Machine startup has not completed successfully- [Solved]

[chocolate]

I broke my Raspberry Pi’s SD holder. I brought my Raspberry PI to a Radio Shack to test a TFT monitor. When I brought the Raspberry PI out of a car, I dropped it, and a piece of the SD holder broke off. I fixed the SD holder with some cardboard and a little bit of gorilla glue, but RISC OS wont boot up properly. I’m getting a error at start up: Machine startup has not completed successfully: ‘The disc drive is empty’. I've tested other operating systems (Raspbian, Puppy, and Plan 9 from Bell Lab) and they all work. I've tried multiple SD cards; and a SD card with a micro-SD slot, but RISC OS keeps on booting with the same error. Is there a way to fix my Raspberry PI so I don't get this error message?


Edit: If I press cancel, I boot into the desktop. If I type *SDIODevices nothing shows up. I was able to boot into RISC OS before I dropped my Raspberry PI.

[DavidS]

YES:
Most likely the cardboard just is not holding the SD Card in tightly enough to hold down the inserted switch. See:
Machine startup has not completed
in the ROOL forum. There are some working solutions there.

[chocolate]

I've created that thread a while ago on the ROOL forums. I tried some of those solutions, but they seem not to work for me. I'm unable to solder a microsd slot at this time, so I have to try another solution. I'll try clamping again, and see if that works. Is there anything else I could try?


Edit: I tried re-clamping, but that doesn't work. I'm still getting the error. I might try cutting a giftcard and glue it into place, but I don't know if it will work or not.

Edit 2: I've decided to cut a old giftcard anyway and see if that works.

[AMcS]

Machine startup has not completed successfully: ‘The disc drive is empty’

I presume that's a message coming from the OS (RISC OS ?). In which case the card MUST be readable (to some degree). If you try boot the machine *without* any SD Card fitted is the same error given (if so it's something on the SOC firmware and far below RISC OS - if not then it's likely RISC OS).

There's a leaf spring like metal strip on the Pi circuit board at the inner edge of the SD socket that closes (presumably completing a circuit) when the SD card is inserted - if that signals that a card is in place perhaps RISC OS (or more likely the SD I/O module) checks for this and even if the card is in contact (and should be active) it checks the status of the switch and issues the warning if it doesn't find the switch shut ?

If that appears to be the case then perhaps just clamping the strip with something non-conductive would do the trick

That might explain why other OSes work (presumably they don't check this).

Hope that helps.

[AMcS]

A further thought (and as I don't have a Pi Circuit diagram the following may or may not be valid) - it may be that the leaf also applies power to the SD card - but removes it when the card is ejected - it may be an idea to ONLY clamp the leaf spring when the card is fully in place - and release it before the card is ejected (just in case it corrupts things on ejection).

[Burngate]

A further thought (and as I don't have a Pi Circuit diagram the following may or may not be valid) - it may be that the leaf also applies power to the SD card - but removes it when the card is ejected - it may be an idea to ONLY clamp the leaf spring when the card is fully in place - and release it before the card is ejected (just in case it corrupts things on ejection).

Circuit diagram: http://www.raspberrypi.org/wp-content/u ... .2_027.pdf
So no, that switch doesn't interrupt the SD card power

If you power a Pi without a card, nothing happens - no error message, no screen output, nothing

So to have got that error message, the card must have been read - in fact the RISC OS ROM image must have been read, since that is what has provided the message.

Some time ago, in the Linux world, someone had a similar-ish problem - their Pi consistently stopped booting part way through.
I had a theory that at that point Linux changed the mode in which it was reading the card, and that for some reason that mode wasn't working, even though the default mode was.
I tried researching this but all I managed to do was to stop the Pi from ever seeing the card.

[chocolate]

I tried the plastic card trick and it still shows the error. I even tried using clamps (wood clamps and a fridge clamp), but no avail.


Edit: Good news. I found out that some of the gorilla glue was preventing the back back spring from moving. I removed the glue entirely and it booted up properly. Now, I'll have to glue the gift-card back and give it a try. It might not be working 100% but it is progress.

Edit 2: I'm going to give this a 2nd try in a little bit after the glue dries.

[chocolate]

I found that cardboard sucks up the glue much better so I went back to using cardboard. At first it was a hit and miss until I plugged in my 16gb SD card one more time and it started working.


Thanks for your replies. Now its time for me to muck around my new favorite operating system.

[AMcS]

Circuit diagram: http://www.raspberrypi.org/wp-content/u ... .2_027.pdf

Thanks for that link Burngate, it's much appreciated.

[AMcS]

Some time ago, in the Linux world, someone had a similar-ish problem - their Pi consistently stopped booting part way through.
I had a theory that at that point Linux changed the mode in which it was reading the card, and that for some reason that mode wasn't working, even though the default mode was.
I tried researching this but all I managed to do was to stop the Pi from ever seeing the card.

Just on the off chance this is useful, my brother got a RPi just a few days ago - and I burnt a RISC OS image onto an SDHC card. (The image was the same image I used on mine successfully). When we inserted it into his machine and powered up it got the EXACT same error message - and it consistently failed to boot.

Although the card looked well seated - I pushed it gently in further. Applied the power and the PI booted without issue (and has done so ever since). I suspect if the card is not perfectly in contact with the end leaf spring and the contact pads this precipitates the problem.

[MrMoke]

Just had the same issue myself after several successful boots in the past with same SD card. Assuming card was corrupted I reimaged, no luck. Read this thread in the forum and cleaned contacts on SD card, didn't boot at so I paid particular attention to the 'seating' of the SD Card. Bingo! That did it. What I think the issue was in my case, is that I have had my RPi sitting with the Rasbian distro SD card inserted for a week or more, been too busy to play with it . Anyway, I am guessing that having the card in place 'loosened' the leaft spring contact at the back of the card, it doesn't feel as firm when inserting a card anymore. So I am going to make sure I store my Pi with cards removed from now on and will post here if I experience further issues.