HP Elitebook Power Supply

Thread Starter

trader007

Joined Feb 27, 2010
249
One of the bigger burst electrolytics is more likely a symptom than the cause, if you miss a dud error-amp sampling cap you'll just destroy all the new parts you put in, its usually quite small, say 1 to 47uF, the smaller values I always replace with non-electrolytics that can't dry out, these days multilayer ceramic chip caps are readily available in the higher values.

Its also worth remembering that resistors favourite failure mode is to go high or open - if you have one like that in the voltage sampling divider - your luck won't change anytime soon!
can you make my hope any less grand? :D

I don't know if im going to take the time to test every resistor and smd cap... if im going to spend that much more time on this ill just scrap the project and buy a new supply.
 

ian field

Joined Oct 27, 2012
6,536
can you make my hope any less grand? :D

I don't know if im going to take the time to test every resistor and smd cap... if im going to spend that much more time on this ill just scrap the project and buy a new supply.
I'd leave any plastic encapsulated SMD electrolytics to very last, Ceramic SMD caps usually show physical signs such as cracked through or a tiny burn mark.

You can use a bit of intelligence checking out the resistors, if there's a TL431 programmable zener driving a feedback opto, finding the voltage sampling divider is dead easy.

You can do a certain amount of checking of resistors in-situ, you might get a lower reading than the markings indicate because of other stuff that conspires to be in parallel - but any resistor reading higher than marked is definitely faulty. Test both ways - a parallel semiconductor junction can give a low reading one way round, swap the probes around and you might just get the true value.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
thanks for the info. I do follow, and I'll try to find the resistor(s) around the divider... The problem is this whole thing has been encased in thermal glue... its just a pain in the ass to get to anything. I will try though, I would hate to get good information like this and not even try to apply it.
 

ian field

Joined Oct 27, 2012
6,536
thanks for the info. I do follow, and I'll try to find the resistor(s) around the divider... The problem is this whole thing has been encased in thermal glue... its just a pain in the ass to get to anything. I will try though, I would hate to get good information like this and not even try to apply it.
With laptops there isn't always a voltage divider on the output, many designs use a more precise regulator inside the laptop itself.

If the power brick doesn't have to have spot on regulation, they sometimes regulate on the voltage developed by the aux winding that feeds the chopper chip once its started up. The capacitor on this rail is particularly the error-amp voltage sampling cap that I mentiones, the cap has to smooth Vcc for the chopper chip *AND* sample the regulation voltage.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
I'd leave any plastic encapsulated SMD electrolytics to very last, Ceramic SMD caps usually show physical signs such as cracked through or a tiny burn mark.

You can use a bit of intelligence checking out the resistors, if there's a TL431 programmable zener driving a feedback opto, finding the voltage sampling divider is dead easy.

You can do a certain amount of checking of resistors in-situ, you might get a lower reading than the markings indicate because of other stuff that conspires to be in parallel - but any resistor reading higher than marked is definitely faulty. Test both ways - a parallel semiconductor junction can give a low reading one way round, swap the probes around and you might just get the true value.
well, you might have been right. I tried to look for the resistors, but again everything is so buried in dried glue i felt like i might actually do more damage trying to pry it all out.

I just put the parts I had in it, plugged it in and bingo... it worked. for 10 minutes. then the fuse and regulator blew again. F it, im done with this stupid thing.

sad, because I think it was close to repaired... but oh well. I have another power supply I am working on, for a samsung monitor, but I can get a new one for $25... I doubt I will even try to troubleshoot when the parts will likely be half the cost of a new board anyway. Just replace the whole thing and be done with it.
 

alfacliff

Joined Dec 13, 2013
2,458
since the input filter cap shorted, did you check the input rectifier bridge? a shorted cap would have put a large strain on it.
I have fixed a lot of switching supplies like this, and it is often one of the small electrolytic caps in the supply for the pwm chip. they use a capacitive voltage devider to supply power to the chip till the thing comes on completely.
cliff
 

RamaD

Joined Dec 4, 2009
328
i am keen to all of that ian, but thank you anyway for the input.

When I can, I just replace with the exact part that was there from the manufacturer. In this case, Digikey had the exact brand and model cap in stock. What made me believe it might just have been the cap was its MTBF rating- only 5000 hours. The datasheet said the cap was designed for high frequency switching and the ratings were made based on these conditions.

So, if you figure that to days, its only 208 days worth of charging on average that these power supplies can technically do!! I have had mine for 2 years, and its likely been in charge mode well over 5000 hours already... I did think about buying a different brand cap with a higher lifespan, but it also needed to fit in a VERY tight space, so it was just easier to buy the original cap. This cap was $5! one cap!
The 5000 hours, is not as bad as you have mentioned.
It is not MTBF, but endurance or life of the part (to the specified capacitance derating) at the specified temperature.
Below that temperature, life doubles for every 10 deg decrease in temperature. For example, if the cap is spec'd at 5000 hours and 105 deg C, and operating it at 55 deg C (including the heat caused by cap), the life is 160,000 hours, which is a respectable 18 year life!
Prices seem to vary a lot in these caps and could get get similiar parts from different mfrs at much lower price!
 

Thread Starter

trader007

Joined Feb 27, 2010
249
The 5000 hours, is not as bad as you have mentioned.
It is not MTBF, but endurance or life of the part (to the specified capacitance derating) at the specified temperature.
Below that temperature, life doubles for every 10 deg decrease in temperature. For example, if the cap is spec'd at 5000 hours and 105 deg C, and operating it at 55 deg C (including the heat caused by cap), the life is 160,000 hours, which is a respectable 18 year life!
Prices seem to vary a lot in these caps and could get get similiar parts from different mfrs at much lower price!
Thank you so much for explaining that. I had no idea. Makes a hell of a lot more sense now.
 

ian field

Joined Oct 27, 2012
6,536
well, you might have been right. I tried to look for the resistors, but again everything is so buried in dried glue i felt like i might actually do more damage trying to pry it all out.

.
That dried glue might actually be the cause of the problem.

If its gone brittle its fairly easy to chip away - but by then its usually gone corrosive, you might not like what you find underneath!
 

Thread Starter

trader007

Joined Feb 27, 2010
249
i dont even know why they use the stuff. i guess they figure it will keep parts from cracking their solder connections loose, but like you said it seems to do more harm then good.

anyway during my 3rd round of troubleshooting and sourcing parts, i found on ebay someone just put up brand new OEM supplies for my laptop for $30 shipped. at that price, its just better to buy a whole new one, even though I already sank $30 into parts for the broken one :D

New supply has been working fine. It's a slightly different model it looks like, with a slightly different casing... but the same wattage output and same overall size. This one runs much cooler though... like, barely warm at all when charging. I think my original one had a problem for a long time, because it always ran really warm. That, or it was just an older revision with lower quality parts inside...
 

ian field

Joined Oct 27, 2012
6,536
i dont even know why they use the stuff. i guess they figure it will keep parts from cracking their solder connections loose, but like you said it seems to do more harm then good.

...
When I made my living repairing CRT monitors for a living, that glue was a regular.

Sometimes I found component leads eaten right through. Once the glue got in its brittle corrosive state it could cause leakage paths - even carbonised arcing over in high voltage sections.
 

Thread Starter

trader007

Joined Feb 27, 2010
249
When I made my living repairing CRT monitors for a living, that glue was a regular.

Sometimes I found component leads eaten right through. Once the glue got in its brittle corrosive state it could cause leakage paths - even carbonised arcing over in high voltage sections.
why do they use it then? its like they want people to buy new stuff after 5 years or something...
 
Top