*HP 2510p Owners Lounge*

Discussion in 'HP Business Class Notebooks' started by master blaster, Feb 13, 2009.

Thread Status:
Not open for further replies.
  1. User Retired 2

    User Retired 2 Notebook Nobel Laureate NBR Reviewer

    Reputations:
    4,127
    Messages:
    7,860
    Likes Received:
    10
    Trophy Points:
    0
    Try fan_40pct in the revised INFO: Almost silent fan using a custom ACPI dsdt table.

    Modified HP 2510P BIOS

    DISCLAIMER: Use these modified bios at your own risk!!

    These modified bios have been beta tested to confirm they do not brick the system after flash. However you will be solely responsible for any damage to your computer system or loss of data that results from flashing or using of a system with a patched BIOS. These BIOS are not tested or approved by HP.


    BIOS^1
    Win7
    SLIC 2.1
    no whitelistdual-IDA
    quiet_fan​
    pcie_ASPM
    fan_30pctfan_40pct
    F.10 (264)
    +​
    +​
    +​
    -​
    +​
    F.10 (269)
    +​
    +​
    +​
    +​
    -​
    +​
    F.10 (268)
    +​
    +​
    +​
    -​
    +​
    +​
    F.0F (262)
    +​
    +​
    -​
    -​
    -​
    F.0F (264)
    +​
    +​
    +​
    -​
    +​
    F.0F (269)
    +​
    +​
    +​
    +​
    -​
    +​
    F.0F (268)
    +​
    +​
    +​
    -​
    +​
    +​
    ^1 (x) indicates number of bytes modified in original 2510P HP bios. F.0F is based off TTAV134's SLIC2.1+no_whitelist bios from here
    YELLOW = recommended


    * SLIC 2.1: for Win7 usage.

    * no whitelist: whitelist checks skipped so can use non-HP wifi/WWAN cards as explained here.

    * dual-IDA: Disables the EIST-lock bit. This means then Throttlestop (TS) can do a dual-IDA overclock, gaining an extra multiplier and 10%-12.5% performance from any U7xxx/L7xxx 2510P. TS also unlocks extra overvolting range from 0.9V to 1.0V to provide higher maximum setfsb overclockability. L7xxx also gain a lower SLFM voltage floor for greater undervolting ability using TS. If you don't use Throttlestop then your system will run with the speed and power consumption as it did prior to flashing the dual-IDA capable bios.

    * quiet_fan: exactly the same as quiet_fan's DSDT modifications but more conveniently located in the bios so doesn't require a DSDT override in Windows registry or Linux kernel. Can still use quiet_fan's DSDT override if run multiple OSs requiring different fan load profiles.

    * pcie_ASPM: offers better power saving and eradicates a Linux dmesg warning as described.

    Installation

    Windows users

    Download package of interest in the Version column on the right, double-click, extract to a desktop folder then run the HPQFlash file and follow instructions. The supplied HPQFlash is modified by TTAV134@mydigitallife to allow successful flashing of a modified bios.

    If HPQFlash doesn't do anything and you see a LOG file containing an entry WMI BIOS Flash function is not supported on this system, then install HP Quicklaunch on your system and try again OR perform a DOS based install like for Linux users below.

    Linux users

    Download and install modified rompaq from here. Download bios file of interest from above, extract rom.bin from the ROM.cab and copy it as 68MSP.bin on your DOS bootable USB thumbdrive. Reboot using the USB drive and install bios using the rompaq utility.

    Bootup error after flash

    If you get a message on bootup Your bios failed to complete the update. please ensure that you follow the correct procedures and try again. This may require two sequential bios updates to complete the process then apply mike60's fix here of flashing back to F.0E, flashing the non-modded F.0F then flashing a modded F.0F bios.

    Which modded bios am I using and why?

    I'm using SLIC21+nowhitelist+dual-IDA+fan_30pct with a 5100AGN wifi card. I've previously been using setfsb to overclock my U7600-1.2 to 1.6Ghz (178Mhz@x9) and created the dual-IDA mod in the hope to take the overclock further.

    Unfortunately my CPU is unstable at anything beyond > 1.66Ghz. Dual-lIDA mode's extra voltage not giving extension there. I can run Throttlestop, click EIST to use single-IDA x10 mode so some single-core apps will run at 1780Mhz (178@x10). The dual-IDA bios mod was not necessary to run my single-IDA mode though it doesn't hurt to have it available and other users can get benefit from this. My Win7 WEI CPU score increased 0.2 points using single-IDA mode. The 178@x9 gave some intermittent video hangs on overclock, so I've gone to a rock-stable 160@x10 dual-IDA overclock instead.

    fan_30pct is also now in my bios. No need for my DSDT override in XP or Linux and get a nice and quiet system. For the more heavily loaded Win7 I'm using the fan_40pct DSDT override to avoid fan yoyoing b/w 30% and 55%.

    Customising BIOS DSDT fan speed entries further (Advanced Users)

    The place to tweak the fan speed settings further in the F.0F/F.10 rom.bin file is at:

    @0x714B2: fan speed table bytes matching quiet_fan DSDT fan speed table entries
    @0x6EC50: recommended checksum correction byte to use

    If you decrease the fan speed table entry values, then add the difference of their sum from their original sum total to the correction byte. This will ensure the total checksum of the DSDT table is zero. Otherwise Win7 will fail to boot with some ACPI/BIOS is not compatible type error.

    Flash gone wrong? Use Emergency Recovery. You must use a USB floppy drive. A USB thumbdrive doesn't work.


    Changelog

    (09-12-11) new 2510P bios: complete F.10 mod, added pcie_ASPM to previous F.0F and fixed bug in F.0F-fan30pct bios.
    (10-29-10) fan speeds altered as recommended by e&i
     
    Last edited by a moderator: May 8, 2015
    schdrag likes this.
  2. Evilandi666

    Evilandi666 Notebook Enthusiast

    Reputations:
    0
    Messages:
    27
    Likes Received:
    0
    Trophy Points:
    5
    Will try this later on - that would be so great if that works.

    Do you have any idea why the vpro disabling doesn't work with the newest Bios? Maybe there is a posibility to disable it directly with the bios? Does the disabling like you described it work with this changed bioses?
     
  3. PowerToTheUsers

    PowerToTheUsers Newbie

    Reputations:
    0
    Messages:
    7
    Likes Received:
    0
    Trophy Points:
    5
    That didn't work unfortunately...
    Thanks for the image. I tried to connect the SSD and the original harddrive again, to no avail... I ordered a usb-to-zif enclosure so I can at least test the SSD and original HDD: I hope I didn't break them.

    On a side note: There's a PCIe-slot for a 3G-card. Can I put a PCIe-SSD in it? Did someone else try that before: does that work?
     
  4. Borh

    Borh Notebook Enthusiast

    Reputations:
    0
    Messages:
    20
    Likes Received:
    0
    Trophy Points:
    5
  5. User Retired 2

    User Retired 2 Notebook Nobel Laureate NBR Reviewer

    Reputations:
    4,127
    Messages:
    7,860
    Likes Received:
    10
    Trophy Points:
    0
    Ensure you lift the ZIF stiffener, insert the cable, then press down the stiffener as shown here. Requires minimal effort when done that way.

    It's not possible to use mPCIe SSDs. They use netbook-specific slots which have sata pins on them. The HP WWAN slot only has USB data transfer pins and the wifi slot has pci-e pins. So neither has sata pins on it.

    If you want a SSD can go either 1.8" ZIF one OR a 2.5" sata one using a 9.5mm sata-to-pata caddy like advised below. In the 2.5" configuration the existing 1.8" ZIF HDD can have pin1+2 jumpered to run as slave to act as an additional storage space.

    Fenvi 9.5mm sata-to-pata caddy is the recommended ebay caddy

    Yes. That is the product I tested, but I would recommend the newer Fenvi 9.5mm sata-to-pata caddy pictured below. It has the same internals BUT now has a stronger metal chassis AND the faceplate is attached by clips. A user briefly showed this caddy's faceplate can be removed via clips, clips that look to align correctly to do a faceplate swap. We are waiting in this other thread for someone to demonstrate the 9.5mm faceplate swap in detail.

    [​IMG] [​IMG] [​IMG] [​IMG]

    I've have also been advised that setting a logic=1 on pin33 (pin in the upper right corner) of the Marvell chip sets it to run as slave though I haven't confirmed this. ebay caddy owners may wish to try it?? Make sure you check that pin33 isn't grounded before trying this as would otherwise short the system.
     
  6. thinkaddict

    thinkaddict Newbie

    Reputations:
    0
    Messages:
    3
    Likes Received:
    0
    Trophy Points:
    5
    Hi Jedis of the Hack ;)

    I've got a EliteBook 2530P with a C2D SL9600, french located.

    I think that the thermal ingenering of this subnotebook has been botched. On the keyboard I can feel 8-10°C of difference between his left and right sides. That's sucks !!!

    Before having enought money for come-back to ex-BigBlue, I would like fix as possible my thermal problem.

    I can't force fan on AC/DC by using bios option because I'm running Debian who can't been booted without disabling ACPI in that case (and that I won't !).

    I want to enable (force) lowest fan speed for idle run for any case, I don't care about power consumption 'cause I've got two batteries ...

    So I've disassembled the DSDT table like shown in nando4's linux link.
    But I can't find inside the section about fan's power tweaking (there's no quote like in nando4's DSDT table post).

    You can dl my DSDT table here:
    2shared - download DSDT_2530p-SL9600.dst

    Thanx for your help !

    PS: Sorry for my poor english, I'm french ;)
     
  7. User Retired 2

    User Retired 2 Notebook Nobel Laureate NBR Reviewer

    Reputations:
    4,127
    Messages:
    7,860
    Likes Received:
    10
    Trophy Points:
    0
    INFO: Hardware based undervolting mod

    Thank you for posting some more details on this mod. I was already getting 5.8-6.2W at idle with wifi on, 3 bars brightness. I did a trial VID3+VID2 [shown in next post] and noted that the CPU could only now only run x6 133Mhz FSB (800Mhz) at 0.75V.

    [​IMG]
    What would a VID3+VID2 (0.1V) undervolt gain?

    Idle power dropped about 0.2W. Under full cpu load there was 2W difference.

    I didn't actually like running my system like this. At the stock lowest voltage of 0.85V, I can run my u7600 at x9 160Mhz overclocked (1.44Ghz), which is my DC profile. Indeed, because the CPU can do it's work a lot faster it may spend more time in C3 power saving state, perhaps neutralising the gains at full load.

    Overall, the greatly lowered performance was not something I was prepared to give up what may be questionaby battery life gains.
     

    Attached Files:

    Last edited by a moderator: May 7, 2015
  8. zerGus

    zerGus Newbie

    Reputations:
    0
    Messages:
    6
    Likes Received:
    0
    Trophy Points:
    5
    Hi Nando,

    there is an easier way. All of the VID lands except VID1 are soldered through to the back of the motherboard. It was pretty easy to connect them with conductive paint. The only problem is getting access to the back of the motherboard, which required taking the whole laptop apart :eek:

    The photo shows the connection between VID2 and VID3, which allows voltages down to 0.75v with RMClock. Afterwards I went mad and connected VID0 to VID2 and VID3 which allows 0.7125v and undervolts the default LFM VID down to 0.7375v (without RMClock). I've had no problems so-far, but haven't done any significant testing.

    I've seen RMClock show an idle power draw of under 6W, so I'd say that it saves just over 0.5W in near-idle conditions.

    I'll do some stability testing at various VIDs and write it up a little better when I get a chance.

    best Regards,
    zerGus
     

    Attached Files:

  9. User Retired 2

    User Retired 2 Notebook Nobel Laureate NBR Reviewer

    Reputations:
    4,127
    Messages:
    7,860
    Likes Received:
    10
    Trophy Points:
    0
    delete 10char
     
  10. zerGus

    zerGus Newbie

    Reputations:
    0
    Messages:
    6
    Likes Received:
    0
    Trophy Points:
    5
    Hi Nando,

    my understanding of the logic is the same as yours: a logical 1 (Vcc) is dominant. Connecting VID0, VID2 and VID 3 is interesting because it gives a good variety of available voltages within the VID range of 0.85 to 0.9v.

    the pinout of the processor can be found on the Intel Datasheet, link to the (pinout identical) 45 nm version attached.
    http://edc.intel.com/Link.aspx?id=1484

    The relevant bit is on the attached table (careful! the BACK of the motherboard is a mirror image of the table). VID0 through VID3 are all nicely adjacent to each other...

    I've started testing and have a couple of impressions:
    - At 10x and 0.875v I have no stability issues. At 10x and 0.8625v I get an immediate BSOD, a good sign that the CPU is probably actually getting 0.7125v instead of the 0.8625v requested.
    - At 6x / 7x and 0.8625v (0.7125v actual) I have no stability issues running 2 instances of CPU burn for about an hour.

    The CPU temperature drops down to about 51C at idle. Room temperature is currently 27C here (South of France).

    My next step is to do some power consumption testing under load and at various VIDs to see if the power saving is actually worth it.
     

    Attached Files:

Loading...
Thread Status:
Not open for further replies.

Share This Page