RGH Stock nand boots but xell won't

[shinji257]

I'm trying to install the coolrunner and get xell going so I can get the cpukey but for some reason it's failing to boot. On stock nand the coolrunner green light flashes once then boots fine. That's good. When I flash the glitched/xell image it fails to glitch and boot it. The green light just keeps flashing and never finishes. Any ideas?

I already installed the capacitor because during our dry run with the stock nand it took a bit to get a solid boot from it.

I know you need pics so here they are. They were taken with my Galaxy S2 using macro mode.

I attached what I think will probably be the key ones but I also uploaded them on my personal website as there were more pics (attachment limit 5). I apologize for the blurriness of the pictures. Hope they are sufficient. If not then I'll redo the pictures.

More Pics: http://www.robertpendell.com/rgh/

---------- Post added at 20:36 ---------- Previous post was at 20:33 ----------

For some reason the pictures don't seem to have actually attached even though I know I uploaded them. You should be able to see them in the link supplied above though. Sorry about that!

 

[BL4K3Y]

First of all, I suggest re-aligning the long QSB and re-soldering STBY_CLK.

Secondly, how did you build your XeLL (ECC image)?

Did your NAND have any bad blocks?

 

[shinji257]

There were no bad blocks. When dumping the nand there were no errors. Dumped 3 times and all 3 matched. I built xell using the 360 multi builder which has been mentioned on this forum already. Both 0.4 and 0.7.

I'll have another go at the CPU_RST wire and see if that works but I thought the long QSB was fine because I could dump the nand ok without error.

For reference my box is a Jasper 256MB board.

---------- Post added at 20:49 ---------- Previous post was at 20:48 ----------

Oops. Meant STBY_CLK as you said.

 

[BL4K3Y]

You could try Bestpig's ECC Generator or Python to build your ECC image.

I am yet to have a failure when using the Python command line.

 

[Martin C]

Also try setting the CR to 'slim' - I had a Jasper which wouldn't boot on 'phat'.

 

[shinji257]

You could try Bestpig's ECC Generator or Python to build your ECC image.

I am yet to have a failure when using the Python command line.

Thanks for the tip. I'll give that shot as soon as I'm able.

Also try setting the CR to 'slim' - I had a Jasper which wouldn't boot on 'phat'.

Tried that early on and after installing the capacitor. It didn't work. I had noticed the reverse working on some Trinity/Slim boards so that's why I had tried it.

FYI - For some reason I lost nand access (again). My box has been cursed I think and it doesn't want to be modified.

I actually have a co-worker working on the soldering stuff (with my eye checking the work) so I'll bug him to fix it up. He had to repair the 3v3 wire because it broke off last night.

 

[Martin C]

...which tells me that your soldering is probably at fault somewhere. Possibly a broken or bridged wire?

 

[shinji257]

We couldn't regain access to the nand again so we are going to change our game plan to use headers instead. I'll just need to wait for the parts to come in. We are clearing out the existing solder from the header areas so that it's prepped for the install later. Thanks for the assist.

P.S. - If this works out then I'll probably just use headers for any future install. It might just be the more sane thing to do.

 

[x360ccm]

why is there 2 people or more doing this.

it is a one man job.

 

[shinji257]

I know but I don't really have time to do the soldering and he wanted to learn anyways. The other reason is that he doesn't want to do any of the software half of things so I'm taking care of that part. The hardware mod was his idea (I never was going to go that far previously) and I'm letting him work the board for me. The xbox is mine however. Besides he owes me for other things I've done for him.

P.S. - It is only two people.

 

[shinji257]

Alright. Got the parts and finally did the pin header install (should of just done that from the start). It boots xell when in slim only. It doesn't get that far on phat. Even on slim it takes 29 seconds. Cap point not bridged. This is a new coolrunner rev b -- took out the old one to start fresh on pin header. Is that too long? How long should it take to do a glitch on a Jasper board?

---------- Post added at 14:18 ---------- Previous post was at 14:12 ----------

EDIT: Gone one that's 10 seconds so it seems to be a bit inconsistent.

 

[Hackerjac]

boot time depends on alot of things, wire routing for a start
if its too long time is up to you
i did one BB jasper and that booted within 10 sec every time, diden't do anything special to it

 

[shinji257]

Scratch the bit about it booting on slim mode only. I just had 2 successes with the coolrunner switched to phat and with similarly inconsistent boot times.

---------- Post added at 14:26 ---------- Previous post was at 14:26 ----------

@Hackerjac so given the routing (it's the same still as was posted in the opening post) do you have any suggestions?

 

[JayDee]

The exposed wire to pll_bypass looks awfully long imo (if it still looks like that, i would fix it!)

Ground+pll_bypass looks bridged on the (old?) CR and many of the solder points look very dry.
Get your friend to use flux next time, or get a scrap board and train yourself

 

[shinji257]

Oh we did a new CR board. That particular pic will be old. Here is a new one of the CR board. I'll get him to re-route the PLL_BYPASS wire and see if that helps but all wire lengths are standard from the kit. The only added wire here is that we ran STBY_CLK to the alt point on the underside of the board.

http://www.robertpendell.com/rgh/20120207_143430.jpg

---------- Post added at 14:47 ---------- Previous post was at 14:43 ----------

P.S. I deleted the old QSB and the former CR pics from the server to prevent confusion.

 

[shinji257]

PLL_BYPASS re-routed and we found that the solder point on the underside was loose after the fact so we redid it as well. Now it runs along with CPU_RST. Boot times are still inconsistent.

 

[DayvG]

Try routing your wires like this. (Martin C's Advice)

Edit: For clarity the CPU_RST runs topside along the board, then underneath.

 

[Martin C]

(CPU_RST is soldered to a standard copper ring with goes THROUGH the board, so just solder it topside )

 

[flymods]

If you lost access to r/w to the nand you can try and short out pins 44-47 on the flash chip which will reset it. Here is a link that may help you out.

Some of your solder joints look cold, if the guy doing the solder work doesn't know how to use flux and properly heat up both metals to form a good joint you may want someone else working on the system. Also the green wire going to the bottom of the board has too much bare wire exposed at the end and could possibly bridge on something, you may want to cut that down.

One good method for forming a better solder joint that you or your friend can practice is to take a tinned wire and put flux on the actual bare wire end. Then take a small soldering tool and spread some flux on the pad or area you want to solder the wire to. Flux tells the solder where to go so if you do it this way both pieces will bond better when you add the solder.

 

[DayvG]

(CPU_RST is soldered to a standard copper ring with goes THROUGH the board, so just solder it topside )

You know what, I did think about that whilst actually doing it, but didn't Just in case. I actually saw it done that way on one of the "lesser" reliable and helpful forums whilst researching the Cygnos for Matt2416 as his wiring was all over the place. Would you believe it, some more Golden Advice from Martin........do you ever take a day off??? :biggrin: