Daystar Universal PowerCache Clone?

[tt]

It has been mentioned before, sort of sparked by JDW's find, but I thought I would add this topic to kick-off discussion of possibly reverse engineering the board or other similar boards.

Here is a lower-resolution version of the original x-ray image I have. It was automatically composited by the x-ray equipment, and is not the best stitching, but it is pretty clear overall.



Here is a link to high-res version: http://ge.tt/87Pfkt8?c If someone can upload it somewhere and repost a link here for future reference, it would be appreciated. The original file size is 7MB and too big to attach here.

 

[Trash80toHP_Mini]

Very thanks for posting that! :approve:

 

[Unknown_K]

I would think cloning that board would still cost more then buying a used one on ebay correct? Cloning works when demand (and there has to be quite a bit of a demand) is enough to recoup costs of manufacturing plus profit.

 

[IamSpartacus]

That's pretty cool to study. You can even see the internals of the chips. One thing's for sure, it would take an enormous amount of time to diagram.

 

[tt]

I would think cloning that board would still cost more then buying a used one on ebay correct?

I was thinking more along the lines of redesigning it to be more like the CPU plug-in type that was made for the SE/30. I could potentially see a copy of that board being made with donor chips from a Universal PowerCache. A lot of work, yes, but maybe there are better ways of doing it, like doing a slight redesign with modern components, etc.

 

[trag]

I would think cloning that board would still cost more then buying a used one on ebay correct? Cloning works when demand (and there has to be quite a bit of a demand) is enough to recoup costs of manufacturing plus profit.

The 68030 chip alone costs far more than a PowerCache usually goes for. Repackaging existing PowerCache's is certainly possible, but does not increase the supply of upgrades available.

But, if you're going to repackage a PowerCache anyway, why not just turn it into a PDS card which will plug directly into the SE/30 socket and support one or two additional cards with the connectors positioned in the most convenient locations?

 

[Bunsen]

But, if you're going to repackage a PowerCache anyway, why not just turn it into a PDS card which will plug directly into the SE/30 socket and support one or two additional cards with the connectors positioned in the most convenient locations?

One would need to either re-engineer the PowerCache to suit the '030 PDS *, or clone the Daystar SE/30-to-IIci PDS adapter ** and(/or) incorporate that into the clone ... yeah?

* and what, in theory at least, would be involved there? ***

** or, as noted elsewhere, the much simpler IIcx CPU socket adapter.

*** cause that would rule.

 

[Trash80toHP_Mini]

I never thought this was about doing an AncientTech clone, I thought it was regarding an electrically/instruction set compatible clone using current technology. Sort of like a ColorMac version of BMOW's hack . . . on a card!

That way it would be a lot less expensive and MUUUUUUCH FASTER!!!!!!!!!!! }

 

[Bunsen]

an electrically/instruction set compatible clone using current technology.

Yeah, and we've been down that lonesome road before. Short version: Coldfire and/or FPGA, and a couple of pro code/board monkeys for a couple of years.

a ColorMac version of BMOW's hack

A what ... of what?

... oh, bigmessofwires' hack. Well, sort of, only not.

 

[Trash80toHP_Mini]

If not. Why not? :?:

 

[ojfd]

Cloning PCB could be done, but what are you going to do about chips that are now in unobtanium? And who's going to program all those GAL's, even if you manage to buy some old stock?

Cheers,

ojfd

 

[trag]

Cloning PCB could be done, but what are you going to do about chips that are now in unobtanium? And who's going to program all those GAL's, even if you manage to buy some old stock?

Lattice is still selling the GAL chips. And they have a development system available on their site for free with which to create the logic/program files.

However, you are correct in spirit. The tough part would be figuring out exactly what logic equations are implemented by the GALs.

However, once that is known, making duplicates of the chips does not present any substantial challenges. There are at least a couple of members on the list with programmers which can write to GALs.

But, again, 68030 chips are expensive if bought new and hard to find used.