DE10-Nano Died After Accidently Unplugging SD Card

[twinkles76]

I accidently unplugged the sd card with the power on the de10 and now the system won't boot with either the mister sd card or the original sd card that came with the device. However, i have connected the de10 to my pc via usb and it is recognised in the device manager so it seems that the de10 is not totally destroyed.

Has anyone an idea as to what component may be damaged/able to be replaced ?

thanks in advance

[Bas]

Did you verify that the cards are both still 100% certainly ok?

[twinkles76]

yep sd cards are fine as they work in my other mister.

[beamrider]

Are you sure it's died? I thought mine had died a few times as it came on to a black screen, but when I unplugged from the power for 30 seconds and restarted, it worked okay.

Also prepare a new SD card and see if it boots with that.

[Neocaron]

I'd never say never, but it's very unlikely that it would die from this. 99% unlikely to be precise ^^
It's could be corruption of the data on both your sd cards or a problem with the sd cards themselves but your Mister should be fine.

[emmadness]

I would do the whole process from scratch: MicroSD correctly formatted and start the setup from scratch.

As you have been told, and I share, it looks like only the files were damaged.

The other day it was my turn to get scared since I turned on MiSTer and it did not give a video signal nor did the IO lights blink; it was that the MicroSD was connected to the PC ¬¬

Luck.

[Bas]

You may be able to get a reading on the state of your DE10 by attaching a USB to serial cable and reading the serial console output. It's likely that will tell you what the system thinks is the problem before it's willing to give you networking or a picture on your display. The DE10 Nano manual has instructions on how to do this.

[Malor]

At least in theory, you shouldn't be able to kill your hardware just by unplugging the SD card. They're designed to be removable, so you should not have been able to, say, put any wrong voltages anywhere. What probably did happen is that the filesystem was corrupted by the sudden removal; the Mister software is doing some really bizarre stuff with a loopback ext4 filesystem on a base vfat partition. That's pretty fragile, and data loss after sudden removal wouldn't shock me in any way.

If you rewrite your Mister card from scratch, and then plug it in correctly, your board should boot fine. Make sure you're not putting the card in upside down or something.

Now, if you actually damaged the SD socket, that could break the board, but you didn't do that, did you?

[amstan]

At least in theory, you shouldn't be able to kill your hardware just by unplugging the SD card.

Static could be a big one. I see there area some ESD diodes on the schematic but they're marked "Do Not Installl". Static on the other pins (maybe as OP was holding the board) could be another problem.

[amstan]

However, i have connected the de10 to my pc via usb and it is recognised in the device manager so it seems that the de10 is not totally destroyed.

Yeah, that just means one of the chips (an FTDI FT232, USB to Serial) still shows signs of life. The majority of your board could be dead (like if it was cut in half) and this would still work.

See chip labeled "UART-to-USB Controller": https://rocketboards.org/foswiki/pub/Do ... 1486394232

[twinkles76]

thanks for all the input on this

The sd card being corrupted is is definitely not the reason it will not boot.

in reply to Bas: after about 4 hours i worked out how to program the fpga chip from pc using JTAG. I'm not susre how to get any serial readings but it would appear that the fpga works fine ( uploaded various .sof files with no errors) and it is only the sd card function that is not working.

in reply to amstan : yes i looked at the schematic and to me ( no idea what im doing) it looks as the sd card is wired directly to the fpga chip.

I do know of other people accidently doing the same thing with the sd card ( unplugging with power on) and having the same problem, so if we could solve this it would be a great help to the community.

[twinkles76]

This problem is now fixed. It would appear that flashing of the fpga via jtag ( see manual) and then reinserting the linux/mister sd card fixes this problem.

[beamrider]

Good to hear!

Perhaps switching off before inserting/removing things going forward is a good idea...

(speaking of which I'll have to look out for a good PSU with a switch on it)

[twinkles76]

just in case someone else has the same problem:

to flash fpga on a windows machine

You will need

Quartus Prime: https://www.intel.com/content/www/us/en ... 53221.html

DE 10 Manual : https://www.mouser.co.uk/ds/2/598/E10-N ... 108056.pdf

DE 10 cd rom files and sd card image: https://www.terasic.com.tw/cgi-bin/page ... 6&PartNo=4

USB blaster driver: https://www.terasic.com.tw/wiki/Altera_ ... structions

1. disconnect any cables from de10.
2. close the loop of the JTAG chain (alows programming) by bridging the two pins as shown on page 13 of the manual.
3. Set msel dipswitches to 10010 ( the on position is 0 and the off position is 1 ) as shown on page 107 of the manual.
4. install usb-blaster driver ( these may be done as part of the quartus installation)
5. Attach mini usb to J13 and the other end to pc.
6. connect power cable to de10.
7. follow the instructions on page 14 of the manual.

If done correctly the leds on the de10 should be flashing.

8. disconnect de10 and set MSEL dip switches to 010100.
9. write sd card image to sd card
10. Plug in sd card and connect hdmi to monitor
11. power on system.

You should now have the Linux desktop on your screen!

[amstan]

This problem is now fixed. It would appear that flashing of the fpga via jtag ( see manual) and then reinserting the linux/mister sd card fixes this problem.

That's cool, I'm glad you fixed it!

But hold on, this doesn't make sense (ESD damage wouldn't get fixed with reflashing). What was broken then?
Just flashing a new image on the FPGA (something that happens to volatile storage) shouldn't change behavior (like your sd card not booting) that persists across unpowers.

8. disconnect de10 and set MSEL dip switches to 010100.

I see you twiddled with the boot switches when doing this, perhaps when you removed the SD card when the problem started you also bumped the switches? And you following this instructions put them back to the proper place?

(speaking of which I'll have to look out for a good PSU with a switch on it)

Careful with those, some are known to have quite the resistance leading to voltage drops (enough to make something like an mt32-pi get concerned that it's not getting exactly 5V anymore).

[Malor]

If changes to the FPGA are permanent until rewritten, it might have been put into a state that prevented a system boot. Doing the reflash via JTAG would put it back to factory defaults.

Since I don't know details about how the system actually boots, I can't even guess what the specific change might be, but it doesn't seem too weird to me that a bad circuit/config on the FPGA could mess up the ARM side enough to prevent Linux from starting. It would be vaguely related to having a bad BIOS flash on a motherboard; with a bad flash, you might not be able to boot the board to fix it.

[twinkles76]

If changes to the FPGA are permanent until rewritten, it might have been put into a state that prevented a system boot. Doing the reflash via JTAG would put it back to factory defaults.

Since I don't know details about how the system actually boots, I can't even guess what the specific change might be, but it doesn't seem too weird to me that a bad circuit/config on the FPGA could mess up the ARM side enough to prevent Linux from starting. It would be vaguely related to having a bad BIOS flash on a motherboard; with a bad flash, you might not be able to boot the board to fix it.

That sounds like a perfect explanation to me

[dave18]

Nothing is permanent on the FPGA, it is RAM and loses its contents when powered off. The dip switches determine how the FPGA is configured at boot, with 2 of the options being program from flash memory or from a file on the SD Card. It sounds much more likely the a dip switch was knocked and the De10 was no longer booting from SD Card. I think the default program in flash just blinks an led on the board and I can't even remember what some of the other dip combinations might do.

[dave18]

BTW, you can configure the FPGA over USB without changing the dip switches from the SD Card boot position. Sounds like what you did was maybe write to the flash memory as well (which is non-volatile) so if you do have the dip switches in the wrong place going forward the De10 would boot into whichever core you flashed rather than the blinking led (although not sure what would actually display without the Linux sub-system running)

[futurepr0n]

I was trying to follow your instructions as my nano is sort of messed too and I wanted to reflash - when you say:

2. close the loop of the JTAG chain (alows programming) by bridging the two pins as shown on page 13 of the manual.
3. Set msel dipswitches to 10010 ( the on position is 0 and the off position is 1 ) as shown on page 107 of the manual.

I am looking on Page 13 of the manual, are you saying you need to solder?
is this what I need to get?

headerpin.jpg (7.43 KiB) Viewed 7403 times

" The J8 header is not installed,
so users need to solder a 2.54mm 2 x 5 male pin header if it is necessary"

If I pick up one of those, do I need to bridge any specific pins? or that piece being soldered into each of their rows is good enough??


Can you possibly help show some pictures of how you did this? The MSEL switch position I can figure out, but this bridge through me for a loop

If I understand correctly, would I remove the bridge at step 8? 8. disconnect de10 and set MSEL dip switches to 010100.

Then you are saying to flash the sd in step 9 with the XFCE linux image from terasic right? So I can omit that step if I have another sd card with the image already?

My issue is not the same, but I am wondering if these steps will fix whatever is problematic with my nano. I get HDMI and can boot into xfce no issue, but it will not boot mister os.


Thank you

[seastalker]

Thank you twinkles76 for this thread! I JUST had my one and only MiSTer appear to 'die' and this thread gives me hope. I'll notate and share my troubleshooting in hopes my efforts can help 'futurepr0n' and others that seem to have hardware issues.
My 'stack': DE10 Nano, Blisster, I/O board Ver 6.1
No case yet, and I use the alternative Northpanda power supply (5V, 3A) solution Todd's Nerd Cave suggests in his video called "MiSTer FPGA - A better power supply with built-in switch!"

Days ago I played through an arcade game from start to finish. Nothing was touched nor moved since then. Today I turned on my Mister and had no HDMI picture. I turned on my connected CRT and no picture there. Uh oh.

I don't THINK I removed the SD card with the power on. I also just checked if any dip switches are out of place but all four are "running away from the heatsink" as intended. With the IO board installed, it would be hard to accidentally hit the switches, especially in the other direction.

I unplugged all cables except the power lead that connects the DE-10 to the Blisster. I tried reconnecting to a different HDMI port, different ethernet cable, different HDMI cable, and used my Raspberry Pi's 5V 3.5A power supply - same result. I had a backup Mister install card - same. I reflashed it with the Mister Fusion image and still no luck. The only GOOD news is that even though I get no picture, at least my TV recognizes that SOMETHING is plugged into that port when powered on.

I then noticed the main power connection is a little wonky and was flickering on and off in it's new temp location. I don't know if this was doing this where it initially was but it's doing it now. I powered off and gently wiggled it in the power jack and sometimes there is a slight click sound and some give as if the pin in the jack was movable and it's NOT a tight fit in the 'well'.

I then remembered that power goes into the Blisster and not directly into the DE10 and it's the Blisster with the wobbly power input! I separated the Blisster and IO board and went straight with just a DE10 Nano, hdmi and power barebones test. SAME. The Nano's power jack feels less wobbly than the Blisster but not a perfect barrel fit either. Power light goes on fine and my tv input knows when it is getting power.

Looks like I may need to try flashing the Nano as per twinkles76 and when I get to the relevant stage, I'll try to help futurepr0n with the asked question: [I'll have to look but you may not need that part IF by 'bridging' it is just soldering some points together if the pins aren't present- just haven't gotten that far yet.]

Will keep all posted...

[seastalker]

As a followup, following the steps above to reflash, I was hoping to also assist futurepr0n but find myself now in the same boat with the same question;

Adding the pin header ALONE seems to me would do nothing nor 'bridging' - just transform pcb holes into pins. I saw page 13 and thought the details could be more specific too. It is clear in the following diagrams so its a shame that we are getting lost by the start. Page 13 barely shows which pins MIGHT be the ones to bridge, but it SEEMS like 7 and 8 if the square is indeed pin 1.

