Quick Brown Box?
Pheuque
wondering if given the extreme dearth of REU options out there weather
or not the QBB could be open sourced at this point to provide some
solution for those who want ram expansion for their old Commies.
For that matter, what about the GEORAM, or CMD RAMdisks?
Cameron Kaiser
>For that matter, what about the GEORAM
The geoRAM is a very simple design and would lend itself well to being
cloned. It's just a paged RAM array, but access is fast and the design is
very extensible.
--
Cameron Kaiser * cka...@floodgap.com * posting with a Commodore 128
personal page: http://www.cameronkaiser.com/
** Computer Workshops: games, productivity software and more for C64/128! **
** http://www.armory.com/%7Espectre/cwi/ **
Kyle J Cardoza
saying:
> Pheuque <Ratt...@gmail.com> writes:
>
>>For that matter, what about the GEORAM
>
> The geoRAM is a very simple design and would lend itself well to being
> cloned. It's just a paged RAM array, but access is fast and the design is
> very extensible.
>
I'm not a hardware engineer, and I've no idea what the GeoRAM is like
inside, but I do know that RAM chips of the size that would be used
(512k-2MB?) would be dirt cheap.
It'd be cooler if someone could clone the Super 1750 Clone -- no boosted
psu needed, and REU compatible, IIRC.
--
Faith does not, in fact, move mountains; it can't even cure diarrhoea, or
clear up acne. It certainly won't protect anyone from his own stupidity.
Joseph Fenn
I for one still use 3 of the 256 K boxes and one 64K box on 2 different
128's. They work fine and I recently replaced the internal Lithium
batteries with the larger 3V lithiums by duracell.
Dont know if this answers your question or not.
Joe
Nicolas Welte
> Pheuque <Ratt...@gmail.com> writes:
>
>> For that matter, what about the GEORAM
>
> The geoRAM is a very simple design and would lend itself well to being
> cloned. It's just a paged RAM array, but access is fast and the design is
> very extensible.
True. I built a 2MB battery backed GEORAM clone, and I call it the NeoRAM.
The design is built from scratch, based on the documented GEORAM register
mapping, and I only made a small change to improve usage with other
cartridges in a cartridge port expander.
There is a thread or two in the German Forum-64 web board about it:
http://www.forum64.de/wbb2/thread.php?threadid=11770
http://www.forum64.de/wbb2/thread.php?threadid=16292
Also, I had the prototype and the first real unit on display on last
weekend's classic computing exhibit in Stuttgart, Germany:
http://www.classic-computing.de/
There's an entry and some pictures on the blog.c128.net site of the unit:
http://www.c128.net/album/cc2007/CIMG0109.JPG
I think I will offer PCBs, maybe kits, maybe kits with the RAM ICs already
installed. No completed units, I have no time for it.
Nicolas
Pheuque
> It'd be cooler if someone could clone the Super 1750 Clone -- no boosted
> psu needed, and REU compatible, IIRC.
The super 1750 "Clone" wasn't actually a clone at all. They tore the
MMU chip out of real 17XX class REU's and when C= perished, I think
they got a limited surplus of MMU chips. They put them on their own
cartridges with newer ram chips that were denser and lower in power.
The heart of real 17xx series beat with in every Super 1750 clone.
This is why they are no more. All the collective wisdom of the 8bit
hacking community combined couldn't find a suitable replacement for
the 17xx's MMU chip that could replicate it's ability to copy up to
and from C='s 64K of directly accessible memory in a single operation.
Find DMA chip that can do that, and I'm sure we'll see TRUE 1750
clones.
In fact... IIRC, with a bit of twiddling on how the 1750 address
registers worked, it would have been theoretically possible to have a
16MB of REU by just fully utilizing the REU Bank selection bytes. But
without a suitable DMA chip. it's just a pipe dream.
Pheuque
> Cameron Kaiser wrote:
> > Pheuque <Ratte...@gmail.com> writes:
>
> >> For that matter, what about the GEORAM
>
> > The geoRAM is a very simple design and would lend itself well to being
> > cloned. It's just a paged RAM array, but access is fast and the design is
> > very extensible.
>
> True. I built a 2MB battery backed GEORAM clone, and I call it the NeoRAM.
> The design is built from scratch, based on the documented GEORAM register
> mapping, and I only made a small change to improve usage with other
> cartridges in a cartridge port expander.
>
>
> Also, I had the prototype and the first real unit on display on last
> weekend's classic computing exhibit in Stuttgart, Germany:http://www.classic-computing.de/
> There's an entry and some pictures on the blog.c128.net site of the unit:http://www.c128.net/album/cc2007/CIMG0109.JPG
>
> I think I will offer PCBs, maybe kits, maybe kits with the RAM ICs already
> installed. No completed units, I have no time for it.
