I am would like to solder a new SPI Flash onto my 8800GT due to the fact that it's current SPI Flash chip is no big enough to hold Mac firmware. I was wondering if there was anyone on this forum who has successfully done this? I have a AOYUE 968 rework station. Where can I buy the SPI Flash that is 1MB in size and compatible with the video card and nvflash? How do I go about programming the new SPI Flash?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Soldering a new SPI Flash
- Thread starter heisenburg
- Start date
- Sort by reaction score
If your card isn't reference one:
or very close to it, even larger chip won't probably help without ROM modifications.
You can get proprer EEPROM on mouser or another such parts supplier.
If you'll want to flash card with blank EEPROM, you need to do it on a PC, using nvflash. Most of large enough, 2.7 - 3.6 V SOIC-8 SST, ATMEL chips should work with it. Mac Pro won't boot with card with blank EEPROM. Been here, done that.
or very close to it, even larger chip won't probably help without ROM modifications.
You can get proprer EEPROM on mouser or another such parts supplier.
If you'll want to flash card with blank EEPROM, you need to do it on a PC, using nvflash. Most of large enough, 2.7 - 3.6 V SOIC-8 SST, ATMEL chips should work with it. Mac Pro won't boot with card with blank EEPROM. Been here, done that.
Yes, my card has exactly the same cooler as the one pictured. I bought some AT25F1024N-10SI-2.7 chips. Do you think that will work?
I always flash my Mac cards in a pc anyway so I'm not worry about the no boot problem
I said 2.7 - 3.6 V (flexible voltage) not without a cause. I don't remember to see chips on video cards using other than this voltage range. And I'm always buying such if I need one (rarily).
AT25F1024 - without N-something – was/is supported by nvflash. And it's 2.7 - 3.6 V. AFAIK higher voltage is used in programming mode, so using chip with voltage fixed on lowest supported would be probably no go. But don't quote me on this.
MacVidCards should know better than me. Maybe he will chime in...
Last edited:
I just got the new chips in the mail the other day. I was able to successfully remove the old eeprom of the 8800GT, but I had to max out the temperature on the Aoyue 968. I tried soldering on the new one, but it just would not stick even with the hottest heat setting. Am I doing something wrong?
Probably a ground plane sinking enough heat away that you're unable to desolder the component.
Not sure what equipment you have, or how you're using it, but you might want to turn the air flow down and take more time to see if that will get the joints to melt. Past that, you're looking at using multiple heat sources, such as a hot plate (preheater) beneath the EEPROM.
If you got the old one off, that's 90% of the battle.
GO to Radio Shack and drop $10 on a tiny, super cheapie solder gun.
Point is to have a fine point. Grab some super fine solder while you are at it. The Pb free stuff is better for environment but requires more BTUs to melt so get the gold old deadly stuff if you want to make your life easier.
Fancy reflow stuff is nice, but $12 at RS is money well spent.
GO to Radio Shack and drop $10 on a tiny, super cheapie solder gun.
Point is to have a fine point. Grab some super fine solder while you are at it. The Pb free stuff is better for environment but requires more BTUs to melt so get the gold old deadly stuff if you want to make your life easier.
Fancy reflow stuff is nice, but $12 at RS is money well spent.
Depends on how you remove it.
Cutting the legs from the package (assumes it's not leg-less such as BGA or MLF packaging) for example will do so without the need for heat until you go to desolder the legs. Most of it seems done before ever trying to remove the legs from the pads. Now imagine that too much heat is being wicked away by a ground plane, but enough is applied to the pad long enough that the pads lift off.
Not recommended as there's no heat control at all. Too much risk of damaging the board or components, especially with SMD parts.
For proper soldering, you either want a proper temp controlled iron or RF based design (Curie Point, such as Metcal/OKi). These types tend to be better at thermal recovery when using the correct tip for the job vs. a firestick (stick irons earned this nickname for a reason
).I do get the point of one time use, but if the OP knows someone with a proper iron, they'd be better off letting that person do it.
And if they're interested in working with electronics on a regular basis, invest in a decent iron to begin with (no need to break the bank to do this in the US). Hakko FX-888 is $80.27 on amazon, and includes free shipping.
My reasoning for mentioning this, is that the OP already has a hot air station.
Worst tip geometry out there. Proper chisel geometry to fit the pad size is far superior (better for transferring the heat to the joint).
Yes on the diameter recommendation (0.15" - 0.02"). And Yes, lead free does require more heat (one of the reasons getting the existing part off has been a challenge).
The logic behind ROHS is crap. It's not more dangerous to the user, nor does it release the lead as they reasoned (that would be a battery that's using nothing but lead for the plates, not an alloy, which is what leaded solder is, typically 60/40 or 63/37).
For a one time use, OK. For regular use, not in the least is this correct. Too much risk of damage to the boards, components, and operator (burns), even with a skilled operator.
