minimig resets itself when using workbench

[juanjerez]

I did the update_all recently. it happens to me a lot that when I am using workbench I try to run some game or program the core restarts by itself for no apparent reason. I have tried to delete the minimig configuration files again but the problem persists. This has been happening to me since recently. Has anyone else had something similar happen to them? What could be the reason?

[rcade]

The simulated Amiga resets, or the whole core?

[juanjerez]

the whole core

[Chris23235]

I did the update_all recently. it happens to me a lot that when I am using workbench I try to run some game or program the core restarts by itself for no apparent reason. I have tried to delete the minimig configuration files again but the problem persists. This has been happening to me since recently. Has anyone else had something similar happen to them? What could be the reason?

Sounds like a power supply issue. Try changing the power supply to see if it persists.

[proteque]

This sounds very familiar. But my problems are hitting a bit harder. I made a post about something similar today. however it also happens to AO486 for me. This is the post I made in the newbie-forum. Is it the same?

After an update the mister started crashing when I loaded cores with diskfiles (hard disk files). Not sure where the size limit goes for the problem but lets use minimig as an example: minimig with a 100mb image ran for ... a little while. but with 10 gig image it crashes before it is done loading AmigaOS (about halfways).

What happen is that the mister crashed (completely) and then reboots into the menu. I have been logged in through ssh when it happens and the linux part crashes too.

My first thought was the SD card so I bought a new one and downloaded mrfusion and started over from scratch. Tried with new known working disk images and copied them over. but still the same. the mister crashes. I then ran a memory check. everything fine with the memory as well. I then tried having the disk images on a usb disk. same problem. I have ran out of options to try.

I also tested another PSU just in case something happened to that one and it rebooted due to not enough power.

[LamerDeluxe]

This sounds very familiar. But my problems are hitting a bit harder. I made a post about something similar today. however it also happens to AO486 for me. This is the post I made in the newbie-forum. Is it the same?

After an update the mister started crashing when I loaded cores with diskfiles (hard disk files). Not sure where the size limit goes for the problem but lets use minimig as an example: minimig with a 100mb image ran for ... a little while. but with 10 gig image it crashes before it is done loading AmigaOS (about halfways).

The limit might be 4GB, that is a common limit.

[desin24]

Backup your config folder
delete your old *.cfg file
start the core and generate a new config

hope it helps

[Flinto]

I am also having a very similar issue. I am using an unofficial I/O board that is electronically identical to the official 6.1. I get frequent resets when using the latest minimig core (24.04.09), this issue is not present when I go back to the previous revision of the minimig core (23.12.04). The issue seems to happen almost everytime, on cue, when I load a game from within a hdf file. Some of my larger hdf files do not even load at all, before the reset occurs. I've not noticed any issue with any other core. I would assume that it's a software issue but it seems strange that few others are experiencing this as well.

[limi]

Random resets generally indicate a power supply that isn’t delivering enough power (amps/voltage).

The Minimig core is one of the most well-tested (since it is one of the first FPGA cores and part of why MiSTer is named what it is), and works well.

[Flinto]

It's a meanwell 4amp psu, but the same issue also occurs with the supplied 2amp psu.
Also, going back just 1 revision in the minimig core solves the issue for me.
The resets are also not random. They actually occur at very precise moments.

[Flinto]

Removing the I/O board also solves the problem

[Chris23235]

Removing the I/O board also solves the problem

A 2amp PSU might be to weak for supplying the DE10-nano and the I/O board (depends on the connected peripherals). If the issue goes away after detaching the I/O board it is either faulty or a power supply issue. I would put my money still on the power supply and not the I/O board, but that's guessing at this point. What do you have connected to the MiSTer?

[limi]

Also, going back just 1 revision in the minimig core solves the issue for me.

Interesting. I had a multi-hour gaming session on the Amiga with some friends just two days ago, and had no issues. Will check which Amiga core version I am on when I get back home.

[FPGA64]

when 24 bit output was added some non official IO Boards were causing reboots. I wonder if this is the case her. One of the switches on the De10 had to be thrown to tell the Mian Mister it didnt have a 24 bit capable Io card attached. This was only for the non official boards

[Retro-Nerd]

5V, 2A is more than enough for the MIster with (official) analog I/O board and plenty of pads/keyboards. Must be a indeed a problem with unofficial boards.

[Chris23235]

5V, 2A is more than enough for the MIster with (official) analog I/O board and plenty of pads/keyboards. Must be a indeed a problem with unofficial boards.

I agree it should be enough, when the user only connects pads and keyboards. Question is, if a SSD is connected? At this point the 2A could become a problem. But if this was reported with other boards, the possibility that it has something to do with the board is high.

[Flinto]

OK, I've made some strange progress on this. I have repeated this this twice and got the same results both times.

I decided to measure the voltage to see if I could see a voltage drop when the core resets. I didn't see much of a noticeable voltage drop but what I did discover is that when measuring the voltage, the core does not reset as frequently.

It seems that the act of measuring the voltage causes less frequent resets. I know this because one of the tests I normally do is to load a large hdf file, around 32gb. This file will fail to load 100% of the time on the latest minimig core, and will load successfully 100% of the time with the previous minimig core.

However, if I load this large hdf file while measuring the voltage, the hdf file itself will load successfully. The core still resets, however it is now resetting sometime after the hdf file has loaded.

My electrical knowledge isn't enough to understand what is happening, but could it be that measuring the voltage is acting as if it is some sort of pull up resistor?

[limi]

Could you clarify if you are using the official I/O and USB boards? Or clones? (Just to be able to rule that out as a possible cause)

[Retro-Nerd]

SSDs are fine too, I use an external Samsung on my Mister. An external HDD though draws a lot of power.

[Flinto]

Could you clarify if you are using the official I/O and USB boards? Or clones? (Just to be able to rule that out as a possible cause)

For this particular setup they are clone boards.

[FPGA64]

That could mean a badly made Ali Clone. I think they have to be treated as Unofficial and suspect

[rcade]

32GB!? I have never made one larger than 4GB, because the file system is FAT32 and generally the Amiga was limited anyway (with original file systems, of course). I guess you are mounting something that large over SMB?

[rcade]

Could you clarify if you are using the official I/O and USB boards? Or clones? (Just to be able to rule that out as a possible cause)

For this particular setup they are clone boards.

I thought it was just a specification and anyone can make them? What makes something official vs. clone?

[FPGA64]

Could you clarify if you are using the official I/O and USB boards? Or clones? (Just to be able to rule that out as a possible cause)

For this particular setup they are clone boards.

I thought it was just a specification and anyone can make them? What makes something official vs. clone?

Alot of the Ali Clones are very badly made, missing headers and with abysmal soldering. As such they are often causing issues.

[limi]

32GB!? I have never made one larger than 4GB, because the file system is FAT32

MiSTer uses exFAT, which handles much larger files.

[limi]

Interesting. I had a multi-hour gaming session on the Amiga with some friends just two days ago, and had no issues. Will check which Amiga core version I am on when I get back home.

I am on the latest core, Minimig_20240409.rbf, and have no issues running 10GB HDFs, including Workbench and apps. Official analog I/O board and USB board, the standard triple stack.

Since it works when you disconnect the I/O board, I’d definitely suspect that being the issue.

[Flinto]

That could mean a badly made Ali Clone. I think they have to be treated as Unofficial and suspect

It's not ali Express! It's very well made to the exact specification of the official one. Its unofficial as it is made it fit inside an Amiga 500 case

[Flinto]

32GB!? I have never made one larger than 4GB, because the file system is FAT32 and generally the Amiga was limited anyway (with original file systems, of course). I guess you are mounting something that large over SMB?

No. Exfat. It's my Amiga 1200 setup made into a hdf. Its 32gb because that's the size of the sd card. The 4gb limitation is easily solved with a 3.2 rom from hyperion.

[Flinto]

Interesting. I had a multi-hour gaming session on the Amiga with some friends just two days ago, and had no issues. Will check which Amiga core version I am on when I get back home.

I am on the latest core, Minimig_20240409.rbf, and have no issues running 10GB HDFs, including Workbench and apps. Official analog I/O board and USB board, the standard triple stack.

Since it works when you disconnect the I/O board, I’d definitely suspect that being the issue.

Yes the issue is more specifically somewhere in the 40 pin header. I'm going to test the connections 1 by 1 as soon as my dupont leads arrive

[Flinto]

I've fixed the issue. It was an issue with the 40 pin header. It was connected via a short ide cable. Replacing the cable with 40 x short dupont connectors has fixed it. All working now.