VRM modify ICC_MAX to run high TDC OEM cpu
- Thread starter RolloZ170
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Let me know if you need better pictures. I think imgur subsamples. There is no information on HP Z8 VRM but someone was saying he was planning to z8 earlier in this thread.
i guess MPS MP2955A or TI TPS53xxx (or OEM 53678 / SN1701021 )
edit:
i wait for anyone to confirm:
your motherboard is the only one with blank VRM controller chip.
Anyone have a old copy of the X11SPM BIOS 4.5 or have i idea of the string to mod in th enew 4.7 version?
I got pictures from another z8. This one had label on it and it say mp2955a.
Album — Postimages
So would this work? I see people did this before in this long thread. I have no soldering experience though. I am not sure how I can reach into such a tiny spot and solder pins.
you just need to trace the SDS(35) & SCL(36) pin from MP2955A to the 4pin header(PJ1100) for correct conntection.I got pictures from another z8. This one had label on it and it say mp2955a.
Album — Postimages
postimg.cc
So would this work? I see people did this before in this long thread. I have no soldering experience though. I am not sure how I can reach into such a tiny spot and solder pins.
Ok I am a completely illiterate when it comes to this. I have no electronics experience but I read through the post in this thread.
First of all, I have MP2955A . Is this good enough to enable 255a TDC or will the system be still underpowered because Z8 has weak VRMs?
After confirming that, next step for me to find the pins by using voltmeter. Again I have no idea how this is done but I assume I will do some sort of connectivity test by putting one probe on the pin and one probe on the mp2955a chip legs. I have some cheap voltmeters but only used it for high voltage scenarios, not sure what voltmeter to have and how to use it.
After finding the pins, I need to solder pins. I have not soldering before and I have no tools so I have no idea how hard this would be for novice. It seems very hard to me to be able to put a pin there and then solder it. One hand with solder, other hand with soldering iron, I assume I will need someone to hold the pin down. Can this be even done in such a tight spot?
After that, I need to connect to the pins via MCP2221a , usb cable, and pinout cable. Then use the C# program RolloZ provided in the first post to configure the VRM chip to support higher TDC without providing up the workstation but power is plugged in so there is power on the motherboard.
Then, do I need to modify the BIOS as well? Is there any other step I need to take?
First of all, I have MP2955A . Is this good enough to enable 255a TDC or will the system be still underpowered because Z8 has weak VRMs?
After confirming that, next step for me to find the pins by using voltmeter. Again I have no idea how this is done but I assume I will do some sort of connectivity test by putting one probe on the pin and one probe on the mp2955a chip legs. I have some cheap voltmeters but only used it for high voltage scenarios, not sure what voltmeter to have and how to use it.
After finding the pins, I need to solder pins. I have not soldering before and I have no tools so I have no idea how hard this would be for novice. It seems very hard to me to be able to put a pin there and then solder it. One hand with solder, other hand with soldering iron, I assume I will need someone to hold the pin down. Can this be even done in such a tight spot?
After that, I need to connect to the pins via MCP2221a , usb cable, and pinout cable. Then use the C# program RolloZ provided in the first post to configure the VRM chip to support higher TDC without providing up the workstation but power is plugged in so there is power on the motherboard.
Then, do I need to modify the BIOS as well? Is there any other step I need to take?
if only 4x 50A yes, to few.
multimeter: the beep function.
no, first apply a drop to the pad, and solder the cable, then left hand cable, right hand the cold side of soldering iron.
if 205W cpu supported there is no add. TDP block. just needs a BIOS with 2nd gen support.
Hello, I tried to find the connector for HPE Proliant DL160 Gen10, which is undiscovered so far in this thread.
tl; dr: It turns out DL160/DL180 has the J266 header just as DL360/DL380. (See this post)
I found a suspicious 6-pin holes J137 connected to TPS53678 pin 18/19 (see the picture below).
However, I have no idea what pins they are (or whether they are SCL/SDA at all) since I do not have the datasheet.
Below is some additional info:
- The orange pin does not seem to be connected to any of the TPS53678 pins
- When powered off, the three pins(blue, red, orange) connect to GND at like 60Ω (weird?)
- When powered on, they are all at around +1V from GND (weird2 ?)
- There are two TPS53678's (+three TPS53658) on the board, and there is a separate 6pin holes for the other one. The corresponding pins in the two connectors do not seem to be connected(except for 100-200Ω connection through GND)
Does anyone have idea what these pins are, or what is going on?
tl; dr: It turns out DL160/DL180 has the J266 header just as DL360/DL380. (See this post)
I found a suspicious 6-pin holes J137 connected to TPS53678 pin 18/19 (see the picture below).
However, I have no idea what pins they are (or whether they are SCL/SDA at all) since I do not have the datasheet.
Below is some additional info:
- The orange pin does not seem to be connected to any of the TPS53678 pins
- When powered off, the three pins(blue, red, orange) connect to GND at like 60Ω (weird?)
- When powered on, they are all at around +1V from GND (weird2 ?)
- There are two TPS53678's (+three TPS53658) on the board, and there is a separate 6pin holes for the other one. The corresponding pins in the two connectors do not seem to be connected(except for 100-200Ω connection through GND)
Does anyone have idea what these pins are, or what is going on?
Last edited:
For DL160/DL180 Gen10
I just found the same J266 pin behind the Smart Storage Battery. (or the proc#2-side & fan-side corner if it's not installed)
Confirmed working.
Some more info about DL160/DL180 Gen10:
The output of the mod suggests their default TDC limit is 215A whereas DL360/380 (and most other boards) has 228A. Indeed, my DL160 at first could not run a pair of 8173M which has a TDC of 228A.
Also, they officially support CPUs only up to 150W TDP(whereas DL360/DL380 up to the flagship 205W), so it might not be very recommended to do this mod on DL160/180.
The output of the mod suggests their default TDC limit is 215A whereas DL360/380 (and most other boards) has 228A. Indeed, my DL160 at first could not run a pair of 8173M which has a TDC of 228A.
Also, they officially support CPUs only up to 150W TDP(whereas DL360/DL380 up to the flagship 205W), so it might not be very recommended to do this mod on DL160/180.
Code:
MCP2221a VRM mod utility (build 13.07.2024).
The MCP2221a device is connected.
ReadBlock( Page00,cmdAD,2)=0178
TPS53678 found: starting modd
ReadBlock( Page00,cmdDA,2) D7BA
WriteBlock( Page00,cmdDA,FF 00)=OK
ReadBlock( Page00,cmdDA,2)=FF00
WriteBlock( Page00,cmd11)=OK
TPS53678/TPS53679 modd successfulHI Rollo me again, just quick one for reference, when we are writing these commands, are they being written to some physical hardware chip or are they living somewhere in the bios? I say this as its easier for me to write new bios then complete the mod.
written in the physical VRM controller chip (TPS53679/8,MP2955A,PXE1610C)
if in the BIOS you need to mod BIOS every new BIOS update.
I tried modifying on ASUS WS C621 SAGE and it worked. It now supports dual 8175M.
The pinout Rollo provided is correct.
SCL -- I2C8SCL
SDA -- I2C8SDA
GND.
Instead of MCP2221a, I utilized the raspberry pi that I had, which supports i2c. I decompiled Rollo's C# executable, rewrite it in Python and run the Python script over ssh. Updated both PXE1610Cs without issues.
Below is the Python implementation, should be working on raspi or any other similar devices.
Yumanlou/raspi_iccmax_FF
Thanks @RolloZ170 — standing on your shoulders here.
Pinout & prior testing made it possible to confirm this on the WS C621E Sage.
Last edited: