FireWire 800 throttling down to 400 speeds: Say It Ain't So!

It isn't true.

The FW400 port is independent from the FW800 port, otherwise it would make almost no sense to have two different ports for one bus.

If you have FW400 device on the FW800 port, and a FW800 device connected to the FW400 device, you will only get FW400 speed for the FW800 device.

Note: These situations may or may not be applicable to yours. My personal suggestion is to test it out yourself.

If you prefer just to read a short version, this would be the only paragraph to read - unfortunately, yes, the speeds throttles down a significant amount. The long version follows ... (I'm trying to dig up an amazon user sharing his experiences with FW, will post a follow-up update when I get it.)

Robert Mohns, FCP user, did some research on this issue specifically related to FCP ...

This is actually a very long-standing issue with video capture to Firewire hard drives. It seems to be voodoo, in that there's not much way to tell whether a setup will work or not until you try it.
Apple themselves have a few things to say on this topic:
Final Cut Pro: What Kind of Hard Drive Should I Buy notes:
If you choose to use a FireWire hard drive, make sure it meets these requirements:
The drive is the only device on the FireWire bus.
The drive's speed is at least 7200 RPM.
The drive can sustain minimum data rates appropriate for your video format (3.7 MB/s for DV).
The drive uses a modern, high-performance bridge chip such as the Oxford 911 or 922.
The drive uses its own power supply, and is not powered from the FireWire bus.
Final Cut Pro: Dropped Frames and Loss of Audio Sync lists many possible causes of capture problems.
Finally, there's an interesting overview of new features in FireWire 800. It includes reduced arbitration delays (crucial to running data and video on the same bus), and the ability to connect the chain back to your Mac so if one end unplugs the data still flows.
The most reliable solution is to capture to an internal drive. The next best solution is to have the DV device on a dedicated bus, and the FireWire hard drive on a separate bus. For a Power Mac or Mac Pro, you can add a PCI firewire card for short money. PowerBook G4 (15"/17") users can purchase a FireWire CardBus card. I'm not sure what the options for the new ExpressCard MacBook Pros are. iBooks, MacBooks, and iMacs are pretty much up a creek as far as expansion options go.

Click to expand...

The general consensus - generally on the same bus. If memory serves me correct, I have not heard/read a positive report yet.


Other referring links:
Link 1
Link 2

Update 1:
The Amazon Link, a veteran mac user's experiences


Update 2:
Firewire BUS, Early 2008 Mac Pro, This is specific for the aforementioned Mac Pro, it may be similar to your Power Mac too.
Apple's Documentation for Firewire states, "The PowerMac G5 computer has one FireWire 800 port based on IEEE 1394b and two FireWire 400 IEEE 1394a ports. The three FireWire ports are all on the same FireWire bus. They share a single power supply that can provide up to 15 watts total. The FireWire 800 port and the FireWire 400 ports provide a regulated power of approximately 25 VDC."

1

The Amazon review is quite interesting. I've not read or heard about other people having difficulties with Apple FW architecture. I've used daisy chained FW 400 drives on iMacs for the past several years and had no problems. The Amazon reviewer seems to indicate that there is a problem with FW 800 when used in this capacity. Does anyone else have any experience with this?

rtrt, you were posting at the exact (well, almost exact) moment I was updating my post with more results. Thanks for helping out. I wouldn't have needed to dig that up if I knew you had it at hand 😉

Here's another link to a user who tested out what happened when one introduces a mix setting.

For what is worth ... here's another user supporting the fact that FW800 throttles to 400 ... This report (e-mail correspondence) is from Leica Users Group.

I vaguely recall Rob from BareFeats doing a similar test years ago. If anyone knows of the link, do share. That said, my memory may be in question too. Forgive me k? 😉

Just note one thing when/if you are going to perform your own experiment. Ensure that the hdd you are testing is capable of achieving 100 MB/s sustain large transfer (I'm ignoring burst speeds and random read/write since most of us use FW to work with large files) since that is the ceiling cap of FW800. Also good to note the chipset used.

For me, I prefer practical real-life usage kind of test to synthetic benchmark testing.

According to this ... in short - NO throttling of speeds as spinnerlys indicated

mscriv said:

The Amazon review is quite interesting. I've not read or heard about other people having difficulties with Apple FW architecture. I've used daisy chained FW 400 drives on iMacs for the past several years and had no problems. The Amazon reviewer seems to indicate that there is a problem with FW 800 when used in this capacity. Does anyone else have any experience with this?

Click to expand...

This other thread shows no throttling of speeds.

The post (near the end of the thread) by Article Accelerator (September 30, 2009 01:59PM) is very informative. I'll just do a compete quote here for reading conveniences for all ...
(note: i have not verified the links working in the quote that follows)

[www.firstglimpsemag.com]
"Finally, bilingual mode permits a mixed environment of FireWire 400 and 800 devices, while ensuring that each runs at its optimal speed"

[72.14.207.104]
"FireWire beta will allow all devices on the bus to operate at their maximum speeds, even in bilingual mode. This has been made possible through the concept of beta clouds. Beta mode devices cluster together on a logical level in what is called a cloud. These clouds operate as one block, inside which performance is beta quality, i.e. 800 Mbits/sec. The border nodes of these clouds connect legacy devices with the beta clouds, and they operate 'as usual'. The result is that each device can operate at its highest speed, delivering the fastest throughput overall."

[itcert.ccsn.net] (PDF)
"The concept of beta clouds allows FireWire to keep all devices on the bus operating at their maximum speeds. Devices are clustered together on a logical level in a cloud. The cloud operates as a single block at 800 Mb/sec. The borders of these clouds are where legacy devices are connected, operating normally. Each device operates at its highest speed."

[www.faculty.iu-bremen.de]
"So, if a user has only new IEEE 1394b devices and links, the bus will automatically operate in beta mode (S800, S1600, S3200). For example, when a user plugs in legacy IEEE 1394 devices, the bus automatically recognizes these and will cluster these devices together on a logical level. The same will happen with the beta devices (B clouds), ensuring that both types of devices run at their highest speeds"

[ In all cases, emphasis mine ]

(Just to be completely clear, the FireWire "bus" is defined here: [www.1394ta.org] i.e. the "bus" is the cable(s) plus linked/interposed nodes or devices. It is not the same as a port or channel.)

Finally, note michaelb's results above. BTW, his results match mine: On a single bus connected to a FireWire 800 port on the host computer, two daisy-chained external drives--one using a FirewIre 400 port and the other using a FireWire 800 port--each transfer at their normal maximum rates. These results are sustained regardless of the order the drives are connected in on the daisy chain bus and even when both drives are operating simultaneously (subject to total bandwidth limitations obviously).

It's important to understand that FireWire has three automatically negotiated and activated modes of operation: legacy (1394a, S400), beta (1394b, S800), and bilingual (mixed). The ability to mix devices of different speed devices (nodes) on a single FireWire bus yet have each operate at their maximum rated speeds is a distinct advantage of the FireWire protocol, especially in comparison to USB.

Click to expand...