What is the closest model for a streaming use case? Sequential QD=1 sounds close, but if streaming videos stored on a USB drive on a tablet/phone, you're not running at peak throughput even at QD=1. That's also the kind of case where power consumption and battery life matter a lot.
I don't know if that's the kind of task you were referring to with "rush to idle".
I mean like performance per watt, so even if a drive consumes more power doing a certain task it maybe finishes a lot quicker and the total amount of power consumed could actually be less. Perhaps figures like this could be interesting to include in the future.
While maximum power draw for certain tasks, it would be helpful to be able to see a normalized figure to be able to compare efficiency. It may even be necessary to have a test that has a specific set of usual tasks, heavy, medium and light like you do with the SSD reviews to see how the characteristics change depending on load.
You (and plenty of the commenters above) are being exceptionally ignorant, and didn't read the article properly where this is all explained. There is a HARD CAP on the power draw available to a USB2 port of 2.5 W. There is a HARD CAP available to a USB3 port of 4.5W.
This means that if you take one of the drives in the list that exceeds 2.5W under some conditions. it will (more or less randomly) work then fail when connected to a standard USB2 port --- essentially it will exhibit random (and thus very hard to debug) failure, especially since it will always mount without a problem. Even if you're connecting it to USB, if you're connecting to a hub you may have the exact same problem --- the drive always appears to mount, then sporadically fails when a sequence of random reads or writes exceeds the current available (given that some other device on the hub is also using power).
THAT is why the important number for these devices is the worst-case peak number, not an average or the total power consumption or anything like that.
While you have a valid point about non type C connectors, it was not the only one in the article as the implication I read was how the power of a storage stick can adversely affect the life of a portable device which would make efficiency a major factor. No-one has mentioned removing the above information, only adding to it, which can only help.
How about USB hard drives? I know most won't power up because they need more than 500 mA on USB2 but it would be interesting to see exactly how much power is used. I've gotten USB hard drives to work with phones and tablets over OTG but only with a powered hub or a charging hub that allows a battery pack to be connected over micro-USB.
I can power a 320GB 7200 rpm USB 3.0 hdd from my 2012 Nexus 7 and Galaxy S5 with a USB 2.0 OTG cable. One would assume that if they can supply power across the bus they should be able to power anything that runs on a PC, 500mA is 500mA whether from a PC or phone.
That's weird, I've never been able to get 500 GB or 1 TB USB HDD working over OTG. The drive spins up, mounts for a second and then disconnects, probably because the peak current draw exceeds 500 mA. I looked at a bunch of spec sheets from WD and Seagate and all their drives had peak current in the 1 A or higher range. They work fine when plugged into powered hubs though.
This explains so much!!!! I was having random disconnects with a USB 3.0 to M.2 Enclosure when using the front USB 3.0 ports on my Desktop. When I use the rear ports it works perfectly and I get 400 MB/s sequential read/writes... Normal USB thumb drives and powered drives work fine in front but not USB powered enclosures. I knew something was wrong with my front ports but this is the reason.
Interesting. I have a SIIG USB 3.1 enclosure that I put a Patriot Blast 480GB SSD in. The performance is very nice in that it gets very close tot he limits of SATA itself. The included cable has a second host port connector to enlist more power where needed but it seems to work fine on 3.0 on a single plug.
Is anyone developing native USB 3.1 controllers for SSD class flash memory that would exceed the SATA envelope? This would be ideal for consoles like the PS4 if the USB host was upgraded to 3.1 in later model.
An exciting article from 2016. I read it yesterday and found very useful. I am designing a new product - a special type of Monitoring Meter. Am using PIC32MM0256GPM064 in it as the main microcontroller. It comes with a USB OTG module. I then want to save data in the Meter to a USB flash drive (USB2.0). However, my meter cannot provide more than 100mA current to the USB flash drive. So, I was looking for a low power flash drive. You article shows that 128GB flash drive that you tested could be suitable. Have tested some of the many flash pen drives I have and measured the current drawn by using Windows Device Manager for the Rot Hub. One Data Traveler flash drive of 8GB showed 100mA. All others are showing 200 to 500mA. Is there a way to have a USB flash drive that would draw less than 100mA as drawn by HID devices. Would be grateful for your comments. I am also thinking of an option where I provide two Micro USB sockets on the Meter. One to power up the meter, say from a Mobile Phone charging Power Bank (with only the +5V and GND connected and another Micro AB socket where the flash drive is connected. Then I hope the flash drive would be supplied with power by the Power bank and data would sent by my Meter. Is this possible? Again would thank you very much for your comments and advice. ggrUK
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
20 Comments
Back to Article
InterClaw - Friday, March 18, 2016 - link
How about a graph comparing total power consumption for a specific task, like "rush to idle".icrf - Friday, March 18, 2016 - link
What is the closest model for a streaming use case? Sequential QD=1 sounds close, but if streaming videos stored on a USB drive on a tablet/phone, you're not running at peak throughput even at QD=1. That's also the kind of case where power consumption and battery life matter a lot.I don't know if that's the kind of task you were referring to with "rush to idle".
InterClaw - Friday, March 18, 2016 - link
I mean like performance per watt, so even if a drive consumes more power doing a certain task it maybe finishes a lot quicker and the total amount of power consumed could actually be less. Perhaps figures like this could be interesting to include in the future.ISGoodwin - Friday, March 18, 2016 - link
While maximum power draw for certain tasks, it would be helpful to be able to see a normalized figure to be able to compare efficiency. It may even be necessary to have a test that has a specific set of usual tasks, heavy, medium and light like you do with the SSD reviews to see how the characteristics change depending on load.jabber - Friday, March 18, 2016 - link
Have to say power consumption is the last concern I have re. USB sticks. Some folks have far too much time on their hands.name99 - Friday, March 18, 2016 - link
You (and plenty of the commenters above) are being exceptionally ignorant, and didn't read the article properly where this is all explained.There is a HARD CAP on the power draw available to a USB2 port of 2.5 W. There is a HARD CAP available to a USB3 port of 4.5W.