>
> Nicolas
That is AWESOME!
How much do you think a kit would go for?
Also, can the board take a ROM? I don't know you if you remember, but
CMD had a product called a the RAMdrive. It was a marketed as a Junior
version of the RAMlink, but was really almost the exact same thing as
your NEOram. A battery backed GEOram, from 512K up to 2 MB. The only
difference between a GEOram and the RAMdrive, was the RAMdrive also
included a CMD file system compatible RD-DOS with JiffyDOS in it's
ROM, and I think it might have had a reset, and swap button it similar
the other CMD products.
Even without all that, I'm sure you'll have a problem keeping these
kits in stock if you can keep them under $50 USD.
The only C= thing I still search for on Ebay are REU's.
Cameron Kaiser
>>It'd be cooler if someone could clone the Super 1750 Clone -- no boosted
>>psu needed, and REU compatible, IIRC.
>The super 1750 "Clone" wasn't actually a clone at all. They tore the
>MMU chip out of real 17XX class REU's
Small pedantic nitpick: the REC isn't an MMU, as it doesn't change the
memory layout of the host computer. It is best described as a fast
device for doing DMA copies. When swapping stuff back and forth the
distinction blurs, I'll grant, but it's not an MMU in the traditional sense.
Nitpick off. :)
>Find DMA chip that can do that, and I'm sure we'll see TRUE 1750
>clones.
>In fact... IIRC, with a bit of twiddling on how the 1750 address
>registers worked, it would have been theoretically possible to have a
>16MB of REU by just fully utilizing the REU Bank selection bytes. But
>without a suitable DMA chip. it's just a pipe dream.
I'm surprised no one has successfully made an FPGA version of the REC.
With the register mapping so well documented, it would not seem like a
difficult project, but I don't claim to be an expert on FPGA cloning.
Pasi Ojala
> I'm surprised no one has successfully made an FPGA version of the REC.
Someone started that in 1999, I still have the first skeleton VHDL
rec.vhdl in my home directory.
> With the register mapping so well documented, it would not seem like a
> difficult project, but I don't claim to be an expert on FPGA cloning.
It would not be difficult per se, but you need to have a suitable board
for the FPGA with the right card-edge to attach to C64 and some memory
on board, unless you use one with enough internal memory for testing
purposes (but then the FPGA may be too expensive).
What I have lately thought about having a fast enough micro-
controller or DSP to simply do the whole thing by polling the
module port.. :-)
-Pasi
--
/It's good to have them with me, the others; it's crucial, really --
and what I always thought I wanted: a grand coven of the wise,
the enduring, the ancient, and the careless young./
-- Lestat in "The Queen of the Damned"
Pheuque
You are correct!
I couldn't remember what the damn chip was called, just what it did.
LOL! (It's been a while since I actually had 17xx REU.)
Thanks for correcting me, and lighting up those last few old C=
neurons that haven't been busy in a while.
The last time I did any serious code was when I was diddling with the
REC registers in GEOS to do weird fake MMU stuff. Over a decade ago.
On Oct 1, 10:46 pm, Cameron Kaiser <ckai...@floodgap.com> wrote:
> Small pedantic nitpick: the REC isn't an MMU, as it doesn't change the
> memory layout of the host computer. It is best described as a fast
> device for doing DMA copies. When swapping stuff back and forth the
> distinction blurs, I'll grant, but it's not an MMU in the traditional sense.
>
> Nitpick off. :)
I don't think the FPGA can actually pull off DMA stuff. It was
discussed a while back, but some one doing a lot of FPGA stuff at the
time basically said you needed a different kind of programmable device
to do the DMA block moves that the old REC did.
It would also be prohibitively expensive.
I'm surprised no one has found a cheaper chip in something like a
GAMEBOY that could do the job.
> I'm surprised no one has successfully made an FPGA version of the REC.
> With the register mapping so well documented, it would not seem like a
> difficult project, but I don't claim to be an expert on FPGA cloning.
>
> --
> Cameron Kaiser * ckai...@floodgap.com * posting with a Commodore 128
Craig Taylor
> I'm surprised no one has found a cheaper chip in something like a
> GAMEBOY that could do the job.
Just out of curiousity - what's to prevent a higher speed micro
controller, say a 8mhz PIC from interpreting the REC lines, asserting
the busy signal and doing the operation itself? It might not have the
same speed but from the C64's perspective it would be almost
identicial (and probably could be fairly close to the same speed).
- Craig Taylor
http://www.ctalkobt.net (under construction)
Pheuque
No one who knows how... has bothered. LOL!
Pasi Ojala
> Just out of curiousity - what's to prevent a higher speed micro
> controller, say a 8mhz PIC from interpreting the REC lines, asserting
> the busy signal and doing the operation itself?
You still have to be synced to the 2 MHz DRAM cycles, so I doubt
8 MHz would be enough, even if PIC could do 1 instruction per cycle
(does it?).
I was thinking about a 48 MHz DSP that executes one instruction
per cycle. Unfortunately those that I have easy access to do not
have enough I/O. Something like 16(addr)+8(data)+3(r/w,/cs1,phi)
would be needed.
Nicolas Welte
> That is AWESOME!
> How much do you think a kit would go for?
about 25 - 30 euro, maybe 35 including shipping and the cartridge enclosure.
> Also, can the board take a ROM? I don't know you if you remember, but
> CMD had a product called a the RAMdrive. It was a marketed as a Junior
> version of the RAMlink, but was really almost the exact same thing as
> your NEOram. A battery backed GEOram, from 512K up to 2 MB. The only
> difference between a GEOram and the RAMdrive, was the RAMdrive also
> included a CMD file system compatible RD-DOS with JiffyDOS in it's
> ROM, and I think it might have had a reset, and swap button it similar
> the other CMD products.
The RAMDRIVE seems to be much more advanded. I feel we will never have new
products with the software and firmware quality level of CMD again ...
My device is more like the BBGRAM, but with more modern components and less
power consumption. Should be pretty compatible as well. My kit will come
without software, it should be supported by every software that can use
GEORAM and BBGRAM (maybe with minor fixes for RAM size).
> Even without all that, I'm sure you'll have a problem keeping these
> kits in stock if you can keep them under $50 USD.
I never planned to make this a good seller, I wanted to make a useful kit
that people can assemble themselves. I already failed because I had to
include SMD RAM chips that are not so easy to solder for everybody.
Nicolas
Craig Taylor
True that you would have to be synced to the dram cycles but that
shouldn't be too problematic I'm thinking - if you're not on a cycle
boundary then there's a small time delay - oops. I don't think it
needs to be time equivalent to the 1750 REC - just close.
With the busy signal asserted the whole time you're only waiting on
the RAM chips to come back with their RDY signal I believe so it's
more of a busy-wait -- you don't necessarily need to wait on the clock
signal....
(Disclaimer: It's been some time since I've reviewed the RAM accesses
inside the C64/128 and I'm no electronics guru so interpret the above
in that context and feel free to correct me.)
- Craig Taylor
http://www.ctalkobt.net/ (under construction)
Mark McDougall
> I don't think the FPGA can actually pull off DMA stuff. It was
> discussed a while back, but some one doing a lot of FPGA stuff at the
> time basically said you needed a different kind of programmable device
> to do the DMA block moves that the old REC did.
> It would also be prohibitively expensive.
Are you kidding??? You really think a programmable logic device that can be
clocked at upwards of 400MHz and can handle 2.5Gbps PCIe couldn't implement
a simple DMA engine on a bus running at a few MHz??? LOL!
I know next to nothing about the REU, but after reading this...
<http://ftp.giga.or.at/pub/c64/library/reu_programming.html>
...I can't see why it wouldn't be trivial to implement in an FPGA.
The problem with interfacing an FPGA with old 5V logic is expense - you need
1.8/2.6/3.3V rails (ie. supplies) as well as buffering on the I/O signals.
However there's a few families of the larger CPLDs that are 5V tolerant that
may even be large enough to implement the REU.
Regards,
--
| Mark McDougall | "Electrical Engineers do it
| <http://members.iinet.net.au/~msmcdoug> | with less resistance!"
Dragon To Be
Douglas
:)
Brian Ketterling
Nicolas Welte <welte...@freenet.de> wrote:
> I never planned to make this a good seller, I wanted to make a useful kit
> that people can assemble themselves. I already failed because I had to
> include SMD RAM chips that are not so easy to solder for everybody.
I don't know whether it's actually usable information -- maybe just a
trivium -- but years ago, an American electronic hobbyist magazine had a
project that included some SMDs (I think one chip and a few resistors).
These were used in the design to help introduce people to surface-mount
technology. The reader could order a kit that included the board and parts,
plus a little package of indium paste solder, which melts at a very low
temperature. The instructions were to apply a spot of the solder to each
pad with a toothpick, set the device in place in the paste, then float the
board in a pan of hot water until the solder flowed.
