Print

[bmwcyclist]

Even so, I would try removing the terminators, or disable the BlueSCSI terminators so there is only one set in use.

You can examine the expansion card ROM from service mode.

Terminators removed.

Slot 1 Test 1, BlueSCSI plugged in but no external power

007F BBFF 007F.... etc.

Slot 1 Test 2, BlueSCSI plugged in with power on.

007F EF7F 007F.... ect.

[sigma7]

Slot 1 Test 1, BlueSCSI plugged in but no external power

When the Lisa CPU is reading an expansion card's ROM, in most (and perhaps all) cases, external devices attached to the expansion card are not involved (so it shouldn't matter if there is a BlueSCSI attached, or a parallel drive/printer attached to a dual parallel card, etc.)

007F BBFF 007F.... etc.
...
007F EF7F 007F.... ect.

... that indicates a problem... the SCSI card ROM shouldn't be reading as 7F FF 7F, nor as 7F 7F 7F.

And perhaps it indicates two problems, since a blank ROM would be FF FF FF and if everything else is ok, the ROM data should be repeatable.

The problem with the high bit may be due to a poor contact at the edge connector... sliding the card up and down slightly after installation can help "wipe" the contacts and improve the connection.

If the high bit is still unreliable, try the other slots to see if it is slot dependent.

If you have another 2716 EPROM (from eg. a Dual Parallel card) with known contents, I suggest you install that in the SCSI card to see if service mode can read that ROM properly.

[bmwcyclist]

... that indicates a problem... the SCSI card ROM shouldn't be reading as 7F FF 7F, nor as 7F 7F 7F.

And perhaps it indicates two problems, since a blank ROM would be FF FF FF and if everything else is ok, the ROM data should be repeatable.

The problem with the high bit may be due to a poor contact at the edge connector... sliding the card up and down slightly after installation can help "wipe" the contacts and improve the connection.

If the high bit is still unreliable, try the other slots to see if it is slot dependent.

If you have another 2716 EPROM (from eg. a Dual Parallel card) with known contents, I suggest you install that in the SCSI card to see if service mode can read that ROM properly.

Ok. so cleaned all of the slots with DeoxIT and the card. All look bright and clean.

Insterted card BROKE plastic clip and I was not strong-arming it. Does anyone have a 3D print file for these?

Ran test 3 times, results in attached photos.

Still not happy about booting with the card installed.

[sigma7]

Ran test 3 times, results in attached photos.

Try the other slots to see if there is any difference (maybe you already did, but I think the pics show slot 2 only). The problem may be the card or it may be the slot.

edit: earlier on you had tried slot 1, so try slot 3 if you haven't yet. And/or if you have another expansion card (eg. Dual Parallel) that would be useful to check the slots are in working order.

Trying a known good EPROM (from some other device) in the SCSI card will also be helpful to determine if it is the card or the EPROM that is the issue.

After all these years, the yellow cams on the ZIF sockets do break pretty easily. If you are extremely careful to slide and hold the card all the way in before closing the ZIF you can minimize the chance of breaking a cam.

As mentioned later-on in the thread "Sun Remarketing SCSI card: A cautionary tale", I've been experimenting with 3D printed replacement cams. I had posted there about trying GF Nylon SLS -- I've concluded it wasn't the right material to use, so ordered and have received the Aluminum SLS variety, which I think will be the winner, more to come.

[bmwcyclist]

I have tried all of the slots, no good.

I do not have a ROM chip or other Lisa cards.

Any sources to recommend?

[sigma7]

Any sources to recommend?

Since you mentioned friends with EPROM programmers, I suggest you ask one of them to read the ROM from the SCSI card.

The first ten bytes should be 00 1B 00 00 00 00 00 00 00 00, to set the card ID to 1B, and the counter/pointers to 0. I think the rest can remain blank (FF's).

Depending on what they find, you could ask VintageMicros to resolve the problem, or just reprogram the part, or program a new one if your friend has some on hand - it is a 2716.

[bmwcyclist]

The first ten bytes should be 00 1B 00 00 00 00 00 00 00 00, to set the card ID to 1B, and the counter/pointers to 0. I think the rest can remain blank (FF's).

Depending on what they find, you could ask VintageMicros to resolve the problem, or just reprogram the part, or program a new one if your friend has some on hand - it is a 2716.

Do I understand correctly that the only data on the ROM is 00 1B 00 00 00 00 00 00 00 00 and the rest is blank??

Therefore, I don't need any special data to make a new ROM?

[AlexTheCat123]

Yep, that's all you need. There's no actual code; the 1B is there to identify the card to MacWorks as a SCSI card, and MacWorks handles the rest. I've attached a file that you can burn straight into a 2732 if that makes things easier.

[bmwcyclist]

Yep, that's all you need. There's no actual code; the 1B is there to identify the card to MacWorks as a SCSI card, and MacWorks handles the rest. I've attached a file that you can burn straight into a 2732 if that makes things easier.

Thank You!

[stepleton]

Yep, that's all you need. There's no actual code; the 1B is there to identify the card to MacWorks as a SCSI card, and MacWorks handles the rest. I've attached a file that you can burn straight into a 2732 if that makes things easier.

Component golf challenge: you're MacGyver, the 2732 is missing, and you don't have an EPROM burner anyway. You need to reproduce the ROM with TTL ICs, discrete transistors, and/or passives --- or at least (it sounds like) you need something that returns 00 1B 00 00 00 00 00 00 00 00 for addresses 0..9. How do you do it? Note that use of a 74LS68[2/4/5/7/8] e.g. or similar is basically cheating :-) Highest marks for lowest BOM cost.

[AlexTheCat123]

Component golf challenge: you're MacGyver, the 2732 is missing, and you don't have an EPROM burner anyway. You need to reproduce the ROM with TTL ICs, discrete transistors, and/or passives --- or at least (it sounds like) you need something that returns 00 1B 00 00 00 00 00 00 00 00 for addresses 0..9. How do you do it? Note that use of a 74LS68[2/4/5/7/8] e.g. or similar is basically cheating :-) Highest marks for lowest BOM cost.

Here's what I came up with. Not sure if it's the most optimized design possible, but it's something! It just consists of a 74LS04 hex inverter, 74LS08 quad 2-input AND gate, and a 74LS245 bus transceiver, and will produce the pattern 00 1B 00 00 00 00 00 00 00 00 00 00 00 00 00 00 repeatedly as you increment through the ROM's address space.



The top circuit in the Logisim screenshot is the design expressed in terms of gates, and the stuff on the bottom is the schematic using the actual TTL logic chips. The real design uses more gates than the one on the top because of the LS245's active-low enable pin that isn't represented in the top design.


To avoid having to add another chip, I omitted the RD signal that's used on the EPROM's /G pin and I'm only using /SLn for chip selection. I'm operating under the assumption that nothing will ever try to write to anything in the lower half of the card's address space since the only thing there is the EPROM, and this should hopefully be a safe assumption.



The parts cost comes out to $2.81, less than a single 2732 from most online sources, so not too bad!

[bmwcyclist]

So...

Is it a 2732 or 2716??

I see lots of variants, -b, etc Should any version work?

program a new one if your friend has some on hand - it is a 2716.

Component golf challenge: you're MacGyver, the 2732 is missing, and you don't have an EPROM burner anyway....

[AlexTheCat123]

Either one will work. The 2732 is twice the capacity of the 2716, so you'd just duplicate the ROM image to fill the entirety of the ROM if you used the 2732. I believe the ROM file I sent you is 4K though, so it's the proper size for a 2732 to begin with.

And yeah, any variant should work!

[Lisa2]

I just recieved a new Sun Remarketing Lisa SCSI card from Vintage Micros.
The card looks great but it is giving me a Card Error 92 on bootup.

bmwcyclist,
I am sorry you are having issues with this board from Vintage Micros.  Normally these boards are very reliable, you should not have deal with this much trouble to get it going.

On behalf of Vintage Micros, I will repair or replace your defective SCSI card.  Please PM me and we will get you fixed up.

Thank you,
Rick

[bmwcyclist]

Rick, thank you for the offer!

John is currently working on a CPU card for me, so I don't want to bother him just now, and I don't mind troubleshooting the SCSI card, I am learning a great deal.


I have been in IT and IS since the mid 1990's but always on the networking and security side, never low-level hardware or development.

But if I don't get it working after the ROM experiments I will get in contact with you. Thanks!!!

 

I just recieved a new Sun Remarketing Lisa SCSI card from Vintage Micros.
The card looks great but it is giving me a Card Error 92 on bootup.

bmwcyclist,
I am sorry you are having issues with this board from Vintage Micros.  Normally these boards are very reliable, you should not have deal with this much trouble to get it going.

On behalf of Vintage Micros, I will repair or replace your defective SCSI card.  Please PM me and we will get you fixed up.

Thank you,
Rick