This means that if you take one of the drives in the list that exceeds 2.5W under some conditions. it will (more or less randomly) work then fail when connected to a standard USB2 port --- essentially it will exhibit random (and thus very hard to debug) failure, especially since it will always mount without a problem. Even if you're connecting it to USB, if you're connecting to a hub you may have the exact same problem --- the drive always appears to mount, then sporadically fails when a sequence of random reads or writes exceeds the current available (given that some other device on the hub is also using power).
THAT is why the important number for these devices is the worst-case peak number, not an average or the total power consumption or anything like that.
ISGoodwin - Friday, March 18, 2016 - link
While you have a valid point about non type C connectors, it was not the only one in the article as the implication I read was how the power of a storage stick can adversely affect the life of a portable device which would make efficiency a major factor. No-one has mentioned removing the above information, only adding to it, which can only help.InterClaw - Friday, March 18, 2016 - link
Ideally, the drive should enter a low performance mode under those circumstances so as to not encounter failures. But perhaps that's wishful thinking.krutou - Saturday, March 19, 2016 - link
Based on name99 and dano_spumoni's comments, this article should have been named and focued on "Peak Power Consumption Measurement..."I've had similar connection issues when running a portable SSD via front panel USB.
Murloc - Sunday, March 20, 2016 - link
just because you're not connecting external HDs to your phoneserendip - Friday, March 18, 2016 - link
How about USB hard drives? I know most won't power up because they need more than 500 mA on USB2 but it would be interesting to see exactly how much power is used. I've gotten USB hard drives to work with phones and tablets over OTG but only with a powered hub or a charging hub that allows a battery pack to be connected over micro-USB.jbrizz - Sunday, March 20, 2016 - link
I can power a 320GB 7200 rpm USB 3.0 hdd from my 2012 Nexus 7 and Galaxy S5 with a USB 2.0 OTG cable. One would assume that if they can supply power across the bus they should be able to power anything that runs on a PC, 500mA is 500mA whether from a PC or phone.serendip - Monday, March 21, 2016 - link
That's weird, I've never been able to get 500 GB or 1 TB USB HDD working over OTG. The drive spins up, mounts for a second and then disconnects, probably because the peak current draw exceeds 500 mA. I looked at a bunch of spec sheets from WD and Seagate and all their drives had peak current in the 1 A or higher range. They work fine when plugged into powered hubs though.jasonelmore - Saturday, March 19, 2016 - link
$200 for this thing?!?! what are they smoking.. I'll just make a arduino version for $35.... Thanks for the idea tho.dano_spumoni - Saturday, March 19, 2016 - link
This explains so much!!!! I was having random disconnects with a USB 3.0 to M.2 Enclosure when using the front USB 3.0 ports on my Desktop. When I use the rear ports it works perfectly and I get 400 MB/s sequential read/writes... Normal USB thumb drives and powered drives work fine in front but not USB powered enclosures. I knew something was wrong with my front ports but this is the reason.alin - Saturday, March 19, 2016 - link
There are still people out there who care about power consumption for flash devices? Hmm. Very well then... :)woggs - Saturday, March 26, 2016 - link
I bought one of these to play with... And changed my mind. It's on ebay if anyone wants it.http://my.ebay.com/ws/eBayISAPI.dll?MyEbay&gbh...
woggs - Tuesday, April 19, 2016 - link
Still on ebay...http://www.ebay.com/itm/141961691354?ssPageName=ST...
epobirs - Sunday, March 27, 2016 - link
Interesting. I have a SIIG USB 3.1 enclosure that I put a Patriot Blast 480GB SSD in. The performance is very nice in that it gets very close tot he limits of SATA itself. The included cable has a second host port connector to enlist more power where needed but it seems to work fine on 3.0 on a single plug.Is anyone developing native USB 3.1 controllers for SSD class flash memory that would exceed the SATA envelope? This would be ideal for consoles like the PS4 if the USB host was upgraded to 3.1 in later model.
ggruk - Monday, April 19, 2021 - link
Hi GaneshAn exciting article from 2016. I read it yesterday and found very useful.
I am designing a new product - a special type of Monitoring Meter. Am using PIC32MM0256GPM064 in it as the main microcontroller. It comes with a USB OTG module. I then want to save data in the Meter to a USB flash drive (USB2.0). However, my meter cannot provide more than 100mA current to the USB flash drive. So, I was looking for a low power flash drive. You article shows that 128GB flash drive that you tested could be suitable. Have tested some of the many flash pen drives I have and measured the current drawn by using Windows Device Manager for the Rot Hub. One Data Traveler flash drive of 8GB showed 100mA. All others are showing 200 to 500mA. Is there a way to have a USB flash drive that would draw less than 100mA as drawn by HID devices. Would be grateful for your comments.
I am also thinking of an option where I provide two Micro USB sockets on the Meter. One to power up the meter, say from a Mobile Phone charging Power Bank (with only the +5V and GND connected and another Micro AB socket where the flash drive is connected. Then I hope the flash drive would be supplied with power by the Power bank and data would sent by my Meter. Is this possible?
Again would thank you very much for your comments and advice. ggrUK