I didn't build that project, so I don't know if there were any difficulties
in the process.
Brian
--
Nicolas Welte
> I don't know whether it's actually usable information -- maybe just a
> trivium -- but years ago, an American electronic hobbyist magazine had a
> project that included some SMDs (I think one chip and a few resistors).
> These were used in the design to help introduce people to surface-mount
> technology. The reader could order a kit that included the board and parts,
> plus a little package of indium paste solder, which melts at a very low
> temperature. The instructions were to apply a spot of the solder to each
> pad with a toothpick, set the device in place in the paste, then float the
> board in a pan of hot water until the solder flowed.
Actually my SMD components are not *THAT* hard to solder. It's still
possible with a normal soldering iron, maybe the only thing you'll need is
0.5mm solder so you can apply a smaller amount of solder per pin.
Nicolas
Alan Jones
wrote:
>On Sep 30, 4:15 pm, Kyle J Cardoza <ad...@zetachannel.com> wrote:
>
>> It'd be cooler if someone could clone the Super 1750 Clone -- no boosted
>> psu needed, and REU compatible, IIRC.
>
>The super 1750 "Clone" wasn't actually a clone at all. They tore the
>MMU chip out of real 17XX class REU's and when C= perished, I think
>they got a limited surplus of MMU chips. They put them on their own
>cartridges with newer ram chips that were denser and lower in power.
>
>The heart of real 17xx series beat with in every Super 1750 clone.
>
>This is why they are no more. All the collective wisdom of the 8bit
>hacking community combined couldn't find a suitable replacement for
>the 17xx's MMU chip that could replicate it's ability to copy up to
>and from C='s 64K of directly accessible memory in a single operation.
How many REC chips were actually produced or sold?
Alan
Anssi Saari
> The problem with interfacing an FPGA with old 5V logic is expense -
> you need 1.8/2.6/3.3V rails (ie. supplies) as well as buffering on the
> I/O signals.
Yes, I actually ran accross a company called Entrepoint in UK. They
make component replacement FPGA boards for 32/36/40 pin DIL packages.
Those have a Xilinx FPGA, also level shifters, regulators and
oscillator are on board. Nice, but the things cost 90 USD or 70 euros
each...
> However there's a few families of the larger CPLDs that are 5V
> tolerant that may even be large enough to implement the REU.
Actel also seems to still be making their antifuse FPGAs which have 5V
operation, i.e. power and I/O. ISSI still makes 5V SRAM and DRAM
chips. In theory then one could do a 5V only REU replacement.
Mark McDougall
> Yes, I actually ran accross a company called Entrepoint in UK. They
> make component replacement FPGA boards for 32/36/40 pin DIL packages.
> Those have a Xilinx FPGA, also level shifters, regulators and
> oscillator are on board. Nice, but the things cost 90 USD or 70 euros
> each...
fpgaarcade is doing similar with a CPLD for common custom ICs.
> Actel also seems to still be making their antifuse FPGAs which have 5V
> operation, i.e. power and I/O. ISSI still makes 5V SRAM and DRAM
> chips. In theory then one could do a 5V only REU replacement.
Altera MAX3000 & MAX-II parts are compatible with 5V I/O, and use a single
supply rail for the core. I've recently designed a MAX3000 into a 'retro'
project but am looking at the MAX-II parts for future designs.
In fact, our 1541 emulation may end up in a CPLD. When we get the design
to a point where we can design the board we'll probably look at other C64
boards to prototype as well... REU would be fun...
Regards,
--
Mark McDougall, Engineer
Virtual Logic Pty Ltd, <http://www.vl.com.au>
21-25 King St, Rockdale, 2216
Ph: +612-9599-3255 Fax: +612-9599-3266
Macintosh Dragon
<47043c12$0$3618$5a62...@per-qv1-newsreader-01.iinet.net.au>,
Mark McDougall <ma...@vl.com.au> wrote:
> Anssi Saari wrote:
>
> > Yes, I actually ran accross a company called Entrepoint in UK. They
> > make component replacement FPGA boards for 32/36/40 pin DIL packages.
> > Those have a Xilinx FPGA, also level shifters, regulators and
> > oscillator are on board. Nice, but the things cost 90 USD or 70 euros
> > each...
>
> fpgaarcade is doing similar with a CPLD for common custom ICs.
>
> > Actel also seems to still be making their antifuse FPGAs which have 5V
> > operation, i.e. power and I/O. ISSI still makes 5V SRAM and DRAM
> > chips. In theory then one could do a 5V only REU replacement.
