Zitat von jjxaker im Beitrag #5People do not seek easy ways))) just had to initially remove the GUID for TPU firmware. in the end for this controller will not automatically update the firmware. Just first you need to learn, what controllers on the board and what kind of firmware are responsible for them. it's simple obvious things. Did not need anything to solder, it's stupidity. to restore it was enough to replace the firmware in GUID with the correct version, the system would automatically update it. when enabled, the BIOS makes a reconciliation of the versions each time.
I understand what you are saying, however this was related to the coffee-lake mod research and i needed to have a thread which i can refer to in order to explain issue when updating bios after switching id to sku of z370 and recompiling for coffee lake mod. It can obviously be approached in different way, however as i made the mistake not knowing about tpu automatic upgrade it bricked my board.
Mentioned above, if you somehow end up in this state, any changes in the bios flash do not help. If you are able to determinate the uC core and reprogramm the flash of tpu via TPU programming pins, guss that works too. For me it was much easier to de-solder and determinate if tpu is the actual issue. It was.
Zitat von Dnatwork im Beitrag #72Complete noob here, but wondering about extra power pins. I’ve seen the tape things that swap pins for LGA 771 to work on 775. They have holes to let all the other pins through, with traces to connect the two pins that were swapped.
Could you make something like that to get the power from the original pins in the socket to the extra pins on the i5-8400? It’s a low power chip, so the current draw should not be an issue for the power coming from the board. The problem would be trying to force the current for six cores through the wrong pins on the chip.
Would you need to have separate traces for every pin, or could all the 12v be wired together, and all the GND wired on one line as well? That would spread the current draw across the socket pins to avoid burnout on that side.
That is not needed. Article with measurements will be out in like 48h tops.
@XPLives Simly: NO. Any older board then Skylake has different socket then LGA1151
Change from z270 to z370 is small enough that it can be ignored, at least for the most part. Architecture difference e.g. between Haswell cpus and CFL i3-8100 is big and can not be solved any way near what you suggested. Theoretically possible while making socket re-pinout of like literally 1000's of pins..... And sw mod as well. Not a possibility for any modder with less resources and time then intel :-D - also highly impractical.
Regarding the voltage controll, that is up to a MB manufacturer to decide if the chip they used for VRM can be controlled both from CPU and user input... not everywhere it is possible and it definetelly is not "gone" e.g. z170-pro from asus has manual voltage choices working without an issue.
Hi, i have lit the 8600K on B150 board. NO PIN MASKED, all working. Full stress test shall be conducted and i post it in separate thread. Link gonna be added here. Just posting here to let you guys know about progress and that work is being done on this front.
More info on the [ Discord discussions: https://discord.gg/FgpUVCy #english ] Hopefully we can get the i5-8xxx running and finaly dispute the claims about power delivery etc....
Hi, another board confirmed. ASUS Z170-PRO i3-8100 so far, BCLK overclock as well
Patch works. tested patcher 1.2 & 1.3. Both ok.
WARNING: Flash only bios for your MB! -- Asus MB have TPU modules which causes problems when combining MB with bios which was originally meant for different board. -- More info on that here: [SOLVED] Asus TPU unit firmware GUID
I found the TPU firmware module. Thanks @Mov AX, 0xDEAD for tips. GUID for tested ASUS boards is: 38715E0A_ECD3_4776_8AC4_7D93510A7468 #edit: (tested-board="ASUS Z170-PRO")
_What and why, possibly how:_ What is TPU unit? Q1) Almost all Z*70 ASUS boards have this. The TPU module. It is supposed to enable auto-overclocking etc. More info on asus web or forums. (that to be said based on asus marketing of this chip) --- What i did? Q2) I tried flashing bios with version suggested to enable BCLK overclocking. I did this by flashing the bios memory chip directly. Booted and short notification of "bios upgrade" showed on my screen, after that the board shutdown. I could not get it boot again, not even to light up any led on board etc. --- Why bios re-flash "back" to original version does not work? Q3) IMPORTANT! TPU module starts even before the bios starts loading. Is is enabled in stand-by mode - even before user presses the power button. .
Without TPU working correctly (or at least partially correct) the board does not start. It has some conections to power enable and it holds the whole board in "reset" state. Also it ties to the LPC bus and chipset boot sequence. Bios reflash or rollback does not help the situation because this bios does not even start loading to PCH so it can not re-flash this chip back. --- How did i find GUID and flashed the TPU firmware back to original version? Q4) The MX25L512E memory in this case is routed directly on board. I connected my digital analyzer and reverse-engineered the pinout on board. I read the memory and used this data to find original raw data in bios flash under mentioned GUID. Using the same GUID i found the TPU FW in old bios for this board. De-soldered, programmed and re-soldered the memory to board. (While soldered on board the write-protect is brute-forced being tied to power-lane.) After i flashed correct version of TPU-FW to the MX25L512E and soldered it back with old bios being in the bios-flash memory, the board STARTS NORMALY.
Conclusion: A1) TPU FW has a lot to say in board boot and power-on sequence. A2) Flashing wrong bios (eg. ASUS z170-pro gaming, instead of ASUS z170-pro ) can and most likely will result in bricked board. A3) Board can be resurected from this, but it requires re/de soldering the TPU FW memory. Also, TPU-FW GUID might differ between boards production generations.
Hi, does anybody know what firmware is stored inside the TPU chip ? - Possible which GUID does it have in bios ?
The MX25L512E chips is 512kbit of data (64kB ~ 0x10000) and if it is cleared the board does not boot at all, does not even start the VRM. Board at question for now is [ASUS Z170 PRO].
Zitat von nickywin im Beitrag #18Gigabyte Z170X gaming 7 is ok, thx!
Hi, how did you flash the BIOS? Gigabyte EEPROM is soldered, so no SPI flashing is easily possible. Or did you use modified efiflash to get it flashed to the Board?
1st: it is SPI "serial" flash. Not EEPROM. 2nd: And you can usually contact the memory with adapter like this one : https://goo.gl/DFBA96
50/50change the SPI has grounded CS when board is off so if unlucky you actually have to desolder.
NOTE: This is only for the time when normal flashing of the bios fails.
MSI B150 BAZOOKA PLUS confirmed. In the attachment: 1) i included the bios file with microcode for i3-8100. 2) reverse engineered pinout of this board so you do not have to desolder the bios chip from board. (Mine had MX25L6473E chip, but even with different chip should be the same.) note: I have used "ELNEC MEMprog2" to flash this. 3) Also change in OS after pcie fix. 4) added while board photo for anyone using the bios file to be sure he has the same board
Zitat von elisw im Beitrag #65tried but failed to update the BIOS (which looks a bit messy compared to other manufacturers): UBU cannot update the VBIOS and also it doesn't let you add the Coffee Lake microcodes.
Hi, try the attached BIOS. If you are reflashing without external SPI flash tool, you probably need to rename it back to the manufacturers convention. I updated the microcode and VBIOS manualy using the UBU 1.69.6. -- Note: Whatever you use this file for, you are doing it AT YOUR OWN RISK. -- If you do not know how to recover bricked board, probably do not try it until somebody confirms it is working.
Zitat von rootuser123 im Beitrag #59cpudxe GUID: B03ABACF-A532-5E78-ACA0-B11F765B3AFD cpuspSMI GUID: 116E1ACF-2533-4CC2-820A-BBC10A2AB07C cpuIOdxe GUID: BAE7599F-3C6B-43B7-BDF0-9CE07AA91AA6 amicpufeatures GUID: 10B12ADD-F5E2-CC78-5CA0-B77F76223ACD
Thanks for tip. Tested on ASUS Z170, not working so far. Tried every ASUS Z370 board as a source for these, did not even solve the voltage-init problem. I get the system running in same manner as without the change. Used UEFI Tool 45A to extract and UBU 1.69.6 for replace-body function and rebuild the bios image.
Will test further with other boards and post results in few days.
Zitat von rootuser123 im Beitrag #57@LittleHill Try replacing these SMM and DXE files, maybe they will make PCI-E work.
Can you share the exact GUID please ? (image shows only partial GUID and not all bios files have the "GUID name" in them. (I know where you are trying to go with this, just to make sure that i get the suggested modules right.)
Zitat von LittleHill im Beitrag #48@dsanke so far no success Replaced CpuInitDxe on ASUS Z170-PRO by one from ASUS Z370 A AS. Voltage and PCI-E issues still there.
Does it have GUID FCCCE259-07E1-417A-A755- EOF2CF39295E ？
1) I have taken the FCCCE259-07E1-417A-A755-E0F2CF39295E from Asus Z370 A and have put it into Asus Z170-PRO (to not a successfull try as previously stated,board boots and everything but voltage and PCIE problem remained) 2) Some ASRock boards seems to have different GUID for CpuInitDxe then Asus Z370, for example ASRock Z170 Gaming bios you sent states CpuInitDxe under 62D171CB-78CD-4480-8678-C6A2A797A8DE .. migh look into that. 3) As well there is the PowerMgmtDxe (GUID: F7731B4C-58A2-4DF4-8980-5645D39ECE58 ).. Any ideas?
Hi @shixinyuan can you share the guide-lines to get pcie working on other boards ?
Add the power damage: "because of his power supply is the opposite" How can be any pin on 8700k different then i3-8xxx if they are the same pinout and spec for z370 boards ? I do not understand the reason behind this finding.
@elisw DRAM speed and timing is defined in CPU and the motherboard only wires-up them together. CFL new spec for routing means that Z370 boars with CFL cpus can handle higher speeds but on lower ones like 2400-3000MHz there should be absolutely no issue as the SKL/KBL boards were designed for it. (As long as the ram-type is supported in the original bios)
Edit: tested on MSI B150 bazooka + i3-8100, was not able to set higher speeds then 2666Mhz on it in XMP, default falls on 2133. (ram used: Crucial Ballistix Elite 8GB 3200)
Win-Raid Forum (Drivers - Firmware - Modding) » Search for LittleHill's content 
Thu Apr 19, 2018 2:36 pm