Phase5 Blizzard 2060 SCSI issues

[sledge]

I have a Blizzard 2060 and an Amiga 2000 rev 6.2 computer that I got from my brother last week. Everything seems to work just fine, except that I cannot get any SCSI drive to be formatted.

I have a 50-pol flat cable, with active termination in the end. I have tested several SCSI drives from 170 MB to 2 GB in size, but with no luck. When I load the drive information, from either HDToolBox or the SCSI-tools from Phase5, I get faulty information about manufacturer and size. IBM becomes I@M, and Quantum becomes Qualtum and so on.

Could it be that the cards SCSI is faulty? Or am I not getting the termination right? Any help would be great!

[radzik]

Maybe look at jumpers and read this http://amiga.resource.cx/manual/Blizzard2060.pdf.

[sledge]

Yes, I read the manual. I read it again just to be sure

And still, the drive appears in HDToolbox, but when I read the drive information it's always garbaged (though the information is the same every time, not random garbage). The drive size on some drives are -2300MB, while other says 283MB on a 340MB IBM. And the vendor name is screwed up too. You can control the devices in the SCSI-control program from Phase5, like start and stop the device, low level format and verify data and so on.

Are there any firmware updates for the Blizzard 2060 boards?

[matt020]

Have you done a process of elimination?

Tried with different hard drives - done
Tried with different SCSI cable?
Tried with different SCSI card (got a A2091)?
And so on, and so on...

[zipper]

Are there any firmware updates for the Blizzard 2060 boards?

The only one I've seen is v8.5 (check version 2060scsi.device).

http://phase5.a1k.org/files/2060_rom8_5.zip

[radzik]

I have porblem with Quantum IV 9.1 SCSI 68 pin >>> 50 pin (SCSI on Blizzard 2060), it was impossible to manage it to run. On Oktagon 2008 there was no problem. The best solution I found is: buy new big IDE HD and buy IDE SCSI Adapter. Now may new HD with DVD (both conected by Adpater) works perfect. HD has 8,5 MB/s.

[sledge]

Have you done a process of elimination?

Tried with different hard drives - done
Tried with different SCSI cable?
Tried with different SCSI card (got a A2091)?
And so on, and so on...

This is my progress so far:

* Different hard drives (2GB 68-pin SCSI, 170MB Quantum, 4GB 68-pin SCSI, 340MB IBM SCSI, 64MB CF card with SCSI->IDE converter, 40GB Seagate IDE drive with SCSI->IDE converter)

* 4 different internal SCSI-cables (4 contacts, 3 contacts and a 2 contact. Also a 3 contact with a terminator in the end of the cable)

* External SCSI-drive (not sure if it works ok, since I can't even see the drive in HDToolbox then)

* New SIMM modules

* Termination pwr set to off and terminator set to ON always instead of auto.

* Kickstart 2.04 and 3.1 tested

The card seems to work in general, and it's only the SCSI part that fails on me.

I do not have another SCSI-board.

[sledge]

I ran the "Show config" tool in Workbench. What should the information about the Blizzard board look like there?

[Zac67]

Since you've ruled out a bad cable, I'd very seriously look into the possibility of a bad contact on the board. Double check the SCSI connector (have you tried the external one?), the Glogic chip and the small buffer chips.

It makes a big difference whether every single byte gets garbled up or just every 2nd, 4th or so. If the latter is the case I'd bet on one of the buffers or a contact around there.

[sledge]

I'm not sure how to verify the SMD mounted chips by just looking at them. The connectors pins seems to be OK though.

[rkauer]

Set both Termination and termination power ON.

 Use the cable terminator on the Blizzard side. The Blizzard have a built-in termination, switch it off or auto.

 Also check the +5V voltage level on the HD connector. If lower than 4.8V you have data corruption on transfers and all sorts of misbehaving of the SCSI bus.

[Zac67]

I'm not sure how to verify the SMD mounted chips by just looking at them. The connectors pins seems to be OK though.

You'd need at least a logic tester for that... As rkauer has pointed out, the terminator is worth checking as well: switch it off and replace with an external one - but then again, it'd garble all data transmitted.
Back to the buffers: they're probably 74hct244 or 245. If your soldering skills are adequate, I'd swap them out all the same, can't cost too much.

Found a decent photo - they're 74fct2245
pinout:

Code: [Select]

742245
8-bit 3-state noninverting bus transceiver with integrated 25O series
output resistors.
Enable and direction pins control output enables.

    +----------+             +---------------+
DIR |1  +--+ 20| VCC         |/EN|DIR| A | B |
 A0 |2       19| /EN         |---+---+---+---|
 A1 |3       18| B0          | 1 | X | Z | Z |
 A2 |4       17| B1          | 0 | 0 | B | Z |
 A3 |5  742  16| B2          | 0 | 1 | Z | A |
 A4 |6  245  15| B3          +---------------+
 A5 |7       14| B4
 A6 |8       13| B5
 A7 |9       12| B6
GND |10      11| B7
    +----------+


[sledge]

I have tried 5-6 different SCSI cables with no luck. But today I got my MASOBOSHI-card that I bought from a Ebay-dealer. The card had a SCSI-cable supplied. I tried that SCSI-cable on the 2060-connector. And guess what? Every single drive that has been reporting mumbo jumbo info in HD-toolbox now works perfectly. Everything is fine, and I have formatted a disk and installed OS3.9.

How can it be? I need a beer...

Thanks everyone for all the help. This case is closed!

[sledge]

Ok, I was too fast there.

The card still works, and I can use every single SCSI-drive that I have. Yesterday I wanted to see if it really was a faulty cable that caused the problem. Guess what? Every cable I have works just fine now.

Now it's like there's never been a problem at all.

This makes me wonder if bad caps can cause strange behaviour like this? That it works sometimes, and sometimes don't?

-thomas