>
> Altera MAX3000 & MAX-II parts are compatible with 5V I/O, and use a single
> supply rail for the core. I've recently designed a MAX3000 into a 'retro'
> project but am looking at the MAX-II parts for future designs.
>
> In fact, our 1541 emulation may end up in a CPLD. When we get the design
> to a point where we can design the board we'll probably look at other C64
> boards to prototype as well... REU would be fun...
>
> Regards,
Too bad nobody can do it cheap like this...
http://www.atarimax.com/flashcart/documentation/
Douglas
:)
Mark McDougall
> Too bad nobody can do it cheap like this...
> http://www.atarimax.com/flashcart/documentation/
Unfortunately, it's all about economies of scale. Unless you're making
hundreds of boards, you're going to end up paying a premium for the PCB
and the parts. Given that this is a niche market within a niche hobby,
it's unlikely that someone will take the risk to produce enough boards to
make them as cheap as they could be.
Macintosh Dragon
<4704491e$0$3559$5a62...@per-qv1-newsreader-01.iinet.net.au>,
Mark McDougall <ma...@vl.com.au> wrote:
> Macintosh Dragon wrote:
>
> > Too bad nobody can do it cheap like this...
> > http://www.atarimax.com/flashcart/documentation/
>
> Unfortunately, it's all about economies of scale. Unless you're making
> hundreds of boards, you're going to end up paying a premium for the PCB
> and the parts. Given that this is a niche market within a niche hobby,
> it's unlikely that someone will take the risk to produce enough boards to
> make them as cheap as they could be.
>
> Regards,
From what I've seen on the Atari side, there seems to be a lot of
recycling of materials..cartridge cases, some circuit boards, etc.
I wonder if the same thing could be utilized with common commodore parts
to make new items cheaper?
Douglas
:)
Pasi Ojala
> shouldn't be too problematic I'm thinking - if you're not on a cycle
> boundary then there's a small time delay - oops. I don't think it
> needs to be time equivalent to the 1750 REC - just close.
Because the VIC-II performs the linear-to-multiplexed DRAM addressing,
the address should be available pretty much exactly at the right
clock edge, or is the RAS/CAS system fully combinatorical (just muxes)?
The data bus is more relaxed, but must be valid well before the end
of the DRAM cycle. The R/W signal maybe the most timing-critical.
> With the busy signal asserted the whole time you're only waiting on
> the RAM chips to come back with their RDY signal I believe so it's
> more of a busy-wait -- you don't necessarily need to wait on the clock
> signal....
Because of the RAS/CAS-access provided by the VIC-II, you need
a pretty exact sync, I think.
Craig Taylor
On Oct 7, 5:11 pm, Pasi Ojala <alb...@hopeatilhi.cs.tut.fi> wrote:
> On 2007-10-02, Craig Taylor <ctalk...@gmail.com> wrote:
>
> > True that you would have to be synced to the dram cycles but that
> > shouldn't be too problematic I'm thinking - if you're not on a cycle
> > boundary then there's a small time delay - oops. I don't think it
> > needs to be time equivalent to the 1750 REC - just close.
>
> Because the VIC-II performs the linear-to-multiplexed DRAM addressing,
> the address should be available pretty much exactly at the right
> clock edge, or is the RAS/CAS system fully combinatorical (just muxes)?
Thanks for clarifying - I had not thought about the VIC cycle
stealing. It sounds like if the screen is disabled or if you're only
using the 128's screen then you'll be fine. At least as a "poor man's
1750.
christianlott1
Nicolas Welte
> referring to the slides Welt put up, what is this?:
>
> http://www.c128.net/album/cc2007/Cimg0142a.jpg
>
These are four incarnations of the C2N232 device.
top left: original C2N232 with some hacked additions
bottom left: original C2N232 in a plastic case
top right: C2N232I in a plastic case
bottom right: C2N232I
you can read more about it here:
http://ktverkko.fi/~msmakela/8bit/c2n232/index.en.html
The site mentions some plans to produce a new batch, the C2N232I in the
pictures *is* the new batch.
Nicolas "Welt" Welte
christianlott1
> The site mentions some plans to produce a new batch, the C2N232I in the
> pictures *is* the new batch.
I read about this some months ago. Do you sell to the US?
Nicolas Welte
Of course I do. Price is EUR 25 including shipping, and you can pay with
Paypal :)
Nicolas
christianlott1
Nicolas Welte
> is this your correct email address to send paypal?:
>
> http://freenet-homepage.de/x1541/x1541/firstcontact.html
>
No, but you can mail me on that address and ask for paypal details :)
The address I post this message from does also work :D
Nicolas