PDA

View Full Version : CSPPC SCSI Failure :-(



fitzsteve
1st February 2012, 23:15
Just wondering if it's common for SCSI to fail on these cards?

I was getting some lockups the last few days and then yesterday I had UNIT1 fail error message and then my partition stopped booting.

It was still booting of my CF with the SCSI partitions showing up, upon running games I had some issues again and then today it seems to have bitten the dust totally :(

I tested another SCSI-IDE bridge and the partitions were not picked up at all on first boot and after refitting the cables it started to boot and locked up.

After removing the SCSI cables and booting of CF everything is normal and the IDE HDD is fine as I hooked it up to my Deneb with a USB-IDE adapter (very cool by the way) and all my partitions are working perfect :cool:

I'm hoping it's just a faulty cable, I've ordered another cable of eBay as I only had the one, I hope that's all it is, but wondered if it's common for the SCSI to die on these?

The last user had the whole bus overclocked to 64mhz, I've been running at 55mhz as I didn't have any spare 50mhz crystals, maybe over time it's killed the chip?

:(

cosmicfrog
2nd February 2012, 01:35
Hope its just a cable problem M8 and not dieing hardware

mfilos
2nd February 2012, 07:41
I had these issues from time to time on failing partitions (most common problem was some parity error).
Turned out to be the 16-bit bus that I manually selected for better speed.
Once I switched it to manual 8-bit mode it never gave me an error again.
Sadly this means that I'm only running atm on ~19MB/s when I was on ~29MB/s before :(

Maybe it depends on the drive connected as well. I seriously doubt it's a problem with your SCSI controller. More frequent are problems of Hard Drives/Cables/Termination/Firmware upgrades (for SCSI Bridges).

stachu100
2nd February 2012, 07:57
Firmware upgrades (for SCSI Bridges).
That is why I'm using mechanical HDD's with real SCSI interface.

mfilos
2nd February 2012, 08:02
True mate I second that...
...but can you find me an UW SCSI 2.5" HD to fit in my custom 5.25 panel? :D
Only way to do that, 2.5" IDE HD, 2.5-->3.5 adapter, ACard UW SCSI2IDE Bridge :)

merlinkv
2nd February 2012, 08:05
I had these issues from time to time on failing partitions (most common problem was some parity error).
Turned out to be the 16-bit bus that I manually selected for better speed.
Once I switched it to manual 8-bit mode it never gave me an error again.
Sadly this means that I'm only running atm on ~19MB/s when I was on ~29MB/s before :(

Maybe it depends on the drive connected as well. I seriously doubt it's a problem with your SCSI controller. More frequent are problems of Hard Drives/Cables/Termination/Firmware upgrades (for SCSI Bridges).

I have only problerms with IDs >7 my two CSPPCs 44.71.... don't work with these IDs, they only work fine with IDs between 0-6 :blink:

Another issue I detected is how much longer is the cable...>2 meters and I get a lot of erros.... Yes, two meters :lol:, this is because I'm using some HDDs as external case units. Now my cables are 1.8m longer and works without issues.

stachu100
2nd February 2012, 08:14
I have only problerms with IDs >7 my two CSPPCs 44.71.... don't work with these IDs, they only work fine with IDs between 0-6 :blink:
CSPPC controller ID is 7 by default...
And 8 and above are for wide SCSI.
Added:
Is everything ok on 44.69?

---------- Post added at 09:14 ---------- Previous post was at 09:13 ----------

@mfilos - True. :-)
I'm just not using custom panels. ;-)

fitzsteve
2nd February 2012, 08:57
Thanks for the reassurance guys. I will update once my new cable arrives :thumbsup:

merlinkv
2nd February 2012, 09:45
CSPPC controller ID is 7 by default...
And 8 and above are for wide SCSI.
Added:
Is everything ok on 44.69?

Yes I'm using Acard UW adapters but don't work on 44.69 or 44.71, really don't needed to work with IDs 8-15 but is an issue.

BLTCON0
2nd February 2012, 10:29
Another issue I detected is how much longer is the cable...>2 meters and I get a lot of erros.... Yes, two meters :lol:, this is because I'm using some HDDs as external case units. Now my cables are 1.8m longer and works without issues.

For SE Ultra SCSI (20 Mhz bus - doesn't matter narrow or wide) the maximum total cable length (internal + external) is only 1.5 metres, so at 2 metres you were quite a bit beyond specifications!
You still are at 1.8 metres but of course good quality cables can allow going a bit beyond specs :thumbsup:

merlinkv
2nd February 2012, 11:10
For SE Ultra SCSI (20 Mhz bus - doesn't matter narrow or wide) the maximum total cable length (internal + external) is only 1.5 metres, so at 2 metres you were quite a bit beyond specifications!
You still are at 1.8 metres but of course good quality cables can allow going a bit beyond specs :thumbsup:

Hi mate :)

Yes I'm using HQ cables, but after I readed your comment I will reduce my config to around 1.4m :thumbsup:

Thanks for the info :)

mfilos
2nd February 2012, 11:57
Before installing GREX 4000 I had the latest non-PCI firmware (44.69) on my CSPPC but since GREX requires the new one (44.71) I dunno if it also involves the SCSI compatibility as well.

I'll try for fun the 8-->15 ID to see if it makes any difference.
Also I tried alternative ACard firmwares (3.89 and 3.84) without any difference.

There was exactly a similar issue over A.org (http://www.amiga.org/forums/showthread.php?t=54626) that was solved just with a firmware upgrade while mine hasn't :)
Well.... can't be sure 100% with the Amiga anyway lol.

D.B
2nd February 2012, 12:52
For SE Ultra SCSI (20 Mhz bus - doesn't matter narrow or wide) the maximum total cable length (internal + external) is only 1.5 metres, so at 2 metres you were quite a bit beyond specifications!
You still are at 1.8 metres but of course good quality cables can allow going a bit beyond specs :thumbsup:

Hi mate :)

Yes I'm using HQ cables, but after I readed your comment I will reduce my config to around 1.4m :thumbsup:

Thanks for the info :)

A cable made up of Twisted Pairs is another way of helping stop corruption over longer cables.

http://www.qvs.com/prodimages/SCU160-4T_LR.jpg

@ fitzsteve

Sounds like the card is half fried already & probably due to the Syncronous overclock. Get a 50Mhz Crystal back in there asap & have a go at reflashing the card if all appears good for the 50Mhz Crystal being in place.

Wish you luck. :thumbsup:

fitzsteve
4th February 2012, 19:02
Good news guys, I replaced the cable with a nice rounded one and it seems to be working again :)

mfilos
4th February 2012, 19:11
Awesome news Steve mate :)

fitzsteve
4th February 2012, 19:26
Awesome news Steve mate :)

Thanks bro, I just played ShogoMAD to celebrate, I forgot how cr@p it was :lol:

Now SCSI is working I might grab my Radeon 9250 and test out OS4.1 Classic...

johnim
4th February 2012, 19:32
nice 1 steve even the cables degrade over time

glad it was only a cable

fitzsteve
4th February 2012, 19:35
nice 1 steve even the cables degrade over time

glad it was only a cable

Me too mate :bowdown:

Cheers!

D.B
4th February 2012, 21:19
Good news guys, I replaced the cable with a nice rounded one and it seems to be working again :)

Thank goodness for that. About time you had a break.

Still consider running the card Async on the X-tals though. Just to be sure. :thumbsup: