-
Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.
Cloning IIsi ROMs
- Thread starter Quadraman
- Start date
The difficulty with this project is that it requires that one have a circuit board made. That's not difficult in itself, but it does require a substantial up front investment to keep the unit price down to something reasonable.
Prototype circuit boards can be bought in twos and threes for about $30 each. That's expensive and it won't work for the Mac II family ROM module. The ROM module is .047" - .050" thick, and the proto services limit one to .063" thick. If you order a run of 100 - 200 the unit price drops to $3 - $4 and .050" board is available but the total expenditure increases to $400 - $600.
One cannot simply hack existing ROM modules because the candidate modules did not use a ROM chip with a form factor which is readily available today. In other words, the flash chips, into which you would pour the IIsi (or IIci, or IIfx) ROM code do not fit on the SE/30 ROM modules.
One could also make boards oneself using something like the Toner Transfer Method of photo-resist (little up front cost), but if you panelize a ROM design to fit on a 8" X 10" panel or similar (about 20 ROMs per panel), you'll have about 1000 holes to drill by hand, which must be within a few mils of the correct position. Sure, you could build a computer controlled drill...Someone even converted an old HP pen plotter into a automatic drill. But by the time you get that working, you could have just bought professionally made boards. And cutting the panel with nice straight edges, at least along the card connector edge is non-trivial unless one spends another couple hundred dollars on a nice shear.
Even if you are willing to drill the boards by hand how are you going to connect the front and back traces through the vias? Soldering a wire jumper in 50 holes per ROM module would quickly become a pain. There are techniques for coating the holes with conductive ink and then electroplating copper into them, but the setup to electroplate copper is another several hundred dollar investment which makes commercial board production look good.
Note, if you want more than 20 boards, drilling by hand gets a little easier, because one can stencil one board for drilling, carefully stack and job multiple boards and then drill them all in the same stack. You could probably drill six to eight boards this way if you thought there was that large a need for ROM modules. But you'd still have the via coating issue.
I have a circuit board layout for the ROM module drawn (or mostly drawn, can't remember which) which would take four 32 pin PLCC flash chips, but I've never thought it was worth the expense to have it manufactured. I also have a panelized version which puts twenty (eighteen?) of them on an 8 X 10 panel. But once I really looked into what it would take to make them, I dropped that idea.
ROMs will sell at, what, about $20 - $30 probably? One would have to count on selling about 40 of them for it to be worthwhile. I'm not sure the market is even that big.
Now if someone wanted to take it on for the love of the project, that would be justifiable. Otherwise, it's rather expensive.
Or if someone could show where my analysis is mistaken or incomplete, that would be even better. It would be really nice to have a source of 32 bit clean ROM for the SE/30 which doesn't rely on cannibalizing another old machine.
Prototype circuit boards can be bought in twos and threes for about $30 each. That's expensive and it won't work for the Mac II family ROM module. The ROM module is .047" - .050" thick, and the proto services limit one to .063" thick. If you order a run of 100 - 200 the unit price drops to $3 - $4 and .050" board is available but the total expenditure increases to $400 - $600.
One cannot simply hack existing ROM modules because the candidate modules did not use a ROM chip with a form factor which is readily available today. In other words, the flash chips, into which you would pour the IIsi (or IIci, or IIfx) ROM code do not fit on the SE/30 ROM modules.
One could also make boards oneself using something like the Toner Transfer Method of photo-resist (little up front cost), but if you panelize a ROM design to fit on a 8" X 10" panel or similar (about 20 ROMs per panel), you'll have about 1000 holes to drill by hand, which must be within a few mils of the correct position. Sure, you could build a computer controlled drill...Someone even converted an old HP pen plotter into a automatic drill. But by the time you get that working, you could have just bought professionally made boards. And cutting the panel with nice straight edges, at least along the card connector edge is non-trivial unless one spends another couple hundred dollars on a nice shear.
Even if you are willing to drill the boards by hand how are you going to connect the front and back traces through the vias? Soldering a wire jumper in 50 holes per ROM module would quickly become a pain. There are techniques for coating the holes with conductive ink and then electroplating copper into them, but the setup to electroplate copper is another several hundred dollar investment which makes commercial board production look good.
Note, if you want more than 20 boards, drilling by hand gets a little easier, because one can stencil one board for drilling, carefully stack and job multiple boards and then drill them all in the same stack. You could probably drill six to eight boards this way if you thought there was that large a need for ROM modules. But you'd still have the via coating issue.
I have a circuit board layout for the ROM module drawn (or mostly drawn, can't remember which) which would take four 32 pin PLCC flash chips, but I've never thought it was worth the expense to have it manufactured. I also have a panelized version which puts twenty (eighteen?) of them on an 8 X 10 panel. But once I really looked into what it would take to make them, I dropped that idea.
ROMs will sell at, what, about $20 - $30 probably? One would have to count on selling about 40 of them for it to be worthwhile. I'm not sure the market is even that big.
Now if someone wanted to take it on for the love of the project, that would be justifiable. Otherwise, it's rather expensive.
Or if someone could show where my analysis is mistaken or incomplete, that would be even better. It would be really nice to have a source of 32 bit clean ROM for the SE/30 which doesn't rely on cannibalizing another old machine.
I found a cheaper circuit board prototyping service which makes this project feasible. Here's the link:
http://www.seeedstudio.com/depot/fusion-pcb-2-layer-10cm10cm-max-10pcs-p-396.html?cPath=64_12
Ten (up to) 4" X 4" boards for $40 which brings the price per board down to $4. They support 1.2mm thickness which is .047" which is what we need for old SIMMs.
Apparently there's some value in hanging out in the aquarium LED forums, which is where I found a reference to this prototyping service.
http://www.seeedstudio.com/depot/fusion-pcb-2-layer-10cm10cm-max-10pcs-p-396.html?cPath=64_12
Ten (up to) 4" X 4" boards for $40 which brings the price per board down to $4. They support 1.2mm thickness which is .047" which is what we need for old SIMMs.
Apparently there's some value in hanging out in the aquarium LED forums, which is where I found a reference to this prototyping service.
as a much more active member of hobby electronics than here I find trag's post a little bit too negative, 1 its not THAT bad and 2 no one said this had to be on par with dell's production either
making the designs would IMO be much more painful than making the physical product, if you do it as a project, and post it online then any yahoo with a radio shack,25 bucks, and a weekend could make their own
making the designs would IMO be much more painful than making the physical product, if you do it as a project, and post it online then any yahoo with a radio shack,25 bucks, and a weekend could make their own
Yep. Affordable PCB prototypes makes a *lot* of projects more feasible. When it was $30 per boad with minimum $120 it was much less practicle than when it's $4 per board with minimum of $40.Interesting. Seems like that would make a number of things more feasible, eh trag?
Notice that the size is limitied to 4" X 4", but that's big enough to make any of the SIMMs or DIMMs we might want. It might be a bit short for a PDS or NuBus card. I'm not sure, I'd need to look at the connector dimensions again. But, for complicated PDS or NuBus cards one would like want to go four layers instead of two. Still, I think it would be workable to do an SE/30 PDS adapter card (to Daystar upgrade) in two layers, maybe. If the traces are short.
Similar threads
