0

Setup and Theory

I'm testing my USB connection from my Laptop to an external USB 3.0 drive and vice versa by copying a ~3 Gigabyte folder back and forth. Natively my laptop does only have USB 2.0 ports (up to 480 Mbit/s), so I bought an adapter to plug into its Thunderbolt 1 port and get USB 3.0 that way.

In theory this should work: USB 3.0 can provide up to 5000 Mbit/s and Thunderbolt 1 up to 10'000 Mbit/s (on each of the two channels). So the setup laptop - adapter - hard drive should get my USB 3.0 speeds, I think...

Setup:

Laptop ------> Adapter ------> External Hard Drive
          |               |
     Thunderbolt 1      USB 3.0 

Actual Copy times

My experience differs though! The measured times to copy that ~3 Gigabyte folder are as follows:

  • External hard drive to Laptop 24 seconds.
  • Laptop to External hard drive 45 seconds.

I therfore calculate bitrates of 980 resp. 520 Mbit/s, which is roughly 20 resp. 10 % of the maximal bitrate of USB 3.0. If I plug a docking station to this adapter and connect hard drive to it and, in addition, an external screen, the transmission times go up, i.e. bitrates plummet (also, the screen is laggy):

  • Both directions 77 seconds - so approx. 307 Mbit/s.

Setup:

                                                 DVI
Laptop ------> Adapter ------> Docking Station ------> Screen
          |               |                    ------> Ext. Hard Drive
     Thunderbolt 1     USB 3.0                    |
                                               USB 3.0

Of course this is way faster than USB 2.0, when I connect the external hard drive directly to the laptop:

  • External hard drive to Laptop 86 seconds.
  • Laptop to External hard drive 117 seconds.

So that's about 270 Mbit/s resp. 200 Mbit/s, or about 60 resp. 40 % of the max. speed of USB 2.0.

Setup:

Laptop -----> External Hard Drive
         |
      USB 2.0

Questions

  1. Is my assumption, that I should get USB 3.0 speeds with the setup Laptop - Thunderbolt 1 > Adapter - USB 3.0 > Device, true?

  2. Are these USB 3.0 transmission speeds normal?

  3. Are these USB 2.0 transmission speeds normal?

  4. If no once or twice: Why? Which device is responsible for these results?

  • If you take a golf cart on the highway should you expect to drive at highway speeds? If this external drive is a rotating disk drive (rather than an SSD) then there is a physical limitation on the read/write speeds of that drive relating to its mechanics and firmware. For SSDs it is also possible that they will perform well under 5 Gbit/s, but due to the characteristics of the flash memory and the drive firmware. – Romen Sep 10 at 15:33
  • Alright, thank you @Romen for your quick answer! Well, your statement should have been obvious to me...! So, these speeds are normal. I shall test this with a SSD/Flash Drive again and see whether my results are different. – CJLM Sep 10 at 15:36
  • Your assumption #1 is flawed. You're trying to measure the (averaged) speed of a copy, not the bus (USB) speed. The copy operation has moments of idle bus, so that reduces the average "speed" you calculate. You need an oscilloscope to measure the actual data transfer speed. – sawdust Sep 10 at 19:05
  • @Romen The results with a USB 3.0 memory stick are the same or slower... But thank you nevertheless! – CJLM Sep 13 at 13:41
  • @sawdust Thank you for your answer. That might be, but still for my day-to-day file transfer I'm interested in the average time it takes per Byte. So I guess, I just don't unterstand bitrates yet... – CJLM Sep 13 at 13:45
1

Your results will strongly depend on what kind of fragmented files are in your folder.

Forget your stopwatch "method" and get ATTObench, CrystalDiskMark, or any other reputable disk benchmark. Linux also has corresponding benchmarks. The benchmarks will do various file length and calculate all transfer times for you.

First, bench your HDD on any good desktop PC, 64-bit OS, with native xHCI controller for USB. You will be surprised that the internal disk inside your external enclosure might be performing asymmetrically on Reads and Writes, and drastically below the raw bus data rate. Please use actual MBYTES/s, and not bit rates, because bit-byte conversion might have an overhead related to data encoding schema. This will be your reference point.

Then try the same with your TB-USB adapter hack.

  • Damn that's an advanced answer! Thank you. Exactly what I expect from the members of this forum! :D I shall try your real method again later today and will be back with results worthy of this term. – CJLM Sep 12 at 3:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.