Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

EFI Boot Menu

persianz8 said:
perhaps it is my logic board...but i'm soo pissed that i have to spend $1000 for another logic board when it was from a simple software update...i mean how can a software update affect hardware like this?
Click to expand...
Assuming the update included a firmware update, it's possible that a glitch occured during the process (i.e power fluctuation such as a brown-out or outage).

If not, then the partition tables could have been damaged (not the case though, from what you've posted). In such cases, you'd still be able to load the Snow Leopard install disk, and reinstall the OS, then restore the data from a backup (or restore the entire drive from a clone).

In your case, Apple should support you on this, but I've a feeling they won't, as there's no way to prove it (firmware update = FAIL, even though it's rather obvious). Apple doesn't have a way for the geniuses to restore the firmware to a board (remove the ROM chip, reprogram, and resolder, as it's not a DIP package with exposed pins you can use a clip to reprogram).

So they have to replace the board. If under warranty, you get it for free. If not, you "eat" the cost of a board. :(

Sorry, as this really sucks. :(
 
hyram said:
It can be done. The files don't match exactly as the EFI keeps some of it's info in volume 4, so this changes somewhat. The other volumes are static.

I'm sending you a PM.
Click to expand...
For some reason, I'm remembering the ROM is FBGA. Is it something else package wise (easier to do that doesn't involve solder/desolder and adapters)?
 
Nano, It has to be removed. Parallel flash devices don't support the clip-a-chip type methods for flashing.
 
hyram said:
Nano, It has to be removed. Parallel flash devices don't support the clip-a-chip type methods for flashing.
Click to expand...
This is what I'm familiar with (example). But I was hoping someone had come up with a clip that would eliminate desoldering (I do think it's technically possible).
 
nanofrog said:
This is what I'm familiar with (example). But I was hoping someone had come up with a clip that would eliminate desoldering (I do think it's technically possible).
Click to expand...

The clip-a-chip works well for serial parts that work off of 3.6-5V. When you move below this voltage they get a little fussy due to loading. Even more so for the parallel parts, it is just too much of a load to reliably read/write and there can also be contention with CS's and WE/RE enables. Most manufactures have moved to JTAG, B-scan or some other ISP for factory flashing. And most systems have the code to re-flash. The problem arrises when something goes amiss and the device gets horked.... Like has happened to the OP. It can only truely be re-programmed by removing the device and sticking it in a programmer.

MCUMall stuff (what you linked) is good. I've got one of their programmers.

hyram
 
hyram said:
The clip-a-chip works well for serial parts that work off of 3.6-5V. When you move below this voltage they get a little fussy due to loading. Even more so for the parallel parts, it is just too much of a load to reliably read/write and there can also be contention with CS's and WE/RE enables. Most manufactures have moved to JTAG, B-scan or some other ISP for factory flashing. And most systems have the code to re-flash. The problem arrises when something goes amiss and the device gets horked.... Like has happened to the OP. It can only truely be re-programmed by removing the device and sticking it in a programmer.

MCUMall stuff (what you linked) is good. I've got one of their programmers.

hyram
Click to expand...
I hadn't thought loading would be an issue with the newer programmers (USB, not parallel port). Particularly the Xeltek's, as they can go down to 1.2VDC. That's why I was wondering if anyone had created a different adapter.

Few adapters I encounter are clip types (primarilly DIP). The most recent package that I've seen is the SOIC8 IIRC. The rest are of a similar type to that linked previously (PCB adapter that the part has to be snapped in). They're great for engineering proto work, and the large units (multi-gang) for production. They're a bit on the labor intensive side for repair work though, not to mention higher risk of damaging the component or board if over-heated (yes, I've lifted traces a time or two).
 
nanofrog said:
I hadn't thought loading would be an issue with the newer programmers ...
Click to expand...

