / internet (writing from PC at work, so got nothing with me).
If you modify the ROM, you better change the values for NTSC and add a switch on LK4 (Port B, bit 4 of the 8255) which is used to select the CRTC timing.
The 6128 FW dissasembly is available here : http://cpctech.cpc-live.com/docs/os.asm (http://cpctech.cpc-live.com/docs/os.asm)
Table for CTRC init are at address 05C5 for 50Hz, and 05D5 for 60Hz
That's a good idea, but since Radu already got an X-MEM it's more easy to use the switch already being on the X-MEM instead of adding another switch to the CPC. However on the long term I guess your idea is the best solution. :)
I was wonder how overclocking look on other platforms. Recently I read about for example Amiga 500, speed up (only procesor) by 100%, it accelerates only from 5% to 30%, depending on type of memory, because processor must wait for response. It's disappointing. But change processor into more resistant seem to be not necessary, because 68k 7MHz overclocked to 14MHz with an ordinary heat sink has a lower temperature than the same 68k 7MHz working without a cooler. I suppose that same would be in CPC. Of course, exist a lot turbo cards for Amigas.
Atari Mega STE in standard have switch 8/16 MHz acceleration and there is a 100% speed up. I read about similar alterations to ordinary ST, but I don't know details.
To Commodore you can buy Chameleon 64 - 20MHz. But actualy it's separate computer, not turbo card and it's not cheap, but performance is impressive even can emulate Spectrum 128k with ULA+. Apple already in 1986, did IIGS (a 16bit version of Apple II) almost 3 times faster than original, and in 1988 - IIc (8bit) but 4 times faster. Polish fans of 8bit Atari resently trying to do something similar as Apple. I read about testing prototype turbo card with 16bit versions of 6502. Almost 100% compatible, but faster only with new fresh writen software.
Russians and Ukrainians already in 90's made ZX Spectrum clones 100% faster, and later even 24MHz. As I read they only tweaks processor and memory and nothing else. This is apparently easiest way to keep synchronization and working of all connected things. They improve graphics too, but it nothing common with acceleration. Only Amstrads still in The Stone Age.
Maybe we should collect money and announce the prize, for someone who effectively accelerate Amstrad at least 100% ? ;)
Because otherwise it again just talk on forum what we can do, and again will end the conclusion that too expensive, too difficult, too slow, and it's not worth risking damage to beloved equipment. And the only thing what rest will be to use RSX FASTSCREEN and do not forget DEFINT in BASIC, or compilers. Because what for trying if we have emulators that so easy to accelerate. ;)
Well, acceleration is a complex problem, overclocking the CPU is the first step. Although in those days of FPGA, this is the less hack solution and let you experiment with more interesting things, as adding special video modes or embedding a playcity or a blitter or ...
Getting a CPC so faster as the CPC fast mode that Tobiflex made for the C-One or the CPC-Trex without an FPGA, it is going to mean a lot of changes inside the CPC. So many, that you should be better redesign the motherboard.
Of course, we can take a deeper look to other machines for getting inspiration.
For example, the Amiga way is more sensible that you make look. Amiga has two kinds of RAM:
1.- Chip RAM is the RAM used by the custom chips (mainly sound and graphics) and shared with the CPU. There is an IC that signal with device has priority over the RAM, the cpu or the custom chips, usually are the custom chips.
2.- Fast RAM: This RAM is used in exclusively by the CPU and is not shared with any other device. In our CPCs, we lack of this Fast RAM because the GA stop the Z80 even when is using the expansion RAM.
You only need to add a Fast RAM expansion to your Amiga, nothing else and the CPU will go 100% faster executing code in the Fast RAM, only will be stop (read as "running to the same speed than before") when access to Chip RAM. And of course, you can get a lot of patches to move the AmigaOS and its variables to Fast RAM.
In case you need more speed, the supercharged Amigas use expansions that replace the custom chips, like Graphics or Sound cards. Although with those expansions, the Amiga looks less Amiga with every step :P
STs have a memory system similar to the Amiga ones and a lot of these techniques can be shared.
In the case of the C64, the Chameleon is the new thing, but it's nothing more than a FPGA board, even can works without a C64 :P and if you port the CPC-Trex to it, then you will see a CPC in it, in the same way that happened with the C-One.
If you look deeper, you will discover the old SuperCPU (that it's now emulated in the beta versions of Vice) that add a 65816, a 16 bit version of the 6502 (the equivalent to the Z800 or Z280 in the Zilog world) used in Apple IIgs and Snes, running to 20 MHz and with 16 MBs of RAM. But only software made for the 65816 will make full use of this expansion.
But in the C64 world there is a few other interesting things, first one is the REU, a RAM expansion that uses DMA transfers (copy blocks of RAM without using the CPU). Then you have the C128, where the CPU goes to double speed, although during the C64 mode you can only use the extra speed while the video chip is not building the screen. Of course, the prototype C65 is even a better option for getting a superfast c64... if you can find one finished :P
But at the end of the day in the C64 world, the people only wants a way of acceleration the floppy drive XDDDD
Tthe Apple II has cpu acceleratos since the first day, even there is a wikipedia page about it (http://en.wikipedia.org/wiki/Apple_II_accelerators), and Apple used one of these solutions for making its Apple IIc (really pretty machine :)) goes to 4 MHz (a lot from of original 1 MHz :P ). A few of these solutions look CPC friendly.
I never considered the IIgs part of the II family, too much improved in the graphical and sound side, although it's compatible and has similar acceleration solutions as its older brother.
The Atari 8 bit computer guys solution looks very similar to the C64 SuperCPU... these guys always with the same inferiority complex about the C64... always that i read an A800 forum, automatically i think in Iznogoud xDDDD
The russian speccy clones are not so faster, at least the 90s ones, but there is a lot of nice speccy clones as Peter Plus's Sprinter that go really fast. The Sprinter has an old FPGA as video chip and include superlovely graphical modes... it's sad that this machine is practically impossible to get those days.
The superbasic speccy architecture (practically a z80 + ram) helps a lot to having very few limitations and you can add things like DMA chips from Zilog for accelerating the instructions that use repetition (LDIR, LDDR, OTIR, CPIR, INIR, ...). And except in one case, a lot of those graphical improvements are based in interlace modes, that means can not be used without get blind xDDD
The only difficult part in the speccy hardware was its ULA (our GA + CRTC), but it was documented by Bob Smith in a book a few years ago, and since then, the speccy has converted in one of the most FPGAble machine in the world and a lot of those FPGA implementations add extra things (more speed, new graphical modes, ...).
There is a LOT (http://tarjan.uw.hu/zxclones_en.htm) of fun expansions for ZXs, one that i would like to test add an MSX video chip in the expansion bus and this is another project totally portable to the CPC, that it would help to bring a few MSX, Coleco and Sega SG-100 games to the CPC ;)
Another interesting fast Z80 project is the V6Z80P (http://www.retroleum.co.uk/electronics-articles/v6z80p/), one of my favourites because is an Amiga 500 with a Z80 at 16 MHz and an FPGA chip where the audio and video are syntethized, an a CPC mode could be possible.
The BBC is another superinteresting machine to look, for example its Tube expansion where you can connect extra CPUs. And those days we could put there an ARM and send code to be processed in the ARM and returned to the host machine. That is another possible solution for accelerate the CPC, even the ARM could be running a Z80 emulator, jejeje.
I could speak about the Dragon/Coco world and other more rare machines. But i only going to comment two more, because this is boring and everybody get the general idea.
First one is the MSX, a machine really easy to accelerate the Z80 because the video chip uses is own RAM, and the fast Z80 only needs to add extra waits when is sending bytes to the video chip. You can find MSXs with Z80B or the Turbo-R that use the R800, an improved Z80 clone that goes at 7 MHz but it's more or less equivalent to a 28 MHz Z80. Of course, there is a lot of Graphical and Sound expansions for MSX, even they have one expansion that add the Master System video chip, a nice improvement over the old VDP.
And the last, the PCW, you can take a look to the Sprinter board, and accelerator board that goes in the Z80 socket and add a faster Z80, extra RAM and shadow RAM. The code in the extra RAM goes to the full speed of the new z80, the shadow ram is a copy of the PCW internal RAM and every thing is written there, it's sent to the PCW internal RAM (those writes and the I/O access goes to the normal PCW speed).
Quote from: ZbyniuR on 20:47, 29 September 14
Only Amstrads still in The Stone Age.
Well, the only way of changing that is working to fix it, not waiting to somebody does for you. TotO and me wanted a CTC and 2xYMZs, and if we were waiting to somebody for doing it, we will continue waiting :P
Quote from: ZbyniuR on 20:47, 29 September 14
Maybe we should collect money and announce the prize, for someone who effectively accelerate Amstrad at least 100% ? ;)
Because otherwise it again just talk on forum what we can do, and again will end the conclusion that too expensive, too difficult, too slow, and it's not worth risking damage to beloved equipment. And the only thing what rest will be to use RSX FASTSCREEN and do not forget DEFINT in BASIC, or compilers. Because what for trying if we have emulators that so easy to accelerate. ;)
Sure speak is free and painless, work is hard and take time :P Accelerate a CPC is easy, replace the CPC RAM for a Dual Port RAM and you can cut the Gate Array pin that stop the Z80. Now the Z80 will go to 4 MHz and the GA doesn't disturb it, then buy a 16 MHz or faster Z80, replace the old one and connect the 16 MHz system clock to it and that is all. You will get a 400% faster CPC.
My bank details are... xDDD
Your post look 1000% overclocked!!! ;D
:-X :o
Quote from: SyX
-----
" Accelerate a CPC is easy, replace the CPC RAM for a Dual Port RAM and you can cut the Gate Array pin that stop the Z80. Now the Z80 will go to 4 MHz and the GA doesn't disturb it, then buy a 16 MHz or faster Z80, replace the old one and connect the 16 MHz system clock to it and that is all. You will get a 400% faster CPC."
-----
If it's so easy. Has anyone done it? Was it presented somewhere?
I guess is that you can add some circuits that will give CPC extra skills, but I'm afraid that number of people that will be interested in writing for this program will be negligible. So I think that the priority should be to keep compatibility and ability to run old programs. Except that faster.
If someone, some day will build a new motherboard, here's my wishlist:
1. Suitable for old casing 6128 or 464 and uses old keyboard.
2. Works as CPC PLUS, but faster.
3. SD card as an external storage with support for folders and DSK.
4. USB mouse and pads.
5. It can display images from a ramdisk.
6. Digiblaster.
7. Displays normal PAL.
Quote from: ZbyniuR on 14:36, 30 September 14
... but I'm afraid that number of people that will be interested in writing for this program will be negligible. ...
I will support a faster CPC or compatible computer.
I dunno how to patch the FW.
i would need your help ppl :D
increasing the BIT RATE of the HxC FLOPPY EMULATOR by 50% doesn"t work @6Mhz on mine....maybe i do something wrong :o
Well, you can try different bit rates. Somehow it should be working.
tried 250, 300 and 500 ! same situation.
IIRC the CPC uses 250 right?
So... can you set it to 375 ??
(I can't. Have no working monitor).
only 250, 300 and 500 can be set on my HxC.
Oh bad! Maybe the next firmware update could have 375 too?
in the menu beside de 250,300 and 500 bitrate i have also the option to set the birate from the HFE file.
would this be a solution ?
did you manage to patch the FW for 6Mhz ? :)
Thanks TFM !
never cared about it. Did a FutureOS for 6 MHz. But still I have no monitor at my CPC yet. As soon as I have a new monitor I can take a look at the Firmware though.
Are you ok to have a firmware patch which enables the CRTC for 6 MHz when switching to "60 Hz", that would need you to integrate a switch into the CPC with two wires to the PCB at LK 4.
Or do you like a patch which is always active?
well, both solutions are fine but for long therm, even if i don"t like to another switch, the switching only @ "60 Hz" is better.
Thanks TFM ;) ;) ;) !!!
I alway wonder why the cpc-s have so many switches ! Now i understand :o
Btw, it is possible to ask like for i.e. the cpc6128+ at start, to choose between 4 Mhz or 6 Mhz FW ?
Sure. Everything is doable, it depends how much time s.o. want's to put in.
Thanks TFM !!!!
;D ;D ;D
Quote from: TotO on 09:16, 06 March 13
On A500, this method (doubling the clock) allow a non sinificative gain. (around 15%)
Here, it's exactly the same... +50% speed allow 7% gain... Because, in both case, the RAM is not enough faster to take advantage on this hack.
You may have problem with the PPI too... Because, it will not work faster to drive properly the peripherals.
I am hoping with the GODIL spartin packages that RAM issue will be less of an issue, if it can be gotten to work at all... I intend to use the memory inside the FPGA as kind of 'fastRAM' like the Amiga but then would need normal speed transfers with new instructions into the CPC RAM for I/O. I suspect even theoretically, the CPU could be like a dual CPU design too where an internal CPU will have a turbo mode for the internal fastRAM, and while it is doing that, it can output NOPs (remember your perfect project) to the outside World until it comes back out of turboMode. Likely I wouldn't actually have 2 cores, but make the 1 core switch between normal CPC Z80 mode and internal turbo mode. I'm not even sure 'yet' how much RAM i can get out of the GODIL after the core is in place...
A simple question...
Is someone had already tried to overclock a GX4000? (40MHz to 60MHz) :-\
Quote from: TotO on 14:08, 23 September 16
A simple question...
Is someone had already tried to overclock a GX4000? (40MHz to 60MHz) :-\
Ok, this will not help much. But I think to remember that a 6128 Plus can run at 5, but not at 6 MHz. But I never tried it myself, because I don't like to solder them (Plus or GX), it's harder to exchange the crystal compared to the CPC6128.
But IMHO this experiment remains to be done. :D
Why not at 6 MHz? The ASIC crash using a 60MHz clock? ;D
Quote from: TotO on 20:53, 23 September 16
Why not at 6 MHz? The ASIC crash using a 60MHz clock? ;D
As I remember yes. But you should give it a try and test different crystals. I would do by myself actually, but only got my CPC6128 at my side - and a black cat. 8)
OK, I will see a day... ;)