HP dc7800 bad caps

[c_hegge]

Re: HP dc7800 bad caps

I'd actually be going with the Nichicon LG polymer. The specs would be superior to the original and polys will last longer than any electrolytic, regardless of what the endurance lifetime is. Remember, the endurance lifetime is what it lasts running at it's maximum rated voltage, ripple and temperature. In reality, though, it won't be under anywhere near as much stress, and so will easily last untill the PC is well and truly obsolete.

[jakah]

Re: HP dc7800 bad caps

Thanks for the tip about picking a cap with higher voltage! The cap that actually has gone visibly bad is the 1500 uF cap, so if I would pick polymer caps for the 470 uF caps I guess I should pick a polymer replacement for the 1500 uF aswell. I noticed that only replacing the 1500 uF cap makes the system stable. The system runs very cool anyway, so I think good regular electrolyte caps will give it a long lifetime.

What I just noticed though is that where some of my machines have 680 uF Rubies, some machines have 680 uF 10V OST RLX caps instead (see picture). I found this replacement for them:
http://se.farnell.com/panasonic/eeuf...0uf/dp/1848385

[jakah]

Re: HP dc7800 bad caps

My lab assistant have recapped about 10 machines now, and they all run stable after recap

Now I have a question for you. Next to the failing 1500 uF OST capacitor (EC13), there is a "slot" for another capacitor (EC14). Can I put another cap in there? Is there any way of telling if that cap will end up in parallell with the EC13 cap? See attached picture. If I get no resistance with a multimeter between e.g the "+" holes of EC14 and EC13, they must be???

[momaka]

Re: HP dc7800 bad caps

If I get no resistance with a multimeter between e.g the "+" holes of EC14 and EC13, they must be???

Define "no resistance" because it can mean either:
- Open circuit (infinite resistance - multimeter shows OL or 1 on the left side of the screen)
or
- short circuit (2 Ohms and less)

If you get a short circuit or a very low resistance less than 2 Ohms between the "+" pads on the 2 capacitor spots, then the capacitors are in parallel.

If you get an open circuit between the "+" pads on the capacitor spots, then they are not in parallel. If that's the case, then see if the cap is connected to any of the other rails (3.3v, 5v, 12v... etc.) by checking if the "+" for that cap has continuity to that rail.
Since there are no voltages on the motherboard above 12v, 16v caps should work for any empty spot. If you want to go with a lower voltage cap, then you need to measure the voltage on that spot so you'll know how much you can lower the voltage rating of the cap.

[PCBONEZ]

Re: HP dc7800 bad caps

To be in parallel BOTH the + and - pads must be connected. [Zero resistance.]
.
Also if there is a coil between them the resistance might look like zero [because coils have very low resistance] and they aren't actually in parallel.
You will run into this a lot when working with filter caps because coils [inductors] are common in filters.
.

[jakah]

Re: HP dc7800 bad caps

Thanks for that information! I'll test the resistance...
There is only one single 1500 uF capacitor there on the board. What do you think it is used for? The system gets lots of disk I/O errors when it fails.

Jakob

[apabhilash]

Re: HP dc7800 bad caps

Hi, I am also Having the Same prob with 4 of my DC7800.

The machines will not start in the morning without 3-4 restarts, then it will start. Also machine hangs sometime with blue screen .I check the motherboard and fount the OST caps are bulged. So i am planing to change all ost caps from those motherboards (1500uf 6.3v & 1000uf 10v ).It it enough or i have to change some more.

Before 3-4 months i had to change a set of 1000uf 6.3 from 8 of my HP dx7400 just after the warranty.

I don't know why this HP's caps are dying just after the warranty.


Regards
Abhi

I found this when I was troubleshooting a bunch of HP dc7800 machines in my computer lab that had started to behave strange; a bad OST cap close to the memory slots. The board has a mix of caps from Rubycon, Sanyo, some green ones where I can't read the brand, and one OST (and some labeled FL...is that a name of a brand or a particular series of one???).

I haven't recapped yet, but planning to. Anyone else seen bad caps in those machines?

Regards,
Jakob

[Uranium-235]

Re: HP dc7800 bad caps

The Sanyo (green ones around the CPU VRM) are the some of the best, as well as the ones with the K stamp are to. However they put a crappy cap in the VRM for something else (looks like RAM VRM) and it dosen't matter how good caps are anywhere else.

Add that to the fact these are SFF, heat becomes an issue, and crap caps + heat = death. Hell even good caps have a temp limit and they will bloat if they get too hot. If yours are SFF. Clean the crap paste off the heatsink and CPU (with alcahol or thermal paste cleaner chemical if you want to get fancy), apply some arctic silver 5, try to spread it THIN, get all the dust out, and hope for the best

[bluerobboman]

Re: HP dc7800 bad caps

Hi Chaps, Just a follow on this thread I have 3 dc 7800 Towers which all have started to behave "strangly" My problem was with the faster PC6400 RAM installed the graphics would get corrupted and hence the machine would hang up completly! Visually the only capacitor which had bloated was that OST (1500uF 6.3v) one near the RAM slots see jakah fig 1. It was my belief that the higher clocking ram places more current demand on the supply to the RAM and hence the faster RAM which demanded more current would fail whereas the slower RAM which had less curent demand would work! Well upon removal and test of the aforementioned cap revieled that it's capacitance had dropped from the nominal 1500uF to roughly 890 uF and it's ESR had risen from a nominal 15mohm to roughly 1.6 ohm (cold). Upon replacement and retest the PC worked quite happily with the PC 6400 RAM and all was fixed. So, in summery, if you suspect the RAM don't condem it! As there is a much cheaper fix!! Incidentaly REV A03 boards have two (1500uF) capacitors fitted whearas the REV04 Boards have only one with a few other caps having been removed for "Penny pinching purposes!" Which makes this problem more prevelent on REV04 boards and above! Just a note I see that people are replacing all sorts of capacitors on these boards which leads me to ask why? Just because a certain make has had problems and only in certain value /voltage situations and applications I hasten to add! Why replace them? Surely you stand more chance of damaging something in the process than the possible benefits you stand to gain. Another point in the previous comments above certain brands which have a few milliohms more will make no difference for the motherboard's performance as the ESR will have had to risen to ohms and the capacitance value to have dropped significantly for problems to emerge. The main parameters to be aware of are, of course, the temperature and operating hours, I tend to always fit 105 degC caps and long working life caps as I believe this is good practice. I know these PC's are a pain to remove the motherboard however I would suggest measuring capacitance values and ESR's of suspected components before blanket replacement, this keep's costs down as well

[momaka]

Re: HP dc7800 bad caps

Just a note I see that people are replacing all sorts of capacitors on these boards which leads me to ask why? Just because a certain make has had problems and only in certain value /voltage situations and applications I hasten to add! Why replace them? Surely you stand more chance of damaging something in the process than the possible benefits you stand to gain.

It all depends on who you are repairing the board for. If it's for myself, I sometimes get lazy and do the same thing as you - just replace only the caps that are bad.
Of course if you are repairing this for someone else (and especially if they are located far away), you may not want to have to open the computer a second time just to change a few caps again and have the process repeat over and over until all the caps have been changed. The problem with bad capacitor brands is their consistency (or rather lack of it). They *will* in most cases fail predictably like they did in that same spot on these DC7800 PCs. However, it's the fact that they can also fail unpredictably in other areas that makes it worthwhile to change them all at once. The only excuse not to do that is if you really have bad soldering skills and think you will mess up the board - in that case I agree with you: just changed the caps that are bad and maybe the ones that you don't think you'll mess up.

Another point in the previous comments above certain brands which have a few milliohms more will make no difference for the motherboard's performance as the ESR will have had to risen to ohms and the capacitance value to have dropped significantly for problems to emerge.

Wrong.

First, the performance of the motherboard may get affected but you just may not even notice it until it's too late. Data corruption is one of those instances.

Second, it all depends on the circuit the caps are in. Some circuits are more forgiving, others are not. The bucking circuit for the CPU voltage (the VRM) usually is since you have many caps in parallel and also because most VRMs are multi-phase. However, sometimes the motherboard's NB may also employ a bucking circuit for its voltage regulation and most of the time it's a single phase - with this type you have to be careful. If the caps go completely bad, the MOSFETs may burn out too. I have a Shuttle motherboard that did that. Even worse, when the upper MOSFET shorted on mine, it also took out the PWM controller. I have yet to find a way to fix that.

I know these PC's are a pain to remove the motherboard however I would suggest measuring capacitance values and ESR's of suspected components before blanket replacement, this keep's costs down as well

If my time is worth even as low as $5/hour, taking out the motherboard out more than once will offset any costs. Even worse, if you bought replacement caps only for the ones that have failed, you would need to do that again as more caps fail, so you'll be paying for shipping twice as well.

[Astronutty]

Re: HP dc7800 bad caps

I just got 40 of these HP 7800s, and they each have two OST-RLA 6.4v 1500μF that are swollen. The two caps are of different form factors, one tall and skinny and one short and fat, but with the same ratings. Is it okay to replace them both with https://www.badcaps.net/store/produc...roducts_id=147 ?

They each also have two UCC KZG, but I would like to pretend they are okay.

Steve

[momaka]

Re: HP dc7800 bad caps

^ Yes, those Nichicon HZs should be fine.

If the KZGs are in hot areas (near the CPU, NB/SB, or other high stress places), I would say replace them too. Two more caps shouldn't be that big of a deal.