I was thinking there might be subtle differences between them and just wanted to make it as clear as possible what I was doing.

Did you check the other file to see of it is indeed corrupted? No errors were displayed when I read it.

Anyway, I’ve managed to get another read on MX25U4033E, this time with the adapter. Here is that new file: https://ufile.io/bb070qhs (524.3kB/512kB)


I’m going to try and get some separate dumps of Gigadevice 25D10BT using the programs you sent me in Windows. I’ll update you when I’ve got them (if I get it working)!

Also, I want to thank you for being so helpful. It’s not fixed yet, but your help has been invaluable so far. :)


Live updates:

* I’ve successfully installed the driver.
* Running CH341A programm v1.34 - I managed to find 25D10 (minus the BT) and have read it. File added below.
* Going to try out another program.
* Using CH341AFree. Upon using the Detect button, it displays ‘Chip Info: C8401110’. Not sure what that means, but I’ve read the chip and created another dump. File below.
* Running CH341 Programmer v1.1.1.32...it’s in Russian! Using Google Translate, I pressed the button labelled ‘Identify Chip’ - it’s displaying the chip as 1 MBIT GD25Q10. Managed to find the read chip and and save file buttons - fortunately GUI layouts are relatively similar, despite the language barrier. File below.
* Tried to run Colibri, but despite showing in the taskbar, it’s not responding (as seen in task manager). No windows coming up, except for the Windows one asking for confirmation to run.
* Using AsProgrammer. Had to select CH341a under the Hardware tab, but managed to find the chip. It listed this info:

Current programmer: CH341
ID(9F): C84011
ID(90): C810
ID(AB): 10
ID(15): FFFF

ID(9F) is the same as what CH341AFree showed (minus the 10, which is also ID-AB), so that’s interesting.

I went to save it, but it wanted to save as a ‘.’ file, aka no file extension, so I just added .bin when saving and it seemed to save just fine with that extension. File below.

* Lastly, Colibri. I’m still struggling to get it to work. I’ll update if I do, but I’m not hopeful.


Ch341a AsProgrammer_1.4.1 - https://ufile.io/hb0dr05z
CH341A programm v1.34 - https://ufile.io/lxmu86s8
CH341A v1.31(1.4) (CH341AFree) - https://ufile.io/objyi4bd
CH341 Programmer v1.1.32 - https://ufile.io/005kn7da
Colibri_program 1.0.1.61-CH341A - Not available yet

So that’s 4/5 new dump files from separate programs for chip 25D10BT as requested. Also the dump file from MX25U4033E using the adapter. Hope this is what you wanted :)

@TheCM - Yes, I looked at the file you sent before, as I mentioned. No errors means nothing, you wont see any errors in hex editor with any file, corrupted or not, so I'm not sure what you expected to see error-wise
I just wanted proper 1.8V adapter dump to comapre with, to see if same/dif, etc so we could at least be sure of it looking "OK" as we could Thanks for the new dump, I will check it out.

You're welcome! Hopefully we can fix!

MX25U4033E = vBIOS dump for GP106 GFX Card (1060/1050 + few non-standards - TDP 120W Max), and both files match, so maybe your power level dropped from 3V down to 1.5-2V low enough for read without adapter to be OK
Neither of the vBIOS inside the main BIOS match this, they are both for GP107 card (1050 only + some other non-standard - TDP 75W max). What is your actual card model?
And, if you have the system apart, and have thermal paste handy, might be a good time to repaste the GFX Card, and while you do give me the actual ID printed on the silicon or PCB around the silicon.
Since both could be 1050, maybe it's best you check silicon to confirm, in case you should have GP107 vBIOS in there? Well, I guess even if you check, we can't tell if it should be 106/107 since both can = 1050 card - Not sure what to do there, I guess you can try the other two from BIOS at some point

CH341 Programmer v1.1.1.32, change the language, last menu up top

Save without extension or any random extension is fine. Not sure what's up with Colibri not working for you? What OS are you using? Try disable anti-virus and close any anti spy/adware programs, in case they are causing issue.
Please package all files into a single zip, thanks

@Lost_N_BIOS

In terms of errors, I was expecting something to come up in the terminal for that chip. However, it recognised and read it just fine. Maybe you’re right about the power level dropping...I don’t know.

