laptop hp dv6-3132 dead, no lights at all

[piernov]

Re: laptop hp dv6-3132 dead, no lights at all

First you will have to remove them one at a time to see which one is shorted.

I can't find anything with those markings so they might be actually different.
The schematics list some Renesas parts (RJK03B9DPA and RJK03D3DPA) so you could replace all the MOSFETs onboard with the Renesas parts. It is recommended to replace all of them and not just 1 in order to keep the phases balanced. Also it is recommended to replace the controller (PU1) too in case it may have died.

Also if one of the high-side MOSFET is dead, the CPU might be dead too. Btw that's why you have to disconnect everything from the board when injecting voltage: if a high-side MOSFET is dead, the voltage you're injecting on the +VIN line goes directly to the CPU.

[starling22]

Re: laptop hp dv6-3132 dead, no lights at all

As I find it odd that those 8 chips should go. Of course at this stage I do not know if more components are dead. These eight did get hot when about 10W power or more was applied.

You have me a bit concerned about the cpu I should have dismantled this earlier as well. I understand that if I had applied external power to the location where PD23 is, say on a good motherboard, with my method the cpu might be damaged.

In my case the lowest voltage setting I would have applied is 2.7V both for my 'low power' supply and 'high power' supply. What I was thinking when I had located a shorting condition was following:

With 2.7V and current limit this would be safe start as the voltage would not rise very much on the board. The experiment seem to support the thinking as even with 0.9 A going in hardly only fraction of a volt developed.

However with the high power supply the situation is not soo good, in my mind I was thinking it was safe under 3.3V but that is only because I have just been working on small microprocessor using 3.3V.

Looks like I may well be learning the hard way.

[starling22]

Re: laptop hp dv6-3132 dead, no lights at all

Been slightly sidetracked fixing up lighting and making room in my workshop in preparation for attempting SMC work. I had an old Intel motherboard about 10 years old I guess. Starting practicing removing components.

First go removing a small 8 pin device, I had too fast air stream as I blew away a small component when the chip came away. I may have shaken my hand when grabbing the chip with my tweezers. Removing components I will be able to do. Soldering them back again will take me some time to practice as I notice that my hands shake a bit much for comfort.

A good friend and very good electrical/electronics engineer did not take long to identify the MOSfets as being Fairchild after I sent him a couple of pictures I took last night doing my best effort to get markings visible. What I thought was the chip name appears to have nothing to do with the name. I got Fairchild correct. I made a big mistake not attaching a couple of poor visibility but very close up pictures to an earlier post as I am sure somebody would have put me right. I am learning in my old age.

Message from my friend:

Fairchild Mosfet in a DFN8 package.
You have the wrong part numbers!
First one probably is FDMS0310
Second one - FDMS7692

I have looked on the internet and got the data sheets, also they can be purchased easily enough, not exactly cheap. I will obtain 5 of each after I have rechecked my pictures.

Following is a reminder to myself, when I read the archived thread in the future, to use my now working hp 6284A power supply which can supply 20V/4A and fully variable 0V to 20V (with over run to 24V) current limit can be set from 40 mA to 4 A. I resurrected it today.

By the way I have spent some time identifying the electrolytic capacitors on an ASUS P5P41D motherboard, I think I counted 19 and of those at least 4 has bulging top and the board gave up the ghost a couple of years ago. I have 5 replacement caps but reading on this Forum I think they all should be replaced if I think of getting it back to work again.

[piernov]

Re: laptop hp dv6-3132 dead, no lights at all

In that case you can replace all the MOSFET at once, and then before powering up the board check resistance between +VIN line and VCore, and between VCore and ground, *without CPU*. It should be at least 10kohms I think.

Those kind of PSU are great but I often find 4A not to be enough. I've got a DPS5015 DC-DC converter which can do up to 50V/15A, and I feed it with a 48V/8A PSU. I often go up to 8A current limit on low-voltage high-power lines.

For desktop motherboard, I never replace all the caps, only the ones that are physically damaged and their friends (ie. caps with the same specs and on the same line as the dead ones). Works for me in most cases. Also it depends on the brand/series of the caps, some are known to fail. (just went replacing 4 KZG and 4 KZJ on a Dell MB, every other caps were fine) Also I don't really bother matching the caps, I only make sure the voltage is high enough and the capacitance is around the same or higher, and obviously that the caps fit in. I do use low-ESR/high ripple current caps when needed though. Well… Do what is the easiest and what works for you.

[starling22]

Re: laptop hp dv6-3132 dead, no lights at all

RS here in Australia carry the FDMS7692 MOSFET but not the FDMS0310 but they do have another FDMS037N08B which in most respects seems more rugged. I am tempted to buy that. Buying from RS is good insurance against getting low quality from internet.

Capacitance and min Vgs (Gate-Source) voltage is approx 1V higher I do wonder if that is important enough to dismiss it. I will attach data for both.

[piernov]

Re: laptop hp dv6-3132 dead, no lights at all

Capacitance is an important factor because it will greatly affect the time the transistor takes to switch between one state to another. The timings are described in "Switching Characteristics" section. You often want to have similar or lower timings here.
That's what happen when you have a "beefy" MOSFET, since it has to be "bigger" to be able to handle the voltage and the current, you get a greater gate capacitance, so it is slower.

This MOSFET won't do. I will roughly explain how you can see this, this is probably the wrong way to do it but well…ask a real electronics engineering if you really want to understand how this works.
Let's say this buck converter works at 300kHz. With a 50% duty cycle (half time on, half time off), this means that we have roughly 1.5µs of "on state" and 1.5µs of "off state". A 50% duty cycle would also mean that the output voltage is roughly 50% of the input voltage, ie. from 19V to 9.5V. However our CPU can work as low as 0.9V, which is around 5% of the input voltage, so we need a 5% duty cycle, ie. 0.165µs of "on state" and 3.135µs of "off state".
If the MOSFET already takes 0.130µs (max td(on)+tr) to fully turn on, that's no good, even if the controller should compensate for the latency.
Well… that's a bad explanation but you should get the idea why those timings are important.
And since the turn-off time is higher than the turn-on time on MOSFETS, for low-side MOSFET (which more or less has the reverse behaviour than the high-side MOSFET) the timings must be even better.

About the gate treshold voltage, it has its importance too. The MOSFET drivers inside the buck controller work at 5V, so the gate threshold voltage of the MOSFET must be lower than that, and the lower the better. If the margin is too tight, the MOSFET may be noticeably slower. Remember that the datasheet lists timings at a gate voltage of 10V, but here it will be driven at 5V so it will be slower.

[starling22]

Re: laptop hp dv6-3132 dead, no lights at all

Ok, I understand perfectly. The risk is too high that something will go wrong and it will be wasted effort. I will have enough trouble trying to solder in the ICs, this exercise will do as far as creating problems this time.

I will buy one set off the internet and the other from RS-Australia.

If I stumble over a website(s) that look at these things in depth I will listen in out of curiosity.

[starling22]

Re: laptop hp dv6-3132 dead, no lights at all

While waiting on replacement MOSfets I wondered about on the net and took an interest in trying to see if I could locate a "boardview" for the hp dv6-3132 laptop.

It seemed harder than expected. I have downloaded "Openboardview" and a random boardview. Playing around with this.

Yesterday I believe I got very close to finding a boardview, I made a note too somewhere but failed to find it again today.

google: "QUANTA LX6-LX7 DA0LX6MB6H1 HP liverpool boardview"

The term above does give some promises in that:

http://laptopserviz.bg/enview/boardview.html?p=2

Failed to find a boardview as the search system seemed to fail me.