MacToTheFuture SE/30 10/100 Ethernet card

[jamesmilne]

The first boards have arrived, and my board wizard friend Mark, who designed it, has assembled the first prototype!

We’ll likely simplify the board further. No need for a socketed EPROM if we can program the firmware via the computer, for example, allowing us to use a smaller SMD EPROM.

He is in London, Ontario and I’m in London, England, so once he has finished his initial tests he’ll send it over to me and I can get started on the driver.

 

[Von]

Very cool. Is the jumper so that you can assign the slot used like the Asante models?

Any plans for a PDS pass-through on top?

 

[360alaska]

Nice,  please let me know when your selling one.

 

[Sunoo]

This is getting exciting. Once you are selling these, I’ll definitely need to grab one.

 

[ants]

I'm in awe! Amazing work - totally count me in for a purchase.

 

[omidimo]

Amazing! Can't wait to see the production run.

This card paired with a PDS pass-thru, and cache slot would be the Ultimate SE/30 card. Just dreaming a little. 

 

[techknight]

Awesome!

For version 2.0, May I suggest removing the flat flex cable? that is a disaster waiting to happen. Rather have a molex style wiring harness. 

 

[jamesmilne]

Yes the jumper is to change the slot address.

We haven’t planned for a passthrough at the moment. It may make the board routing harder. We had to byteswap the data bus for the LAN9218 chip so we’d probably have to add another layer to forward the signals to a passthru slot.

The boards hopefully won’t be too expensive, but we’ll have to think about what we do for producing more. We’ve both got pretty busy jobs and kids to attend to  

But the goal was always to produce them in some kind of quantity.

 

[jamesmilne]

Yeah this is just our first prototype. The flat flex may go, it’ll have to twist a bit when the jack board is mounted to the rear of the chassis. Its quite narrow though, so maybe not too bad.

 

[techknight]

My worry is snagging and ripping the cable. I do it all the time with flat flex which is why I hate it. 

 

[joethezombie]

I second the call for PDS pass-through.   SE/30 freaks tend to have multiple PDS cards shoe-horned inside using right angle hacks or permanent wrong angle Asante modification.  Would hate for the fun to stop at the first upgrade :undecided:  

 

[techknight]

Out of curiosity, are you going to open-source the design? 

 

[K55]

Out of curiosity, are you going to open-source the design? 

Yeah. Otherwise I have the feeling this will produce a few units then end up as dead as the protocahce

 

[Trash80toHP_Mini]

Did Bolle stop shipping the fruits of the ProtoCache1 program? It was alive and well last I heard. :huh:

edit: forgot to say HOORAY! Very, very cool, guys.

 

[joethezombie]

The complete schematics and equations are in the ProtoCache thread.  It's fully open sourced.  Of course, questions for that topic should stay in its respective thread.

 

[Bolle]

Yeah. Otherwise I have the feeling this will produce a few units then end up as dead as the protocahce

I did not see the part where I had to build hundreds of adapters for everyone when taking the information from the thread and putting it down to a board layout.

I went ahead and spent 800€ on parts and PCBs, countless hours assembling and testing every single unit and shipping them out to all over the world and in the end made next to nothing from the units I sold here to forum members. Not saying it wasn't fun and everything but everyone can go ahead, take the schematics from the thread and a big roll of cash to get another batch done.

/rant

Back on topic:

Good job on the ethernet card. If the driver development works out I can't wait to take my hands on one of the cards.

In case you decide to make the schematics public I could imagine the ethernet part would fit very well onto a board with PDS passthrough and cache slot.

By looking at the parts on the ethernet card I would say I could easily get the routing to fit into 4 layers with passthrough and cache slot in the form factor of the bigger Artmix style PowerCache adapter.

 

[jamesmilne]

I'm not against open-sourcing the PCB designs, but I'll talk to my buddy Mark who designed the boards about that.

I will definitely open source the driver code.

However, no point in open-sourcing anything until we get it to work.

We've managed to squeeze the current design into a 2 layer board, which makes the PCBs super-cheap, but going to 4 layers wouldn't be prohibitive. The boards are very small.

I don't want the product to wither & die either, but as Bolle points out making real stuff can be an expensive & time consuming business. Mark and I do this stuff for a living so we're well aware of what's involved. Maybe once we've got it working we can talk to someone else who has an interest in building them commercially.

 

[Bolle]

If it’s two layer only it should be really easy to implement it into a passthrough/cache adapter design.

When funding a production run does not work out at reasonable cost having the schematics available will have someone picking it up sooner or later anyways.

On a side note I personally would not publish the whole board design files. Makes it just too easy for certain folks to knock off a product. If you take the time to go through the schematics, route your own board and try your luck with that... this way you had to do some work at least.

Anyways, before talking business... looking forward to see this thing working. Got any details on the hardware setup by the way?

DeclROM in the big EEPROM, ethernet controller wired up right to the 030 bus and another smaller serial ROM for controller specific things?

Did you follow Apples guidelines on how buffering all used signals should be done?

 

[jamesmilne]

Components:

- LAN9218 controller

- DeclROM in big flash EPROM

- MAC address in little EPROM

- CPLD doing some address decoding for DeclROM and LAN9218

Once we verify that we can program the EPROM via the PDS slot, we'll likely replace the DIP EPROM with a TSOP version, and probably remove the MAC address EPROM and put the MAC address in the main EPROM, then load it into the LAN9218 in the driver.

The board follows the specs in "Designing Cards and Driver for the Macintosh Family". I'll see if I can twist Mark's arm to make it passthrough.