A3640 Rev 3.3 PCBs

[fordav1]

While this is a great idea in theory, what we really need in 2017 is a dedicated 060 board using modern parts that are easily available. Using those other silly 74F1803 parts is really an incredibly bad idea and is going backwards. You should do a proper assessment of what is needed and make a better design. You could implement the silly logic with either some other modern logic or modern CPLD (which is just a big programmable logic chip anyway). You could actually replace almost all of the logic with a single fast CPLD. The PAL code is available and is easily converted to equations and recompiled for whatever chip you want to use using easily available software. You should get someone on-board that has hardware design experience and do it right with independent clock so it can be overclocked or with 2 clocks (CPU 200MHz OSC and the stock 50MHz OSC) so the CPU can be overclocked to 100MHz using the last rev 060 like what was done on the GBA1000. Everyone knows the A3640 was not a really good design (it was a cheap and nasty typical Commodore rush-job using now obsolete technology) so re-making it identically is not the best thing to do in 2017. Also a re-design using modern parts will allow some needed RAM to be added and maybe other features. But certainly adding RAM will fix the biggest let-down with the A3640. A single 128M or 256M RAM chip is only a dollar or two. There's even enough room in an A3000/A4000 case to make the board slightly bigger if necessary. But making an identical copy is IMO not the best use of your time and not really what people want now. I'm sure a lot more people would jump on board (myself included) if you had announced an A3660/256M
So I ask that you please reconsider where you are going with this and improve it rather than just re-make a bad design.

[plasmab]

Originally Posted by fordav1

While this is a great idea in theory, what we really need in 2017 is a dedicated 060 board using modern parts that are easily available. Using those other silly 74F1803 parts is really an incredibly bad idea and is going backwards. You should do a proper assessment of what is needed and make a better design. You could implement the silly logic with either some other modern logic or modern CPLD (which is just a big programmable logic chip anyway). You could actually replace almost all of the logic with a single fast CPLD. The PAL code is available and is easily converted to equations and recompiled for whatever chip you want to use using easily available software. You should get someone on-board that has hardware design experience and do it right with independent clock so it can be overclocked or with 2 clocks (CPU 200MHz OSC and the stock 50MHz OSC) so the CPU can be overclocked to 100MHz using the last rev 060 like what was done on the GBA1000. Everyone knows the A3640 was not a really good design (it was a cheap and nasty typical Commodore rush-job using now obsolete technology) so re-making it identically is not the best thing to do in 2017. Also a re-design using modern parts will allow some needed RAM to be added and maybe other features. But certainly adding RAM will fix the biggest let-down with the A3640. A single 128M or 256M RAM chip is only a dollar or two. There's even enough room in an A3000/A4000 case to make the board slightly bigger if necessary. But making an identical copy is IMO not the best use of your time and not really what people want now. I'm sure a lot more people would jump on board (myself included) if you had announced an A3660/256M

So I ask that you please reconsider where you are going with this and improve it rather than just re-make a bad design.

If you don't like what people are making then make it yourself. Its postings like this that make makers throw in the towel and turn their backs on amiga.


Sent from my iPhone using Tapatalk

[Mr.Flibble]

Yeah, coming in and pissing on someone's hardware project because it goes in a different direction from what you personally want is *really* bad form.

If you want something else then gain the skills and build it, don't come into a thread and basically demand something completely different.

Someone might well design a board similar to what you want, but this thread isn't that.

[Chucky]

if I COULD do stuff like that.. I would. I am by no way a hardwaredesigner.. (that's why GND/VCC planes most likley break ALL rules there is!) but I did this as so many A3640 boards die and most reasons is that the caps leak electrolyte. my first step was actually to make a way to follow traces easy. after this was done I was thinking. HELL WHY NOT DO PCBs?? so we can REPLACE bad boards with new.. that's what I did..

I am now doing a REV 3.3.2 with SLIGHT changes (removing 2 jumpers and possability to choose delayline) after that. it will be a FINAL.. I will then learn how to use eaglem, try to move it there and then just say: hey guys: GO CRAZY..

I guess when my vacation soon is over (2more weeks..) I will soon need to go on another week-businesstrip to gothenburg. slow days at a hotel can do magic...

So.. why I did the 3640.. simply: all data was available (IF there was for ANY other card. I would done that as the 3640 is actually EXTREMLY crappy) and I have this hate/love relationship to this board.


and as plasmab said: not happy? do one yourself..

[Chucky]

and btw "not what people want now" and I look at at my notes that ALL my 30 PCBs went in like 2 DAYS.. and I ordreed more.. and they will also go when I think of my mailbox in amibay and email.. so...

[SpeedGeek]

Originally Posted by plasmab

If you don't like what people are making then make it yourself. Its postings like this that make makers throw in the towel and turn their backs on amiga.


Sent from my iPhone using Tapatalk

+1!

First of all, excellent work Chucky!

Secondly, I originally posted my A3640 speed up mod project here on EAB and moved it over to Amibay exactly for this reason...

Some of the comments were like "It's not worth it because the A3640 don't have any fast RAM", "You can buy a WarpEngine for 100 quid more", etc.

Now, we have the 040-060 adapter and the 64MB fast RAM hack for the A4000...

Where would we be now if such negative comments discouraged the development of these projects?

[matthey]

I believe fordav1 was trying to be constructive and *not* negative. Perhaps there is a bit of frustration for lack of 040/060 big box accelerators thrown in. He has some reasonable suggestions even if upgrades are not always so easy. He also does not understand the value of available and tested replacement boards. The first step is usually the most important in a slow process. Someone else can use the design or maybe even Chucky could figure out how to better support a 68060 on the board with a little help from SpeedGeek and/or Cosmos.

[fordav1]

cool down guys. I knew this would happen. you have ALL jumped the gun and mis-read my post. very disappointing to see people here are not open-minded about suggestions for new hardware development. I'm NOT pissing on someones work. If he wants to re-make a bad design then he is free to do so. I'm only suggesting that he should improve it rather than re-make the same thing, but if he wants to make the same thing go ahead! I'm suggesting that the Amiga community would like to see something better than a 30 year old A3640 design simply re-made as-is. They all went (the 30 PCBs) because people will buy anything for Amigas, virtually everything remade sells no matter what because we need new boards to replace 30 year old boards. That's a given. I'm not the only one suggesting improvements, look at previous posts, there are plenty of people asking for improvements like 060. Sheez!
What we should be doing is thinking to the future. Yes, use the re-made same design as a stepping stone towards a newer improved design, but please don't make 1000's of them and flood the market with another badly designed board. This is exactly why Jens from Individual Computers bought the Apollo accelerator design then killed it. We need the ACA3660-100/256MB that isn't being made by Individual Computers ;-)

[Chucky]

and I also told WHY I did not implement the 060: first batch I wanted to see if the damn thing even WORKED. changing too much and you do not know if changes screwed up or reverseengineer missed something. (like my first version missing some powerconnections)

and as I also told many times: I repair amigas, but do not develope hardware atall, simply I cannot do that. and the community DOES want to see 3640 remade as they want to replace their bad boards in their drawers (I have 13...)

we would LOVE new designs. but 3640 is better than nothing.. right. and I did this for ME actually.. but as I hate closed stuff I give my work to the rest. and the rest have 2 things they can do: 1. Use it 2. "Keep going, nothing to see here"

and interest is there, as told.. I sold my 30 boards in 2 days as people want spares.. (and some just "to have") and I have in 2 days built 3 boards.. now I only have 10 boards left in my drawers...

[Chucky]

and about Jens and killing the apollo.. well.. then why didn't he release any 040/060 yet.. we need it NOW. but nothing happens. that's again why I did this: SOMETHING happens..

Apollo might be crap. but if that data was out.. we would still have 060s to use.. now we have to rely on old boards with no spares and no real good repairinfo.

[fordav1]

Re: seeing if it worked, that's what prototypes are for. You make a very small quantity and fix it to get it working then improve it, make more prototypes, test, improve etc etc. in a loop until you have it working better than the original.
Re: Apollo, I'm talking about the A1200 accelerator, which was known to have serious design issues. Jens probably doesn't do an 060 for big box Amiga's likely due to development costs and the market. It'd probably cost more than an ACA1233-40 and he would have to sell 10000 to make any money and get back his investment. He's already minus several hundred thousand dollars so that's why it won't happen.
You could still continue to re-make the original A3640, but also look towards making a version with improvements. As I said, you should get some people on board that can help with hardware design rather than trying to do it all yourself with an old design.
As for repairs, the A3640 schems are available and so are the PAL images so repairing them shouldn't be a problem. I repair Amiga's too (just spent 4 days on a SMD leaky cap damaged A1200 which is now working fine) and I repair plenty of other things too like arcade PCBs etc. But I know my limitations and I don't get into hardware design without getting help from people that know more.
There are already some parts that are difficult to find for the A3640 so making the same thing but not making a newer design is only frustrating for people that just want to keep their Amigas going but can't because the silly parts aren't available.

[Chucky]

I know.. (and the 4000 version) but it still WORKS.. and if data was out, someone might have corrected the issues.

I think it is soo sad that both Apollo and Blizzard/Cyberstorm data IS there. but kept secret.. instead of doing a kickstarter so we could have the info so we do not have to use slow 030 cpus or use bad 040 cards (or .. old ones that dies one after eachother)

so still again: 3640 is better than nothing.. as so far we have seen nothing (except the TF540 that I hope plasmab someday can finish) so simply: I did not want to wait and needed spares..

[Chucky]

Aaand.. I was prepared of this.. as YES the 3640 is sucky sucky

[plasmab]

Originally Posted by fordav1

Re: seeing if it worked, that's what prototypes are for. You make a very small quantity and fix it to get it working then improve it, make more prototypes, test, improve etc etc. in a loop until you have it working better than the original.
Re: Apollo, I'm talking about the A1200 accelerator, which was known to have serious design issues. Jens probably doesn't do an 060 for big box Amiga's likely due to development costs and the market. It'd probably cost more than an ACA1233-40 and he would have to sell 10000 to make any money and get back his investment. He's already minus several hundred thousand dollars so that's why it won't happen.
You could still continue to re-make the original A3640, but also look towards making a version with improvements. As I said, you should get some people on board that can help with hardware design rather than trying to do it all yourself with an old design.
As for repairs, the A3640 schems are available and so are the PAL images so repairing them shouldn't be a problem. I repair Amiga's too (just spent 4 days on a SMD leaky cap damaged A1200 which is now working fine) and I repair plenty of other things too like arcade PCBs etc. But I know my limitations and I don't get into hardware design without getting help from people that know more.
There are already some parts that are difficult to find for the A3640 so making the same thing but not making a newer design is only frustrating for people that just want to keep their Amigas going but can't because the silly parts aren't available.

Who are you to say what we can and can't release? Make it yourself or remain silent. Please. You are betraying a very low intelligence rating with your posts


Sent from my iPhone using Tapatalk

[fordav1]

yes 3640 is not great. the changes required to drop in an 060 are pretty minimal, just a few traces changed, add regulator etc, not a lot. The built-in 060 design could be added to the 3640 design pretty quickly. Then the weirdo logic can be replaced with more modern parts and the PALs reduced to a CPLD. All in steps and tested before releasing boards in quantities.

[fordav1]

Originally Posted by plasmab

Who are you to say what we can and can't release? Make it yourself or remain silent. Please. You are betraying a very low intelligence rating with your posts

well if you knew "who" I really was you wouldn't say that ;-)

[plasmab]

[QUOTE=fordav1;1176437]

Originally Posted by plasmab

Who are you to say what we can and can't release? Make it yourself or remain silent. Please. You are betraying a very low intelligence rating with your posts



well if you knew "who" I really was you wouldn't say that ;-)

I would. Unless your are David Haynie please refrain from trashing this. Even then you'd have no right.

You have no idea the 1000s of hours put into these projects to have people like you trash them. Do it yourself if you are so good. Then post when you are done

[fordav1]

Your response is amusing. I'm not "trashing" anything. I already said if the OP wants to make the same thing go ahead!
I have *exactly* the idea what it takes to do projects like this. I've spent probably $50k and 20 years of my life as part of a team doing things for a specific section of the retro-community (all open source and publicly available) and it's still going now (probably around 60000 hours so far). I prefer to remain anonymous outside of that environment so my previous activities don't influence responses. Like I said, you would not say that if you knew who I was and what I've done. And no I won't elaborate more. If you read all of my previous posts across this forum, amibay and other forums you should be able to work it out.

[plasmab]

Haha. Talk is cheap. Open source a board. Then we'll talk.

Beating up Chucky for letting people reproduce the A3640 is pretty pathetic. If you have half the skills you claim to then you make the "easy fix" go on.


Sent from my iPhone using Tapatalk

[fordav1]

again, an amusing response. I don't care if you believe me or not. That's not the issue here, and you know it. I'm not "beating up" anyone. You should probably re-read the above posts, you're clearly not fully informed.