Falcon Afterburner - Please Help!

Hardware, coding, music, graphic and various applications

Moderators: Mug UK, moondog/.tSCc., [ProToS], lp, Moderator Team

Guest

Re: Falcon Afterburner - Please Help!

Postby Guest » Sat Jan 19, 2013 5:51 pm

one question dogs me

i see people with dsp problems sometimes
running at higher dsp clocks will need faster refresh sram

so perhaps as faster sram is available ive seen it on ebay.fr

then it should be changed out
wont do any harm to do this anyway...

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Sun Jan 20, 2013 8:36 am

Thanks for the tips on the caps. Lots of soldering to do.
Have had the DSP at 50MHz for years, never had trouble. It has a small heat sink, then all I use it for is JPEG decoding and whatever Apex might use it for. This is on an Afterburner, not CT60.

User avatar
dml
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3472
Joined: Sat Jun 30, 2012 9:33 am

Re: Falcon Afterburner - Please Help!

Postby dml » Sun Jan 20, 2013 9:20 am

Rustynutt wrote:Thanks for the input. We are both talking about the same things.


I wasn't sure I followed the whole explanation so I thought i'd just brain-dump some bits and bobs to see if it tied up :) But it makes more sense that 32MHz is sourced through the expansion port (I thought it only fed Combel and Videl in).

Rustynutt wrote:I want to change the pin arrangement so the 040 on the installed Apple card is getting the same PCLK as it would on the Afterburner without the Apple card.


Yes that sounds like it should work. Interesting to see what happens.

Rustynutt wrote:Skew is negligible on those chips, so will run the AB PCLK to a 1X output on the Apple divider.


I think that's the other difference between 040 and 060 - the 060 is fussy about skew, or phase between B/PCLK where the 040 is less so. Apart from that they should be pin-for-pin and lower voltage and that's it.

Rustynutt wrote:Basically, a test to make sure the Apple "Quad Doubler 25MHz/50MHz" card will work with the AB 040 as an adapter of sorts. Even though it will be the same speed as the Afterburner without the apple card, this would tell me if there is an issue with the resistor arrangement for power and clock signals with the Apple card placed in the Afterburner CPU socket. My gut feeling is, there should not be an issue as this is also how it fits into the Quadras. All guess work lacking the proper test equipment, but that's how Spock arrived at many solutions :).


Will be nice if that works. BTW are there any Mac accelerator boards which adapt 030->040? Just wondering if that could be a source of 'afterburner' type upgrades with a suitable adaptor and software....

Referring to the chips as clock dividers as I'm to lazy to get up and find the proper nomenclature :)

Rustynutt wrote:There is a feature the divider on Apple card includes, not available on the divider on the AB, which holds the CPU clocks in reset for a specified duration to allow the clock signals to obtain phase lock on chip before releasing the signal to the processor. Don't know how critical this is in real world applications with the 040, or maybe the PAL on the Afterburner takes care of it . That being the case, there could be noticeable skew between the BCLK on the AB and signal going out to the 040 after RST and PLL is achieved. If the PAL is doing this as well, might be a problem to detect without a good scope. If not, replacing the divider on the AB with one spec'ed for the 060 timing, and bypassing the divider on the Apple card should take care of that. Another concern is how the COMBEL handles clock division internally, as I would like to put the 060 divider right off the oscillator from the Nemesis, and feed it to the COMBEL and bypass it's 32 MHz clock signal to the expansion pin, DMA and everything else running at 32MHz on the board. The other COMBEL clocks are slow enough that I don't think that would be a problem. For now, will stick with them on the accelerator cards first to verify it works.


I think the man to ask here might be Rodolphe Czuba.

For the CPU/FPU it will likely be ok to have the clock out of phase with the rest of the board since the whole thing is async but performance might be a bit off with severe skew.

For some other chips e.g. DMA the sync probably needs to be maintained on the ouput side of the Combel or things go wrong (?).

Rustynutt wrote:The other Apple card, a DiiMOCache 050, uses the same clock divider as the Afterburner. It also has a bunch logic on the board, assume it is for the additional Cache. Think another mistake I made was referring to that as L2 Cache, isn't that the CPU cache, or is it the term for memory available outside the CPU? Anyway, will play with it some after testing with the other card.


I forget if L2 is the 1st or 2nd cache outside the CPU, but it's definitely outside the CPU - implemented on the main or daughter board.

L1 is either a closer version of that (still external) or it refers to the on-CPU/internal cache. I can't remember now.

In any case having an external cache on an Atari accelerator I think would be a new thing? I don't know of any existing ones that do it. They usually have FastRAM but not external cache. Possibly they don't need it with low wait states to FastRAM? I don't know the answer. Mr Czuba will know for sure.

Rustynutt wrote:What got me started on this 8 years ago was after I purchased the Interconnect 060 adapter, couldn't figure out why it didn't work on the Afterburner. I'd read somewhere that in the first few lines of TOS there is some incompatible code. I took off that direction by having a coworker burn TOS onto a chip I mounted on a card to fit the cartridge port. Forget whom it was at the time, an Atarian from either Poland or Czechoslovakia (Peter S ?) provided a TOS image which was suppose to have the one or two lines of code modified to work. Never got anywhere with that effort, honestly way above my head and understanding. Think this was just after the time Titan kind of dropped out of the picture and began what I recall was website design or something.


I think the only way to get it to work is to modify and burn a new ROM yes. However it's more than the first few lines of TOS needing changed IIRC. You need to do a few things to get rid of 68030-isms and probably the blitter, but all are achievable with patches. It has to be done offline though for the CPU to get through the boot sequence.

I don't know how the AB does it - possibly has some mask rom on the board which allows it to crawl over the nasty bits in ROM, in order to get as far as the AUTO folder and let you do the rest properly in software. I don't know the details here but without something like that the 040 would probably just throw exceptions and hang.


Rustynutt wrote:More recently, when reading about the Milan's requirement for an adapter card that not only provides 3.3V and ties a couple pins together on the 060, there was an additional requirement to delay the CPU from starting during power up and reset functions. This was a function the Interconnect socket does not provide. With the Milan adapter no longer available, tried to implement that circuit on the Interconnect socket without success.


I think without knowing the TOS ROM is correct to start with, it's difficult to be sure of correct feedback on your other experiments. It's very likely somebody already did this patching and has made a file available - or possibly try with an alternative TOS image (EmuTOS? Milan ROM?) which is known to be 040/060 compatible... could improve chances of success.

User avatar
dml
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3472
Joined: Sat Jun 30, 2012 9:33 am

Re: Falcon Afterburner - Please Help!

Postby dml » Sun Jan 20, 2013 9:46 am

I had another idea. If I can get the AB to boot using a modified version of the TK driver, which only applies pure, in-situ patches to the ROM - just enough to get it as far as the AUTO folder or Desktop (but maybe not stable for general use) then you'd have something you know should work and it should be electronics from there onwards?

I haven't looked at the TK patches in a while so I don't know how many hooks to external routines are involved or how far the AB can get without a properly configured PMMU. But it might be worth a look.

[actually this is probably more relevant for 060 - the 040 should behave the same providing it plugs into the AB, since the AB can definitely get past the standard TOS rom init somehow - but the CT60 does something else entirely, the TOS image gets flashed onto it IIRC and I don't know what it does with it after that stage - I don't know if the ROM chip is referenced at all during CT60 bootup]

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Tue Jan 22, 2013 12:12 am

Some years back asked Rodolphe about some of these things, he pretty much felt it a waste of time as the CT63 was available..... To his credit, it does take a lot of effort, and he is busy with other projects.

"I had another idea. If I can get the AB to boot using a modified version of the TK driver, which only applies pure, in-situ patches to the ROM - just enough to get it as far as the AUTO folder or Desktop (but maybe not stable for general use) then you'd have something you know should work and it should be electronics from there onwards?"

Kind of the same lines I've been thinking, once the Fuji appears :)
May be incorrect here, thinking booting the Falcon from the diagnostic cart bypasses TOS. Recall being able to boot with the cart when troubling shooting other modifications. Not sure if this bypassed the cpu entirely, and just performed test on other functions, but it does start. I don't think I tried it with the 060, never the less, have to get the clocks correct first. My AB is in a Wizztronics case, with a NOVA, Expose and all sorts of other stuff packed in. It's going to have to be disassembled onto a bench before I get a chance. Kind of given up on the Milan until resolving this on the AB first. From what I understand, the Milan needs nearly all the IC's replaced to run even at 20MHz :(.

From what I recall, all the 040 upgrades for 030 Macs had the processor on the nubus, kind of like the Amiga.

Hope to at least get the clock pins sorted at work next week, I need a microscope as my eyes can't resolve good any longer! Took keep from posting such long messages (I'm good at "Brain Dumping"), will post once I try some hardware mods.

Thinking too, there is some credence to what simbo says about the caps, how the power comes up on the board, and possibly deteriorating IC's on the board can cause issues too. Of course, will jump past all that at first and try the quick and easy thing. Then look at the power issues later, if I don't give up at all before the SuperVidel arrives :)

Thanks again!

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Tue Jan 22, 2013 7:26 am

Short tid bit. Had a thought to look at the application the Quadra Overdrive was intended for. One of the models it can be installed in is the Centris 610. I was under the assumption that Apple ran the 040 at 2x bus speed, but in those models, both the 040 and main board ran at 20Mhz. The Quadra Overdrive card is a drop in unit, which, as well as replacing the 20MHz 040, added the FPU version of the 040 MHz CPU. Not sure why I didn't read up on this before, what drew my attention was the installation manual stated it worked on both 20 and 25MHz Apples. Also, before arousing my curiosity well before it should have been to even slightly understand all this some time ago, tried it on my Quadra 840AV, where it failed. That model does in fact run the 040 at 40Mhz, thus the Overdrive card was supplying a 2x clock to the CPU there as well.

This confirms what I had thought, with the Afterburner sending a PCLK of of 2X Falcon bus clock, the Apple card was running that clock at 2x, putting a 4x clock to the 040. With the Nemesis, the 040 was getting a clock of 88MHz.

I'll modify the output on the Apple card from the 2x clock to a 1x output. This will give the 040 on the AB a 44MHz clock.
Had to think back a few years of why I was trying to utilize this in the first place, and it was to offset the processor location so as to allow the Expose card and Nova adapter to sit without having to use extremely long raisers, with the 060 socket adapter installed.

I would think that from what I have read about CT63 users getting for clock speeds, there should be no problem feeding it a 44MHz clock.

Fired up my long dormant AB Falcon tonight, looks a go for testing as soon as I can modify the adapter.

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Wed Jan 23, 2013 1:17 am

While reading up on the processor card, stumbled across a link where a NeXT user used the same card to basically do the same modification to his system. This is the same card I have, what is interesting is what he states about the PAL. I don't know enough about the Falcon to know which side of the fence she sits here.

Kudos to the work, looks like he ended up testing with the card, then just built his own "Nemesis" for the NeXT.

http://www.nextcomputers.org/forums/vie ... sc&start=0

Guest

Re: Falcon Afterburner - Please Help!

Postby Guest » Wed Jan 23, 2013 12:23 pm

atarifreaks on ebay is selling a nem for £40 or so,,, has the buffer board also

its not a buy it now he also has a number of other bits there for sale at decent prices most as buy it nows
usual dogs breakfast of 'wires' 'pathetic in RF terms ' adding some mini coax is a must

if anyone needs silver mini coax just pm me ill send you a few meters
Last edited by Guest on Wed Jan 23, 2013 1:17 pm, edited 3 times in total.

Guest

Re: Falcon Afterburner - Please Help!

Postby Guest » Wed Jan 23, 2013 12:31 pm

http://www.ebay.co.uk/itm/C-LAB-Atari-F ... 2327465da9

found the link to ebay nem for sale

£50 inc post {so far no bids }

so get it quick... 6 days left on the auction

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Mon Jan 28, 2013 5:46 am

Been re-reading this thread, and doing my homework. Condensing notes down tonight, found a note sheet from 6 years ago, had the same data on it as what I came up with today. Funny how the mind works. Same mistakes over and over :).

Came across something that caught my attention reading the data sheet for the 88916. On the first page it states the outputs do not satisfy the requirements for the 20 and 25MHz 68040's, and refer to the data sheet for the 88920 Clock Driver for more information. One Afterburner I have came with a 33MHz full version 040, and another purchased from a chap in the UK came with the LC version, 33MHz. My Milan came with a who knows what 040, the heat sink being welded on, and replaced it with a 40MHz 040 shortly after discovering some issues with that machine. Tonight, I decided to sacrifice the original Milan 040 to find out it's designation, but even after chiseling the heat sink off, the adhesive was so thick I couldn't make out the rating. Anyway, might be something for those with earlier versions of both 040 units to at least check up on.

Have tried to understand how the PAL on the Quad Doubler handles the BG BR BB signals. Think somehow it holds off the PCLK during bus access, or a bus master other than the cpu is accessing the bus by tri-stating (is that the proper term?) one of the pins. Ringing out the 040 pin array one to one does not show any discontinuity, so it's either that, or the state is active high (when powered up, haven't looked at it that way yet), or it just causes a high impedance on the pin.

Either way, think cutting power to the PAL will be sufficient to test it again on the Afterburner.
Received another Falcon in the mail Friday, this one with a 14mb ST RAM card, and a GE Soft Eagle Sonic 68030. The main board is clean, never looked to be hacked, spreading it across the work bench within 10 minutes of unpacking :(. With little or no circuitry other than what is needed to fit the PGA 030 onto the Falcon expansion pins, it look promising to stick a fast clock on to "she what she can do".

Four paragraphs, not including this one. I'm getting better....

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Mon Jan 28, 2013 7:09 am

Afterburner with Apple Quad Doubler fitted. Clears ST RAM 6 to 10mm.
You do not have the required permissions to view the files attached to this post.

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Mon Jan 28, 2013 7:26 am

You do not have the required permissions to view the files attached to this post.

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Mon Jan 28, 2013 7:29 am

Afterburner w/DiiMO Cache 50MHz 040 Apple card. Would need a small raiser to clear expansion pins. Card is missing the PLL clock driver and 50MHz can (for now).
You do not have the required permissions to view the files attached to this post.

Guest

Re: Falcon Afterburner - Please Help!

Postby Guest » Tue Jan 29, 2013 1:37 am

i have a ceramic 040 i dont need

ill dig it out
it came from a mac pro pcb

perhaps it can help someone

hey its free +post

i think for your clock diorama you need to include a clock phase change using a hex inverter and a dil switch
this will allow you to invert the phase of each clock 'on bus assert' 'on process clock assert' etc
i would assume the bus asserts antiphase from the process
or you suffer time dilation effects {bi phase skew}
if you double the clock the assert changes with respect to top process level

so if you use a pll you not only suffer more jitter and issues with the chip and its bypass etc {look to improve the caps }
but latent process to bus
unless you phase correct it

you should use two bipass caps on the pll to avoid issues

you always try with bipass to sink the voltage node and antnode to the center of the caps pair range ,,,

sometimes 2 is better than 1 bypass caps

i would use a 4.7nf and a 47nf gives the best of both worlds
if the originals is like 10 forget playing with clocks and its preset ranges till you add more range to the bypass band blocking
etc

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Tue Jan 29, 2013 8:28 am

[quote="simbo2"]i have a ceramic 040 i dont need

ill dig it out
it came from a mac pro pcb

IF it is the 3,3V version, it would be handy to test the 060 adapter socket.

perhaps it can help someone

hey its free +post

i think for your clock diorama you need to include a clock phase change using a hex inverter and a dil switch
this will allow you to invert the phase of each clock 'on bus assert' 'on process clock assert' etc
i would assume the bus asserts antiphase from the process
or you suffer time dilation effects {bi phase skew}
if you double the clock the assert changes with respect to top process level

The clock driver allows for this and keeps skew to nil, unlike an inverter. However, the only part of the system that would require any shifting of the clock is the B clock on the 060, which is handled by the 060 adapter.

so if you use a pll you not only suffer more jitter and issues with the chip and its bypass etc {look to improve the caps }
but latent process to bus
unless you phase correct it

you should use two bipass caps on the pll to avoid issues

you always try with bipass to sink the voltage node and antnode to the center of the caps pair range ,,,

sometimes 2 is better than 1 bypass caps

i would use a 4.7nf and a 47nf gives the best of both worlds
if the originals is like 10 forget playing with clocks and its preset ranges till you add more range to the bypass band blocking
etc

Power for the clock drivers already exist, I will be removing the slower drivers and replace with higher spected units to the same pad.

Guest

Re: Falcon Afterburner - Please Help!

Postby Guest » Tue Jan 29, 2013 10:37 pm

change out the bypass cap 10nf or 100nf for 4.7 and 47nf this is all you need to do once you change it out
looking at the datasheet the bypass bandstop works better with two caps and not 1 nominal value

youll need to add one also under the ic
remove the smd ceramic and add as above under the ic

and replace the yellow block electrolytic with a bigger value or two

two caps are better than one here

Rustynutt
Captain Atari
Captain Atari
Posts: 380
Joined: Wed Mar 21, 2012 7:38 am
Location: Oregon
Contact:

Re: Falcon Afterburner - Please Help!

