Re : Error 0xBFFF800B in MAX with firewire cameras

[shb]

Hello,

MAX lists all the cameras connected by FireWire. But when I click on one to show the configuration, MAX is busy for several seconds and then reports this in a dialogue:

Error 0xBFFF800B
The session for your device could not be configured

The camera is shown with a red cross now. Sometimes the camera disappears from MAX and from the Windows Device Manager some seconds after this failure. It reappears after reconnecting, but it againt fails to show the properties. Rebooting does also not help.

There are seven identical cameras connected, two of them work, all the remaining react as described. It does not matter on which FireWire port or on which FireWire controller the cameras are connected. It also does not help when only one of the failing cameras is connected.

I followed the tipps of the first reply in this thread:

I deleted the xml, iid and icd located in "Documents and Settings/All Users/Shared Documents/National Instruments/NI-IMAQdx/Data", but the behaviour remains the same. Two cameras work

The compatibility test with CameraValidator.exe (see this KB) prints "Unhanded Exception". For the working cameras it passes and generates the report file.

Where else could be the problem?

Greetings,

shb

[shb]

Some Details:

OS:   Windows 7, 32bit, SP1

IMAQdx:3.9.1

MAX: 5.1.0f0

An equal system is already running at a customer. It is the same computer with the same operation system (except of the language, it is English here), the same PCI cards, the same software... I do not see any difference. And I did not have this problem on the other computer.

[toto26]

Hi,

Did you tried your cameras individualy, I mean, only one camera connected at a time?

Are all of your cameras connected via a hub, or directly to your PC board?

When you use so many cameras, you have to give enough power to all the cameras.

Regards

[BlueCheese]

Are the cameras identical? What models are they? Is it possible some of them have different firmware revisions than the others? Maybe the camera vendor can tell you based on the serial numbers if they have different firmware revisions.

Eric

[shb]

Thanks for your questions. I answer them below.

toto26:

• Yes, I tried with only one camera connected. I disconnected all and connected only one camera. It did not work.
• All cameras are connected directly to the pc board.
• When only one camera is connected, then there should be enough power. But it still fails. (And sometimes it runs with all cameras connected. See below.)

Eric:

• The cameras are identical.
• It is very unrealistic that the firmware is different because the serial numbers are continuous. But I can ask the vendor.
• (I have forgotten to note the model, sorry.)

When I change the driver of the FireWire controllers to "OHCI-compliant 1394 Host controller" or "OHCI-compliant Texas Instruments 1394-Host controller", I can see all cameras and acquire some images. But the maximum rate of the camera is somewhat random. It should be 400 MBit/s but can be all of 100, 200 or 400. Sometimes I can just do some snaps and then a dialogue tells me that the bandwidth is too small. And sometimes the cameras still disappear.

(Because of an earlier problem I must stay on "OHCI-compliant 1394 Host controller (legacy)" anyway.)

I just realised that the Version of NI-IMAQdx is different between the two systems. The old one has version 3.8.1, the new one 3.9.1 (Driver set August 2011 and February 2012). Should I try to downgrade?

Greetings,

shb

[shb]

NI-IMAQdx 3.9.1 was the problem indeed. After uninstalling it (and what depends on it...) I installed version 3.8.1 and now it works for 6 cameras. I'll report a bug to NI and add the CAR no in this thread.

The problem of the last camera I still did not find,the camera behaves as before. I will try on later and report when I am successful.

What I have also tested is if the problem is the same on a controller with a different chipset than Texas Instruments (with NI-IMAQ-dx 3.9.1). One driver is different, but the cameras on the controller behave as the others.

Does anybody know how to downgrade NI-IMAQdx without uninstalling everything depending on it? I must do the downgrade on the development computer too.

Greetings,

shb

[shb]

I revoke what I wrote about NI-IMAQdx. I am now back at the exact same problem with the old version. Six cameras working was probably only a side effect of changing the drivers. While trying to get the last camera running, the problem reappeared after a reboot.

Some more details:

• The camera model is DMK 41AF02 of Imaging Source.
• With the non-legacy FireWire controller drivers ("Texas Instruments OHCI-compliant...", "VIA OHCI-compliant..." or "OHCI-compliant 1394-Host controller"):
• the cameras not working (with the legacy driver) show minimized maximum bit rate in Max (200 Mbits most time).
• after acquiring, the reserved bandwidth is not given free. I can acquire three times. And three cameras is the maximum number of cameras that can acquire at the same time on one controller with these settings.

Any hints what I could try next?

Greetings,

shb

[shb]

Some more details:

• All the cameras on the current machine do have the same firmware version.
• The firmware of the cameras has changed between the last system and the current one. But the DCAM protocol did not change. (This protocol is used by the drivers.)
• One camera needs about 250 mA of current. There are maximum 3 cameras on one FireWire controller. One firewire controller provides up to 1.5 A of current, so the half load is used only.

Greetings,

shb

[shb · Accepted solution]

Finally the problem is found. The cables were the problem. 5 of the 7 identical cables seem not to work at 400 Mbps. I have tried different combinations of connections, but until today I mainly changed at the controllers and not at the cameras. So every camera has always been connected to the same cable most time.

(The cables reached only to some cameras at the camera end, at the controller end I could try every combination.)

The non-legacy drivers seem to have detected the failing at 400 Mbps and limited the rate to 200 Mbps therefore (or 100 sometimes). But I have thought it was a false reading of the capabilities of the camera. The legacy driver did not switch to a lower rate, so communication failed.

For me some riddles remain. Why do some cables work and some not? And why did downgrading the NI-IMAQdx driver temporary help?

Anyway, ...

... the source of the problem is found. This is the main thing.

A big thank to those all who have answered or have given some thoughts to the problem. (Especially thanks goes to Imaging Sources because they have always replied very quickly and competent.)

Final greetings,

shb

who can finally go back to programming