This laptop has a GTX 1060 (6GB). Definitely not a 1050.

The repair shop I took the laptop to told me there wasn’t anything wrong with the BIOS, however I’m pretty sure they only checked the main BIOS chip and not the one for the EC FW (I think I flashed the wrong model by accident). They also told me there was an issue with the graphics card, which I had suspected. Specifically that it’s likely gotten too hot and melted the contact points?

It was the main reason I attempted to update the BIOS in the first place (although this was before I got it checked at the repair shop).

I think the issue with the graphics card is a separate issue as it booted to Windows just fine before I tried to update the BIOS. There were issues with the GPU then though.

I could try repasting it, but it seems risky. Right now it’s best we focus on trying to sort out the EC chip to hopefully get it to POST. In fact, I was planning on dealing with it after.



Looking at this image again, we know that:

Chip #1 (MX25L6473F) is likely the main BIOS chip.
Chip #2 (GD25D10BT) might be the EC/KBC chip.
Chip #3 (3940S-A) is a linear fan driver chip.
Chip #4 (MXIC 25U4033E) is the vBIOS chip.

Can you confirm that 25D10 is indeed the EC FW chip? (Potentially from looking at the files provided below) However, you did say this before:

Zitat

---

this is 2114A FW, so does not look like EC FW, contents do not look similar nor does naming scheme match. So this is some other items FW chip, or wrong contents flashed into chip etc.

---

If this isn’t the EC FW chip, what is?


I’m not sure what changed, but Colibri is working now. Maybe a couple of restarts and a day later fixed it? Either way, I’ve got another dump of that chip. Again, it’s recognising it as GD25Q10 and not GD25D10. According to this datasheet, that’s just the product series.

Here’s that .ZIP of all five separate dumps for GD25D10 (now including the Colibri dump): https://ufile.io/2kwm5w7m


Update:

I forgot to mention - the repair shop replaced the BIOS chip when checking it (at no cost) and I just read from it using the tools you sent me. They told me the hex code was copied over from the previous chip, but after checking for myself, it was completely empty! I created a backup just in case and then rewrote the updated BIOS.ALL file from the download portal.

Of course, I didn’t expect this to fix anything as it was already in this state before I sent it off to them. However, if we can verfiy which chip is EC FW 100%, I can write N85HPTR1.08 to that EC chip and this should hopefully fix my issue. I’m just waiting on confirmation from you.

@TheCM - Re-pasting the GPU is same as re-pasting a CPU, only you need to carefully alternate screw tightening as you go if it's a bare chip unlike CPU, generally that is done with CPU mount too but it's far more critical if chip is bare silicon so you don't crush a corner
If/Since it's been overheating, this is highly suggested, the old paste under there now is likely crap to begin with, too thickly applied, and is now baked through and old. Unless the repair shop re-pasted it when they looked?
Still, odd to see three GP107 vBIOS in BIOS only, and chip dump off the board is GP106, maybe the repair guy switched your card for a dead card and that's why he mentioned graphics specifically (so when you found it was dead you maybe wouldn't notice the switch)?
Or maybe this due to wrong BIOS package you linked me too? I see this only in the stock BIOS package is where I'm looking. What did he say when he gave it back to you about fixing it, was it working then, or he said he couldn't fix etc, I forget?

Send email off to XMG and ask them which chip is EC FW chip, or where it's located on the board. No, I cannot confirm 25D10 is the EC FW chip, I can only confirm it's not or it's current contents are not EC FW.
Open EC FW in HEX, you will see UUITE Tech near start, this is type of many EC FW. If you then look at current contents of 25D10 chip, this looks nothing like the EC FW.
Nothing under memory, you checked, I forget?

