search

jbilander / sdbox

A parallel to sd-card project for the Amiga
GNU General Public License v3.0
94 stars 14 forks source link

Unable to load #2

Open BleuLlama opened 3 years ago

BleuLlama commented 3 years ago

The sd card is a 4 gig card formatted FAT using the "SD Card Formatter" application on my Mac. I have tried other cards with no success either.

I'm sure it's something obvious i'm doing wrong... any suggestions?

jbilander commented 3 years ago

Hi,

You can try and format with FAT and a smaller partition, try different cluster size when formatting. I've tried with a 2 GB partition on a 8 GB card and it worked for me, but you can try with even smaller, and if that doesn't work try FAT32. I formatted in Windows though after cleaning the sd-card with diskpart. There is a thread on the a1k.org forum on how to make the SD-Box more compatible by modifying the sd-card-adapter with pull-ups. It is a german forum but it is ok to write in English. Also you can verify that you have a working parallel-port by using Keirf's excellent Amiga Test Kit, boot from the ATK-adf https://github.com/keirf/Amiga-Stuff/releases .Write the adf to a floppy or If you have a gotek just put the adf on a usb stick and boot from it. It can check your CIA's timings and in one of the menus you have instructions howto make a dongle to test the parallel-port to verify it works correctly. Another thing to check that your Nano has a 16 MHz oscillator. It is needed since the timing is based on that. I've heard some Nano V3 clones sometimes has a 8 MHz oscillator and that will not work with the SD-Box. Hope this helps! /J

BleuLlama commented 3 years ago

I tried that card with other formats, and none were successful. I just found a 1 gig card, and it showed the same results. (Fat32) I've verified my parallel port with the Amiga Test Kit, with the parallel port loopback adapter for it. Yes, my nano is 16mhz. (Thank you for the suggestions, etc!)

I also just tried the 0.3 binaries here, to no avail. :(

I'll look into that thread you mentioned...

kendickson commented 3 years ago

Hi BleuLlama. Have you been able to make any progress? I've got the same problem – mount succeeds (or, at least, doesn't give any error), but then trying to change directory I get "Unable to load sd0:: Error 115".

BleuLlama commented 3 years ago

I was never successful with it, and ended up getting distracted by another hobby.

On Sun, Dec 6, 2020 at 12:20 AM Ken Dickson notifications@github.com wrote:

Hi BleuLlama. Have you been able to make any progress? I've got the same problem – mount succeeds (or, at least, doesn't give any error), but then trying to change directory I get "Unable to load sd0:: Error 115".

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/jbilander/sdbox/issues/2#issuecomment-739456725, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAQ7GH2KE4XI7UJCVD6NSLSTMICVANCNFSM4SJADGOQ .

-- Scott Lawrence yorgle@gmail.com

SukkoPera commented 3 years ago

Very same problem here. I have even tried a 128MB (Yes, MB!) SD cart with just a single FAT16-formatted partition, no luck :(.

I have noticed that the ACT led never ever blinks at all, so it looks like the SD card is never selected.

jbilander commented 3 years ago

To troubleshoot I would try and rule out if this is your Amiga, your SD-Box or your microSD-card causing the problem. If you have a friend with a known working SD-Box then try your card in his/her SD-Box. This way you will know if your formatted card is compatible or not. Next step is try your SD-Box in a known working Amiga running SD-Box meaning unplug the good working SD-Box on that machine and try yours in that same machine. This way you will rule out if anything is wrong with your SD-Box. If it doesn't work then you will have to start troubleshoot your SD-Box, check voltage levels, check no dry soldering joints, check for continuity between pins, verify the Nano, verify the SD-Card adapter etc. If your SD-Box works in another Amiga but not in yours, chances are your configuration wasn't done correctly or your CIA-chips or parallel-port isn't working correctly. To verify that your Amiga's parallel-port is working you can make a dongle and use Amiga Test Kit to test it together with verifying the CIA-timings. Some things to verify before that:

To mount card manually in wb 1.3 in a shell-prompt type this:

cd Devs
mount SD0: from SD0
cd SD0:

After hitting enter in the last step the activity LED should start blinking. If successful it shows an icon on the workbench with the volume name of the microSD-card. In order to see folder and files as icons in wb they need to have corresponding *.info files of correct type with the same name as the file/folder to appear in a window.

I use a Netac 8 GB microSD HC card formatted with a 2 GB FAT partition in Windows. I used diskpart in a windows cmd-prompt to clean it before creating the 2GB FAT partition in diskmgmt.msc with default values. I haven't had any problem at all with this card using it in my SD-Box with my Amigas.

https://i.imgur.com/KVXiJ9K.jpg

Hope it helps, /J

SukkoPera commented 3 years ago

I have no friends with other SDBox's, but I can build another one. I will definitely investigate and do more tests, but what I can say at the moment is that the Nano, SD adapter, SD card and parallel port are all working good on their own, flashing was successful and the clock is 16 MHz.

I can add that I'm testing on an A500 rev.8A with a TF530 card and KS 3.1. I used the SD0 file from the sdbox/amiga directory, which seems to match the

Anyway, I think that since there's 3 of us here having the exact same problem, there must be something going on. Could you please verify that the spisd.device contained in sdbox_drivers_v0.4.zip is a "good" one? I tried to recompile it, but the binary I built on one machine crashes the amiga as soon as it's loaded, on another machine the cross-compiler won't build (I think because I have GCC 10). I saw that some guys around are still using v0.3, but I haven't tried it yet.

jbilander commented 3 years ago

Hi SukkoPera, yes I verified v0.4 before I put it out on github, but I haven't done much testing. I think it should work with a TF530 installed. I have only tried with a TF534 but the E/VPA/VMA logic is the same between the two I think. The TF generates the E-CLK on posedge which I think is slightly wrong, it should be on the falling edge but it should work anyway. What you can try is put back in the plain 68000 CPU, that should definitely eliminate any E-CLK/VMA/VPA problems although you only get like ~83 KB/s transfer speed. I will send you an email with a prepared adf that I know works and you can troubleshoot further with...

SukkoPera commented 3 years ago

OK, thanks to the ADF you sent me I made some progress: it turns out that if you use the mount sd0: from devs:sd0 syntax, this mountlist is not adequate, as it needs to start with SD0: and to end with a #. This was probably my misunderstanding. Anyway I have replaced it with the one that was on your floppy and at least now I can see the Activity led light up and blink a few times when I try to access the SD0: device, but then the dreaded bad number show up again.

I have checked all pins with a scope and I think the problem is that there is no activity at all on the MISO/MOSI pins. I can see the clock pulses on SCK and the CS line going low but nothing on MOSI at least (measured right on D11). Could it be an issue in the firmware? I'm going to build another board with another Nano in the meantime.

Thanks for the prompt help btw!

jbilander commented 3 years ago

Yes, this is all documented here https://github.com/jbilander/sdbox/tree/master/sd If I recall correctly the SD0: and the # needs to be removed in wb 3.1 (on KS 3.1) in order to use the double-click method to mount, but in wb 1.3 it is a different story because the double-click doesn't work and you need to manually mount via command line. The SD0: just indicates the start of the section in the mountlist and the # indicates the end, and is needed as a separator when you have a mountlist with many entries.

I got bad numbers too with one of my other microSD cards. It may be an indication that the card you are trying with doesn't work for some reason. I don't think it is a firmware error if the main.hex was flashed successfully. I mean the firmware can probably be improved but it should work with some cards as is.

Best of luck with the other board.

SukkoPera commented 3 years ago

I've tried 3 different SD cards:

But no luck :(. Even with the new board, I can't see any activity on the MISO/MOSI lines :(.

SukkoPera commented 3 years ago

I think I've found a solution. I need to make some more tests when I get home, but I feel optimistic :).

This should actually give a nice boost in compatibility with more SD cards.

SukkoPera commented 3 years ago

So, here we go: thanks to this topic on the Arduino forum, in particular to lepoloni's post, I tried to put a 10k pull-up resistor on the MISO line before it reaches R4 and it did the trick on both of my boards! My 1 GB card is now recognized 100% of the time and seems to be working fine. The 4 GB one seems to have too big a filesystem while the 128 MB one gives I/O errors, but I'll keep trying with this one.

Here are some pictures to help locating the right spots for soldering the resistor (sorry but I still need to clean the boards after all the building and tentative bodges):

IMG_20210126_191740 IMG_20210126_191730

After looking at the schematics of the module, I thought that R4, which is supposed to be there for protection, actually has no utility at all, since it doesn't look like the buffer is suddenly going to reverse its direction. It just makes it harder for the SD to drive the MISO line, so I thought that removing the pull-up and bridging R4 altogether could be sufficient. This seems to work on the board with the larger SD module, while the smaller one still needs the pull-up, which is very odd as the basic circuit seems to be the same.

IMG_20210126_195014

Anyway, hope this helps, I'll keep doing more tests. Thanks @jbilander for your support!

PS: The board is my own modification of @jbilander's one. I made it mainly for my personal use (I like edge connectors, 0805 and SOT-23) but it's available here if anybody is interested.

PS2: I also took the time to report the Amiga 68k toolchain build error. The issue contains a link to the patch I applied to m4 in order to get it to build. Have a look there if you are having the same problem. BTW, I still haven't tested the spisd.device I finally managed to build.

jbilander commented 3 years ago

Very nice, glad you got it working. Yes I think pull-ups on the SD Card Adapter is what they did and talked about in the SD-Box thread on the german forum a1k.org too. I mentioned it in my first comment here but that thread seems to now be behind a registration wall unfortunately, so very good that you documented and shared your findings here. Many Thanks 👍 Maybe that 4 GB card of yours will work if you format it with a single smaller partition like 2 GB FAT or so.

SukkoPera commented 3 years ago

After more testing, I decided to reinstall the pull-up resistor on the larger board too, it makes it more reliable to get any card to mount at the first attempt.

I got the 128 MB card working, seems like fat95 likes partitions with type 0b (W95 FAT32) better, I think I was using 06 (FAT16) before. I also got a 32 GB card working by making a single 2 GB partition on it (Maybe larger ones work too, I didn't try any other size). It takes a while to mount, but it works. I don't want to erase the 4 GB card so I guess I'll stop here with my tests.

Finally, I tested the spisd.device I built and it works fine.

I think all of the conclusions from this issue should be added to the docs, as it will be easier for new users to find there.

kendickson commented 3 years ago

Thanks for all of the investigation SukkoPera, and for the additional feedback, jbilander. I'll add a 10k resistor to my build (great photos – very helpful!) and see if it's the trick I was missing as well. Cheers.

Rezamjj commented 1 year ago

Hi, I have the second version of this board and I have exactly the same problem (error 115). What should I do? Please help me.

jbilander commented 1 year ago

Hi, I have the second version of this board and I have exactly the same problem (error 115). What should I do? Please help me.

For the rev 2 board use the V2 driver and firmware available here, best if you compile/build them yourself and you will get the latest bugfixed driver.

https://github.com/jbilander/SDBox-v2 https://github.com/jbilander/amiga-par-to-spi-adapter