Doornkaat wrote on 2023-06-04, 19:27:

I'm also confused. Regularly soldered SMD electrolytics with a regular iron withput preheating the board or anything. No issues so far.🤷‍♂️
Edit: Are you by any chance not using flux or are you using an iron with too low temperature, shevalier?

Solder a few smd electrolytes on an unnecessary board.
Diameters 5, 8 and 10 mm. The latter will be in power circuits and at ground/power plane.
Then break them off by rotating, as I drew above.
Look at the result.
PS. I also thought for a long time that everything is OK.
Lack of contact across the entire pad results in increased ESR&ESL and no mechanical strength.

shevalier wrote on 2023-06-05, 02:41:

Solder a few smd electrolytes on an unnecessary board. Diameters 5, 8 and 10 mm. The latter will be in power circuits and at gro […]
Show full quote

Solder a few smd electrolytes on an unnecessary board.
Diameters 5, 8 and 10 mm. The latter will be in power circuits and at ground/power plane.
Then break them off by rotating, as I drew above.
Look at the result.
PS. I also thought for a long time that everything is OK.
Lack of contact across the entire pad results in increased ESR&ESL and no mechanical strength.

Have you considered that the root cause might actually be your particular soldering technique?
Everything that you wrote is based on your own experience, which does not constitute enough data/evidence to draw a pertinent conclusion, wouldn't you say?

It's like me saying that I can't play the cello, therefore nobody can. Just something to think about... 😉

bloodem wrote on 2023-06-05, 09:16:

Have you considered that the root cause might actually be your particular soldering technique? Everything that you wrote is base […]
Show full quote

shevalier wrote on 2023-06-05, 02:41:

Solder a few smd electrolytes on an unnecessary board. Diameters 5, 8 and 10 mm. The latter will be in power circuits and at gro […]
Show full quote

Solder a few smd electrolytes on an unnecessary board.
Diameters 5, 8 and 10 mm. The latter will be in power circuits and at ground/power plane.
Then break them off by rotating, as I drew above.
Look at the result.
PS. I also thought for a long time that everything is OK.
Lack of contact across the entire pad results in increased ESR&ESL and no mechanical strength.

Have you considered that the root cause might actually be your particular soldering technique?
Everything that you wrote is based on your own experience, which does not constitute enough data/evidence to draw a pertinent conclusion, wouldn't you say?

It's like me saying that I can't play the cello, therefore nobody can. Just something to think about... 😉

Dude, just rotate a couple of capacitors with needle nose pliers and see for yourself.

shevalier wrote on 2023-06-05, 14:31:

Dude, just rotate a couple of capacitors with needle nose pliers and see for yourself.

"Dude", I've been doing this for 20+ years. I don't need to rotate anything at this point.
What you are describing is something that can indeed happen when using an improper/suboptimal SMD soldering technique.

This is the first time I heard such a thing.

If you heat both the pad AND the capacitor leg properly with the iron, the solder should also flow under the leg as well. If that doesn't happen you are probably not applying enough heat to the joint. You can also pre-tin the pads and apply braid before soldering to have a thin layer of solder between the pad and the leg that will melt and mix with the solder you add later.

And of course, flux helps this process.

bloodem wrote on 2023-06-05, 16:31:

I've been doing this for 20+ years.

I envy.
To Your youth.

bloodem wrote on 2023-06-05, 16:31:

What you are describing is something that can indeed happen when using an improper/suboptimal SMD soldering technique.

It is possible to solder SMD electrolyte with solder iron tip
1. If pads on the board have shape for hand soldering with an increased length for touching the tip. (For reflow soldering, the pads are smaller in size to prevent the part from floating.)
2.1. the pad has a thermal relief at the plane.
2.2 or capacitor in the signal circuit.

Or if you have an inductive 125 watt Metcal.
Otherwise, use preheating. If you care about the result.

Tip for beginners - do not solder SMD electrolytes with a soldering iron

TheMobRules wrote on 2023-06-05, 18:11:

This is the first time I heard such a thing.

https://youtu.be/iFhyP_QwtMc?t=1352
22:30
Correct and complete solder reflow.
The fact that the capacitor is held on the PCB does not mean that it is soldered well.

