Using 1394b over CAT5 TSB41 / TSB17 / TSB81

Klaus KÃ¼hn

Other Parts Discussed in Thread: TSB41BA3A, TSB43AB23, TSB41BA3D, TSB81BA3E, TSB41AB1

Hello,

We have for 1000s of devices used the following network TSB43AB23 <-> cable <-> (port 1) TSB41BA3A * mode 8, pmc mode, bmode (port 2) -> TSB17BA1 <-> cat5 cable crossed <-> TSB17BA1 -> (port 2) TSB41BA3A * mode 8, bmode, pmc mode (port 1) <-> TSB41AB -> say a disk drive with a dedicated link/controller.

*- TSB41BA3A in later experiments was replaced with TSB41BA3D

However, if we change the last phy (tsb41ab) to any 1394b capable controller such as TSB81BA3E , communications stop. In Windows, the device is shown as "disconnected" or only detected for a short while.

I have tried with other modes (0,12,), BMODE=0, etc. to no avail.

Any ideas why this is?

over 12 years ago

0 Elias Villegas M. over 12 years ago

TI__Guru* 76130 points

Hello Klaus,

Are you using the 81BA3E as a drop replacement of the 41BA3D without changing anything?

Can you share your schematics?

Regards.

0 Klaus KÃ¼hn over 12 years ago in reply to Elias Villegas M.

Prodigy 40 points

4666.1394_cat5_TI_forum_schematic.pdf

I attached the schematic of the TSB41 / TSB17 combination (note we replaced TSB41BA3A with D version also). We use the "noise rejection" circuitry on the TSB17 as recommended by the data sheet. By the way, why is it impossible to search for the TSB17BA1 on ti.com ?

I am not using the TSB81 as a drop-in replacement, I am using a commercial firewire to 2.5" disk converter, such as ONNTO AT-S111 (which uses TSB81 and initio). (My own design uses LSI FW843 & OXUF943SE which acts similar). Again, it seems any 1394b device attached is problematic.

Hope it helps, best regards

0 Elias Villegas M. over 12 years ago in reply to Klaus KÃ¼hn

TI__Guru* 76130 points

Hello Klaus,

Can you confirm that all the devices are speed compatible?

This sounds like a driver issue, the host is probably trying to communicate with the USB 3.0 HDD at a higher speed than can be supported over the S100B connections in between.

Regards.

0 Klaus KÃ¼hn over 12 years ago in reply to Elias Villegas M.

Prodigy 40 points

If the last device is a TSB41AB1 (1394a only), then it works fine, even though of course the throughput never gets beyond S100.

Any 1394b device connected and it will not work.

I don't see where USB3.0 is relevant. A harddrive attached will usually use PATA or SATA.

I played a bit with drivers. Linux drivers won't connect, Windows drivers won't connect, Unibrain Windows drivers curiously crashes the entire PC when the last device (1394b) is attached.

0 Elias Villegas M. over 12 years ago in reply to Klaus KÃ¼hn

TI__Guru* 76130 points

Hello Klaus,

Forget the USB 3.0, I've got confused.

Talking with another engineer, she found the following:

The 1394 OHCI will read the self-ids from all devices on the 1394 bus at S100 speed, but the driver may not be smart enough to force the connection between the 1394B HDD and the TSB41BA3x device to S100 speeds. We ran into this issue when we tested optical/CAT5 systems.

Regards.