| TG68 Core In Cyclone II Questions | page 1 2 3 4 5 6 7 8 9 10
|
|---|
|
|---|
Igor Majstorovic
| |
Posts 49
15 Apr 2012 10:30
| Again some questions...
So someone verify this.
On original MC68K as, lds, uds and rw are in active low state but in TG68 core is active high state so I changed that to:oTG68_ASn <= not as_s;
oTG68_UDSn <= not uds_s;
oTG68_LDSn <= not lds_s;
oTG68_RW <= not rw_s; dtack is active low on Mc68K and on Tg68 is active high. So my board overheats on alvt devices(voltage level translation) because of this and when I have done some changes regarding to this accelerator board stops overheating. Can someone confirm that those active high and low states difference could be a problem.
| |
Jakob Eriksson
Sweden
| | (Moderator)
Posts 1097
15 Apr 2012 20:04
| Don't know but I think that you are the first to actually interface the TG68 to something external. In other words, I would NOT be surprised if the TG68 got it wrong compared to a real 68000. You are doing pioneering work.
Also awesome to see that you are still working on it, good luck! :-)
| |
Igor Majstorovic
| |
Posts 49
15 Apr 2012 20:57
| Yes I work on this project for about 10 hours per day.
10 Days ago managed to get core working, and my accelerator worked, Amiga 600 detected new accelerator card but there was some problems with synchronization so I m trying to do every signal step by step now. Also have no idea what I have done to start core so need to start over :) Today managed to start core, now I m stuck with R/W and data but I have some ideas :) I ll never give up I must say, I finished every project in my life I started and the same will be with this one I just know, but it is hard... My work will be published so anyone can continue.
| |
Jakob Eriksson
Sweden
| | (Moderator)
Posts 1097
15 Apr 2012 21:31
| You are an inspiration to me. I have a similarly very hard project (but software) and when it feels too hard I think of you and Thomas.
| |
Marcel Verdaasdonk
Netherlands
| |
Posts 3974
16 Apr 2012 00:18
| What frequency are you running the core at and what speed is the system bus on that specific A600 system?
| |
Lord Aga
| |
Posts 129
16 Apr 2012 10:18
| Go Igor :)
We all have some difficult tasks in our lives, but we must never give up.
Let us keep inspiring each other !
| |
Deep Sub Micron
Germany
| | (MX-Board Owner)
Posts 566
16 Apr 2012 13:33
| I think these signals need to be highZ when BR, BG and BGack signals are used. If there is an alternate master on same bus that is driving these pins to opposite direction then it explains the hot chips.
| |
Jakob Eriksson
Sweden
| | (Moderator)
Posts 1097
16 Apr 2012 15:26
| highZ means tristate?
| |
Megol .
| |
Posts 672
16 Apr 2012 15:35
| Jakob Eriksson wrote:
|
highZ means tristate?
|
Tristate is a trademark for some company but yes it's high Z.
| |
Matt Hey
USA
| |
Posts 726
17 Jun 2012 14:34
| wawa tk wrote:
|
tg68 does already work on an Amiga since a long time and it was done by tobias himself.
EXTERNAL LINK
and here is another thread about tg68k equipped minimig on chameleon (c64 extension by jens schoenfeld)
tobias opens his sources in the first post there.
EXTERNAL LINK
unfortunately the forum is in German.
|
Good references. I see Majsta has already found these threads even if the help he received was minimal. wawa tk wrote:
|
i wonder if this would not be more beneficial to support rather tobias and/or deep sub micron directly?
|
In what way? Do you mean financial incentive? igor majstorovic wrote:
|
No they don't want to help, either way 20 days ago I have managed to get accelerator up and running so my concept is proven. I need to have few more test and then I ll publish more information's. After that Amiga open source scene will be started for first time. There is nothing I want to hide regarding this project. Sorry for oftopic :) |
You are not off topic here ;). Do you have the TG68 booting code now? P.S. I wish the geniuses out there would learn how to teach and document at a reasonable comprehension level. My best professors in college had Masters degrees while professors with a PhD often taught over most students heads and spent their time on research instead of teaching.
| |
Igor Majstorovic
| |
Posts 49
17 Jun 2012 19:00
| TG68 core will boot if you provide dtack,reset and clk signals. Problem with my design was just because when I attach accelerator board system halted because I didn't disable old cpu and there was two processors on the system and nothing was working. There are several ways to disable MC68K and more informations about that you can find on my websit EXTERNAL LINK Also core runs stable on 7.09Mhz for now and in next few days I ll start experimenting with higher speeds. Also there are much faster cores and I ll use TG68 just to see how high I can go with it. Booting codes, schematics and everything will be posted on website when more testings are done. Right now I need to order more PCBs and more cyclone FPGAs because there was lot of modifications regarding BGACK, BR and BG signals. Because those connections are tinny it is hard to work on modified boards so need to order new ones. Also big problem there was analog and digital GND but I think I managed to get that done. Anyhow you have to understand few things here. There are only few people who can help but they don't want and if they answered to me on few simple questions this all could be done in two months. But I m not angry about that I m angry on people who were giving me wrong information's and misleading me and I have spend hundreds of working hours to find out that their information's had purpose to stop my project. Either way only way someone with honest intentions with appreciation of my work could help here is by donations because it is very hard to continue regarding lack of funds dedicated for finishing this project but definitely there is huge progress with this.
| |
Wawa Tk
Germany
| |
Posts 581
17 Jun 2012 20:05
| Matt Hey wrote:
|
In what way? Do you mean financial incentive?
|
not anymore, unless they prove trusty and provide valid results first. i meant rather cooperation.
| |
Wawa Tk
Germany
| |
Posts 581
17 Jun 2012 20:09
| igor majstorovic wrote:
| | Also there are much faster cores and I ll use TG68 just to see how high I can go with it. |
care to elaborate? never heard of any faster 68k core apart of that hypothetical natami 050..
