K3405 mosfet pdf

[cap noob duh]

[cap noob duh]

Re: K3405 mosfet pdf

so the one we'd need is: 2SK3405ZJ ?
- I'm assuming ZJ because the 'feet' are flat against surface.

question: Why don't all motherboards have polymer caps after the MOSFET?
- again, I assume that it's because the IC's don't have a high frequency output; that right?

thanks.


oooo to add, so we electrolyte folk are stuck having to buy old style psu's and not the fast switching ones, right?
- and ofcourse 100% polymer boards are stuck having to buy fast switching psu's...

[cap noob duh]

Re: K3405 mosfet pdf

found a nice writeup on VRD / VRM
https://cdn.badcaps-static.com/pdfs/...ea4b1f54da.pdf

[cap noob duh]

Re: K3405 mosfet pdf

Think I got it.
K3405 switching times mentioned in nano seconds, so it can 'move'.
and the Bulk capacitors are just to balance,
and there must be a bunch of surface mount polymer / high-frequency caps.

[cap noob duh]

Re: K3405 mosfet pdf

although, my last post does not explain all the 'big' caps on the dell.
- why so many?
- why not ANY on some of the new boards?

[370forlife]

Re: K3405 mosfet pdf

so the one we'd need is: 2SK3405ZJ ?
- I'm assuming ZJ because the 'feet' are flat against surface.

question: Why don't all motherboards have polymer caps after the MOSFET?
- again, I assume that it's because the IC's don't have a high frequency output; that right?

thanks.


oooo to add, so we electrolyte folk are stuck having to buy old style psu's and not the fast switching ones, right?
- and ofcourse 100% polymer boards are stuck having to buy fast switching psu's...

You can do a polymod to a motherboard where you replace the electrolytics with polymer capacitors. Generally you can use 1/2 the capacitance, as ESR is much more important in the VRM than bulk capacitance.

As for the psu's, I'm not sure what you mean. There is no such thing as a high frequency psu and a low frequency psu. Any ATX psu can be used with a ATX motherboard.

[cap noob duh]

Re: K3405 mosfet pdf

added:
Thanks for the reply, and I assume you mean the mod is for after cpu vrm only.

I think my confusion was with the electrolyte and polymer choices.
Is it possible that electrolyte replacements were offered at the start of this huge cap problem because polymers just weren't available? or, because no-one had yet done any polymer-replacement testing?

as for the frequency, i've poked around these pdf's:
http://www.irf.com/technical-info/papers.html
- in the DC-DC section
and found that they are/were playing with frequencies from 120k to 400k hz.

(my dell has one of the solder-in vrm's and probably has the 20v MOSFETs)
so from what I've read, I agree that, or I can understand that, electrolytic caps are pretty useless at those frequencies.

(plus I've read mention of fast switching psu's; but they may have been referring to the cpu's vrm, and I just misunderstood.)
- so then, what is the frequency they use with in an ATX psu?
- I figure it must be same principal as VRM or VRD; so they can get the many voltages.
plus, the electrolytics crap-out around or before 1khz(I may be off a bit on this.)

[370forlife]

Re: K3405 mosfet pdf

added:
Thanks for the reply, and I assume you mean the mod is for after cpu vrm only.

I think my confusion was with the electrolyte and polymer choices.
Is it possible that electrolyte replacements were offered at the start of this huge cap problem because polymers just weren't available? or, because no-one had yet done any polymer-replacement testing?

as for the frequency, i've poked around these pdf's:
http://www.irf.com/technical-info/papers.html
- in the DC-DC section


(plus I've read mention of fast switching psu's; but they may have been referring to the cpu's vrm, and I just misunderstood.)
- so then, what is the frequency they use with in an ATX psu?

AFAIK all atx psu's put out the same power, no different frequency's there. I believe the DC-DC applications you are looking at with different frequency's are something to do with the VRM's or the new DC-DC conversion psu's use, where the transformer only puts out 12v and the rectifiers only rectify 12v, and then convert some of that 12v to 5v and 3.3v. Much more efficient that way.

As for solid caps in older computers, they were MUCH more expensive then and were not manufactured in such quantities as today. There were some high end servers and desktops that used them, but other than that electrolytics were cheaper and in vaster quantities.

[cap noob duh]

Re: K3405 mosfet pdf

cool, ty.

[cap noob duh]

Re: K3405 mosfet pdf

the terminology i was missing was Linear vs. SMPS(switched).
my power supply:


it is SMPS with active PF. cool.
it kinda failed on the 140,000 hrs. life though. hehe

so, due to its being a switched psu, I could experiment with polymers in it, one day; if all is well understood by me.

[370forlife]

Re: K3405 mosfet pdf

Ahh, linear vs SMPS. Yes, computers use SMPS's as they are a much more compact design.

As for using polymers in them, it would require a total redesign of the PI filter. The problem is that the PI filter in the psu is "tuned" to the certain ESR of the caps used. Differing from this ESR causes ripple to increase, so you should always replace with caps closest to the original ESR ratings of the old caps. Another problem with the Polymer caps is the UF. Bulk capacitance is important in a psu, if you used low UF in it, the voltages may look fine on the multimeter, but it will actually fall dramatically when encountering even a light load so quickly a multimeter wouldn't pick it up. Basically, the lower UF caps you use, the less "umpf" that the psu would have when more juice was pulled from it.

If you want to mod a psu to do better on things like voltage regulation you can add higher capacitance capacitors and higher amperage rated schottky's. Ripple is mostly controlled by the ESR, so just keep it as close as possible to the ratings of the old caps.

[cap noob duh]

Re: K3405 mosfet pdf

oops and oh and thank you.

I just ordered MCZ and GC for the power supply...
I'll have to check specs now on the old ones.
- 5 of 11 were bust, so I had to remove them anyways.

i'm guessing the cable length from mobo to psu is enough to negate the effects of the better quality caps, as seen by the psu; re: the tuning.

so much to learn, so little beer...

I may be stuck with the caps i've ordered though, hope it don't go boom!

[cap noob duh]

Re: K3405 mosfet pdf

nah, the Fuhjyyu caps I can only find ESR for the TMR type, and most are TNR
http://www.fuhjyyu.com.tw/pro.htm

as for Rubycon, they have sent that section to Sun Elec. and they don't have any info.

Cross fingers and pray? lol yeeesh.

when you say ripple will increase, how bad are we talking?
the TMR 2200 uf @ 10v is listed at 0.039 ESR,
and the Samxon GC are listed @ 0.009 ESR,
and I bought Rubycon MCZ, oh man.
(ding, light came on,,, now I know what the poster meant when he said GA and HZ only after cpu vrm, as it can destablize an inferior psu.)
- ok ESR really matters...

[cap noob duh]

Re: K3405 mosfet pdf

I've read that leaving long leads on capacitors can increase ESR, some say it isn't enough to worry about, when trying to keep ESR low.
I have the space, would long leads help? insulating them is easy with bits of insulation from skinned 16 gauge wire.

[370forlife]

Re: K3405 mosfet pdf

Depends on what psu it is. The increase may not be much, but if the psu is already close to ATX spec for ripple with the original caps, you might want to try something else. I always use the 16v and 10v, 3300uf 10mm x 30mm samxon caps from topcat, they seem to be around the same ESR that is typically used in psu's.

[cap noob duh]

Re: K3405 mosfet pdf

psu is: http://www.vantecusa.com/front/product/pro_list/42

I'll do some for googling on it, and if it turns up as 'critical' on enough sites,(design spec links would be great) i'll be relying on local caps.(I live in a big city though)

I already have an account going with badcaps.net and shipping does not make a couple of caps worth it, from any supplier.
- also the local stuff is good enough at 0.045 ESR ratings.
I may also wait for the dell supply to show signs of strain before redoing it, if the ESR is really that critical. It's running so trying to avoid damage without proper oscilloscope setup may just make things bad. And again, high grade caps are not sounding like an issue.

thanks for all your help on this.

[370forlife]

Re: K3405 mosfet pdf

Ahh, the vantec is a CWT 420W. It should be fine ripple wise. I believe most of those did pretty good in testing.

As for dell psu's, depends on who the OEM is. If it is a delta, you probably won't ever need to recap it. A Hipro you might, they may have teapo or Arcon on the secondary. Lite-on probably won't ever need to be recapped either.

[cap noob duh]

Re: K3405 mosfet pdf

cool !

I'm much less concerned now, plus
did more reading at:
http://techtrain.microchip.com/webse.../Archived.aspx
- so far i've read, get low ESR and High ripple.(and thats' in the 'effects of components in smps' seminar.
(may not be referencing setup and test parameters yet.)

and found a list of supposed true manufacturers:
http://whirlpool.net.au/wiki/?tag=PSU_Manufacturers

the dell has Ltec caps and one green one i can't see the name of, but not sanyo looking.

added: don't feel like taking dell out again, to check the make just yet.

[370forlife]

Re: K3405 mosfet pdf

If it's ltec, it's delta. Don't worry about it. Delta's are by far one of the best psu manufacturer's out there (also the world's largest,) they make very good psu's.

[cap noob duh]

Re: K3405 mosfet pdf

double good news today, this is nice.
so I should be getting new caps soon for psu and mobo(original computer) and
with lower required specs for a psu, i can hold off on stocking for the dell psu and get them local.

it has been a pleasure, and thank you so much 370forlife !!

(i'm gonna keep reading and learning, it's the only thing i've been missing in my computer-help skills. After that it'll be on to mosfet checking, cause who knows, i may find a blown one in the original computer. hehe)