MiSTer Intermittent Bootup

[kfreiberg]

Hi All,

I've tried searching google but couldn't find anyone else having this issue with their MiSTer.

My MiSTer boots up intermittently. Sometimes it boots up perfectly 3 or 4 times in a row, then the next time it won't. I know it boots up correctly when I see the "amber" LED light up when I turn it on. If the "amber" LED doesn't come on after a few seconds, I know it didn't boot up correctly and nothing will come up on the display.

Items that I have noticed help it boot are to re-seat the SD card, or unplug the power cord and plug it back in. But it almost seems random that these things work, they don't always work.

I have tried this with different SD cards with a working image on them. I have also tried having nothing plugged into the MiSTer upon the boot (i.e. no USB hub or HDMI connection, etc.) It still boots up randomly.

I would estimate that my MiSTer boots up correctly 75% of the time. When it does boot correctly, it works flawlessly for all cores until I power it off.

Anyone else have an issue like this? Or any suggestions on what might be the problem or how to fix this?

Here is my setup:

DE10-Nano Kit
SanDisk 256GB Ultra MicroSDXC (also tried a smaller SD card with a working image)
atolla Aluminum 5-Port USB 3.0 Hub with 4 USB 3.0 Data Ports and 1 USB Smart Charging Port, USB Splitter with 5V/3A Power Adapter and Individual Switches
DC Splitter + Power Switch - From MiSTer Addons
PCB Fan Plate and Heatsink - From MiSTer Addons
XSD 2.5 128MB SDRAM Module - From MiSTer Addons


Please let me know. Thank you!

[wasteground]

Do you notice if the DC power switch gets warm/hot when it works after running for 15+ minutes or so? I had this issue with my new MiSTer, and just ended up removing the power switch from the equation and everything has been fine ever since (now I just pull apart the DC plug from the power supply and the splitter - I'll replace the power switch eventually!).

There's an article here that might be interesting: https://www.retrorgb.com/mister-power-s ... rning.html

[Brettster]

I also have this issue, had it with an official IO board and usb 1.2 and also with Antonio's case / pcb
It also happens with 2 different psu's ones a 4amp mean well and the other is the official de10 psu.
I just assumed it was normal

[kfreiberg]

Do you notice if the DC power switch gets warm/hot when it works after running for 15+ minutes or so? I had this issue with my new MiSTer, and just ended up removing the power switch from the equation and everything has been fine ever since (now I just pull apart the DC plug from the power supply and the splitter - I'll replace the power switch eventually!).

There's an article here that might be interesting: https://www.retrorgb.com/mister-power-s ... rning.html

Thanks for the info! I played for about an hour or so last night and the power switch was not warm at all, I'd say it was room temp. I will try bypassing the splitter/power switch to see if that improves anything.

[kfreiberg]

After further testing, I found that bypassing the DC Splitter/Power Switch did not fix anything. The MiSTer still boots up intermittently with the original power connector. I also put a multi meter on the connection and I got a constant 5V DC from both the original power cord and with the DC splitter connected.

So I'm back to square one. Still getting intermittent bootup.

Anyone else have any ideas/fixes?

[ExCyber]

This is kind of a long shot, but I'd check the DIP switches and make sure that they're firmly flipped to the default setting (see page 11 in the DE-10 Nano User Manual). If one of them isn't making good contact, the FPGA might be trying to load a bitstream from the wrong place.

[grizzly]

check the DIP switches

And maybe flip them on/off 2-3 times (if some grime/crap have gotten in) then make sure they are at the right setting after.
Probably not this tough but isn't rocket science and is pretty fast to do so why not.

[DevStarr]

I am currently having this same issue. Dip switches are in proper configuration, and I tried cycling them all a few times and no luck.

Sometimes I have to cycle the in-line power switch 10-15 times before it will boot, but when it does finally boot it works flawlessly. I have tried multiple in-line switches, and also bypassing the switch completely. I am using a 5V/4A meanwell adapter.

[kfreiberg]

I am currently having this same issue. Dip switches are in proper configuration, and I tried cycling them all a few times and no luck.

Sometimes I have to cycle the in-line power switch 10-15 times before it will boot, but when it does finally boot it works flawlessly. I have tried multiple in-line switches, and also bypassing the switch completely. I am using a 5V/4A meanwell adapter.

I have never resolved this issue. I have to powercycle my MiSTer 1-5 times usually before it boots properly.

[Chris23235]

To me it sounds like the cause could be a not properly seated SD card, this would explain, why it works sometimes when you reseat the SD card. SD card slots are fragile so I try to not pop out the card anymore and transfer everything via FTP (more convenient too).

Could also be something with the card, have you already tried another card?
Nevermind, I just read you tried different cards already.

[DevStarr]

I am starting to think the issue lies within the DE10-Nano board, as I have removed all other components and reformatted 3 different SD cards without any change in the issue. MiSTer works great, but still takes anywhere from 1-30 power cycles to properly boot and it seems completely random.

I have not yet measured the power supply output, but it is a Meanwell 5V/4A adapter that I have seen recommended frequently, model #: GST25B05-P1J.

When the DE10-Nano does not boot I get the Blue "3.3V Power" LED and the entire row of green LEDs is dimly lit:

No Boot.jpg

When the DE10-Nano does boot properly I get the Blue "3.3V Power" LED, the Yellow "CONF_DONE" light, and two of the Green "User" LEDs will be flashing bright:

Boot.jpg

Unfortunately I cannot stomach purchasing another DE10-Nano board for $170 at the moment, but if anyone has any suggestions as to how I can get this board to boot consistently I would greatly appreciate it. I don't know if maybe there is some way to reset the board to its factory state that would be within my ability, but I am open to all suggestions.

[jca]

Did you try to move the DIP switches a few times and making sure they are in the right position when you are done? I had a similar problem, I wiggled each switch 4 times to be sure they were back to where they are supposed to be and everything went back to normal. Don't forget to do it with power off.

[DevStarr]

Did you try to move the DIP switches a few times and making sure they are in the right position when you are done? I had a similar problem, I wiggled each switch 4 times to be sure they were back to where they are supposed to be and everything went back to normal. Don't forget to do it with power off.

I did exercise the DIP switches multiple times and I believe they are in the correct position, no joy.

[jca]

Sorry to hear that. It could also be the SD card socket itself but it does not looks easy to check that.

[grizzly]

I would probably have tried using some paper/plastic in the sd card slot meaning between the sd card and the metal backing so it maybe press against the connections a little bit harder.

[ByteMavericks]

Just a thought: what happens to the LEDs when no sd card is present? That might help diagnose whether it could be sd card/socket related or not

[DevStarr]

Just a thought: what happens to the LEDs when no sd card is present? That might help diagnose whether it could be sd card/socket related or not

When there is no SD Card present the LEDs are identical to a failed boot, except the "HPS User" LED remains dimly lit.

When SD Card is present the HPS User LED is momentarily dimly lit and then either: extinguishes during a failed boot or flashes bright followed by dim pulsing during a succesful boot.

No SDcard.jpg

[Sigismond0]

What happens if you power up with HDMI disconnected? Could be getting votagee flowing back on that, causing power issues? I know a lot of people have put in CEC strippers to solve HDMI-related power strippers. My MiSTer can't be plugged into my HDMI switch because it draws power somehow and lights up the MiSTer even when off.

[DevStarr]

What happens if you power up with HDMI disconnected? Could be getting votagee flowing back on that, causing power issues? I know a lot of people have put in CEC strippers to solve HDMI-related power strippers. My MiSTer can't be plugged into my HDMI switch because it draws power somehow and lights up the MiSTer even when off.

There is no change with HDMI disconnected. DE10-Nano still boots intermittently with the same LED indications displayed.

[moldov]

Hey guys,

I'm having the similar issue now. So let me explain my findings and experience, hopefully that will help someone. Looks like our boards are faulty and I will explain below why.

I ordered my first board back in January 2020, Terasic sent it very fast and being impatient I received my board and plugged it in. For initial boot I used linux SD card included in the DE10-Nano package. The system booted up and I saw linux desktop on my TV screen. I switched off and left the board for a while until I prepared MiSTer FPGA SD card.

Then I tried MiSTer SD card but my attempt to boot was unsuccessful. I was trying different SD cards, different versions of MiSTer installation - no luck. So I went back to original linux SD card and system booted up successfully. But after couple of attempts it started to re-boot without any actions from my side. So I suspected the SD card as I saw some information saying that De-10 Nano is quite picky on SD cards.

There were lots of searching through internet and forums - lots of different information and similar symptoms - but no solution at all. I was trying different SD cards, SD card readers, images, PSU (I have precious lab PSU), cleaning contacts with alcohol - no luck.

So after couple of weeks of quite deep and proper troubleshooting I decided to go back to Terasic. I wrote them e-mail explaining my issues and steps had been done, they told me to send the board back.

It took a bit to fix the board and Terasic replied me back with the following quote "Your board has been fixed after welding the FPGA to the board again. See attached the screenshot of the test result. " - as the chip is BGA type, looks like soldering was the issue there.

Then I received my board back and was playing games without any issues for a long time. There were some hang ups during boot up but I thought they were because of my slow and cheap SD card as I had slowdowns in game loadings especially with ZIPed rom lists and NEO-GEO roms.

Month ago I sent the board to friend of mine and he reported me back, saying that board didn't boot up from the first time, it was incosistent though and he expected ESD as he was connecting and reconnecting sound and power cables. I expected slow SD card and told him to swap with another. He swapped and said that games are being loaded much faster now. However boot up issues still in place and getting more often.

Conclusion. I see this intermittent issue become consistent among the users and looks like it is manufacturing process faults with BGA soldering. I was using my board very accurately even installed heat sink with industry grade thermopaste and fan, this particular Altera doesn't requires an additional cooling though.

Here is hardware I use:

- DE10-Nano from Terasic with original PSU
- 128 MB SDRAM board from Misteraddons
- IO Board from Aliexpress (resistors nominals on RGB lines were 200 Ohm (common issue with IO boards from Ali) I changed them with 240 Ohm as it should be ) - the board is very good quality. So I don't expect the board. However I asked my friend to remove the board and try with HDMI output only to have more consistent picture.
- Cheap USB hub for gamepads (the good one is coming) was swapped with OTG cable during troubleshooting.

I don't expect the boards to be faulty as I've been using them since DE-10 Nano was fixed first time and there were no issues in the beginning.

So after few more steps of troubleshooting I will report that issue back to Terasic and will let you know based on their reply. They were quite customer oriented first time, so I hope they will do fix my board again.

[aberu]

Thanks for keep us up to date @moldov - It seems to me that you just got a faulty board. Here's a correction though:

this particular Altera doesn't requires an additional cooling though.

If you are using cores that use higher bandwidth, like the ao486 core, you will definitely want a heatsink. It doesn't require it, but it is definitely a very good idea to use a heatsink.

[moldov]

That was that bloody Power Switch which SmokeMonster was talking about here https://www.retrorgb.com/mister-power-s ... rning.html

Initially, when I didn't have IO Board WITH cooler, the power was enough even through this power switch. Then I installed IO Board with 12V cooler being powered from 5V, so looks like that creates additional quite big power consumption.

Friend of mine confirmed that he's got no issues once he has removed that bloody switch.

So it looks like conincedence but all the things I listed below can give you the clue while troubleshooting your boards:

- Remove Power Switch, especially if you have installed IO Board with additional cooler.
- Check your SD Card, it can create slowdowns if the quality of the card is not good.
- BGA soldering below the IC chip can be bad (that happened to me)

Regards.

[manji1]

I had this same problem and it was driving me crazy. Removed the janky power switch and now everything works perfectly. Thanks so much to the posts in this thread