USB speeds not up to much

  • Thread starter Thread starter Deleted member 68495
  • Start date Start date
D

Deleted member 68495

Guest
I have recently bought a 125GB CFexpress card (Sandisk) and a Sandisk card reader. I've plugged it all into a USB 3.0 port (which, although slower than more modern ones, I am assured is good for 5GB/s) and tried to download some image files to my hard drive from it but I am very disappointed to achieve speeds of only 43MB/s, which is a far cry from the theoretically possible speed of 1250MB/s, this is only twice what I can get from one of my SD cards. Something is definitely awry, what am I doing wrong?

I might add that I am not using the cable that came with the reader as it is USB C at both ends and I don't have a USB C socket on my PC.

EDIT: I used the supplied cable to connect the reader to my wife's Surface GO 2 which does have a USB C socket on it and I managed to get speeds up to 312MB/s (I suspect it's not top speed because the Surface GO 2 is not a fast machine). Are there different cables? I thought a USB cable was a USB cable.
 
Last edited by a moderator:
I have recently bought a 125GB CFexpress card (Sandisk) and a Sandisk card reader. I've plugged it all into a USB 3.0 port (which, although slower than more modern ones, I am assured is good for 5GB/s) and tried to download some image files to my hard drive from it but I am very disappointed to achieve speeds of only 43MB/s, which is a far cry from the theoretically possible speed of 1250MB/s, this is only twice what I can get from one of my SD cards. Something is definitely awry, what am I doing wrong?

I might add that I am not using the cable that came with the reader as it is USB C at both ends and I don't have a USB C socket on my PC.
Does your USB C to USB A cable have blue colour plastic in the A end? IIRC blue colour indicates usb 3 rated. If not that me be the bottleneck.

Alternatively, get a USB 3 rated C to A adaptor ( I use them.

Thirdly, are there fake CFexpress cards?
 
These speeds are always optimistic and depend on many things along the chain, e.g the usb might be sharing its bandwidth with other ports on the machine and even if you could get the full speed could your hdd/sdd write the data fast enough? My 3.5 inch hdd drives top out at 50MB per second write between each other
 
Further to my post above, closer inspection of my PC reveals that I had a USB C port on the front all the time but had never noticed it (PC is on the floor underneath the desk, in a dark cupboard) and using the supplied cable, I've managed to get the speed up to between 200 and 300 MB/s. Still not really anywhere near the claimed speed -- by a long way -- but much better than it was. So, as you may have expected, it was finger-trouble after all.

Note to BoxBrownie: I don't think it can be a fake as it has come in a very nice package, all sealed up and even has a code for a free years download of RescuePro software which would be generous indeed for a counterfeiter.

I can't give it all a proper test as I haven't been able to use the card in my D850 because it is in the process of being replaced under warranty (due to an intermittent focussing problem).
 
OK, I've worked it all out now.

A) Wrong cable, apparently USB cables are not all alike.
B) Claimed speeds are only relevant if the reading or writing device is capable of those high speeds and in my PC's case, they are not. Best case is drive C: which is an SSD, second best is my 2TB backup drive and worst is my working, 1TB, drive which performs rather poorly on UserBenchmark test software.
C) So it seems while my card reader and my new CFexpress card are the bees knees, the rest of my kit is not. Heigh hoe, at least it's all running about 300MB/s which is a damn sight faster than the 25MB/s that my SD cards can manage.
 
OK, I've worked it all out now.

A) Wrong cable, apparently USB cables are not all alike.
B) Claimed speeds are only relevant if the reading or writing device is capable of those high speeds and in my PC's case, they are not. Best case is drive C: which is an SSD, second best is my 2TB backup drive and worst is my working, 1TB, drive which performs rather poorly on UserBenchmark test software.
C) So it seems while my card reader and my new CFexpress card are the bees knees, the rest of my kit is not. Heigh hoe, at least it's all running about 300MB/s which is a damn sight faster than the 25MB/s that my SD cards can manage.


Are you using an built in card reader to get 25MB/s ?
They are often slow. Have you tried the SD cards on a good USB C reader? Should get at least double that (if the card is capable).

300MB/s is quite good though, if it is sustained
 
You're confusing Bits and Bytes a little. Basic USB 3 is 5Gb/s, not 5GB/s, so your 300MB/s is 2.4Gb/s which is not too bad IMHO.

Yes, I wrote it down wrong, it is Giga bits, not bytes. However, 5Gb/s is still a speed of around 625 MB/s so still twice as much as I'm getting from my data transfer.

As an ex-computer engineer (from many years ago), I get really annoyed with companies using bits instead of bytes so as to make things look so much faster than they really are. While computers work in machine language, i.e. bits, they are programmed, and registers work, in multiples of 8-bit bytes and a figure quoted in bits is somewhat disingenuous as it doesn't really reveal the true speed of transmission of useful data. Unfortunately it is a trend that is not going to go away as any company quoting in bytes will make it's products look slow compare to the products of companies that show data transmission speeds in bits. I blame the phone industry as I don't recall seeing this trend before mobile phones turned 'smart'.
 
Last edited by a moderator:
I've found that getting 50% of the rated speed over a USB link is generally very good.

Using bits-per-second for data comms speeds has always been the rule, albeit with some extra confusion between bps and baud. My IT career goes back far enough that I remember acoustic couplers and modems being 300bps. (It actually goes back further than that to the days before microprocessors.)

The one that annoys me (and just about everyone else) is disk capacity megabytes being 10^6 rather than 2^20.
 
Last edited:
It's the lack of consistency in "standards" that bugs me more than any one particular instance. Oh, and the confusion people have with thinking that a b is the same as a B or an m is the same as an M (or even a mju!)
 
Oh yes, I always laugh (hollowly) when someone describes a link as being in milli-bits per second.
 
  • Like
Reactions: Nod
Back
Top