DV9000- Turns on- Shuts off
Symptom 1: Unit runs for a while then turns off
Symptom 2: Unit turns on, turns off immediately
This unit arrives after another tech had attempted repairs. Numerous screws missing from base and incorrect screws used in reassembly. Complete tear-down and disassembly.
After powering the unit up, I noticed that the heat sink copper was discolored [darkened] which is a sign of high heat. Upon close inspection I noticed a kink in the heatshrink material covering the wires for the fan wires, so I strip back a portion of the insulation and the red wire was broken, and the black wire was only 2 strands. Yellow and White wires were OK. Rather than replacing the fan (which is in short supply, and costly) I repaired the broken wires with solder and insulating them so that they would not short each other or the frame of the HS fan, which is negative ground.
Screw also missing on the exterior frame that holds the heatsink downward onto the base, this screw is important because it also keeps the heatsink pressed against the GPU allowing for the heat to be pulled away from the chip to be vented out from the unit. This unit was also upgraded to the modified custom copper shim to further assist in heat reduction.
Reason of varying symptoms:
This model may not require the fan to reply back to P.O.S.T. -so it would run, until it got too hot to continue. The ground (fans black wire) did not complete the circuit for the 'non-working' temp sensor to report that it was overheating. The CPU core thermal sensor eventually reported overheating before the fan could, or CPU thermal sensor would shut it down to protect the CPU. Then when the unit was immediately turn back on (while still over temp) it would shut off, since no correction was made, the info was reported to CMOS and would shut off.
This test should also be in your diagnostic 'toolkit' so that you also test the fans when you have an immediate shutdown, or runs & shutdown problems in this series [DV6000/9000/F700]. Also see the specs on MAX8734 power chip so you can identify this cause, since it do have a SHUTDOWN circuit [pin 6].
Cause: The screw had gone through the base and was smashing/pressing the wires again metal post, effecting shorting positive terminal of Fan lead coming from motherboard connector, and broken, no power reached the fan. The black wire [GND] was nearly broken so the fan circuit was not complete, as well as the speed (YEL) and temp sensor (WHI) circuit because there was no power reaching the fan.
The Max8734 is one of the major players in distributing voltages to sections of the motherboard.
MAX8734AEEI+ product name:
High-Efficiency, Quad-Output, Main Power-Supply Controllers for Notebook Computers
Symptom 2: Unit turns on, turns off immediately
This unit arrives after another tech had attempted repairs. Numerous screws missing from base and incorrect screws used in reassembly. Complete tear-down and disassembly.After powering the unit up, I noticed that the heat sink copper was discolored [darkened] which is a sign of high heat. Upon close inspection I noticed a kink in the heatshrink material covering the wires for the fan wires, so I strip back a portion of the insulation and the red wire was broken, and the black wire was only 2 strands. Yellow and White wires were OK. Rather than replacing the fan (which is in short supply, and costly) I repaired the broken wires with solder and insulating them so that they would not short each other or the frame of the HS fan, which is negative ground.
Screw also missing on the exterior frame that holds the heatsink downward onto the base, this screw is important because it also keeps the heatsink pressed against the GPU allowing for the heat to be pulled away from the chip to be vented out from the unit. This unit was also upgraded to the modified custom copper shim to further assist in heat reduction.
Reason of varying symptoms:
This model may not require the fan to reply back to P.O.S.T. -so it would run, until it got too hot to continue. The ground (fans black wire) did not complete the circuit for the 'non-working' temp sensor to report that it was overheating. The CPU core thermal sensor eventually reported overheating before the fan could, or CPU thermal sensor would shut it down to protect the CPU. Then when the unit was immediately turn back on (while still over temp) it would shut off, since no correction was made, the info was reported to CMOS and would shut off.
This test should also be in your diagnostic 'toolkit' so that you also test the fans when you have an immediate shutdown, or runs & shutdown problems in this series [DV6000/9000/F700]. Also see the specs on MAX8734 power chip so you can identify this cause, since it do have a SHUTDOWN circuit [pin 6].
Cause: The screw had gone through the base and was smashing/pressing the wires again metal post, effecting shorting positive terminal of Fan lead coming from motherboard connector, and broken, no power reached the fan. The black wire [GND] was nearly broken so the fan circuit was not complete, as well as the speed (YEL) and temp sensor (WHI) circuit because there was no power reaching the fan.
The Max8734 is one of the major players in distributing voltages to sections of the motherboard.
MAX8734AEEI+ product name:
High-Efficiency, Quad-Output, Main Power-Supply Controllers for Notebook Computers
Hi Mayo,
ReplyDeleteIm a big fan of your work. I studied electronic engineering at university and now i am trying to work self employed as a computer technician. Any help would be gratly appreciated and you can count me in to buy the DVD and maybe the reflow kit once its all ready. (would you post to London, UK?)
I am currently looking at a dv2000 that would switch on, (fan, LEDs and beeps - no video) and then shut down after a 3 seconds.
I initially thought it was a gpu related issue (hp and nvidia plus the no graphics) so i reflowed the gpu.
once the reflow is done, board will not power up at all. only signs of life is the charging light when battery is plugged in otherwise, no response at all.
on further inspection, i notice that the max8734a is running very hot. does this indicate a fault in the max8734a or a short in the components near it or could the problem originate even further back to the mosfets at the dc jack connection? could reflowing cause this symptom?
Thanks in advance.
Habib
How hot is hot? I usually keep a cheap automotive infrared thermometer around to measure chips like this. A typical symptom of failure of these maxim IC's is running hot (>50C). Replacing these can be very difficult as you can't be 100% sure everything is making contact (very hard to check for continuity on traces that small). I usually use a hot air pistol and just reflow the new IC back into place. There are other methods but thats worked for me.
ReplyDeleteUsing a thermocouple thermometer, the temperature reads more than 50C but i unplug the power before the temperature gets even higher and possibly damages something else. Please keep in mind that this symptom of dead/non-responsive motherboard and overheating Max8734a started/noticed after reflowing the gpu. I use and IRDA station for the reflowing and i have done many before successfully so i dont think its the technique. Though i am open to any suggestions and constructive criticism. I can see two scenarios:
ReplyDelete1-reflowing creates a short in the bga of the gpu this translates to power problem
2-reflowing fixes one problem and uncovers another
I have read on another of mayos blogs that the dv2000 mainly has power not gpu related issues.
What do you think?
Habib: generally speaking on the DV2000 issues they are related to bad solder on BGA and shorted FETS (especially the AO4422's. Which leads to no power, shut downs and no video.
ReplyDeletere: questions -
1: BGA reflows that fail can lead to totally dea dunits, this is casued when the solder balls pool together under the GPU. And who knows what area may be shorted.
2: reflows can generate another problem, but in your case if the video never shows up, it may now be two problems rather than revealing another.
Try another reflow with a water based flux to coat the solderpads and assist in separating pooled solder that cannot be seen. I have recovered several bad reflows that way.
Mayo,
ReplyDeleteThanks for the idea, i just ordered some water-based flux and will give reflowing a go once it comes in. Will let you know how that goes.
Also, for everyone elses info, this dv2000 presented what looked like a gpu fault so i jumped to reflow. Might have done better to go through the basics first. Only later did i notice that the CMOS battery was flat ~1V. This might have something do do with the original fault or not, but still highlights that i should have systematically gone through the diagnosis carefully. lesson learnt.
Thanks again Mayo
Thanks for mentioning the CMOS battery! I forgot to tell everyone that on all the DV Series especially the early ones DV21xx,DV60xx, etc... to check that first.
ReplyDeleteThey don't last 10 years like the specs say!
Got a DV6423 on the desk now with a CMOS issue. It causes the laptop to try to read values from CMOS, and the values are empty, so it restarts, tries to read again. A cold reset of the CMOS will load the default values from BIOS and may start, but when the unit is turned off, the values are lost again.
A tip on suspecting the CMOS battery is the Time/Date is incorrect in BIOS (and Windows if you get to it).
Will post a YouTube video soon on why flux is important in the reflow process (perhaps as early as WED 5/11).
ReplyDeleteWednesday was hectic, went to Chicago area for personal leave, but make that this Wednesday for the flux/oxidation video. I keep having Jr. techs telling me that it is not important, but NASA & milspec assemblies require it, so I continue to do as I was trained. I cannot undertand why/how it could be better without.
ReplyDeleteOn some nVidia reflows you have to be careful of the GPU and hot glue used. If the GPU is anchored by small black dots on the corners, those are merely guides, and Do Not Anchor the GPU. If the hotglue is red and covers most of the GPU except 2 openings top/bottom or sides, that glue DOES anchor the GPU.
ReplyDeleteIf the GPU is not anchored by the glue it can become unlevel and no contact made at some points under it. There is no real recovery if that happens and a reball is necessary. If you do not have the tools to level and properly space the solderballs to the correct height, anchor it, and reflow again. It is a tedious task, and requires a precision level that few can attain.
