H120 CF mod - enable UDMA on USB bridge

[dobbley]

Please add VID=0x0e21, PID=0x0500 (iAudio M3) to the tool for testing. The X5 "side port" ID in turn should be removed. This goes to the OTG chip, not the ISD300. I could add it myself, but I don't know what's needed to build the tool.

Changes made in wiki (0.9.2)

[amiconn]

dobbley: Thanks for adding the M3 PID. I was now able to test the M3 as well. It also has UDMA enabled by default, as its successors.

[bluebrother]

I've added a diff to the wiki page that allows building on linux (at least it did for me). The tool needs to be run as root, otherwise it won't find the bridge. It worked fine for me -- speed increase is similar to the values stated in the wiki.

[markanini]

I wanna try the USB DMA mod on my H120. Do I connect these points to get around write protect?

[DTSyX]

If I may 'reopen' this thread.

I have no experience with such operations so it's a bit diffcult for me to see what dobbley actually did (http://forums.rockbox.org/index.php/topic,20402.msg144038.html#msg144038). And also as far as I understand, even though he seem to agree with dreamlayers (http://forums.rockbox.org/index.php/topic,20402.msg144036.html#msg144036) I think he did something else.
So for me, it's a bit hard to actually understand what he really was doing, at least not good enough to feel encouraged enough to try it myself.

Dreamlayers says: "Above the IC is a circular silvery pad with the letters WP below it.  I can't see what it connects to, but I suspect it connects to the write protect pin, and if you ground it, you should be able to write to the EEPROM (normally, via USB and the ISD-300)."


While dobbley states: "So then took a deep breath, grounded one end of the wire and put the other in place at the EEPROM side of the resistor."


To me, the silvery WP pad and the EEPROM side of the resistor are not the same thing.


So, I ask the ones who actually succeeded in this task, mainly dobbley, could you please post a version of that picture that is referred to in this thread marking the element(s) which have to be grounded?

It would be very helfpul and appreciated!


Because trial and error is not something I want to apply on my loved H140... :-)


Thank you very much!

[dobbley]

The picture from markanini is what I did. If you look at the 24LCO1 schematic that dreamlayers posted the link to, you can see that pin 7 is the write protect pin. This chip is the one between those resistors and the USB connector. You can see that pin 7 (count round from pin 1 with the dot) is connected to R18. Pin 7 is also connected to the WP pad underneath the chip, so electrically they are the same thing.

The reason for using the point on the resistor is that it is a lot easier to get to than the WP pad if you've just taken the bottom end off the unit. It's still very fiddly as you need to be very careful the wire does not move and short anything. It may be easier to use a different ground point you can keep attached (e.g. chassis screw holes, check with a DMM), as you need a spare hand.

I would recommend getting the link in place and removing it with the power off to avoid shorting anything, and don't have any coffee beforehand! Be particularly careful to avoid the closest side of those resistors as that is VCC which you definitely do NOT want to short to ground. This is the reason for using insulated wire with the bare minimum exposed. I did this on a H120, so I can't guarantee it is the same on a H140.

[DTSyX]

dobbley, thanks a lot for clearing that up!
 
I was wondering why WP didn't "play a role" in markanini's picture when it comes to "connecting" or grounding. Now I understand, especially because I had the wrong orientation of the EEPROM in my head when looking at the picture of the board (I just assumed it was the same as in the pdf when it's rotated by 180° as you point out).
 
Now I feel more comfortable when planning to upgrade my loved iRiver.

So, thanks again a lot, dobbley. That was very helpful for me.
 
I'll post here if and when I succeeded (or failed) with my H140. But that might take a while. Today I did some transfer rate tests with the temporary solution. Depending of the tool to measure the rates it's around 60% higher for writing and 50-100% higher for reading. When testing with an actual, ~3,5GB file, it's around 60% faster for reading and writing, so all in all worth trying to make it permanent.

[DTSyX]

Yeah! I actually did it!
 
Thanks to dobbley's clarification I felt sure enough to pull it off. When you look at the pictures and the crappy setup you'll understand, that I propably would not have tried it if I hadn't a backup H140 in my closet... :-)
 
The "cable" you see on the picture isn't actually a cable but that thing that is often used to hold together rolled up cables when you buy them. I chose that thing because the metal inside is quite rigid, so I would feel resistance when I bend it at one end by almost 90° and use that hook to get hold of "the other side" of R18, the side that is faced away and dobbley said would be the side away from VCC.

It took me some time and many pictures, that I zoomed in to get a better mental picture how the feeling in my hand of the resistance when pulling the hook against the back of (hopefully) R18 would match the actual situation. I have to admit, that I probably never really saw the hook as clearly as I would have wanted, but in the end I feld secure enough to connect the player to the power, switch it on and write on the EEPROM.
 
The other end of the "cable" touches (as proposed by dobbley) the metal part of one of the screw holes for grounding.
 
As for the sequence of the setup:
 
I removed the accu and the hdd I normally use, in oder to not damage them in case something went wrong. Looking at the setup, you'll understand why :-). I tried without hdd, but apparently that doesn't work. So I used the old, original hdd that has still an old version of Rockbox on it.
 
I connected the power cord (but did not yet put the plug into the socket!) and the usb connector to the H140 in order to not to have to move the player around (involuntarily) once the "cable" was in place. Also in the final setup the player was kind of stuck under the window so it wouldn't move when I hit the "on" (play) button. I also connected the other usb connector to the pc.
 
Once this setup was in place and I felt comfortable with what I saw and felt in my finger tips, I plugged in the power connector, switched on the player and executed the isd300_udma.exe. No sparks, no smoke, no fire. I could enter the EEPROM writing mode and change the setting to 4. All went fine, and now I can access / load my H140 via USB about 60% faster.
 
Thanks to all that made this hack possible! And a special thanks again to dobbley for giving me the final confidence I needed to pull that off!
 
Sorry for the setup picture. I had to delete some picture information in order to get below the allowed file size without reducing any more of the picture quality or excerpt in order for it to be as "understandable" as possible. And don't mind the file names. They only really make sense to me. :-)
 
Please consult the picture from this post (http://forums.rockbox.org/index.php/topic,20402.msg186903.html#msg186903) for further and more detailed visual information.



[Edit: December 6th, 2014]

Today, almost 2 years after "hacking" my first (and still running) H140, I "hacked" my spare (and still unused) H140 using my own "manual" (above). It worked again. Only this time I used a vice with rubber-protected jaws to fixate the player which made the job so much easier and faster. I didn't see any more than last time though... :-)

Btw., as I didn't mention it before and it is a crucial part of the operation: I didn't "switch off" the player after the successful operation, but I removed the hdd/player from the windows system using the USB-icon and then just pulled the power connector from the wall socket. It's brute force, I know. The HDD should (and did) survive this one time stunt. Of course, you better use a spare hdd (the one that came with the player) instead of your big, expensive hdd you just bought.

I still wonder why iRiver didn't activate UDMA in the first place, or at least made the write-protection more user-friendly.
Too bad it's "only" UDMA4, not 6....

[/Edit]