XU4 USB output 4.8V

Post Reply
DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

XU4 USB output 4.8V

Post by DHGE » Fri Aug 25, 2017 2:47 am

Today I started with a new XU4 to run under OpenMediaVault.
I tested the same Harddisk in an identical setup on a Raspberry Pi3.

I attached a 4000GB Seagate Backup Plus Portable Drive STDR4000200 2.5" to the USB3 port(s) of my XU4

After 15 to 20 seconds a file transfer stops.
I get uas_eh_abort_handle in the logs - reading the forum this is a symptom of undervoltage.
4.8V.JPG
4.8V.JPG (393.74 KiB) Viewed 3039 times
Even with nothing attached I get at best 4.9V at the USB3 ports. I have measured as low as 4.6V.
The discs consumes less than 0.6A.

I have a Hardkernel 5V 6A PSU.

Unfortunately I can not test another PSU here since I do not have on with this plug.

I expect the Voltage to be 5V. On the Raspi it is 5.05V

What do you think: Did I receive a faulty XU4 or 6A-PSU?

phaseshifter
Posts: 3535
Joined: Fri May 08, 2015 9:12 am
languages_spoken: english
ODROIDs: U-2,U3+,,XU-3,,XU3-LITE,,XU-4
C1+,,C-2,,,
N-1,,N-2,...other odroid acc`s as well
Has thanked: 45 times
Been thanked: 45 times
Contact:

Re: XU4 USB output 4.8V

Post by phaseshifter » Fri Aug 25, 2017 4:49 am

i dont know what image you are using but kernel 4.9 fixed many xu4 usb 3 port issues


monitor pin 1 of the 30 pinn header con 10 to see if the voltage there is low...
https://wiki.odroid.com/odroid-xu4/hard ... connectors

if you find this to be ok then something else may be problematic...

check the output voltage of the power supply stand alone ...???

if you are still having issues you can try this
viewtopic.php?f=95&t=15302#p99467

if you still have issues you can try this
viewtopic.php?f=99&t=25813&p=188269&hil ... ng#p188269
Build It And They Will Come...Be Bold And Mighty Forces Will Come To Your Aid..!!!

User avatar
odroid
Site Admin
Posts: 34109
Joined: Fri Feb 22, 2013 11:14 pm
languages_spoken: English, Korean, Japanese
ODROIDs: ODROID
Has thanked: 666 times
Been thanked: 611 times
Contact:

Re: XU4 USB output 4.8V

Post by odroid » Fri Aug 25, 2017 5:02 am

Update the kernel first with following steps.
https://wiki.odroid.com/odroid-xu4/os_i ... st-upgrade

Please also note those USB voltage/ampere meters are not accurate and the voltage/current has almost 5~10% tolerance error.

DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

Re: XU4 USB output 4.8V

Post by DHGE » Fri Aug 25, 2017 6:17 am

@phaseshifter

tried to clean the contacts

Linux xu4 4.9.37-odroidxu4 #4 SMP PREEMPT Wed Jul 26 13:07:46 PDT 2017 armv7l GNU/Linux

PSU: 5.1+V
Pin 1 dito, when trying to measure it with the harddisk attached my XU4Q died :x

@odroid
your link is about Lineage OS??

So now I am here with a dead XU4.
Should I send it back to Hardkernel?

User avatar
odroid
Site Admin
Posts: 34109
Joined: Fri Feb 22, 2013 11:14 pm
languages_spoken: English, Korean, Japanese
ODROIDs: ODROID
Has thanked: 666 times
Been thanked: 611 times
Contact:

Re: XU4 USB output 4.8V

Post by odroid » Fri Aug 25, 2017 7:10 am

I've corrected the link.

Remove all the external devices and try to boot again.
Let me know the status of red and blue LEDs.
How did you fry the XU4 board?
Normally, it has been almost impossible to repair XU4 board.

adeuring
Posts: 46
Joined: Sun Jul 09, 2017 9:50 pm
languages_spoken: english
Has thanked: 0
Been thanked: 1 time
Contact:

Re: XU4 USB output 4.8V

Post by adeuring » Fri Aug 25, 2017 9:46 am

This forum has several threads about power issues in conjunction with hard disks connected to a USB3 port; this new question is perhaps a good occasion to share my own experiences that led me to three conclusions:
  • Use a decent power supply.
  • Connect the 5V pins of the SATA connector directly with the with the power supply.
  • Ensure that that power cabling is good.
The setup: An XU4Q, its shifter shield, a prototype board with some I2C toys and a hard disk connected via USB3/SATA bridge. (Sorry, I can't even remember the brand name. But it simply works.)

I can't comment on the power supply recommended by HK – I live in Germany and the German distributor does not sell HK's power supplies,I believe, so I started with a "wall wart" from Mean Well, type GS25E05-P1J. Declared as a 5V, 4A supply. This worked mostly with the usual cabling but not very well: The disks (I tried two 2.5" models, a Hitachi Z7K320 and a 2TB Seagate whose exact name I can't read ATM since it's mounted below the XU4) made quite often "strange noise" that indicated "undervoltage". Occasionally the XU4 would refuse to boot (Ubuntu installed on the disk) because it even could not read the partition table.

So I measured the voltage at the 5V pin of the shifter shield and saw sometimes pretty low values, while the voltage at the power jack of the XU4 looked better. I can't remember specific values, but below are results from a test I ran today.

My conclusion: Avoid the voltage drop that occurred somewhere on the XU4 board, probably in the "reverse power supply protection", and connect the the GND and 5V pins of the SATA connector directly with the the power supply. I soldered some screw terminals onto a prototype board, soldered them together to two blocks (using 0.75mm² wire), to connect 5V and GND wires between power supply, XU4 and the SATA connector of the USB/SATA bridge.

That worked pretty good, but there were still problems: the first turned out to be the screw terminals who did not make very good contact to the wires, so I replaced them with Wago 236 terminals. This setup worked really fine under "normal conditions", i.e. not a too heavy load fot the disk and the processor, but with some higher loads the disks still tended to make noise that indicates undervoltage.

Next step: I replaced the "wall wart" power supply with a Mean Well RT-50C that I had lying around. This seems to work quite well, but I intend to replace it with a somewhat larger model. It arrived today (Mean Well RD-85A; might even be able to power two XU4s and two disks), but I did not try it yet.

Now for the test I ran today. The test was triggered by this thread, aside from DHGE's question itself by phaseshifter's recommendation to check the power supply voltage at different points.

So I disconnected the 5V and GND cables that connect the disk directly with the RT-50C power supply from the terminal block and measured the voltage at the terminal block, the 5V pin of the shifter shield and at the disconnected cables of the SATA connector, i.e, the voltage supplied via the Odroid itself. I used a DM620 multimeter which has the nice feature to record minimum and maximum values.

The test sequence was very simple:
- let the XU4 boot the Ubuntu minimal image installed on the disk (with some services like Apache installed),
- press the "reset min/max values" button of the DMM when uboot messages are visible at the UART
- write down the min/max voltage when the login prompt is displayed,
- run this test three times for the voltage at the terminal block, at the 5V pin of 40-pin connector of the shifter shield and for the voltage at the SATA connector.

The results:

1. Terminal block:

Code: Select all

min     max
5.01V   5.20V
5.00V   5.21V
5.00V   5.21V
2. Shifter shield:

Code: Select all

min     max
3.71V   5.21V
3.77V   5.15V
3.80 V  5.10V
2. SATA connector:

Code: Select all

min     max
3.80V   5.06V
3.72V   5.06V
3.79V   5.10V
Then I reconnected the wires from the SATA connector to the terminal block again and measured the voltage at the 5V pin of the shifter shield:

Code: Select all

min     max
4.96V   5.17V
4.94V   5.20V
4.95V   5.21V
Conclusion: There is a considerable voltage drop on the XU4 board that makes it unsuitable to power a hard disk. SSDs should work better, but problems might arise there too.

Note, BTW, that the boot process never got stuck, even though the disk was at least shortly underpowered during each of the 9 boots where the XU4 provided power to the disk. But this is clearly not a reliable setup.

And finally a warning: The setup I am currently using – an industrial power supply not fitted into a proper housing – is something for the "home lab" but not suitable for the living room. I would not let get children (or unsuspecting adults ;) near to a barely shielded 220V terminal.

The power supplies sold by HK may do their job to power hard disks in general, but the XU4 itself should not be involved in the current path between power supply and disk.

Another, living room compatible, option would be to use a USB3 hub with it its own power supply. I tried that, but got one with a a too small power supply.

Finally, I wondered if it might help to connect a really fat capacitor, 10 mF or so, to the power pins of the USB3 connector. After all, a disk needs a higher current not constantly. It would be a cheaper and easier to implement. But I don't know how the XU4 board would cope with the current sucked by the capacitor when the board is powered on. But the image of a an old-fashioned little barrel full of chemicals helping a nifty modern high density SMT board is somewhat attractive to me ;)

DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

Re: XU4 USB output 4.8V

Post by DHGE » Fri Aug 25, 2017 4:08 pm

Wow - thanks A LOT for your effort!
adeuring wrote:This forum has several threads about power issues in conjunction with hard disks connected to a USB3 port;
  • Use a decent power supply.
  • Connect the 5V pins of the SATA connector directly with the with the power supply.
  • Ensure that that power cabling is good.
The setup...

I can't comment on the power supply recommended by HK – I live in Germany and the German distributor does not sell HK's power supplies,I believe, so I started with a "wall wart" from Mean Well, type GS25E05-P1J. Declared as a 5V, 4A supply. This worked mostly with the usual cabling but not very well: The disks (I tried two 2.5" models, a Hitachi Z7K320 and a 2TB Seagate whose exact name I can't read ATM since it's mounted below the XU4) made quite often "strange noise" that indicated "undervoltage". Occasionally the XU4 would refuse to boot (Ubuntu installed on the disk) because it even could not read the partition table.
...

The test sequence was very simple:
- let the XU4 boot the Ubuntu minimal image installed on the disk (with some services like Apache installed),
- press the "reset min/max values" button of the DMM when uboot messages are visible at the UART
- write down the min/max voltage when the login prompt is displayed,
- run this test three times for the voltage at the terminal block, at the 5V pin of 40-pin connector of the shifter shield and for the voltage at the SATA connector.

The results:

1. Terminal block:

Code: Select all

min     max
5.01V   5.20V
2. Shifter shield:

Code: Select all

min     max
3.71V   5.21V
2. SATA connector:

Code: Select all

min     max
3.80V   5.06V
Then I reconnected the wires from the SATA connector to the terminal block again and measured the voltage at the 5V pin of the shifter shield:

Code: Select all

min     max
4.96V   5.17V
Conclusion: There is a considerable voltage drop on the XU4 board that makes it unsuitable to power a hard disk. SSDs should work better, but problems might arise there too.

Note, BTW, that the boot process never got stuck, even though the disk was at least shortly underpowered during each of the 9 boots where the XU4 provided power to the disk. But this is clearly not a reliable setup.
Well my unscientific measures only "saw" 4.6V on the crappy USB-meter but the logs speak (UAS issues) for themselves.
I am not into soldering my just arrived XU4Q.
Do you know what Pollin sells as PSU in their bundles? They do not list part numbers.
odroid wrote:I've corrected the link.

Remove all the external devices and try to boot again.
Let me know the status of red and blue LEDs.
How did you fry the XU4 board?
Thanks - this link leads to an older kernel than I got via OMV website, latest version and up to date.
I also see the boot parameter for disabling UAS. A procedure @tkaiser vented against as this would only kill the detection of symptoms of undervoltage and not go against the root cause.

Only red status LED available any more. I also put back the SD card as boot device but also only the red LED came up.

How ... fry? I do not know. Measuring with my multimeter Pin1 to Pin2 and hitting another pin by accident?

adeuring
Posts: 46
Joined: Sun Jul 09, 2017 9:50 pm
languages_spoken: english
Has thanked: 0
Been thanked: 1 time
Contact:

Re: XU4 USB output 4.8V

Post by adeuring » Fri Aug 25, 2017 7:45 pm

DHGE wrote:
Well my unscientific measures only "saw" 4.6V on the crappy USB-meter but the logs speak (UAS issues) for themselves.
I am not into soldering my just arrived XU4Q. Do you know what Pollin sells as PSU in their bundles? They do not list part numbers.
You don't need to solder anything on the XU4 itself. I did not do that either. What I did was to add wires that connect the power supply directly with the disk.

A "tinker free" solution would be to use a USB3/SATA bridge that has its own power connector and a reliable power supply. I am not sure if such an enclosure exists, since USB3 allows somewhat higher 5V currents that should be sufficient for many 2.5" disks. An alternative would be, as I already wrote, a USB3 hub with a good power supply. And if the hub can switch off the power of the output ports, the switching Mosfets should have really low "on" resistance.

The big challenge is to find a reliable off-the-shelf power supply for a disk enclosure or a USB3 hub. I tried a hub that was sold with a power supply that claims to deliver up to 2A, but that did not solve the problem.

I am meanwhile a bit wary of the little "wall wart" power supplies: Disks tend to need higher peak currents. A larger capacitor could provide that, but you can't squeeze a larger capacitor into these small housings.

Another option might be a "SATA docking station" like this one: https://en.sharkoon.com/product/1686/14057#desc They tend to be sold with a power supply, so the main risk here is that the manufacturer provides one that is too small. I have a similar one and had no problems with it – but I did not yet try it with the XU4, only with laptops. But I don't see a reason why they should not work with the XU4.

Regarding the power supply from Pollin's bundles: I don't know what the sell. It may very well the one from HK – I bought a "plain" XU4Q from Pollin. While the price of the bundles is not bad at all, I simply did not need the case for the XU4 or the EMMC module, and then the bundle price does no longer look so good. And I believe that Pollin does not sell HK's power supplies separately. (But I may very well be wrong here – I am not that good in finding stuff in web shops ;)

DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

Re: XU4 USB output 4.8V

Post by DHGE » Fri Aug 25, 2017 8:19 pm

adeuring wrote: A "tinker free" solution would be to use a USB3/SATA bridge that has its own power connector and a reliable power supply. I am not sure if such an enclosure exists, since USB3 allows somewhat higher 5V currents that should be sufficient for many 2.5" disks. An alternative would be, as I already wrote, a USB3 hub with a good power supply. And if the hub can switch off the power of the output ports, the switching Mosfets should have really low "on" resistance.

Another option might be a "SATA docking station" like this one: https://en.sharkoon.com/product/1686/14057#desc They tend to be sold with a power supply, so the main risk here is that the manufacturer provides one that is too small. I have a similar one and had no problems with it – but I did not yet try it with the XU4, only with laptops. But I don't see a reason why they should not work with the XU4.
Yeah - I also tried a Wetek (recommended for my Vero4k - my only problemfree SBC) powered USB3 hub. Did not help either.
Transferring large files (virtual devices here) was hit and miss, mostly stopping after about 15 to 20 seconds. Computing checksums after the occasionally successful transfer was not better - no surprise here. The good thing here: the checksums that made it were all correct. Something that can not be said for my first Raspberry Pi3 setup with the same Seagate disc. But this I got rectified by getting a new SD card and the "official" 2A Raspberry PSU. I have the official 6A for the XU4Q ...

I have a SATA Dockingstation but no SATA-disks lying aroound. I have put them all into USB housings and use them as rotating backup drives for my 3 PCs for up to 9 years now.

I was thinking about reducing their use with a very low power solution in form of an XU4/OMV-server and syncthing. But the XU4 does not work reliable with USB3 I conclude now.
I did not want to get a 3.5" housing with power supply because of the added power consumption of such a setup.

My XU4Q is dead anyway.
Pollin's listing of the XU4Q came up 3 days after I ordered it in South Korea - I am not sure if I want do deal with Hardkernel considering the courier costs ...
Maybe I should have gotten a HP Proliant for <200 EUR, put in one 2,5" disk in that and from time to time hook another to USB for backup. But this would have been 10+W average and +90 EUR up front and more space/noise.

I got lured by the perspective of a small quiet solution with less than 5W consumption on average.

adeuring
Posts: 46
Joined: Sun Jul 09, 2017 9:50 pm
languages_spoken: english
Has thanked: 0
Been thanked: 1 time
Contact:

Re: XU4 USB output 4.8V

Post by adeuring » Fri Aug 25, 2017 8:46 pm

DHGE wrote:I also see the boot parameter for disabling UAS. A procedure @tkaiser vented against as this would only kill the detection of symptoms of undervoltage and not go against the root cause.

Only red status LED available any more. I also put back the SD card as boot device but also only the red LED came up.

How ... fry? I do not know. Measuring with my multimeter Pin1 to Pin2 and hitting another pin by accident?
Don't panic ;)

Did you perhaps inadvertently hit the boot mode switch?

Are you sure that you inserted the SD card properly?

When the SD card is inserted into a card reader connected to a laptop/PC: Is it readable at all and does it show two partitions, a smalll FAT partitonen with boot stuff and a large ext4 partition?

If not, write the OS image again to the card and start over.

adeuring
Posts: 46
Joined: Sun Jul 09, 2017 9:50 pm
languages_spoken: english
Has thanked: 0
Been thanked: 1 time
Contact:

Re: XU4 USB output 4.8V

Post by adeuring » Fri Aug 25, 2017 9:07 pm

DHGE wrote:Yeah - I also tried a Wetek (recommended for my Vero4k - my only problemfree SBC) powered USB3 hub. Did not help either.
Transferring large files (virtual devices here) was hit and miss, mostly stopping after about 15 to 20 seconds. Computing checksums after the occasionally successful transfer was not better - no surprise here. The good thing here: the checksums that made it were all correct. Something that can not be said for my first Raspberry Pi3 setup with the same Seagate disc. But this I got rectified by getting a new SD card and the "official" 2A Raspberry PSU. I have the official 6A for the XU4Q ...

I have a SATA Dockingstation but no SATA-disks lying aroound. I have put them all into USB housings and use them as rotating backup drives for my 3 PCs for up to 9 years now.
Get a cheap used 2.5" disk from a local dealer or via Ebay's Kleinanzeigen or so for a few Euro and try how the setup works.
I was thinking about reducing their use with a very low power solution in form of an XU4/OMV-server and syncthing. But the XU4 does not work reliable with USB3 I conclude now.
I did not want to get a 3.5" housing with power supply because of the added power consumption of such a setup.
Sorry, now I have start a bit of nitpicking ;): USB3 devices work in general pretty well with the XU4, it's only the USB power supply that can be problematic.
My XU4Q is dead anyway.
See my other post: I am not yet convinced that your XU4 is dead. It should be able to survive at least a short short-circuit.
Pollin's listing of the XU4Q came up 3 days after I ordered it in South Korea - I am not sure if I want do deal with Hardkernel considering the courier costs ...
Maybe I should have gotten a HP Proliant for <200 EUR, put in one 2,5" disk in that and from time to time hook another to USB for backup. But this would have been 10+W average and +90 EUR up front and more space/noise.

I got lured by the perspective of a small quiet solution with less than 5W consumption on average.
Well, I'd still recommend to make some more attempts to get your XU4 running again. But I understand that it might not be the best solution when you are mainly looking for a NAS solution due to the hardware hassles in conjunction with disks.

DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

Re: XU4 USB output 4.8V

Post by DHGE » Fri Aug 25, 2017 9:09 pm

adeuring wrote: Don't panic ;)
Only sad ...

All is well in card land (both cards) and I hit the card switch with the intention to try the SD card.

User avatar
odroid
Site Admin
Posts: 34109
Joined: Fri Feb 22, 2013 11:14 pm
languages_spoken: English, Korean, Japanese
ODROIDs: ODROID
Has thanked: 666 times
Been thanked: 611 times
Contact:

Re: XU4 USB output 4.8V

Post by odroid » Sat Aug 26, 2017 12:47 am

Contact "odroid at hardkernel dot com" with a link of this thread.
She will help your RMA process if you purchased the board from us directly.

BTW, Any cables, wires, connectors and even protection ICs are causing the voltage drops.
It is the main reason why ODROID-HC1 was developed to minimize unknown resistance.

DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

Re: XU4 USB output 4.8V

Post by DHGE » Sat Aug 26, 2017 4:23 am

odroid wrote: BTW, Any cables, wires, connectors and even protection ICs are causing the voltage drops.
It is the main reason why ODROID-HC1 was developed to minimize unknown resistance.
Yes, I ordered one yesterday.

I will start the RMA process.

DHGE
Posts: 12
Joined: Fri Aug 25, 2017 2:23 am
languages_spoken: english german
ODROIDs: XU-4, hc1
Has thanked: 5 times
Been thanked: 0
Contact:

On the ODROID HC1

Post by DHGE » Fri Sep 01, 2017 7:00 am

This device arrived today.

I took the same SD card, liberated the harddisk (see above) from its case and plugged it to the SATA port.

All things that were giving me trouble with my XU4 went smoothly. No errors, file transfer via FTP to a Luks-encrypted partition nearly 80 MB/s.
md5sums OK.
Temperature measured via armbianmonitor maxes out at 80° C (XU4 went higher) and quickly goes down to 53° C after filetransfer and md5summinig.
Idle now:

Code: Select all

Time       big.LITTLE   load %cpu %sys %usr %nice %io %irq   CPU
23:58:34: 2000/ 600MHz  0.02  24%  15%   1%   0%   4%   2% 48.0°C
For the XU4 (if it comes back via RMA process) I bought a 3.5" harddisk (with own PSU) and will see if it fares better than supplying the 2.5" disk via USB3 port.

Post Reply

Return to “Issues”

Who is online

Users browsing this forum: No registered users and 1 guest