HP Compaq 7620s 7 flashes of death.

[KountZERO]

This one has been on the shelf for a while now and thought i'd have another look. These motherboards seem to be plagued with this 7 flash then off problem. There are tons of dead ones on ebay...

I did a bit of digging tonight and found a polish tech forum mentioning the model and fault in question.

http://translate.googleusercontent.c...b2N71W9ZEl-BNQ

What i cant fathom from the translation is, is the TPS51120 IC knackered or are they saying there is a short causing the TPS51120 to trip its protection?

It gets too hot to touch with the battery inserted and power off
Same with just power lead and battery and just power lead.
It does however charge the battery ok.



This appears to be the schematic for the area.

http://pdf1.alldatasheet.com/datashe.../TPS51120.html

Im hoping someone more savvy than me can look through their findings and decipher the cure.

Cheers.

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

I can't find much information on a 7620s. Did you mean 6720s?

Schematics for the6720s (and related models) here:
http://www.mediafire.com/?h909692ir7fab6x

The power diagram is on page 4 and you can see that the battery charger/selector is before the main system power rail so it is normal that battery can still charge.

Page 7 shows the TPS51120 and surrounding circuitry. It has two 100mA LDOs, one for 5V (labeled V5AL) and one for 3.3V (labeled V3AL). They should always be on when power is supplied. A short somewhere on either rail can cause the '51120 to heat up really quickly. You can isolate the V3AL by desoldering PAD555 and then checking for shorts beyond it, or by just plugging in and seeing if the '51120 gets hot again. If it still gets hot with V3AL disconnected, either the '51120 is internally bad or the V5AL is shorted. V5AL only powers U520, U8, and U6, the latter two are for the battery charger so I doubt they're shorted. Check the VREG5 and VREG3 voltages on pins 19 and 21 of the '51120, with V3AL disconnected. They should be at 5 and 3.3V.

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Argh my bad! 6720s is the beast in question.

Thanks for the schematic, time to get hands dirty.

Will report back, cheers

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

hmm cant find pad555 anywhere on pcb atm. will do some readings tho with things as they are.

Pin `19 = 3.3V
Pin `21 = 4.45V

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

Have you checked the other side of the board? The pad shouldn't be far from pin 19 and C764. But the 3.3V looks OK so the fault is probably somewhere in the V5AL circuit. Post a high resolution, clearly focused picture of the board and we can help identify things beter.

4.45V is indeed too low and shows the regulator being pushed beyond its current limit. Fortunately there's not much on V5AL... check these for shorts by removing these and testing resistance across them (or just seeing if the voltage rises back to ~5V with them out of the circuit):

C765-767
U520
(on page 5
C500
C503
U6
U8

On page 27 of the TPS51120 datasheet there's "VREG5 output voltage vs output current" and extrapolating that graph down to 4.45V makes it look like the short is consuming almost 1A to pull the voltage that low. Anything that reads <10 ohms is likely bad.

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Yeah i checked both sides, even lifted all the heatproof film up and checked there. There seems to be very slight differences on this mobo, i guess it must be a different revision. However its very close to what the schematic shows.

Gonna put a hires image up of both sides tomorrow in natural light as the images i made this eve have too much glare grr.

The readings around U520 are 3.3v and 4.45 so that suggest to me it hasnt failed [yes noob comment]

Will lift Caps tonite tho to begin with n check

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

The readings around U520 are 3.3v and 4.45 so that suggest to me it hasnt failed

If you remove it and the 5V rail comes back up to normal, then it would be the shorted part. Otherwise it's ok. The shorted part should also be getting rather warm, so you could use that as another tip for where to look.

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Lifted the U520 and no change, same voltage and IC still super hot. Now is the time i wish i had an infrared doodad.
Working on the images, need to borrow a higher MP camera to get the descent macro imaging. Not giving up on this yet, have the feeling its going to be something realy simple.

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

An infrared thermometer isn't that expensive... http://www.dealextreme.com/p/1-2-lcd...-c-380-c-35754

If U520 isn't the problem, there isn't much else left to check...
C500
C503
U6
U8

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Lol! i wanted a hotspot and i got one. After lifting all the components above and not getting any change was about to put it back down and do something else. Turned the magnifier light off and a tiny arc from the IC popped out at me. Had a realy close look and one of the legs [pin 22 vbat] was broken right by the chip body. Exacto knife and kynar wire sorted that but on powering again i was treated to a firework display from an LDO fun !

The offending article is marked q2003 and not on the schematic...

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

Did this happen? http://www.elvikom.pl/forum/obrazki/_f/f_0009853.jpg

http://translate.google.com/translat...hl=en&ie=UTF-8

Q2003 is FDS6900AS, it replaces Q514 and Q513 in the branch 3.3V.

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Yeah, that was it exactly! Took out some of the copper so looks like I'm going to have to trace other solder points. It will live again! Lol

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

That image seems to be of the version of the motherboard as well. I cant seem to be able to get an image as good as that one

Checking the dmg again it seem that only a bridge to the other side of the board needs to be repaired. Its a shame i cant find an image of the pcb mask for the area under the LDO.

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

LOL who would have thought i'd find a use for my old nokia n73.

Here is the damage after cleanup. You can see the bridge to the other side, its track came from pin 3, back under the IC. Now to fathom what caused the pyrotecnics before i blow another up lol.

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

Now that the 3.3V rail has basically been "isolated" from everything else, it would be a good idea to check if it's shorted to ground. That would be one way to blow the MOSFET like that.

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

It is shorted to ground, lots of caps to check now >< .My smd soldering is getting a workout tho...

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

If you have a low-voltage current-limited supply you can set it to <3.3V and ~100mA current, connect it to the 3.3V rail (with correct polarity!) and feel for warm parts. If you don't feel any, increase the current limit slowly and try again. You can also check for voltage drops as described here: https://cdn.badcaps-static.com/pdfs/...5b14c897b2.pdf

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Going to get hold of a bench supply asap.

http://www.ebay.co.uk/itm/2906677324...ht_4029wt_1031

This should be sufficient, yes?

I only have an old 12v 5A supply that isnt variable [old CB gear] and the only other supplies i have are of the 500ma - 1A multi-use things. Not good enough tbh.

I guess the more i tinker with broken junk, the more test gear i'll accumulate. This won't beat me lol. But i have had another craptop that belongs to my young nephew dumped on me to rescue - i will get earache if i dont give it a whirl. He's had it 3 weeks and already managed to short out the jack, arcing a nice big hole in the board Ive redone the jack but i think one of the internal planes that surrounds the pin hole is an issue so ill have to find another place to bridge it.

Thanks for the advice so far, i've learnt a lot in the process,I will keep updating.

[b700029]

Re: HP Compaq 7620s 7 flashes of death.

That looks like a bunch of other generic adjustable cc/cv supplies... some may be built safer than others.

you can find a lot of people who review those things here:
http://www.eevblog.com/forum/product...nd-discussion/

[KountZERO]

Re: HP Compaq 7620s 7 flashes of death.

Ok after a little break i looked at the board again. I am starting to loose faith, i cant find the short anywhere ><. I'm now wondering if its going to be under the gpu and the person it belonged to has tried reflowing it, popcorning it in the process. Would that cause the symptoms seen before?

Anyway ill take a look again later, got my partners Acer to look at now. Not a task i'm relishing lol