OK... why no new 68K boards?

[jiffydos]

So, even with PPC boards going for so much money, the plane old blizzard 68K boards are going for almost as much.  Why hasn't anyone created a new 68K board, even just exact copies of the old ones?

[alexh]

Price and availability of MC68060RC50 chips and edge connectors.

Freescale do not make the MC68060RC50 chips anymore and try to encourage people to use a Coldfire in a new design. One of the ways they have encouraged people is to raise the RRP to $373.29 well beyond the price of anything other than Military apps.

The tooling costs alone for the edge connectors in RoHS plastic would be several thousand dollars before you have even made one and then you'd probably have an MOQ (Minimum order Quantity) of about 1000 units.

You'd have to make at least 500 units to break even. When was the last time you saw 500 MC68060RC50 chips for sale?

Add to that the demand. With lots of people already owning an 060 board the total number of people wanting one is much smaller than say for 24-bit Scan doublers or USB controllers.

[DoogUK]

Plus the fact that no-one will sell the rights/code needed to clone existing accelerators...you would have to start from scratch designing a new type..costing a pretty packet in this day and age and as Alex says..there would not be enough sales to warrant it.

[Crom00]

I was in a similar boat lamenting over the lack of CPU cards. Then I discovered UAE and never looked back. I was what you would consider a power user with a couple of Video Toaster Flyer A4000 T's.

Using commodity PCs' With multi-ghz CPU's and gobs of ram, you get a super Amiga with a spending spree $400 Wal Mart PC.

That's the price they're asking for modestly expanded A4000 or A3000 with graphics cards and other amenities on Ebay in the USA. Pretty Crazy... I paid $650 for an A4000 030... in 1995!

Happy with my A1200 and 030/50 as it's used for games and tinkering.
If I want to do modern things like play MP3's etc. I'll use UAE, a Mac or a PC.

A true fully 68k compatible 2ghz Amiga would be great though.

[alexh]

If UAE wasnt so flakey then I'd do that but IMO it is still a bit poop, especially in JIT modes.

Designing the 680x0 to SDRAM logic would not be that hard, especially as Rodolphe has done it recently for the CT060 for Atari Falcon. There are some timing diagrams and some bus wave form captures to look at. Probably get it done in a few days.

Considering how much hard work that Oliver Hannaford-Day put in a few years ago, coupled with the investigation work being done over at EAB I would imagine you could get a lot of information about the technical pitfalls to making an accelerator before you start.

[Belial6]

As far as I can tell, new 68k boards are being produced, and they can run faster than the A500.  You have MiniMig and C-One.

[Jose]

060's can be got second hand for a relatively cheap price. At least for now. I even remember some old Mac 68k cards had a recicled 68k on them.

[alexh]

Belial6 wrote:
As far as I can tell, new 68k boards are being produced, and they can run faster than the A500. You have MiniMig and C-One.

Neither are that fast compared to a bog standard 1230.

Neither support the 020+ instruction set and so are of limited use with "modern" Amiga software designed for accelerators.

Neither have an FPU (not that there is much Amiga software that uses one).

A C-One plus MiniMig FPGA expansion cost a lot more than most second hand Amiga 060 accelerators.

Jose wrote:
060's can be got second hand for a relatively cheap price. At least for now. I even remember some old Mac 68k cards had a recicled 68k on them.

In ones and twos? Yes.

In 100's? No!

[cv643d]

If a single developer can create Minimig I am sure someone could make a new 68k board. Dont listen to all the negative comments.

[Belial6]

You can argue that they are not 'good enough' accelerators, but that doesn't change the fact that they are A500s that can be boosted above the standard speed.  You have to start somewhere.  No doubt that as improvements happen, and new FPGAs come out, speed will increase, and someone will add 020 instructions, and the fact that they cost more doesn't change the fact that they are being produced.

So, maybe the original question should be refined to... Why no 68k boards that are faster and cheaper than used ones?  Personally, if I had the know how to make a new accelerator board today, I wouldn't bother.  Hardware wise, Dennis and the other MiniMig developers have freed us from AInc. and 20 year old hardware.  That is the direction I expect to see push us past current Amiga speeds.

Now, if we could just get AmigaOS/Workbench/Kickstart replaced with an opens source version, the reigns would be off and we could run.

[arnljot]

In the book "On the Edge: the Spectacular Rise and Fall of Commodore" by Brian Bagnall, he writes that Atari didn't want Jay Miner to design a computer around the MC68000 processor since it cost $150 per unit, the management couldn't envision that it would become cheaper.

That's why they founded HiToro to build a computer for tomorrow. As Jay Miner was sure that both the MC68000 and RAM would become cheaper.

For the retro scene, I'm sure this will be the case too. FPGAs will become cheaper. So the MiniMig that Dennis designed will see it's FGPA drop in price. So too if someone made a 020 based FPGA accellerator for the A500.

Now, if someone sufficiently skilled has the time and motivation to pull it of... now that's a completely different matter...

[Darrin]

Belial6 wrote:
You can argue that they are not 'good enough' accelerators, but that doesn't change the fact that they are A500s that can be boosted above the standard speed.  You have to start somewhere.  No doubt that as improvements happen, and new FPGAs come out, speed will increase, and someone will add 020 instructions, and the fact that they cost more doesn't change the fact that they are being produced.

So, maybe the original question should be refined to... Why no 68k boards that are faster and cheaper than used ones?  Personally, if I had the know how to make a new accelerator board today, I wouldn't bother.  Hardware wise, Dennis and the other MiniMig developers have freed us from AInc. and 20 year old hardware.  That is the direction I expect to see push us past current Amiga speeds.

Exactly.  The Minimig and the C-One are new hardware, even if they do cost more than a dirty, 2nd hand A1200.

While we don't have a 68020 or better, the Minimig's 28MHz 68000 adds a bit of oomph when playing Gunship or Wing Commander.  The C-One is also the prototype for the Clone-A so another offering is in the wings (when it's ready).

Now, if we could just get AmigaOS/Workbench/Kickstart replaced with an opens source version, the reigns would be off and we could run.

Freeing Kickstart from Amiga Inc, or at least having them sell Kickstart images on their online shop would be nice.  Copies of Workbench can still be bought "new" or copied from old Amigas for those who need them.

The Minimig v1.1 is now in a state to be manufactured in bulk for Christmas as a retrogaming beauty.  Why Amiga Inc don't put together a package:

Minimig v1.1 with latest PIC flash & PSU
Screwed to the bottom of a PS2 PC Keyboard
PS2 Mouse
SD Card pre-loaded with core and Kickstart 1.3
a selection of pre-loaded ADF games (and more to buy and download from their shop)

and

Jakub's ARM board with:
SD Card pre-loaded with with core, Kickstart 3.1 and Workbench 3.1 (plus a configured WHDload)

Well, that's what I would do if I had a serious interest in running a company and making some money.

[DoogUK]

Some of us like the fact we can still use 20 year old hardware..and the mere fact they can be accelerated way beyond their original abilities is also one of the attractions, 20 years ago no one would have dreamed they would be capable of running PPC speeds.

If they developed those then more modern accelerators could be too...and i would just love to point out to people Hey...my 20 year old miggy is doing ok in the modern world of computing.

This is one of the many reasons people still ask why and when if possible will someone develop new accelerators...but we all know now that will never happen.

[alexh]

cv643d wrote:
If a single developer can create Minimig I am sure someone could make a new 68k board.

MiniMig was created for fun. Not for profit and originally not for others. Dennis spent a lot of his own money which he never got back.

When MiniMig went on sale it was unique, new. It had a much wider appeal than a new 68k board, potentially the entire Amiga retro community. It was a better investment.

Belial6 wrote:
You can argue that they are not 'good enough' accelerators

Don't need to argue. Its fact. A v1.1 MiniMig, even with it's 28MHz(?) 68000 cannot compare on price, performance or compatibility with an even a 14MHz A1200. Form factor and novelty are MiniMig's two saving graces.

Belial6 wrote:
doesn't change the fact that they are A500s that can be boosted above the standard speed.

They are not quite A500's (yet).

Belial6 wrote:
You have to start somewhere.  No doubt that as improvements happen

Yup. Yacqube has already shown it and as long as people keep an interest I imagine developments will progress.

Belial6 wrote:
and new FPGAs come out, speed will increase, and someone will add 020 instructions, and the fact that they cost more doesn't change the fact that they are being produced.

You're not going to make anything close to a 100MHz MC68060 with an FPGA for a LONG time. But yes I think that adding the 020+ instructions and stack format will be possible now. The cache... that is another issue as AFAIK no-one really knows how it was implemented. Trial and error no doubt. UAE does without it so I guess it's not THAT important. I bet Tobias Guebner is already working on it.

[AeroMan]

Quote:
Belial6 wrote:
and new FPGAs come out, speed will increase, and someone will add 020 instructions, and the fact that they cost more doesn't change the fact that they are being produced.

You're not going to make anything close to a 100MHz MC68060 with an FPGA for a LONG time.

I hate to be pessimistic, but bigger FPGAs will probably be all in BGA package, and soldering these beauties is not something easy to do in our backyard
Besides this, they also recquire a multi layer PCB (8 to 10 layers, maybe more) which are expensive and hard to do

In the good hand, I´ve heard some good news from some friends about Freescale.(hehehe) Let´s wait for the future :-D