Maybe not so silly, what are the cost/performance tradeoffs? How close do the IBM or available 50 pin drives come to exceeding the bandwidth of the 50 pin SCSI 1 spec controller on the mobo?

You could reserve the fast/wide bus id's for peripherals that really need the bandwidth, or is there some other overriding i/o concern that I'm missing?

As for cutting the case and the power supply change: why not lobotomize, disembowel, and emasculate a PC, sentencing the empty husk and power supply to forever cow-tow at the feet of a perfectly restored IIfx as a SCSI-Tower attendant?

8-}

Besides, it would cut down on clutter and collect less dust than a bunch of external cases...

Might resolve possible airflow/cooling issues too...

jt

It would be very cool, but it might be more effort a/o cost effective to "think outside the box" rather than trying to expand it.

Can you break the internal cable out the backplane as a workaround for the filter/termination issues? Is the external side of the bus terminated on the mobo or are you missing parts that preclude the use of the onboard controller entirely? If it's functional, the other side has to be filtered already and your terminated "internal drive" could be in the tower at the end of the chain. I'm fairly sure that one ribbon cable/connector on the backplane of each box with only one cable in between shouldn't give you capacitance or termination problems. You might be able to daisy-chain one or maybe even two external devices reliably in between. Pretesting for problems could be done with a temporary cable/connector lashup of stuff you've probably got onhand.

Don't move the mobo or hack the case if you can avoid it. IMHO, fx's were/are much cooler than anything you could hack one into. A SCSI tower you could hook up to any other CPU in house or take out on location.

Sorry about the long post, I got carried away again.

jt

btw: if the case is beat, disregard all of the above and HACK AWAY!

I was suggesting that you ignore the external side and run a modified internal cable (like an external drive's "internal" cable) from the mobo connector to the backpane.

Alternately, find a IIseries case/PS/dead mobo, hack that and stack 'em, or slice off and toss top & bottom panels, reassembling the remainder of lid under remainder of box to make an M-F extension tube w/extra supply and two additional floppy ports to hack.

Now you've got LOTS of room and the original case remains untouched.

Do a few of them and you'll have a Borg Cube with room for an auxiliary power generator 8-)

Sorry, I've just got to get this out of my system...

jt

btw: adorable kid, enjoy!

tip: my little guy thought macpaint was a computer game until elementary school.

Ahh, now that would require the IIfx specific SCSI filter, which I mentioned before and also don't have. The IIfx was a hot machine in it's day but it sure has one unique and quirky SCSI bus!

Drat! That's why I had asked if the internal side of the bus was functional. You could still do the hack on a sacrificial case for JackHammer bus devices and revive the internal bus at a later date.
If you have any IIsi era 'Users or 'World (blech) back issues, look for ads showing the IIsi "extension tube", I don't think that vapor(hard)ware product ever condensed, but it was a great concept.

Exactly which parts are required to make either or both sides of the internal bus functional (together or seperately)? Does anybody have the electronic spec on the parts? Terminators are duck soup to fab with trinkets from C***Shack. Some filters can be easy too (inexpensive).

jt

Could possibly be STANDARD SCSI 1, I didn't think Apple ever implemented that spec correctly.

Do you need both a filter and a special terminator on the external side? Would you need a special terminator for the mobo to make the external side functional?

Unless the '040 card you're considering talks directly to the JackHammer over Nubus like the Rocket, the I/O bottleneck at the mobo level makes most bandwidth concerns irrelevant. I always wanted the Fast SCSI 2 Rocket Daughtercard to get around handshaking at the NuBus level, then that sucker would have kicked even a 950's....

When it comes to ROMS, Zips, DAT's, Syquests or anything I can think of but HDDs/arrays, I 'm not convinced that SCSI 1 bandwidth is really a problem for you and keeping all those devices off HDD or array's bus would probably be a plus.

this is too nuch fun.... back to work again....

jt

No, just the special terminator.

quote:

---

Unless the '040 card you're considering talks directly to the JackHammer over Nubus like the Rocket, the I/O bottleneck at the mobo level makes most bandwidth concerns irrelevant.

---

Thanks for the link, I've got that book!

(leaps to his feet and pulls volume off shelf}

(AHA!)

GTTMFH p.395

"The SCSI DMA IC uses master mode to perform DMA transfers. In master mode, the SCSI DMA IC uses the normal bus-arbitration procedures of the MC68030 to transfer data to and from main memory without assistance from the MC68030."

Except fo interleaved memory, that's the same as the Rocket/Daughtercard setup, too bad it's SCSI 1 and '030 on the fx, but way cool for pre Quadra. I wonder how it compared to Sun's boxes? Can the JackHammer do DMA via the NuBus?

Try to determine the build date of your fx, P.237 of "Troubleshooting Your Macintosh" says: "The internal SCSI filter was provided with IIfx systems shipped prior to March 1990". Maybe you don't need one afterall. I'll bet the second run of DMA IC's or mobo's had a bugfix, maybe check for rev & datecodes on the ASIC's & mobo?

It also says "The SCSI terminator II (also called the glitch eater) is a black terminator pack..." if you haven't googled those terms yet.

I've got the two "Cat-Mac" (?) books and the "Hackintosh" clippings in a box somewhere, maybe one of the troops has them handy?

As for going w/68 pin ROM, make sure it's not going to knock a HDD or Disk Array's bus down a speed mode before you spend any money. Then again, is that why there is such a thing as a 68 pin ROM?

jt

That's what I'm assuming, that if nothing else it will otherwise coexist unobtrusively with an Ultra-SCSI array/environment. The Plextor 20x 68-pin is the only 68-pin CD-ROM I've ever seen and it was OEM'd by IBM for some of their Ultra-SCSI based servers some years back. If it degraded a RAID's thoroughput they probably wouldn't have used it. Which reminds me, I would like to have a two-disk RAID inside the IIfx.

I think two 1/3 height, 68-pin SCSI drives would fit on top of eachother where the second floppy would go. The only problem is, again, power consumption. The IIfx only has one power connector, so I'm going to need a new PSU to use a HDD and the CD-ROM anyway.

This agains bring up the question of whether I should transplant a new PSU (such as one from a Quadra 900/950) into it, or transplant it into a nice ATX case? The IIfx's case isn't exaclty pristine, anyway. I had planned on sanding it down and painting it, if nothing else.

quote:

---

btw: adorable kid, enjoy!

---

thanks for the kinds words on my daughter! She loves my duos especially, the miniature track ball is just the right size for her fingers.

666th poster to the
68K Macintosh Liberation Army Forums
Mod of the Mac II series Forums
Total 68K Macs liberated: 5
Visit my site!

Edited by - cinemafia on 11 Jan 2002 10:41:38

I kinda thought that's why they made 'em that way too, that's a lot of expense to add to a drive spec. for bandwidth it doesn't need.

Power cable "Y" adapters are cheap, modern disks & ROM's are stingier w/watts than fx era half & full heights, run the numbers, that $1.98 Apple thing may work. Don't you just love apple's psu's? How many Dock supplies have you had to fix?

quote:

---

I had planned on sanding it down and painting it

---

Run the numbers, if it's close: it's fused, fry it, they go ALL the time, the other way it's landfill anyway. Drive cables piped in would be ok too, especially w/phony high-voltage stickers for show!

jt

p.s. checked the site, the extension tube might fit in over there.