The loading comes in because you're attempting ISP, it's from the other devices on the board that the device you're trying to program is connected to.
 
hyram said:
The loading comes in because you're attempting ISP, it's from the other devices on the board that the device you're trying to program is connected to.
Click to expand...
As the adapters typically don't exist, so I hadn't tried it (pad attached leads or pins on a proto). Obviously no problems with popping chips in and out, but the desolder step drives me nutz, as if it's not a proto device, there's quite a bit of work involved. Oh well... :rolleyes: :D

But what throws me, is the loading shouldn't be as big an issue, as that's the goal of ISP in order to save time (leave it on the board, as it contains the circuit to enable it to generate it's programming voltage in the device).
 
nanofrog said:
But what throws me, is the loading shouldn't be as big an issue, as that's the goal of ISP in order to save time (leave it on the board, as it contains the circuit to enable it to generate it's programming voltage in the device).
Click to expand...
Mac Pro uses Firmware Hub mode that greatly reduces I/O count compared to full parallel mode but even then ISP has to be designed into the layout if you plan to use it later. If EEPROM is just hooked up directly to the bus drivers then not only your programmer will attempt to power up anything attached to the supply rail but also will be fighting for control with bus drivers. And they are good at what they do. I am not sure why you specifically refer to USB programmers? Maximum supplied power by USB socket is only 2.5W.
 
lbodnar said:
Mac Pro uses Firmware Hub mode that greatly reduces I/O count compared to full parallel mode but even then ISP has to be designed into the layout if you plan to use it later. If EEPROM is just hooked up directly to the bus drivers then not only your programmer will attempt to power up anything attached to the supply rail but also will be fighting for control with bus drivers. And they are good at what they do. I am not sure why you specifically refer to USB programmers? Maximum supplied power by USB socket is only 2.5W.
Click to expand...
I only mentioned USB, as they're newer, and easier to find now. Centronics are still out there, but not as common, when I look around at the models available.

I also made the assumption the ISP design would accomodate it, as there's no way to access the firmware and recover from a bad flash (i.e. power loss). It would make sense to me, that they did design the board to accomodate reprogramming this way as a result.
 
nanofrog said:
I also made the assumption the ISP design would accomodate it, as there's no way to access the firmware and recover from a bad flash (i.e. power loss). It would make sense to me, that they did design the board to accomodate reprogramming this way as a result.
Click to expand...
Maybe on a large scale logic board cost is probably less then $100 so it's cheaper to replace it then pay for technician to troubleshoot and fix it (which at this complexity level needs either luck or lots of time and equipment.)
 
lbodnar said:
Maybe on a large scale logic board cost is probably less then $100 so it's cheaper to replace it then pay for technician to troubleshoot and fix it (which at this complexity level needs either luck or lots of time and equipment.)
Click to expand...
This is quite understandable, as the systems are so complex, the labor spent on diagnostics will usuall exceed the manufacturing costs. Ultimately it's less expensive for the system vendor to just do a swap (Apple in this case).

To me, the issue is, that users who's system has gone out of warranty are having to eat the board cost at retail. ~$800USD last I checked for '08 system boards. The '09's would be a bit easier on them, as it's a main + daughter board configuration. So assuming both boards aren't DOA, only one needs replaced. They don't provide an exchange program in such instances (where non test equipment can determine the fault correctly = failed firmware update in the OP's case). Apple could offer the boards at a reduced rate (still enough that they're not losing anything and even have a modest profit), as there's no provision to allow for a faulty flash to be recovered due to the lack of firmware access.

Given the lack of a recovery method, and the price of the boards, it's a little harsh for non warranty coverage cases. That has to hurt, as I'm accustomed to new DP Harpertown boards being under $600USD for the most expensive, and a few under that (depending on specific features of course).
 
persianz8 said:
thanks to all and your suggestions. I came to the conclusion that the best thing for me is to take the bloody thing to the store and have them replace my logic board which will cost me $893.83 canadian.

but all of your ideas were helpful had I not tried them myself. THANKS TO ALL..
Click to expand...

persianz8,

What was the conclusionto all of this? Just like to know.

hyram
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back