Postby Rustynutt » Wed Jan 30, 2013 8:50 am

After a rather extensive time with the meter, have come to the conculations the PALs on the Afterburner card do not supply PCLK or BCLK to the CPU. There is a 16MHz and 32Mhz feed from the clock divider to two sepereate PALs on the AB, and might be used to determine bus arbration or somthing. Anyway, the clock divider is infact traced to the CPU. This was a releif, as Doug had said, if the PAL were controlling the clock, we could be lost.

Another note, the 15JC PALs on the boards are rated :

tWH Width HIGH 6 6 8 ns
D-type 66.7 50 40 MHz
T-type 62.5 47.6 38.5 MHz
fMAX D-type 83.3 66.6 50 MHz
T-type 76.9 62.5 47.6 MHz
83.3 83.3 62.5 MHz

Where replacement with MACH210A-7
tWH HIGH 3 ns
D-Type 100 MHz
T-Type 91 MHz
fMAX D-Type 133 MHz
T-Type 125 MHz

As long as GE did not lock the PALs, seems simple to increase the speed at what the AB runs at.

More learning before soldering :)
Even reviewed the CT60 schematics for a few hours.

rpineau
Atari Super Hero
Atari Super Hero
Posts: 501
Joined: Wed Jun 29, 2011 6:39 am
Location: California / USA
Contact:

Re: Falcon Afterburner - Please Help!

Postby rpineau » Tue May 28, 2013 3:51 am

After reading this whole post I need to point out a few things :
First of all, thanks all for all the investigation.. I too have an AB040 and this thread is the most useful one I ever read about the AB040.
The expansion port DOESN ' T have 32MHz. the CPUCLK signal is 16MHz as explain in http://dev-docs.atariforge.org/files/Fa ... 1-1992.pdf
So the clock must be doubled on the AB040 to get PCLK .
I ran mine at 40MHz for a long time but kept having SDMA issues (corrupted data on floppy write, confirmed when I used an HxC Floppy Emulator, I don't use SCSI). I might retry to fix this at some point but for now I put a 33MHz oscillator and everything is stable.
I'm a lot more interested in fitting a 68060 to the AB040 (I have one in a drawer that I never used). Has anyone ever got the schematics for the Milan 68060 processor adapter ?
Also, has anybody mapped the AB040 expansion slot (the long one close to the 3 MAX for the DRAM) ? I'll be interested in putting something on it with the full 040 bus accessible (as opposed to the Flacon 16 bits bus), I might try to build a card with a lot of SRAM as for some reason my AB040 never recognized anything more than 1x 16MB SIM no matter what position I put the jumper on JP1 and JP2).

As a side note, there is no more 68030 on my mother board (it got removed by Rodolphe Czuba a long long time ago when I was still living in France).

In any case I'll be happy to help in any way (I have the latest know source code for the AB040 toolkit if someone is also interested in these).

Regards, Rodolphe/
Falcon + AB040 + Eclipse PCI + ATI Rage VGA card + NE2000 Ethernec + HxC Floppy Emulator
MegaSTE 4MB + CosmosEx / 1040 STF for hardware dev
http://www.rti-zone.org/atari.php

ldv-01
Atari maniac
Atari maniac
Posts: 94
Joined: Wed Jun 01, 2011 10:14 pm
Location: Italy

Re: Falcon Afterburner - Please Help!

Postby ldv-01 » Tue May 28, 2013 8:20 am

rpineau wrote:Has anyone ever got the schematics for the Milan 68060 processor adapter ?

Hope this is what you were looking for:

Circuit diagram of the 060 adapter : http://www.uweschneider.de/Download.php?lang=en (Milan section)

General informations on the Milan Computer: http://www.der-ingo.de/en/milanhelp/index.html

rpineau
Atari Super Hero
Atari Super Hero
Posts: 501
Joined: Wed Jun 29, 2011 6:39 am
Location: California / USA
Contact:

Re: Falcon Afterburner - Please Help!

Postby rpineau » Tue May 28, 2013 2:32 pm

Thanks.
That's exactly what I was looking for. I'll see if I can put this in Eagle and see how doable this is.
Regards, Rodolphe
Falcon + AB040 + Eclipse PCI + ATI Rage VGA card + NE2000 Ethernec + HxC Floppy Emulator
MegaSTE 4MB + CosmosEx / 1040 STF for hardware dev
http://www.rti-zone.org/atari.php

rpineau
Atari Super Hero
Atari Super Hero
Posts: 501
Joined: Wed Jun 29, 2011 6:39 am
Location: California / USA
Contact:

Re: Falcon Afterburner - Please Help!

Postby rpineau » Wed May 29, 2013 4:23 am

I took a look at the difference between the 68040 pinout and the 68060. The extra signals are on the inside and the adapter generate these for the 68060 using the 74F157 (non inverting multiplexer) and some of the 68040 signals. Some other signals are simply tied up to VDD via a resistor network (3.3v) like the SNOOP* pin (which is why you can't use the bus snoop function of the 68060 in this case).
The rest of the wiring is pretty much straight through. The issue is to figure out how to sandwich the 2 board and not have VCC (5v) connect to the VDD (3.3V) pins as well as not connection the innermost row of pin from the 68060. I think it would be easier for me to do a flat adapter in Eagle (I already have the 68040 and 68060 in there) with the added logic.
So I'll see what I can do sometimes this week or next weekend.
Regards, Rodolphe
Falcon + AB040 + Eclipse PCI + ATI Rage VGA card + NE2000 Ethernec + HxC Floppy Emulator
MegaSTE 4MB + CosmosEx / 1040 STF for hardware dev
http://www.rti-zone.org/atari.php

User avatar
dml
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3472
Joined: Sat Jun 30, 2012 9:33 am

Re: Falcon Afterburner - Please Help!

Postby dml » Wed May 29, 2013 7:55 pm

rpineau wrote:So I'll see what I can do sometimes this week or next weekend.
Regards, Rodolphe


Please keep us up to date on any progress you make :) Will be interesting to see if a 68060 can function on the AB board.

User avatar
dml
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3472
Joined: Sat Jun 30, 2012 9:33 am

Re: Falcon Afterburner - Please Help!

Postby dml » Wed May 29, 2013 8:01 pm

rpineau wrote:In any case I'll be happy to help in any way (I have the latest know source code for the AB040 toolkit if someone is also interested in these).
Regards, Rodolphe/


Is that based on the original sources or a disassembly of the binary? I don't remember if the sources were public but it seems very likely. Too long for me to remember... in any case I do have the original source here if needed. They can almost certainly be modded for a 68060, and it could be relatively easy if the MMU ops are very close to the 040.

rpineau
Atari Super Hero
Atari Super Hero
Posts: 501
Joined: Wed Jun 29, 2011 6:39 am
Location: California / USA
Contact:

Re: Falcon Afterburner - Please Help!

Postby rpineau » Wed May 29, 2013 8:53 pm

The source I have are the original source sent to me (I don't remember who send them to me, from memory it was the original author but that was a few years ago).
So the source are properly commented. The version I have is 5.08 with the Eclipse fix.

AS to putting a 68060 on the AB040 we will probably need to find a way to get the Motorola 68060_lib code in there as the 68060 is missing some instructions that are software emulated (the source are available from Motorola web site and I have them if they are no longer there).

Regards, Rodolphe
Falcon + AB040 + Eclipse PCI + ATI Rage VGA card + NE2000 Ethernec + HxC Floppy Emulator
MegaSTE 4MB + CosmosEx / 1040 STF for hardware dev
http://www.rti-zone.org/atari.php

User avatar
dml
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 3472
Joined: Sat Jun 30, 2012 9:33 am

Re: Falcon Afterburner - Please Help!

Postby dml » Wed May 29, 2013 9:03 pm

rpineau wrote:The source I have are the original source sent to me (I don't remember who send them to me, from memory it was the original author but that was a few years ago).
So the source are properly commented. The version I have is 5.08 with the Eclipse fix.


I am the original author - but I believe Johan Klockars applied a fix for Eclipse afterwards and he probably provided his version. Johan had the original sources from me at some earlier point.

I don't actually have the Eclipse fix myself as I haven't contacted Johan about it yet, but if you have the source for that I could update my own version from it :-)

rpineau wrote:AS to putting a 68060 on the AB040 we will probably need to find a way to get the Motorola 68060_lib code in there as the 68060 is missing some instructions that are software emulated (the source are available from Motorola web site and I have them if they are no longer there).


I haven't fully investigated this for 68060 but some of the same problems existed on 68040, at least for the FPU. I think the 060 is missing some integer operations (64bit mul/div I think), so yes - it will be necessary. (It might be possible to boot the CPU without it though - they are less often used).


Social Media

     

Return to “Professionals”

Who is online

Users browsing this forum: No registered users and 3 guests