Rebooted and now there's no 5V

Post Reply
ejolson
Posts: 31
Joined: Sat Feb 08, 2020 2:14 am
languages_spoken: english
Has thanked: 7 times
Been thanked: 9 times
Contact:

Rebooted and now there's no 5V

Post by ejolson »

I've been running an Odroid-N2 purchased around February 2020 no monitor, no keyboard, no mouse but only an SD card, network connection and a Sandisk 240GB SSD connected through a Eluteng USB to SATA bridge. I rebooted and it didn't come back up. I pulled the power, connected to a monitor and petite boot comes up when the SPI switch is flipped but there was no 5V on any of the USB ports. Among other things, this means there is no way to control petite boot with a USB keyboard. The red power light does not shine, but the blue sys light does when in petite boot and then starts blinking while waiting for a DHCP response.

I checked the voltage across C158 as suggested in the post

viewtopic.php?p=296172#p296172

and obtained 0.034V instead of 5V. For reference, this also seems to be the same problem discussed in the thread

viewtopic.php?f=181&t=35553

There was never anything connected to GPIO, the system was in the official Odroid-N2 case and powered by the official power supply.

I purchased direct from Odroid less than 6 months ago, is there any warranty?

What are my options?

ejolson
Posts: 31
Joined: Sat Feb 08, 2020 2:14 am
languages_spoken: english
Has thanked: 7 times
Been thanked: 9 times
Contact:

Re: Rebooted and now there's no 5V

Post by ejolson »

I found the FAQ

https://www.hardkernel.com/faq/

where it says the warranty is 4 weeks. That seems to rule out any sort of RMA, so it's on to asking about do-it-yourself repairs.

Would it be possible to feed the needed 5V in through the GPIO header?

I noticed in the schematic that output from the 3V3 supply is used to raise the enable pin on the 5V supply. Is this power sequencing important or could I just wire in a 5V buck converter that has the enable pin already tied high?

In particular, one idea is to draw 12V from the barrel connector to feed a tiny MP2315-based buck-converter module and then send the output to the 5V pin on the GPIO or possibly the 5V side of C158. Alternatively, I could do the same but additionally wire up the enable pin on the buck converter to the 3V3 on the GPIO [edit: actually to R48] to approximate the same power sequencing as the original. Are either of these ideas likely to work?

Do I need to remove any part of the broken 5V power supply that was built into the Odriod-N2 before connecting the external buck converter? Is there an easy place to cut the trace for the 5V circuit leading from the internal supply?

Has anyone done this? Is there a better way to attempt a repair?
Last edited by ejolson on Mon Jul 27, 2020 3:11 am, edited 1 time in total.

mad_ady
Posts: 8316
Joined: Wed Jul 15, 2015 5:00 pm
languages_spoken: english
ODROIDs: XU4, C1+, C2, C4, N1, N2, H2, Go, Go Advance
Location: Bucharest, Romania
Has thanked: 573 times
Been thanked: 434 times
Contact:

Re: Rebooted and now there's no 5V

Post by mad_ady »

These users thanked the author mad_ady for the post:
ejolson (Sun Jul 26, 2020 3:37 am)

ejolson
Posts: 31
Joined: Sat Feb 08, 2020 2:14 am
languages_spoken: english
Has thanked: 7 times
Been thanked: 9 times
Contact:

Re: Rebooted and now there's no 5V

Post by ejolson »

mad_ady wrote:
Sat Jul 25, 2020 3:53 pm
See if this applies to you: https://wiki.odroid.com/odroid-n2/hardware/repairs
Thanks for the link. I already looked at that and understood it applies to 3.3V not working. The problem with mine and the posts I linked appears lack of 5V.

In particular, I can boot from SPI into petite boot, which I don't think would be possible if the 3.3V supply were out.

It's possible I'm mistaken and I definitely appreciate help with this. While it is curious that I get video output and the petite boot prompt without 5V present on the board, it is also frustrating to get that far because neither the SD card nor the network connection work after that.

User avatar
joerg
Posts: 1130
Joined: Tue Apr 01, 2014 2:14 am
languages_spoken: german, english, español
ODROIDs: C1, C1+, C2, N1, N2, C4
Location: Germany
Has thanked: 53 times
Been thanked: 122 times
Contact:

Re: Rebooted and now there's no 5V

Post by joerg »

@ejolson, before you apply external 5V to the 5V header, please check with a meter if there is no short circuit.
As reference, look this part of circuit that makes the VCC5V:
Bildschirmfoto vom 2020-07-25 17-29-18.png
Bildschirmfoto vom 2020-07-25 17-29-18.png (33.56 KiB) Viewed 610 times
If this step down contrioller is broken, it could be that the output part is short circuited. In this case the controller (I think it's designation is U12) must be removed, before you apply external 5V.
The 5V_EN is a GPIO that comes from CPU. If you have a meter you can also check if there is a high level, best at R48. But please be careful not to do a short circuit with the meter probes.
These users thanked the author joerg for the post:
ejolson (Sun Jul 26, 2020 3:36 am)

ejolson
Posts: 31
Joined: Sat Feb 08, 2020 2:14 am
languages_spoken: english
Has thanked: 7 times
Been thanked: 9 times
Contact:

Re: Rebooted and now there's no 5V

Post by ejolson »

joerg wrote:
Sun Jul 26, 2020 1:06 am
If this step down contrioller is broken, it could be that the output part is short circuited. In this case the controller (I think it's designation is U12) must be removed, before you apply external 5V.
The 5V_EN is a GPIO that comes from CPU. If you have a meter you can also check if there is a high level, best at R48. But please be careful not to do a short circuit with the meter probes.
Checking the signal from the CPU is a good idea and for shorts. I'll report back on that soon. It seems I'm also finally understanding the difference between VCC3V3 and VDDAO_3V3.

Do you think I will need to connect the enable from R48 to the external 5V buck converter later?

ejolson
Posts: 31
Joined: Sat Feb 08, 2020 2:14 am
languages_spoken: english
Has thanked: 7 times
Been thanked: 9 times
Contact:

Re: Rebooted and now there's no 5V

Post by ejolson »

Here is an update with measurements. With power off I have
  • 14.42 KOhms with red on VCC5V (pin 2 on GPIO connector) and black on ground.
    • 1.75 KOhms with black on VCC5V (pin 2 on GPIO connector) and red on ground.
    With the system powered on with the SPI switch set running petite boot, the blue light is blinking and
    • 3.35 V appears on VDDAO_3V3 (at the end of R48 toward the edge of the board).
      • 2.52 V appears on 5V_EN (at the other end of R48).
        • 3.33 V appears on VCC3V3 (pin 1 on GPIO connector).
          • 0.042 V appears on VCC5V (pin 2 of GPIO connector).
          Do you think it will work if I feed some 5V into GPIO pin 2?

          Given the soldering equipment available here, it looks like it might be difficult to connect 5V_EN to the external buck converter. I wonder if R48 is the only place to tap it. Would that even be necessary?

          User avatar
          joerg
          Posts: 1130
          Joined: Tue Apr 01, 2014 2:14 am
          languages_spoken: german, english, español
          ODROIDs: C1, C1+, C2, N1, N2, C4
          Location: Germany
          Has thanked: 53 times
          Been thanked: 122 times
          Contact:

          Re: Rebooted and now there's no 5V

          Post by joerg »

          It seems that there is no short circuit on the VCC5V lane. But the 2.52V seems to be odd. When I measure on my N2+ at R48 I see 0V on the 5V_EN side. I wonder if this is a hardware or a software problem that the GPIOH_8 does not drive low. But if there is really no short circuit it might be safe to feed 5V from external power supply.

          ejolson
          Posts: 31
          Joined: Sat Feb 08, 2020 2:14 am
          languages_spoken: english
          Has thanked: 7 times
          Been thanked: 9 times
          Contact:

          Re: Rebooted and now there's no 5V

          Post by ejolson »

          Now I'm confused. Is the 5V_EN active when low or when high?

          From the data sheet
          Monolithic Power wrote: Buck enable. EN is a digital input that turns the buck regulator on or off. When the power supply of the control circuit is ready, drive EN high to turn on the buck regulator, and drive it low to turn off the buck regulator. Connect EN to 3V3 through a pull-up resistor or a resistive voltage divider for automatic start-up. Do NOT float this pin. EN can be used to set USM. When EN is in the range of 1.4 V to 1.8 V, it enters USM. If EN is in the range of 2.6V to 3.6V, it operates in normal mode.
          https://www.monolithicpower.com/en/docu ... t_id/4030/

          It doesn't seem to be quite the right part, but maybe it's close enough for helping to understand how things might work.

          User avatar
          joerg
          Posts: 1130
          Joined: Tue Apr 01, 2014 2:14 am
          languages_spoken: german, english, español
          ODROIDs: C1, C1+, C2, N1, N2, C4
          Location: Germany
          Has thanked: 53 times
          Been thanked: 122 times
          Contact:

          Re: Rebooted and now there's no 5V

          Post by joerg »

          Hello @ejolson, you might be right with that (5V_EN). I measured on my N2+ and it could be that there this part is different. I can't find a schematic for N2+. My N2 is in use and it's difficult for me to measure there. Need to ask Hardkernel hardware guy how it works.

          ejolson
          Posts: 31
          Joined: Sat Feb 08, 2020 2:14 am
          languages_spoken: english
          Has thanked: 7 times
          Been thanked: 9 times
          Contact:

          Re: Rebooted and now there's no 5V

          Post by ejolson »

          joerg wrote:
          Sun Jul 26, 2020 10:18 pm
          Hello @ejolson, you might be right with that (5V_EN). I measured on my N2+ and it could be that there this part is different. I can't find a schematic for N2+. My N2 is in use and it's difficult for me to measure there. Need to ask Hardkernel hardware guy how it works.
          That's a good idea. I agree that 2.52 V seems like a weird value to send to the EN line on the NB679 buck converter. According to the datasheet I found, that's too low to be enabled and too high to enter USM. I'm pretty sure 0V would not lead to any 5V output, so suspect that to be a measurement error. Those resistors are very small!

          ejolson
          Posts: 31
          Joined: Sat Feb 08, 2020 2:14 am
          languages_spoken: english
          Has thanked: 7 times
          Been thanked: 9 times
          Contact:

          Re: Rebooted and now there's no 5V

          Post by ejolson »

          Here is an update. Abandoning all caution I wired up a buck converter to the 12V plug, made sure it was calibrated for 5V under load and the results were
          • If the 5V output of the buck converter is connected to VCC5V on pin 2 of the GPIO header the red led flashes when the Odroid is plugged in and then goes out again. Maybe there's a short circuit when EN_5V is raised.
            • If the N2 is booted into petite boot and the 5V output of the buck converter is hot-plugged to VCC5V on pin 2 of the GPIO header, the screen goes black, though the blue light continues to flash.
            I obviously didn't perform the above tests very many times, but they are repeatable.

            Miraculously, the N2 appears no more broken than before and still boots into petite boot when the buck converter is not attached, so I'm considering what to do next.

            My ideas are
            • Try to remove the on-board 5V converter under the assumption that it shorts out the external buck converter. How would I do this without decent soldering equipment? Do you think a chisel would work?
              • Find a place to cut the EN_5V trace that goes to the on-board NB679 converter so there is no need to pry it off the board. Maybe I could tap the EN_5V signal and hook it up to the external buck converter.
              Before attempting either, it seems reasonable to wait, in case someone replies here with a better idea or at least helps me decide between the above two. Could it happen that one of the Odroid engineers would take an interest in this and provide some advice? Does anyone have some advice?

              ruppi
              Posts: 126
              Joined: Fri Feb 22, 2013 11:13 pm
              languages_spoken: korean
              Location: South Korea
              Has thanked: 0
              Been thanked: 12 times
              Contact:

              Re: Rebooted and now there's no 5V

              Post by ruppi »

              Hi ejoson,

              If you look at your board, it looks like 100% NB679GD has been destroyed.
              Perhaps the internal FET of the NB679GD is destroyed, and if you can remove the relatively easy L2 (inductor) than your NB679GD, you can eliminate the short circuit of the 5V output.

              Also, the EN_5V trace does not cause major problems in the power sequence without the need for a local connection, so the pull-up resistance using VDDAO_3V3 power is also acceptable for the enable pin on the external Buck converter.

              I wish you good results.

              -BR-
              Ruppi
              These users thanked the author ruppi for the post (total 2):
              joerg (Mon Jul 27, 2020 8:17 pm) • ejolson (Tue Jul 28, 2020 8:28 am)

              User avatar
              joerg
              Posts: 1130
              Joined: Tue Apr 01, 2014 2:14 am
              languages_spoken: german, english, español
              ODROIDs: C1, C1+, C2, N1, N2, C4
              Location: Germany
              Has thanked: 53 times
              Been thanked: 122 times
              Contact:

              Re: Rebooted and now there's no 5V

              Post by joerg »

              @ruppi, thank you for clarification. But one more question. I like to know why I measure 0V on the EN_5V? Is it not measured well or is this part different on the N2+? An updated schematic would be fine on the wiki.

              ejolson
              Posts: 31
              Joined: Sat Feb 08, 2020 2:14 am
              languages_spoken: english
              Has thanked: 7 times
              Been thanked: 9 times
              Contact:

              Re: Rebooted and now there's no 5V

              Post by ejolson »

              ruppi wrote:
              Mon Jul 27, 2020 4:40 pm
              Perhaps the internal FET of the NB679GD is destroyed, and if you can remove the relatively easy L2 (inductor) than your NB679GD, you can eliminate the short circuit of the 5V output.
              Without a hot-air station, the inductor looks difficult to remove as well. Since I'm not so interested in salvaging the part, would it be possible to simply cut the wire that I see at the base on each side of it?

              There will be some delay as the 12V battery ran out on the cheap multi-meter I have at home and I have not finished (or started) the coronavirus safety training necessary to go back to the office. It looks more and more like work from home during the coming months anyway.

              Some new batteries should arrive by post in a few days, at which point I'll give it a go to cut the wire. Currently at my disposal is a meager assortment of soldering irons, tweezers, an X-acto knife and a couple strong looking needle-nose pliers. If there weren't so much other stuff in the way, I'd try two soldering irons on each side to lift the NV679GD off the board as suggested in the repair guide for the MOSFET. As it is, I'm hoping the knife will brutally transform VCC5V into an open circuit.

              Does anyone have other suggestions or advice?
              Last edited by ejolson on Tue Jul 28, 2020 11:21 am, edited 1 time in total.

              ruppi
              Posts: 126
              Joined: Fri Feb 22, 2013 11:13 pm
              languages_spoken: korean
              Location: South Korea
              Has thanked: 0
              Been thanked: 12 times
              Contact:

              Re: Rebooted and now there's no 5V

              Post by ruppi »

              joerg wrote:
              Mon Jul 27, 2020 8:17 pm
              @ruppi, thank you for clarification. But one more question. I like to know why I measure 0V on the EN_5V? Is it not measured well or is this part different on the N2+? An updated schematic would be fine on the wiki.
              N2 and N2+ do not differ in power circuits, and perhaps the EN_5V measurement was difficult to measure.
              As far as I know, there is no control of EN_5V in the bootloader or kernel, only full up at 3.3V.

              ruppi
              Posts: 126
              Joined: Fri Feb 22, 2013 11:13 pm
              languages_spoken: korean
              Location: South Korea
              Has thanked: 0
              Been thanked: 12 times
              Contact:

              Re: Rebooted and now there's no 5V

              Post by ruppi »

              ejolson wrote:
              Tue Jul 28, 2020 9:27 am
              ruppi wrote:
              Mon Jul 27, 2020 4:40 pm
              Perhaps the internal FET of the NB679GD is destroyed, and if you can remove the relatively easy L2 (inductor) than your NB679GD, you can eliminate the short circuit of the 5V output.
              Without a hot-air station, the inductor looks difficult to remove as well. Since I'm not so interested in salvaging the part, would it be possible to simply cut he wire that I see at the base on each side of it?

              There will be some delay as the 12V battery ran out on the cheap multi-meter I have at home and I have not finished (or started) the coronavirus safety training necessary to go back to the office. It looks more and more like work from home during the coming months anyway.

              Some new batteries should arrive by post in a few days, at which point I'll give it a go to cut the wire. Currently at my disposal is a meager assortment of soldering irons, tweezers, an X-acto knife and a couple strong looking needle-nose pliers. If there weren't so much other stuff in the way, I'd try two soldering irons on each side to lift the NV679GD off the board as suggested in the repair guide for the MOSFET. As it is, I'm hoping the knife will brutally transform VCC5V into an open circuit.

              Does anyone have other suggestions or advice?
              If you can turn off the wire at the end of the inductor, you can get the same result as removing the inductor.

              ejolson
              Posts: 31
              Joined: Sat Feb 08, 2020 2:14 am
              languages_spoken: english
              Has thanked: 7 times
              Been thanked: 9 times
              Contact:

              Re: Rebooted and now there's no 5V

              Post by ejolson »

              Woohoo! It's working now. I snipped the inductor to be sure there were no shorts. That may have been unnecessary as there was also a problem with how I had originally wired the ground on the external buck converter: Surprisingly the shell of the 12V barrel connector is not fully grounded. The final repair looks like

              Image

              The wires are relatively heavy gauge and suspend the buck converter above the main board more securely than it looks. Note the telltale blob of solder that shows where the white-brown striped wire used to be connected at the far right side of the picture.

              It seems to be working fine. Now I'm trying to decide why it broke and how to avoid additional breakage in the future.
              • Was it just the bathtub curve?
                • Did the SSD draw too much power?
                  • Does the case need ventilation?
                  Even though it unexpectedly broke, I'm still quite impressed with the speed of the hardware and the help to get it working again. Thanks!
                  These users thanked the author ejolson for the post (total 3):
                  tobetter (Thu Jul 30, 2020 12:37 pm) • odroid (Thu Jul 30, 2020 1:16 pm) • joerg (Thu Jul 30, 2020 2:17 pm)

                  mad_ady
                  Posts: 8316
                  Joined: Wed Jul 15, 2015 5:00 pm
                  languages_spoken: english
                  ODROIDs: XU4, C1+, C2, C4, N1, N2, H2, Go, Go Advance
                  Location: Bucharest, Romania
                  Has thanked: 573 times
                  Been thanked: 434 times
                  Contact:

                  Re: Rebooted and now there's no 5V

                  Post by mad_ady »

                  That's great! Hopefully the fix gets documented in the wiki.

                  jandre
                  Posts: 1
                  Joined: Mon Aug 03, 2020 5:05 am
                  languages_spoken: english
                  ODROIDs: N2
                  Has thanked: 0
                  Been thanked: 0
                  Contact:

                  Re: Rebooted and now there's no 5V

                  Post by jandre »

                  Experienced exactly the same hardware fault. Bought my Odroid N2 from a retailer in South Africa in the beginning of April. Plugged the board in one morning and it was dead (5V line that is). Only a mouse + keyboard and HDMI attached. Initially tried the fix with linear voltage regulator: It was a mistake -> Too much current drawn and it overheated... Going to get a buck converter like @ejolson and retry! Hope it works :D

                  Post Reply

                  Return to “Issues”

                  Who is online

                  Users browsing this forum: No registered users and 0 guests