Upgrading a motherboard’s BIOS/UEFI (the hard way)(downtowndougbrown.com)
downtowndougbrown.com
Upgrading a motherboard’s BIOS/UEFI (the hard way)
https://www.downtowndougbrown.com/2021/12/upgrading-a-motherboards-bios-uefi-the-hard-way/
33 comments
I disagree. This seems to me to be the heart of tinkering. We don't always have the tools we need so sometimes improvise and especially since it was a cheap board and he knew the risks it seems perfectly fine to me. If everything needed to be perfect every time a lot of things would never happen. I smiled at that part of the story.
He had a heatsink; just didn't bother to put it on:
Success! I cleaned off the CPU with alcohol because I had been touching it, and fully installed the heatsink/fan.
Success! I cleaned off the CPU with alcohol because I had been touching it, and fully installed the heatsink/fan.
Author here — maybe it was dumb of me to try, but it didn’t come close to a bad temperature. I get the feeling that the BIOS doesn’t stress the CPU very much. Of course, I would never try to boot an OS or let it sit for much longer than I did.
The fact that it had a heatspreader probably made the heating much less rapid than it could've been with a bare-die CPU like an old Athlon or some of the mobile ones.
It's hard to imagine it using SMP, so only one core should be active, and it's not likely to use many of the advanced functional units. I imagine you have a pretty high "thermal margin" in that run state.
Modern CPUs will throttle or halt long before they fry themselves.
They have throttling, and the newer ones are definitely "thermally governed", but without a heatsink to provide thermal mass, I'm not sure if they can react quickly enough to stop going over the limits. That said, a CPU with a heatspreader will have more mass than a bare die.
I repaired a piece of manufacturing equipment in 2018. It was terribly slow, even more so than one would expect from a first gen core i3. The heatsink had spent 8 years, hovering a few millimeters above the cpu. The Intel thermal paste was untouched.
I still remember a GeForce MX440 I bought once. It was stuttering as hell in games. Overclocked it for better performance. Still stuttering.
Looked at it and two of the plastic clips (yeap, no screws back then lol) holding the heatsink+fan were snapped so the heatsink wasn't even touching the die... WTF!
That was pretty hilarious. It worked fine for years after I fixed it, long enough for caps to blow and be replaced.
Basically, they're fucking sturdy.
Looked at it and two of the plastic clips (yeap, no screws back then lol) holding the heatsink+fan were snapped so the heatsink wasn't even touching the die... WTF!
That was pretty hilarious. It worked fine for years after I fixed it, long enough for caps to blow and be replaced.
Basically, they're fucking sturdy.
To add on your anecdata: I've seen an MX440 with a fan glued to the die. Without a heatsink and I don't think that was a thermal glue.
It worked badly, but it didn't die.
What is "quickly enough"?
Isn't the reaction time today somewhere on the order of a few milliseconds? How much could a component overheat in that time? The maximum current, though high, is still limited, after all.
Also, there is "thermally governed", and there is that hard reset once a critical temperature has been reached. I'm admittedly ignorant, but a "extremely short reaction time" with regards to a critical event seems like something that one could reasonably expect.
"I static'ed my component X" is something one hears frequently, but I don't think I've ever read about anyone frying their CPU anytime in the last decade.
Edit: Looking at a somewhat recent paper [1] studying CPU frequency transition latency, even the highest latencies were still within 70µs(!). I think one could reasonably expect a similar or better latency regarding critical protection circuitry.
[1] https://www.cse.iitd.ac.in/~sbansal/col862-lpc/ftalat.pdf
Isn't the reaction time today somewhere on the order of a few milliseconds? How much could a component overheat in that time? The maximum current, though high, is still limited, after all.
Also, there is "thermally governed", and there is that hard reset once a critical temperature has been reached. I'm admittedly ignorant, but a "extremely short reaction time" with regards to a critical event seems like something that one could reasonably expect.
"I static'ed my component X" is something one hears frequently, but I don't think I've ever read about anyone frying their CPU anytime in the last decade.
Edit: Looking at a somewhat recent paper [1] studying CPU frequency transition latency, even the highest latencies were still within 70µs(!). I think one could reasonably expect a similar or better latency regarding critical protection circuitry.
[1] https://www.cse.iitd.ac.in/~sbansal/col862-lpc/ftalat.pdf
My experience is mostly with Intel mobile CPUs, which are mounted as bare die without a heatspreader, and those can jump 20-30C a second even in normal operation with a heatsink attached.
Sure, but that "second" you mention is an eternity compared to the max 30µs (see referenced paper) it takes a CPU to switch from highest to lowest frequency.
Why would you expect critical temperature threshold shutdown latency to take any longer?
Why would you expect critical temperature threshold shutdown latency to take any longer?
Lucky that he was able to use a clip - the power draw of the other components on the board sometimes overwhelms the programmer and causes the voltage to sag. When my BIOS failed I was unable to use a clip and had to desolder (don't have hot air equipment). Ended up pulling off one of the pads under the SOIC8 BIOS chip and had to reroute the pin of the newly programmed chip to another point on the board.
Also props to the writer for mentioning the fact that the CH341A drives the flash chip with 5V rather than 3.3V without modification. Many sites don't mention that and that can destroy a 3.3V chip. I believe the datasheet for my Winbond chip mentioned that the limit was 4.3V or something like that. I only figured that out when I probed the programmer's signals as a sanity check before programming.
Also props to the writer for mentioning the fact that the CH341A drives the flash chip with 5V rather than 3.3V without modification. Many sites don't mention that and that can destroy a 3.3V chip. I believe the datasheet for my Winbond chip mentioned that the limit was 4.3V or something like that. I only figured that out when I probed the programmer's signals as a sanity check before programming.
It used to be the case BIOS were socketed.
Now, there's no socket, no header (not even the footprint for it) and of course no microcontroller-based flash method.
It would cost them literal cents to do, but they don't. It's a shame.
Now, there's no socket, no header (not even the footprint for it) and of course no microcontroller-based flash method.
It would cost them literal cents to do, but they don't. It's a shame.
A socketed chip is hardly any easier to program when the clip works fine. It's also an extra cost added to millions of units which maybe 1-10 customers will ever actually make use of. It's just cheaper to offer a program to mail the boards in to be flashed.
Back when BIOS chips were socketed, I once flashed the wrong BIOS onto my motherboard's chip, so it wouldn't boot. Luckily I also had an older motherboard with the same BIOS chip model. So I booted up the old one, pulled its chip out while it was running, put the new one in and flashed the new board's version - without using any special tools at all.
>when the clip works fine
The clip is a hack. And it also does assume there's a diode to protect the board from back-feeding power, or else you have to turn on the board and do it warm, which has its own issues.
When the flash chip isn't socketed, it would cost literally nothing to provide the footprint for a programming header. Just SPI on a header.
That there isn't one is pretty much the middle finger to their customers.
>It's also an extra cost added to millions of units which maybe 1-10 customers will ever actually make use of. It's just cheaper to offer a program to mail the boards in to be flashed.
Negligible cost for just a header, even less so for a footprint.
It's not about these 1-10 customers, but about local repair shops (or any friendly tech guy) being able to do this. Much better than a mailback.
The clip is a hack. And it also does assume there's a diode to protect the board from back-feeding power, or else you have to turn on the board and do it warm, which has its own issues.
When the flash chip isn't socketed, it would cost literally nothing to provide the footprint for a programming header. Just SPI on a header.
That there isn't one is pretty much the middle finger to their customers.
>It's also an extra cost added to millions of units which maybe 1-10 customers will ever actually make use of. It's just cheaper to offer a program to mail the boards in to be flashed.
Negligible cost for just a header, even less so for a footprint.
It's not about these 1-10 customers, but about local repair shops (or any friendly tech guy) being able to do this. Much better than a mailback.
There are SPI headers for flashing on many many desktop mainboards.
With socketed BIOS chips, You could buy preprogrammed updates… no need to program yourself.
More expensive motherboards do have (presumably) MCU based flash tools so you can upgrade the motherboard via USB without a CPU. But this is a low end A series chipset board, it's going to be designed to be cheap, so extraneous features like that aren't going to be fitted.
There probably is also some sort of "header" to allow programming Flash in the factory, but it's going to be a set of test pads that need a "bed of nails" or similar to use. I suppose you could standardise it and make it easier to field use (others mention tag-connect cables), but it's such a niche use case I'm not sure anyone would spend the time on standardising it.
There probably is also some sort of "header" to allow programming Flash in the factory, but it's going to be a set of test pads that need a "bed of nails" or similar to use. I suppose you could standardise it and make it easier to field use (others mention tag-connect cables), but it's such a niche use case I'm not sure anyone would spend the time on standardising it.
There's a connector called Tag-Connect which is actually only a PCB footprint with holes, so you get it for "free".
The Tag-Connect cable on the other hand is anything but free.
I'm surprised there is no competition
Yep, last time I took a look at the cable price they're pretty outrageous. I suppose they only make money on cables, not connectors - hence the markup. And it seems like the Tag-Connect is patented, which is why nobody has tried to make a clone of it?
That is super cool. Thanks for the reference. The TTL footprint would have saved me a ton of hassle on a recent project.
Obviously there are generic solutions without the PCB alignment holes (and aren't patented), if you need something along those lines.
Although I think clips are roughly the same price if you go that route as well.
Although I think clips are roughly the same price if you go that route as well.
I wonder if it would have been easier to use a JTAG probe to read/write the BIOS. I would be surprised if the motherboard didn't have a JTAG port.
It does have an SPI TPM header, which as far as I was able to tell, does bring out all of the signals I needed (including the BIOS chip select). But it's just all of the same signals in a different form factor. It was actually easier to use the chip clip because I didn't have header pins small enough to plug into it. It would be nice if the motherboard manufacturers would standardize on a header (or footprint) that everyone would use...
I had a similar issue like 5 years ago.
I just called up ASUS and shipped the board to them, I paid the shipping cost and they reprogrammed the BIOS for free, besides shipping.
I just called up ASUS and shipped the board to them, I paid the shipping cost and they reprogrammed the BIOS for free, besides shipping.
unbanned(1)
That was a bit cringeworthy as someone who remembers the old TomsHardware video, and knows how rapidly new CPUs can change temperature/power consumption. You might end up with a damaged CPU, finger, or both. The amount of thermal mass relative to the power dissipation is tiny, and the BIOS definitely isn't going to put the CPU in power-saving mode right out of reset. I strongly recommend never powering on a "bare CPU" without a heatsink on it for anything newer than a (original) Pentium. Not having a working fan on the heatsink is not as bad, since the heatsink has plenty of thermal mass.