[Tephran]

I'm out of ideas, I reseated the connectors again and can feel them pop into place which made no change in the card being fully seen connected in hinv.

I also cannibalized another Octane and installed the MXI/Digital video and fresh drive only to receive the same 'unknown device on GIO bridge' on the newly created drive eliminating the compression connectors on the board being the cause of the issue.

Would the unknown device describe the lack of drivers? I've installed the drivers just as you had on a fresh booterizer install of .22 and also tried .30 and receive the same error.

[weblacky]

While I'm afraid I don't have anything software-wise to add. I would ask that perhaps it's a matter of connection or power short on the boards. Do you have a fairly good multimeter and test probes (leads) that would allow you to do a little poking around?

The question was asked about the ribbons and the connectors. I'd suggest taking the board assembly out of the Octane (gently) and getting it placed in an upright position that you can secure and try to use the multimeter conductivity test to gently probe the backs of the boards where the solder points for the ribbon headers are poking through the PCB rear (I thought they poked through the back as through hole mounts? Correct me if I'm wrong) and touch each test lead to each board (with ribbon cables still in place, untouched) and see if you can prove that the current ribbons cable(s) are in fact connecting both sides at the back solder joints. You may find other (cross-connected) points but don't worry about that at first. You can revisit stuff like that to check for bent pins. Just get positive contact that each ribbon header pin on one PCB full connects to the solder joint on the other board's header pin.

This will allow you to see if a pin's missing or not without removal, if you find one you cannot see connected, then it's worth carefully removing that ribbon connector and examining the pins for bends or otherwise.

After that I'd check the main power rail for that board. I'd try to find the ground plane using a PCB ground point (like a platted screw hole, diode, or grounding cap), then use both diode mode and Ohm mode on the meter to poke around what appear to be main power tracks and check for shorts. It's possible there is a small short on the option board preventing good power to the board.

That's really my only suggestion, to try to (line by line) verify the external connections as best you can (power and ribbons). I double you can do much about the compression connectors, so I'd only probe the connector's back or underside, I'd not risk touching the connector face with anything.

Also if you have boards out, take some pics of the board, connectors, and whatnot for us to examine to see if anything doesn't match what another group member may have.

But perhaps we'll see something, you never know.

Best of luck.

[Raion]

Booterizer, dunno if it installs the necessary subsystems. I could be wrong, but I think that you need to install a specific subsystem for this.

Someone else here correct me if I'm off-base.

[kaigan]

The Booterizer instructions install a pretty typical package set - standard, removing java_dev* and java2_plugin*. That eliminates a number of conflicts and allows the installation of a general-purpose IRIX installation. Beyond that, all the Digital Video Option should need is dmedia_eoe, impactdm, and impactvideo.

As Weblacky said, if you're able to take a couple of pictures of the board, I can compare them against mine and we can see if there's anything obviously out of place or unusual. I'll get a couple of pictures of my Digital Video Option and post them to Silicon Image as reference material soon-ish as well.

[Tephran]

Here are pictures of the card and ribbon cable, if I missed anything let me know and I'll snap more.

[weblacky]

Is it me or does the outermost ribbon look skewed? Could you unscrew boards and check the ribbon connections we cannot see? The connection on the underside of the graphics card where the ribbons starts?

[Tephran]

I'll take more pictures later this afternoon with the card off of the carrier. I don't think the two cards come apart that form the board, the ribbon cable is sandwiched in between the boards.

[kaigan]

I'll take more pictures later this afternoon with the card off of the carrier. I don't think the two cards come apart that form the board, the ribbon cable is sandwiched in between the boards.

The two boards that make up the Digital Video Option do come apart - the top one is even called the "Galileo Mezzanine", so it makes some sense that they'd be able to be separated. I used a gentle amount of force near a couple of the standoffs on the top board and removed it from the two sockets it slides into.

Here are the pictures I've posted up to Silicon Image: http://siliconimage.irixnet.org/index.ph...deo-Option

[weblacky]

Ok that would be good enough then.

Looking at Kaigan’s pictures, specifically Base DVO Board - 030-0921-004 REV A (Rear). You can see the ribbon test points on the base board rear.

This would be the time to (very gently) use a multimeter with a continuity test to check each ribbon from the Base rear solder point to the ribbon head. You can approach this two ways. You could risk a needle probe INTO the black plastic header or you can go from rear to the small printed PCB heads of each ribbon and test there.

Either way, I’d test each connection right to the ribbon, as-is. If you don’t show any signs of an issue, then use the same meter mode to test the graphics board header (that the ribbon connects to).

Failing to find an issue there, I’d consider trying another graphics card in this combo. If that also doesn’t yield results, then you could blame either the XIO port position on your Octane or the digital video card itself.

Also while it’s apart you might try to find the voltage rails and check for a short or anything on the digital video boards themselves.

Since you mentioned you cobbled this Octane together I assume you don’t have another Octane to shove this board set carrier into and check the XIO connector or PSU?

[Tephran]

Ok that would be good enough then.

Looking at Kaigan’s pictures, specifically Base DVO Board - 030-0921-004 REV A (Rear).  You can see the ribbon test points on the base board rear.

This would be the time to (very gently) use a multimeter with a continuity test to check each ribbon from the Base rear solder point to the ribbon head.  You can approach this two ways. You could risk a needle probe INTO the black plastic header or you can go from rear to the small printed PCB heads of each ribbon and test there. 

Either way, I’d test each connection right to the ribbon, as-is.  If you don’t show any signs of an issue, then use the same meter mode to test the graphics board header (that the ribbon connects to).

Failing to find an issue there, I’d consider trying another graphics card in this combo. If that also doesn’t yield results, then you could blame either the XIO port position on your Octane or the digital video card itself. 

Also while it’s apart you might try to find the voltage rails and check for a short or anything on the digital video boards themselves. 

Since you mentioned you cobbled this Octane together I assume you don’t have another Octane to shove this board set carrier into and check the XIO connector or PSU?

I've tried the card combo in another Octane and receive the same issue with the unknown gio device, this was using the same disk that was in the original machine with a fresh install of .22 with the same drivers that Kaigan used. This should eliminate the port being the issue as the problem persisted on both Octanes. I should note, that I have the same issue on an upgrade of .30, now that I found my extra SD card a fresh install of .30 is possible; however, I'm not sure that would prove anything.

If the ribbon cable comes up clean through testing, I might try to source a cheap SI card to see if it's possibly the EXMI that is causing the problem. What perplexing is that hinv sees the card; however, doesn't load the drivers or thinks it's not attached to the EXMI.

I work with guys that are well versed in circuit board repair and witchcraft, I've enlisted their assistance to help test the board out.