I have very thin solder wire that's a perfect fit to 'lace' through that size hole so I tested with a quick 'twist tie' approach. Changed DIP SWITCHES 4 and 5 as required. All seemed well (I am happy to see the DE10 Nano show lights again!) but after hanging on a green 100% progress for a bit, the final status message ultimately is always FAILED. I'm getting the same error as what shows here:
https://community.intel.com/t5/Intel-Qu ... m-p/239538

That link may HELP the answer but I don't know how to do/apply what the 'solution' suggests resolves it.


EDIT: SOLVED! That pin header should not be needed (at least not in my case). You may want to try the same steps again with and without bridging the holes in the JTAG step. If you don't have solder (or that small), try some thin copper wire from old audio cable or maybe a tiny paperclip if it's conductive. I also repeatedly [GENTLY] flipped all dip switches on the Nano repeatedly to make sure dust or anything preventing the proper functioning on each switch was gone. EVEN IF you get a 'FAILED' screen in Quartus, it may do enough to reset some inner workings on the Nano- mine is 100% back. Hope to hear updates from others with the same good news.

[twinkles76]

the pin header pins that need to be bridged are when looking at the top of the de10

psu connector OOOOO
OXXOO

HDMI

[ulrikbrick]

I have this problem too. In late october 2021 I accidently ejected the SD card while plugging a controller in the USB port right next to the SD card slot. Since then I've been getting a 'no signal' from the DE10-Nano to any hdmi diplay I have tried connecting to. I also tried different cards, installs, hdmi cables. I'm not a "soldering guy" but I have nothing to loose so I will try to follow the instructions here. Already I'm quite relieved I am not alone with this

Edit: I can't imagine the awesomeness if someone skilled provided a script that took care of this, but it's probably not doable.

[aeberbach]

just in case someone else has the same problem:
...
You should now have the Linux desktop on your screen!

The above did not help in my case. I have a DE10 Nano that can boot Angstrom or whatever the other XDKFE Linux is called first time, no issues, but Mister Fusion only ever shows the flash screen, reboots, flickers the green LED and then blank screen. It was working fine until one day it would not boot and from that day nothing has made a difference though I have tried multiple microSD cards with multiple Mister Fusion version. Tried to flash various .sof files from the CDROM and the Quartus programmer reported success but no change.

[seastalker]

Since April 10th, was smooth sailing but just had this problem happen again. In my case, I left the power button on the cable in the 'on' position so I could power the Mister on via the Blisster's power button. I suspect at some point this routine created some power issue that made it crap out again. I saved my notes so redoing the flash went ok.

Then, Mister Fusion would work, but my SD cards became write-protected. Worked past that but had to repeat this process several times due to seemingly more power issues. I do note that the PC doesn't recognize the DE10 if you forget to unplug the small usb to usb bridge piece to the Blisster - the DE10 won't power and connect.

I got MiSTer running and freshly installed with the update_all script, and even got it to recognize my regular 512gb card by copying and replacing all but my games folder. ALL WAS WELL. I reassembled my acrylic case and wired it all back and it failed again! So frustrating. Now I can't seem to get the DE10 to be recognized by the PC. I tried swapping usb cables and USB ports (direct on the PC motherboard), and swapped the Northpanda 5v 3amp for my 3.5 amp one I used for a rasp pi3. (That worked to fix one of the previous fails).
So I'm stuck w/o my Mister while I await a suggested 4 amp Meanwell PSU as suggested here:
viewtopic.php?p=58139#p58139

I don't see a lot of talk about the Mister having unreliable power connections that can lead to complicated reprogramming it. Maybe I'm just one of the unlucky ones.

[rhester72]

It sounds very much like you've got a high degree of static electricity buildup in the area where you use your MiSTer.

[seastalker]

Well, if mine crapped out on me, hopefully I can at the very least save others's MiSTer from suffering the same fate. I have both a CRT and a flat screen hooked up to it. It was sitting next to the CRT, but fully enclosed in a nice acrylic case and rubber feet. Could powering the TV on and off shoot static to the side of it? I have other consoles on the same shelves with it.

I'm really concerned now - no combination of power supply, usb cable, nor HDMI cable is showing signs of life (flat tv doesn't indicate at least SOMETHING is plugged in the HDMI like it did before), with one exception:

USB into the UART port on de10 nano shines a red and green light when plugged in the PC and a chime tells me the OS 'sees' it. It shows up listed as 'FT232R USB UART. Nothing on the J13 USB BLASTER port though...Hope I did not accidentally do a late night/fatigued flash attempt with Quartus Prime plugged in the wrong port- is that even possible? No hardware let alone the DE10 shows up in Quartus while plugged into the UART.

If static damaged something, how to even test for that? Any hardware troubleshooting tips?

[JayRockets]

(speaking of which I'll have to look out for a good PSU with a switch on it)

Just get an outlet switch like this: https://a.co/d/4yfGj4T

[seastalker]

**UPDATE: I figured out the problem [in my case]!!! I'm sharing in hopes this helps others in their MiSTer setups. If anyone is in contact with RetroRGB, especially Lu's Retro Source, I hope this serves as a MiSTer hardware safety PSA for all:

My problem ORIGINATED with a purchase decision in my very 1st post in this thread on Apr 08, 2022-

"I use the alternative Northpanda power supply (5V, 3A) solution Todd's Nerd Cave suggests in his video called "MiSTer FPGA - A better power supply with built-in switch!" "

I mean NO disrespect to Todd nor any team of his at Todd's Nerd Cave on youtube. [I thank him... I would not have learned so much about the DE10-Nano and Quartus Prime software if I didn't go on this journey:

THE PROBLEM:
The Northpanda power adapter I got from the video's suggestion and link is indeed a decent 3 amp PSU, but perhaps not a great match for the DE10 Nano, Blisster/USB addon, nor Mister in general (respectfully- in my view). It's not Northpanda's fault, but instead the suggested MicroUSB to 5.5mm x 2.1mm adapter that would be needed to convert to the barrel plug type plug. The video was the end of 2018, and at the time, that adapter may have been the only one readily available. It does NOT make a solid nor reliable connection; my earlier post's mention of 'wiggle' or loose connection happens with my DE10 Nano AND my Blisster board. The incorrectly sized adapter is at fault and NOT the power input jacks on either board.

This loose fit has the potential to damage your power connection pin on your Nano and or add-on board if stress flexed... [The vulnerability of many laptops with power adapter plugged while resting on a curved chair or couch comes to mind]. Worse, even plugging in a new USB controller can jostle the power connection wiring barely a billionth of a millimeter and it's enough to start a many day journey like mine of redoing Mr. Fusion SD cards, or the power glitch can confuse the programming enough on the DE10 to force you to learn how to reflash it with Quartus Prime...multiple times even.

I tried SLIGHTLY stressing and half rotating the culprit adapter and the Nano sprang back to life, was reflashed, retested, and all is well again! I won't be using my MiSTer again until my Meanwell GST25A05-P1J power supply arrives that others suggested in this thread, and is allegedly used in the Multisystem builds at rmcretro:
viewtopic.php?p=58139#p58139

Lessons learned are not to 'Clark Griswold' connect adapters (some have Canakit Rpi3 micro USB to additional adapter with power switch before adding the culprit barrel adapter!) and hopefully the 4 amp(!!) Meanwell one from wall outlet to MiSTer will solve all these issues. Also, it is a great testament to the durability on the Nano and Blisster if nothing else.


PLEASE see this link to the video and the comments below it, especially the great comment by the wonderful FirebrandX and return comments by Todd. Although some comments talk of hypothetical issues with the power setup, I'm here to confirm the actual problems this caused my setup:
https://www.youtube.com/watch?v=ci5liZtbfZ0