FileStream Exception

NETMF & Gadgeteer
1

I’m writing some data to the SD card and always get an exception when I try and flush the data.

I tried the sample code in the beginners book as well to see if the problem was with my code, it still raises the same exception.

PersistentStorage SD = new PersistentStorage("SD");
            SD.MountFileSystem();

            VolumeInfo D = VolumeInfo.GetVolumes()[0];
            if (D.IsFormatted)
            {
                StreamWriter w = new StreamWriter(D.RootDirectory + "\\results.txt", false);
                w.WriteLine("Time tested: " + timeTested);
                w.WriteLine("Average result: " + result);

                w.Flush(); // exception
                w.Close();
            }
            SD.UnmountFileSystem();

http://pastebin.com/h9nQVVXp exception
I have the correct firmware loaded.

2

How about telling us what kind of an exception you are getting? Show us the exception listing from the output window.

How is the sd card formatted?

Have you tried a different sd card?

3

http://pastebin.com/h9nQVVXp << you can see the exception there.

Yes the card is formated to FAT32 with the default cluster size.

No I only have one card, its a Kingston SD-C02G SDC/2GB.

4

Add a Thread.Sleep after you call SD.MountFileSystem. It seems some SD cards take longer to respond than others. I have a 3 second wait (Thread.Sleep(3000)) in my code, I did not do a lot of experimentation to see what the minimum pause time should be though.

5

Toshko, out of curiosity, why are you using paste bin instead posting the exception right here? I think it is easier to other to help you if you pout everything here instead of us need to navigate somewhere else.

You can use the "quotes to keep your exception quoted…like this :slight_smile:

6

Old habits die hard I guess Gus :slight_smile: anyway I’ve already tried adding a delay (read it from another topic) it still raises the exception.

Might there be a problem in the PersistentStorage class? I noticed on the fresh test project for this exception that if I added “GHIElectronics.NETMF.IO” it kept asking that I was missing a class USB_Host from “GHIElectronics.NETMF.USBHost”, when adding that reference and using it still kept saying the same thing. Adding the “GHIElectronics.NETMF.System” reference allowed me to compile.

7
  1. format the card on the PC (try one time with FAT and second FAT32)
  2. step in code and do not run the project. This will tell us it is a delay issue.
  3. if you still have problem then try a different card

4, you can also try to add a 47uF capacitor to the power pin on SD card. We have some seen some cards that go crazy on power spikes!

8

Found the problem, it will only work with FAT and NOT FAT32. Are there any recommended microSD cards that are known to work with FAT32?

9

Anyway we can borrow the card to test on our end?

10

I’ll see what the local post office can do. Here are two images of the card itself.

[url]http://ravenlabs.org/storage/TinyCLR/DSC00527.JPG[/url]
[url]http://ravenlabs.org/storage/TinyCLR/DSC00528.JPG[/url]

All the text on the first link:
SD-C02G TAIWAN
SDC/2GB 51 F©

And the second one:
1005
CP5010Y

11

I believe we tested similar Kingstone Cards and it turned out that these cards withdraw relativity a higher current than other kinds on write operations. and that causes voltage drop on the SD. To solve this issue you need to add a hardware fix or you use another kind of cards like SanDisk.

The hardware fix is by adding an additional 10uF Capacitor in parallel to C10 to resist voltage drop on the SD card.

12

Well, That’s a bummer. I had one of those cards and had the exact same issues with it. Nice to finally have some closure on what the cause of the issue was, though.

13

The real kicker is I didn’t buy a SanDisk because I though they’d be of inferior quality and not work correct :smiley:

14

Hahaha, well, SanDisk, Kingston, PNY and Lexar are the main SD card companies, and you can pretty much trust that they get their stuff straight. The only reason I bought the Kingston was because it was really cheap ($8) and I figured I could get away with it.

15

Yes they are, and they are around the same quality. What is different is the price and the hardware.

I would suggest adding a note on this in the ebook?

16

I would suggest putting this in a notes section of the manual or something. It’s not really the kind of thing that belongs in the ebook.

17

I was very excited to see this thread pop up. I was just about to post the exact same issue. This error has been crippling my system for a while now, and I’ve done exhaustive testing of different SD cards, formatting, FEZ software, even replacing a resistor as suggested in a previous thread (essentially adding a 4.7k R in parallel with R14–a pull down at SD power). None of that, save the recommended FAT format, seemed to do much of anything in the long run.

The most frustrating part of this situation is that I’ve been seeing this failure not immediately, but after various amounts of time between 1 second and nearly 2 hours into my data logging application. Even with Verbatim cards, I was consistently failing around 7 minutes into my execution. For the record, all I’m doing is simple USB Host->SD card writes of not more than a few hundred bytes of data/sec.

I will happily try this second HW mod, though I’m still uneasy, since the Verbatims I got to replace the Kingstons are also failing. I’ve been posting about this a bit on the main GHI USBizi forum, and they said NETMF 4.1 has many fixes for the FAT filesystem. It should be out within a week or two.

I caution those who think the problem is solved by moving to a different SD card. Not to rain on any parades, but I assumed the Verbatims were better when they lasted longer than the ~45 seconds of the Kingston (pre-R14 mod), but on further inspection I’m still getting this CLR_E_FILE_IO exception. Let’s hope the cap change will help.

Thanks.

18

I think the problem is not easy to solve because we have multiple sources to the issue. Remember that each card can have slight different characteristic. If we only had to support one single card of the same type, same FAT format and same size then it will be very easy!

The resistor is needed fro those cards that do not reset properly when power drops to about 1V. You need a resistor to help bringing the voltage lower. I see this a card problem but we have to work around it. You will see this problem ONLY when initializing the card.

The capacitor problem is related to card causing mini power surges. We already have 0.1uF capacitor but then we found out that some cards are too bad! 10uf is needed!! You may see this problem anytime since it is random.

The last possible problem is a bug in the FAT drivers. There has been a lot of little fixes on 4.1 so this may help or solve the problem. Once way to know if this is the source is by using USB memory stick. If you never see problem in USB then the FAT driver is fine and your issue is related to the 2 earlier ones.

19

I’ll test it out with an Apacer 4GB flash stick wen I get home.

20

If you are talking about a USB flash driver, you will not see the issue. It only occurs with SD cards. like the 2GB MicroSD Kingston model.