I m angry on people who were giving me wrong information's and misleading me and I have spend hundreds of working hours to find out that their information's had purpose to stop my project.
|
seriously, something like that happened?
| |
Igor Majstorovic
| |
Posts 49
17 Jun 2012 23:44
| I have some other cores in verilog that are faster but when I start working with them I ll tell you more. Also I have some papers Turkish authors regarding to Mc68K VHDL models and those papers are very interesting. Also there are some cores that are not public but I have them but didn't work with them yet. There was some talking about ao68000 who can work stable at 80Mhz.
| |
Matt Hey
USA
| |
Posts 726
18 Jun 2012 01:05
| wawa tk wrote:
|
Matt Hey wrote:
|
In what way? Do you mean financial incentive?
|
not anymore, unless they prove trusty and provide valid results first. i meant rather cooperation. |
You mean Tobias and Jens working together, communicating and cooperating? Surely we can't do much to help without a few years of learning. We better get started :P. igor majstorovic wrote:
|
I have some other cores in verilog that are faster but when I start working with them I ll tell you more. Also I have some papers Turkish authors regarding to Mc68K VHDL models and those papers are very interesting. Also there are some cores that are not public but I have them but didn't work with them yet. There was some talking about ao68000 who can work stable at 80Mhz.
|
If Jens opens up the sources for the N68050, you could use a 68020+ core running at around 100MHz in a faster Cyclone. I expect a 68000 only core would run several MHz faster but I would rather have a more complete and powerful CPU. The Minimig that is 68000 only is a pain to run newer Amiga software.
| |
Wawa Tk
Germany
| |
Posts 581
18 Jun 2012 06:56
| igor majstorovic wrote:
|
I have some other cores in verilog that are faster but when I start working with them I ll tell you more. Also I have some papers Turkish authors regarding to Mc68K VHDL models and those papers are very interesting. Also there are some cores that are not public but I have them but didn't work with them yet. There was some talking about ao68000 who can work stable at 80Mhz.
|
sounds almost like a balkan espionage plot.
| |
Wawa Tk
Germany
| |
Posts 581
18 Jun 2012 06:59
| Matt Hey wrote:
|
wawa tk wrote:
|
Matt Hey wrote:
|
In what way? Do you mean financial incentive?
|
not anymore, unless they prove trusty and provide valid results first. i meant rather cooperation. |
You mean Tobias and Jens working together, communicating and cooperating?
|
something like that would be perhaps the best thing to happen, but i guess its out of question.
Surely we can't do much to help without a few years of learning. We better get started :P.
|
alas i suspect its too late for me, it was already too late when we were working on that warp mesa stuff few years back. i will remain computing noob forever.
| |
Igor Majstorovic
| |
Posts 49
09 Dec 2012 23:09
| Ok here we go after asking million of stupid questions huge success. Named FPGA accelerator Vampire because that is the only Serbian word used in rest of the world:) I separated my project into few small ones and once all of them are finished I will join them. I think that this was the best approach.
1. Since I didn't have any equipment I realized that I could create one. So Logic Analyzer using my Vampire 600 FPGA accelerator board is created.
EXTERNAL LINK
2. Most important thing was to understand that I first need to create some king of reset routine.
EXTERNAL LINK
3. Vampire 600 starting Boot ROM and this was the first time Amiga motherboard is working with Vampire accelerator. Boot ROM done in ASM compiled and converted to HEX and uploaded to board. EXTERNAL LINK
4. Next stage was disabling old CPU with 3 wire bus arbitration and creating mini CPU who writes dff180 to Address Bus (Amiga color register)
EXTERNAL LINK
Detail explanation EXTERNAL LINK What next, you may ask, lets wait, one year ago I said that I will turn Amiga world upside down
So what next :)
| |
Matt Hey
USA
| |
Posts 726
09 Dec 2012 23:45
| @majsta
Congratulations! You have accomplished much with little and you have persevered where others have failed.
I don't mean to create any bad blood but the name of your accelerator sucks. You should call it the fpga Majic instead ;).
So what next? How about some AmigaOS fpga magic?
| |
Igor Majstorovic
| |
Posts 49
10 Dec 2012 07:49
| Thank you :) Any suggestion is welcome and there is no time for "bad blood" :) Anyway, next stage is to control all of this with mouse, and after that integrating TG68 at basic frequency. Since all of this will be open source there are million possibilities what we could build. Mine work regarding this stops at any level of acceleration and then smarter will continue once I provide them hardware.
| |
|
Welcome to the Natami lounge.