I apologize to @mrfusion92, the topic obviously went the wrong way.

shevalier wrote on 2023-06-05, 18:55:

https://youtu.be/iFhyP_QwtMc?t=1352 22:30 Correct and complete solder reflow. The fact that the capacitor is held on the PCB doe […]
Show full quote

https://youtu.be/iFhyP_QwtMc?t=1352
22:30
Correct and complete solder reflow.
The fact that the capacitor is held on the PCB does not mean that it is soldered well.

Not sure I understand what you mean… Don’t you agree that those two capacitors that he removed at 22:30 have very good solder joints?

shevalier wrote on 2023-06-05, 18:55:

I apologize to @mrfusion92, the topic obviously went the wrong way.

No big deal. It's a public forum.

When somebody will have be able to reply to my last questions I'm sure they will.

mrfusion92 wrote on 2023-06-04, 20:03:

Otherwise I've found this that seems to be a "drop-in replacement".

Honestly I would go with this one, I always prefer an exact replacement when possible. Using MLCC instead is probably OK in this case, but if the exact same brand/series of electrolytic that was there originally is available, you're ruling out any potential changes to how the circuit works.

As for the others: Rubycon YXGs should be still available and are very good for motherboards of that era. But the replacements suggested in that page are fine also.

I decided to work on the P2B first. I soldered the correct mosfet in Q10 and replaced all the caps.

Now with the mosfet the cpu fan spins but still no post. A pci diagnostic card shows no 3.3 voltage, I guess that it is an issue.

What components should I check?

there is no 3.3V on PCI slot to begin with so no problem there
Measure all voltage rails with a multimeter, or even better with a scope to check for excess ripple

Do you mean this particular motherboard doesn't have 3.3v in the PCI slot? Because I'm pretty sure in the P3B-F the voltage led in the diagnostic card lights up.
In the P2B, all the other voltages leds light up correctly. Clock too, and even the reset works when I press the button.

I have only the multimeter, I already checked the 5v line from that mosfet and it was okay, I will check the other voltages too.

mrfusion92 wrote on 2023-08-25, 06:42:

Do you mean this particular motherboard doesn't have 3.3v in the PCI slot?

Im looking at P2B boardview right now, only 5V on PCI slot
Re: Early revision ASUS P2B and Coppermine Pentium 3

rasz_pl wrote on 2023-08-25, 08:03:

Im looking at P2B boardview right now, only 5V on PCI slot
Re: Early revision ASUS P2B and Coppermine Pentium 3

Oh thanks! Good to know.

I see that there are only 1.10 and 1.12 schematics, actually my board's revision is even more earlier of the one of that topic, it is a rev 1.02.

I think I will give up. I checked the voltages, all of them are in good range. I checked pins also in the CPU and RAM slots. Some chip must be bad or maybe some broken traces, I don't have enough knowledge and especially equipment to diagnosticate that.

And sadly also the P3B-F is in the same situation, I replaced all the caps in that too today but the the issue persist. It doesn't post until it "warms up".

I will probably give them away hoping that someone more expert can fix them. Me so sad.

warming up not only heals bad caps but also expands metal - you have a bad solder joint somewhere on p3bf
-can of compressed air and selectively cool areas of the board once its working to check when it crashes
-loupe/microscope and needle to poke at individual chip legs under magnification to see if they are floating in the air
-continue poking with your finger tests

I just noticed the prices for the typical coolant spray (non inflammable), 400ml can, I used to buy for ages have gone through the roof.
A 400ml can "Kontakt 75" costs nearly € 50.00 today. It used to be dead cheap not too long ago.
Troubleshooting a renitent mainboard can cost you a can easily.

whoa 😮 but then again its filled with R-134a and not just Co2/butane/propane, you could probably use it to top up your car AC 😀 and its made in Belgium. https://termopasty.pl/en/produkty/non-flammab … compressed-gas/ comes out slightly cheaper than Belgian one.

Ordinary 400ml butane/propane "compressed gas" can http://esperanza.pl/szczegolyProduktu.php?kat … 500&lan=4&cur=2 is still ~$2.
In a bind you can use $1 gas lighter refill can. Just remember to properly vent your workspace 😀