Hello Sir, I have an intereating issue. I ran out of dv9000 boards to work on and got it into my head to repurpose an v6000 into a dv9000, the only things I havent enabled are the bluetooth port, the firewire port and the ricoh card reader. I reballed the board, and it runs fine, I put a dv9000 bios chip on the board, and added a usb port and second hard drive port. everything works fine except the screen >_< with a dv6000 or v6000 screen it fires up fine. and sometimes it will work with the dv9000 17 inch lcd. but as soon as the bios battery is inserted, and it saves values. the 17 inch screen stops working.
ReplyDeletei know this is a bit out there, and a rather odd thing to be doing. but i figured if anyone understood the differences in these boards it would be you.
I've tried to track down any differences in componenets on the two boards, and I have an example dv9000 board with the go 6150 chip, basically what I'm turning this one into. so far I can find nothing that differentiates between the 15 and 17 inch displays
Peter
East Gable: That undertaking may be fruitless as there are a number of minor changes that can throw your plans awry. The biggest being the Battery terminals are in different locations. I had never considered if the electronics of that section are the same or it too has changed.
ReplyDeleteHi Mayo,
ReplyDeleteThanks for your great coverage on the DV9000 series laptops. I have a clients' one here with a good nVidia chip but the CPU was damaged, replacing the CPU and HSF fixed it (also did the copper shim modification) however the sensor still falsely reports that the CPU is idling at 90 degrees and shuts itself off. I know it's false because the individual cores only report 45 degrees and my thermometer confirms it. I'm wondering, would this be a damaged MAX8734 chip? I can't seem to find any thermal sensors on the mainboard itself. Unless the SMBus is saying it's the CPU overheating when it could be the MOSFETs/caps near it that are heating up, or the Northbridge. I'm not sure. It would be easier to just replace the mainboard, which I will do, but I'm curious as to why this is happening for me.
I'm also considering doing the fan mod you mentioned here to make it run at full speed, or patch the BIOS/ACPI table to tweak fan speed via software. Along with that, I may clip the Pin 6 on the MAX8734 chip so it never shuts down under heat. Dangerous, I know, but the client wants it done cheap and fast and is aware of the risk I've mentioned it to them.
What do you think?
Dan: Never ever cut the Shutdown Pin! That not only prevents shutdown from the GPU faults, but any other source, Shorted USB/IEE1394 or other parts that can literally cause a fire. Though working around the erroneous temp reporting. Something tells me that the nVidia could still be the culprit.
ReplyDeleteDepending on the model number it is still possible for it to control the Fan too! Without the model number I cannot even guess which board layout you have (as there are as many as 25+ design configuration) all of them DV6K/9K/F700 series. Intel and AMD.
Hi Mayo,
DeleteThanks a lot for the quick reply. Yes I thought so, that would be very dangerous! The specific model number of the notebook is DV9521TX (Intel), not sure of the exact HP part number for the mainboard since it's back together for the moment. I suppose it could be the GPU or even the Northbridge (MCP?) causing some fault, I don't have the technical skills to specifically diagnose the issue though let alone repeair it. I could investigate it further but I think I am better off just buying a new mainboard. For your (and everyone else's) information, this is the cheapest one I could find for this specific model Laptop - http://laptops-motherboards.com/hp-pavilion-dv9521tx.htm although for all I know it will be a 256MB graphics memory version instead of this current 512MB.
Another random note is that the fingerprint reader has failed (USB Device Not Recognized) so my guess is that while the old CPU was damaged and replaced, the heat from the CPU (or the GPU before the copper shim mod) damaged other mainboard components.
Thanks again for your fantastic information and for sharing your chronicles with the web!
I have a dv6500 where the power would flash once and then that's it after I did a reflow. I reflowed it a second time and now the laptop won't even power on. Can you tell me why this would be happening.
ReplyDeleteThx
Darren: That's hard to say, I don't know where you got the instructions from. There are many bad videos on Youtube that can 'brick' your mobo. I don't know if you did it correctly, or the proper method. Almost anything could be the problem if instructions were incomplete. There are several steps to my reflow process. Not just reflowing. Some models even have added steps depending on the GPU used.
ReplyDeleteAs previously noted, they have 'solutions' using tea candles, torches, baking in an oven etc. It took me 2 years progressing to the point of a Final Solution where they don't come back. Though they still can fail 10-16 months later if they are overheating again via clogged vents or using on bed or carpet. Other than using on a table or flat surface - they will come back sometime in the future. If the vents are blocked it cannot cool- no matter if you use ArcticSilver (or any other thermal compound), Chinese Shim, or even my modified shim.
I just tried make sure they last as long as the mfg does. So I give them a 60 day warranty. the reason I have not posted the video is that everyone will copy and put it up on YouTube saying it is their 'original' idea after I had to go through a hundred tweeks to get it perfected.