Root cause for 3112 and 3114 hanging on Quicksilver + Digital A/V G4 in "sleep"

[FdB]

embedding the "dlattach" and adding a width= property to the img tag will properly size the photos to be viewed properly

Thanks macstuff. I forget those parameters myself, occasionally.

[(S)ATAman]

Wow, it looks better than the first batch of the original cards!
Congratulations!

[SDG]

I am actually horrified that my bodge job looked better than the first revision of the retail cards. 

Just to add that I have also tested it in a Quicksilver 2002 and it also played very nicely with it, something I thought I would never see. Thanks (S)ATAman for the necessary info.

I am going to have a go at recapping the dud card and see if that works. If it does then I will have a think about the aesthetics of attaching the 29150 but I cannot see a way past using short connecting leads on all the pins at the moment.

[IIO]

not even the sonnet tempo performs perfectly, it also tends to not wake the disks up again after sleep in some 10% of the cases.

[SDG]

Some bad news with the 3112 card in both my QS and DA. While both the cards I modded showed up and did operate with an existing drive in place, trying to install Leopard on a SATA drive attached to the cards led to kernel panics right after getting past the first couple of screens onto selecting a disk to install on.

Not sure how to capture a log when it doing an installation to a fresh drive so I have no idea what is causing the crash apart from trying to detect a drive. This happened with both cards in both machines and I swapped the SATA disk in case there was a hardware issue there too. I still got the black Restart Notice of Death every time.

[(S)ATAman]

Some bad news with the 3112 card in both my QS and DA. While both the cards I modded showed up and did operate with an existing drive in place, trying to install Leopard on a SATA drive attached to the cards led to kernel panics right after getting past the first couple of screens onto selecting a disk to install on.

Not sure how to capture a log when it doing an installation to a fresh drive so I have no idea what is causing the crash apart from trying to detect a drive. This happened with both cards in both machines and I swapped the SATA disk in case there was a hardware issue there too. I still got the black Restart Notice of Death every time.

Not sure, what driver you used. If it was my driver, it DOES have bugs.

I am working on an update, not merely to enable the "unsupported" ROM-s.
If all you did to "mod" the card was the Micrel regulator, you are fine, just please wait for the update.
The update will also enable the hot-swap on all cards, it was disabled by intent on the cards with internal connectors only.

An other update: the hot-swap dialog will be gone forever. The underlying API (KUNC, KernelUser Notification) was deprecated by Apple and even hooks within the OS removed.
They have it even now, but only for themselves.

That means, a driver code in order to be kept functional for a longer time can't have these calls.

Sorry for being late and lazy, it was quite a bad period of time behind us.

[SDG]

I was using the firmware distributed with the SeriTek cards, version 5.1.x as the later 5.3.x was said to be buggier. If you wrote both then no biggie. In any case, I've plenty to be getting on with at the moment so happy to wait for as long as it takes.

[(S)ATAman]

I was using the firmware distributed with the SeriTek cards, version 5.1.x as the later 5.3.x was said to be buggier. If you wrote both then no biggie. In any case, I've plenty to be getting on with at the moment so happy to wait for as long as it takes.

Yep, I am guilty for both. It was eons ago and at that time no proper debugging tools were available like machine-to-machine firewire debugging.
The base for the new version is going to be 9.0, 6, 7, 8 skipped for different reasons.
The base card is going to be 3114, others (3112 and 680 parallel) are to some degree siblings of 3114.

3114 is not very fast, can't make over 90 MB/Sec on reads and 62 MB/Sec on writes. This is with Samsung 840 EVO.

[redstudio]

excellent study work for these 3112.. congratulations to everyone for the work done so far.
one question: would changing the 29150 regulator chip make these cheap 3112 cards work on all macs perfectly or just Quicksilver?
I read somewhere that replacing the chip with an FS8860 33C (which has the same shape as the 1117) would still work. has anyone tried?

[V.Yakob]

has anyone tried?

Yes, I have prepared many such boards, and they work great in QS. I wrote about it here and here.

[redstudio]

yes, I was interested to know if once modified it works on all PCI Macs (MDD for example..

[peeperpc]

yes, I was interested to know if once modified it works on all PCI Macs (MDD for example..

MDD Macs don't need this modification. These cards just need the firmware flashing to work with MDDs.

[redstudio]

if I modify the card for Quicksilver I can use the same card ONLY on Quicksilver or then also on MDD or other Macs or then wouldn't it be fine? modified only works for Quicksilver? the same card I mean. or does it become compatible with all PCI macs? even those that don't need modification?

[V.Yakob]

You probably haven't read the messages on the links I left.
The same card works fine on all PowerMacs.
I tested these cards on PM 9600, PG G3 Minitower, PM G3 B&W, PM G4 QS, PM G4 MDD -- work perfectly.
At one time, when I dealt with all this, I made myself 6 or 7 such boards, and I still have chips for 2x for future projects if they take place.

The last board I made for B&W is still not installed and is waiting for its time to replace another board.
These red boards are ideal for B&W, DA, QS because of the body design.