Sometimes EC FW is not on a chip we can read/write like this, but rather one of the QFP chips (Quad Flat Panel), and can only be written to through an onboard header like COM or SPI header (likely not there right now), or via KB ribbon cable etc.
This kind of info can usually be seen in a schematic or boardview file for the motherboard. It may also be EC FW is on the other side of the board, have you looked at the other side at all yet?
If not, do that now. You may be able to see 80% of it with the keyboard removed, but ideally it would be best to pull out the board so you can see it all. And this way, you may find board model # too, then I can try to find free schematic/boardview and we'll know exactly what's where then. - See below

CH341AFree dump of GD25D10 - You forgot to pick correct chip size
All these are 100% match except for the above, due to that (it multiplied the contents x8 to fill the space) - This is not EC FW /= >> 2114A_FW 2114A_RCFG

I found the schematic EC/KBC chip is >> ITE 8587 and it's 128pin LQFP as I suspected See Appendix B >> B-2 & See also B-49 (UART is how it's accessed for read/write via KB Ribbon cable)
Looks like you can program it through the KB ribbon cable, but I don't know how you do that
http://s472165864.onlinehome.fr/anyware/...%2020170420.pdf

@Lost_N_BIOS

I agree, it definitely could do with the thermal paste being reapplied. I’ll be sure to alternate tightening each screw as you say when I do. As I mentioned before, right now getting it to POST is priority, fixing issues with the GPU can come later. :)

I doubt the repair guys switched the cards, they were only supposed to give us a quote and assess what was wrong. There was already a problem with the GPU before they got it - I just didn’t know what. However, that took almost three weeks as they were waiting on a BIOS chip to arrive - which was a colossal waste of time.

Where are you seeing GP107? Are you referring to this BIOS package from the official site (see Firmware)? That’s the one I have downloaded as I’m pretty confident it’s correct. And since the new chip was blank, the BIOS.ALL file from that is what I have written to the chip. It’s official, so it must be right...surely?

From what I can see, GP107 wouldn’t make any sense as the 1060 doesn’t support that. I can’t think why it would be on there. You say the dump I provided originally is GP106?

Order of events: Bought laptop > Fixed separate issue > Realised there was an issue with the GPU > Updated drivers > Attempted to update EC/KBC BIOS > No longer POSTs - Tried to fix it myself, with your help > Sent it off to the repair shop for a quote > Got it back, allowing us to keep trying > Now.

We were told they used to do this particular repair for GPUs, but it wasn’t reliable and they don’t do it anymore. No, it wasn’t working before or after they got it, only before I tried to update the BIOS.

There aren’t any chips under memory, just info like part numbers and model.

—

That’s both good and bad news. We now know which chip holds the EC/KBC FW, but it’s an entirely different chip type. I hadn’t come across the service manual yet, so thanks for sharing (I only had the User Manual). A lot of useful info in there!

So ITE 8587. What can we try? Is there an equivalent of the CH341a SPI programmer for this chip? You say I can use a KB ribbon cable - would you please provide more info on this? Perhaps I can look and order one. If programming it myself isn’t an option, perhaps I can buy one pre-programmed and just solder it myself? Any other options if not?

I’d really hate to give up here because we can’t interface with this chip. There must be something I can do.

@TheCM - The main reason I mentioned to redo the GFX paste, aside from you mentioning overheating issues, was so that you could read the actual ID off the silicon so we'd know if GP106/GP107
Yes, supposed to do, and what they did, may not always = same/same People are terrible sometimes! How would they know anything wrong with GPU, if they didn't make it boot? That's what caught my eye and made me think that
Plus what you say now, asses only and let you know/quote, why would they need BIOS chip to asses, and they is called re-work for them to swap out chip, usually charge for this and not part of assess/quote.
Additionally, they should have been able to program chip in place, no need to swap chips, so it all sounds fishy to me and I have a VERY LOW level of trust for these kind of people (usually WAY overcharge, don't care about customer data, do minimal to function and collect huge overpriced fees - next)

GP107 is the only vBIOS inside the actual BIOS, x2-3 (can't remember), but this is how vBIOS on card would be updated when needed during BIOS updates, and or this is how card would run if onboard vBIOS
This is why I wanted you to check the silicon directly, and why I mentioned maybe they swapped cards on you after knowing the couldn't fix it and could tell the card was ran hot.
Since they knew you wouldn't get it fixed BIOS-wise easily, so mention card issues, snag-card/swap other dead etc, and hope you never know due to either you replace card in your attempt to fix, or you end up tossing system and getting new etc.
Maybe I'm just paranoid, but since we aren't getting things going, and all this is coming to light and my mind, here is where I am

See bottom at each, for general cards/models using GP10x
GP107 = (1050 only + some other non-standard - TDP 75W max)
https://www.techpowerup.com/gpu-specs/nvidia-gp107.g801

GP106 GFX Card (1060/1050 + few non-standards - TDP 120W Max)
https://www.techpowerup.com/gpu-specs/nvidia-gp106.g797

It may be your card is 100% original, and fine, and vBIOS on card is used only and vBIOS in BIOS is never invoked due to how this model is constructed, I don't know for sure, only can tell you what I see.

As for programming the EC, it's done through UART protocols, using the ribbon cable from the KB. How that is done I've never seen, only randomly discussed, but I'm sure how to in general is out there in google, I know CH341A can sometimes be used.
I found one general example, for another chip, but this will give you the idea and maybe help you locate other info - http://dangerousprototypes.com/docs/Flas...with_Bus_Pirate

Thread here, similar situation, they discuss how to and adapter purchase etc - http://forum.notebookreview.com/threads/...-update.825156/

You may need to RMA the system

* Edit - Since they replaced the BIOS chip, maybe they put it on backwards? Or damaged the pads under it etc? Please show me up close images of any chip you can see they replaced.

@Lost_N_BIOS

Thanks for the info on programming the EC chip. I’ve looked into it a little and it’s all a bit overwhelming. The RT809F programmer and all the necessary equipment seems like it will cost a fair bit (£50-£100). There’s also a site you can buy pre-programmed chips from for about £30 - maybe that’s a solution. It would just involve soldering I think.

For now, I’ve made a post on their forums in hopes someone will be able to help in programming this specific chip.

When I had the laptop up and running, all signs pointed to it being a GTX 1060 (6GB). The specs also back that up. However, I’ll try my hand at repasting the chips along with obtaining that ID at some point. I’ve never personally done it, but I know some who have.

As for the repair shop - they assured me just getting a quote would be free and it was. They decided to replace the chip as it was only cheap and they haven’t charged me for anything. I think it was easier for them to do that to test it. They told me they no longer do that specific GPU repair, so I can’t see why they’d lie about it. I got it back from them just fine. However, you are making me doubt them with the GPU swap theory. Who knows what they could have done?

However, I do think they have a communication problem with their individual teams. They were told it was an issue with the EC FW, not main BIOS. They didn’t touch the ITE chip AFAIK.

I’m still a little confused about the vBIOS/BIOS thing. The main BIOS chip also has vBIOS even though there’s a separate chip for that? That separate chip (MXIC 25U4033E) had GP106, right? But the main BIOS had GP107?

@theCM - I believe you just need an adapter, and can use CH341A. If you have skills, and a hot air re-work station with bottom heating plate, then replacing QFP package is doable, without that and previous experience it would be quite a mess, even if you purchased a cheaper priced hot air gun and bottom heating plate.
EC may be fine, but it's the only thing I can think of here that may be wrong, since you've written in correct stock BIOS and dumped that pre-start to confirm it was written in good. Or.... the GFX card thing I mentioned
What you can try here now that I think about this EC/Wrong flash issue, is program in the BIOS that goes along with the EC you think you flashed, then if it's compatible BIOS it may start the board.
But if the BIOS or EC is not compatible with your model, then it would probably fail anyway

Replacing chip/touching your board with soldering iron should not have been done for a free quote, nor should it have been needed at all even if they wanted to re-program the BIOS.
Right, they don't repair BGA GFX anymore, all the more reason to suspect what I mentioned, give you dead card and have good/working card for next person and charge them $100-200 for the card.
All speculation though, unless you have an image of the insides before you gave it to them?

Yes, correct about the vBIOS/BIOS. Generally they put vBIOS into BIOS, so it can be ran from there and in those cases GFX card usually doesn't have it's own, and or it can be copied from BIOS to vBIOS on card during specially setup flashing sessions
Yes, the chip you dumped has GP106, and the BIOS has GP107, that's why I started asking about confirmation of the card via silicon and what you knew was in there etc.
Maybe they programmed in wrong vBIOS to that chip too, in their hunt to try and find the cause of the issue, which you already told them was EC FW
And if so, then they'd have to have replaced this chip too, since you know they can't program a BIOS without replacing the chip

Here check it out, top in hex editor is what you dumped off the chip, bottom x3 is 3 vBIOS in the actual BIOS

@Lost_N_BIOS

Which adapter specifically? And would I need any other equipment?
When you say replacing QFP package, do you mean replacing the actual chip or the FW on it?

I’m pretty confident this is an EC issue as the problems only started when I attempted to update that. I ran the files from USB and used the DOS command line.

I think I accidentally flashed the EC FW for the XMG A517-M18. The files for that are found here: https://download.schenker-tech.de/package/xmg-a517-xa517m18/. It’s what came up when I searched for my specific laptop, but I didn’t realise there were two results until afterwards.

Maybe try comparing the BIOS for the XMG A517-M18 with the original dump from the BIOS chip? I can’t think why it would be there, but maybe it would explain GP107? I’m not sure on the specs of that model.

I have images of the inside, but not with the heat sinks and fans removed. The image with the labelled SOP8 chips for example. I don’t see any differences between that and what it looks like now. All I can see is that the BIOS chip was replaced.

I don’t understand it either. Maybe they though the chip was busted? Maybe they didn’t have the equipment to program it themselves, so they bought a pre-programmed chip (although it was empty anyway...).

I really hope they haven’t swapped out the GPU. I wouldn’t know how to check even if they had.


I don’t understand why they’d have vBIOS in the BIOS and on a separate dedicated chip in the case of my laptop. In this case, the GFX card does have its own, right? Is it just for efficiency to be able to run from the same chip?

Not sure, it was linked in one of those threads I linked you to above. Ohh, I see @Falkentyne mentioned it here in post #6 but did not provide a link (sorry, I thought he linked it)
Hopefully he will reply here and link you to example item to be used with CH341A to program ITE 8587 via ribbon cable.

Replacing QFP would mean chip replacement, not something you can do with proper tools and previous experience. And even then, you'd still need setup or config to program the chip first

Is the XMG A517-M18 a compatible BIOS for your model, same graphics card at least? if yes, program in that BIOS and see if it starts or not, since you flashed in that models EC FW.

You flashing in EC FW for other model would not have anything to do with GP106/GP107 stuff we're discussing.

Show me your before images, vs current, only the ones showing graphics card, so I can see if it looks 100% same card or not from what little we can see with the heatsinks on.

I explained why vBIOS is in BIOS above at post 203, this is common sometimes, but generally it would match the card that's in the system.
And yes, in your case the vBIOS is stored on the card or board, since you dumped it off some chip.

Zitat von Lost_N_BIOS im Beitrag #205

---

Not sure, it was linked in one of those threads I linked you to above. Ohh, I see @Falkentyne mentioned it here in post #6 but did not provide a link (sorry, I thought he linked it)
Hopefully he will reply here and link you to example item to be used with CH341A to program ITE 8587 via ribbon cable.

Replacing QFP would mean chip replacement, not something you can do with proper tools and previous experience. And even then, you'd still need setup or config to program the chip first

Is the XMG A517-M18 a compatible BIOS for your model, same graphics card at least? if yes, program in that BIOS and see if it starts or not, since you flashed in that models EC FW.

You flashing in EC FW for other model would not have anything to do with GP106/GP107 stuff we're discussing.

Show me your before images, vs current, only the ones showing graphics card, so I can see if it looks 100% same card or not from what little we can see with the heatsinks on.

I explained why vBIOS is in BIOS above at post 203, this is common sometimes, but generally it would match the card that's in the system.
And yes, in your case the vBIOS is stored on the card or board, since you dumped it off some chip.

---

@Falkentyne - Thanks for jumping in to let us know, sad to hear you don't know about programming via ribbon cable though
Yes, I do think this one uses the ribbon cable as a method to program, via I2C method, I see UART and KB ribbon cable referenced in the schematic for this board at the chips location.

I guess his only option will be to try the incorrect BIOS so BIOS and EC match, if that fails then maybe I can try to make this BIOS compatible somehow (maybe it's just vBIOS difference, we shall see I suppose )

Thanks again

Zitat

---

Replacing QFP would mean chip replacement, not something you can do with proper tools and previous experience. And even then, you'd still need setup or config to program the chip first

---

You can buy pre-programmed chips, so surely you would just need to swap them out?

Zitat

---

And messing around with EC's...you can't even desolder those chips...be careful with that.

---

Or maybe that’s a challenge in itself...


From what I can see based off this comparison and this Amazon listing, the only difference is that the CPU is an 8th gen i7, instead of 7th gen. It appears to be the N850EP6 instead of the N850HP6.

Unfortunately, I just tried flashing the BIOS for the M18 to the chip (and verified) to no success. It appeared to even struggle with turning on and still didn’t POST or display anything on-screen. The file I flashed was N8xxEP6.BIN as it was the only file that matched the size. Downloaded from here.

I was suspecting maybe the difference was a 1050 and I thought it might explain GP107, but no, that’s not the case. I don’t know why it’s there.

Zitat

---

Yes, correct about the vBIOS/BIOS. Generally they put vBIOS into BIOS, so it can be ran from there and in those cases GFX card usually doesn't have it's own, and or it can be copied from BIOS to vBIOS on card during specially setup flashing sessions.
^
I explained why vBIOS is in BIOS above at post 203, this is common sometimes, but generally it would match the card that's in the system.
And yes, in your case the vBIOS is stored on the card or board, since you dumped it off some chip.

---

So the vBIOS can be stored on the graphics card also? It can have a dedicated chip and also be part of the main BIOS? I was asking for more info as it’s a bit confusing. You’ve said that the vBIOS chip (25U4033E) dump you received had GP106 but the original BIOS chip (MX25L6473F) dump mentions GP107. What about BIOS.ALL, the updated BIOS file from the download portal?

If you can find someone to program this with everything you need for this system + the EC FW, then yes, that would work... If you had the skills and hardware setup to do the swap out.
Is the BIOS you tested the one that matches the EC FW you programmed in previously? If yes, please wait, tell me what is your CPU model and then I will make you a few BIOS to test. All with proper microcodes, and then one with your current BIOS vBIOS (GP107), and another with GP106 vBIOS

Yes, vBIOS can be stored on a chip on the GFX card itself, or near it, such as the one you dumped that contains the GP106 vBIOS.
Normally, it's one or the other where you can find vBIOS, not both, so this also adds to the confusion. Rarely have I seen both, if ever. Maybe those repair guys flashed incorrect vBIOS to that chip and it was originally something else in there?!?!?
I checked that too before I mentioned any of this, BIOS.ALL also contains only GP107.

@Lost_N_BIOS

My CPU is the i7-7700HQ.

As far as I know, the EC FW I flashed through software was the one for the XMG A517-M18. I ran the file off a USB (ecflash). I linked to the download portal where you can download the matching BIOS for this a couple of times, but here it is again: https://download.schenker-tech.de/package/xmg-a517-xa517m18/. I’m interested in testing out any modified versions.

It’s weird that latest BIOS version (BIOS.ALL) for the regular XMG-A517 contains GP107. That would imply it’s supposed to have that. Why it would when the GPU is a 1060 is beyond me. I don’t know why the repair guys would even touch the vBIOS chip, but who knows?

As for swapping the chips, “hardware setup” is referring to the hot air rework station you mentioned earlier, right? I don’t have one of those but I have a heat gun.


Oh, also:

Zitat

---

* Edit - Since they replaced the BIOS chip, maybe they put it on backwards? Or damaged the pads under it etc? Please show me up close images of any chip you can see they replaced.

---

Here’s an image of the replaced chip. I wish they’d never touched it - the original one was much nicer looking...