EP 61BXA-M sucess story, but...

[Mikeluz]

Hi!

I own a EPoX EP-61BXA-M, that used to have bad caps. Actually, the former owner, had a little trouble with this board, on the VRM, near the CPU slot1... a badcap vented and as a "bonus" a MOSFET near the badcap burned out, rendering the CPU -- a P2 350MHz -- dead.

The badcap was a TAYEH 10v 1000uF, and the MOSFET was a D50N03A. I canibalized a 6.3v 1200uF to put on the place of the 10v 1000uf (didn't have any 10v 1000uF, only a 10v 2200uF, but I didn't tried this one until a later time, further on this post). As I said canibalized a 6.3v 1200uF to put on the place of the 10v 1000uf because this capacitor has the same dimensions as the TAYEH one. I think I can do this, but, any enlightening about the is switch will be welcome. The board works well with this cap, so I guess, I did right, but any warnings/thoughts about long term effects that I could I expect from this replacement, will be much welcome

About the MOSFET, I went to the local electronics shop and I was told that a replacement for the D50N03A is a Philips BUK455 100A, which according to it's datasheet is a N channel, 100V, 26A(max) MOSFET . I was fooled by the MOSFET name BUK455 100A, as I thought this was a 100A MOSFET, but actually it's only 26A(max)!

Well I've replaced the burned MOSFET, and the computer work ok, except, that when I run a P3 600 on it, the MOSFET heats a lot, and eventually when I run P95 the MOSFET actually smokes (this is before I shut down the PSU, to prevent this MOSFET to burn out!) I was lead to the conclusion that something was not right! Further searching on the web to a datasheet of the D50N03A leads to nothing! Except a thread on this forum, where, according to kc8adu
Super Moderator, on this thread https://www.badcaps.net/forum/showth...96&mode=linear, on the 7th post said that the D50N03A (original MOSFET), is a 30v 50A N-channel MOSFET.

So, I guess I was misinformed about the D50N03A replacement and as a consequence I'm running a BUK455 100A (100v, 26A, N-ch MOSFET) bove it's capabilities, when "Primming95" (P95 drains the max current to the CPU, and I think that slot1 P3 600 CPU drains up to 30A, well above the BUK455 100A max. 28A limit, and below the D50N03A 50A limit). I think I should replace the BUK455 100A (26A Idrain) to another N ch MOSFET with a 50A Idrain (according to kc8adu Super Moderator. I think this is MOSFET fault, as the overheating only happens when the CPU load is high, this is when the current drained is high!

What do you think?

Any help from him or any other members will be much appreciated! As a side note I used to think that the overheating of the BUK455 100A MOSFET was capacitor fault, so I switched the 6.3V 1200uF to a 10V 2200V, but this helped nothing. So, after I read the above thread, I think the MOSFET ovearheating problem is due to the 26A limitation of the BUK 455 100A which does not comply with the original (D50N03A) MOSFET specs which were (according to kc8adu Super Moderator), 30V, 50A.

Sorry for the long post, I think I've explained all that I've done, but if any doubt arises, please feel free to ask me!

TYA for any replies.

[kc8adu]

Re: EP 61BXA-M sucess story, but...

i would replace the mosfet with the proper type and finish recapping.
if one cap went all others like it are near death too.this can also cause mosfet overheating.

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Hi! Thanks for the reply!

I'll do that recapping, as soon as I get replacements here at Portugal! Back here is difficult to find Low ESR caps, so I've got the order from outside Portugal.

I've searched further on the web about the original MOSFET and I came across these links (by the order I've opened them

http://www.icpart.com/SALESDETAIL/D50N03A.html

clicked on the PDF icon lead us to:

http://www.alldatasheet.com/view.jsp?Searchword=D50N03A

clicked on the first PDF entry lead us to the PDF itself:

http://www.alldatasheet.com/datashee...50N03-06P.html

Now, what should I do with this info? Shoul I measure the voltages that I came across when P95? Then go back to the characteristics curve and find out which Idrain is drained? I'm not sure about the way to go. Any help would be appreciated on the MOSFET question.

TY for the answer!
Mikeluz

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

My principal problem is not knowing about the charateristics I should look for on the MOSFET, how should I get a replacement? The BUK455100A that I was told that I could replace the D50N30A original MOSFET, looks like does not meet the original specs of the board, and I can't find any conclusive datasheet about the original MOSFET. I'm not sure of what could happen if I try one of those Vishay ones, because they have different Idrains @ the same VDS, VGS, so I guess that I could fry the CPU if I feed it with current above its requirements. So I'm not sure about what the "correct type" I should search for Any enlight about this, from you or another member will be much appreciated.

[Rainbow]

Re: EP 61BXA-M sucess story, but...

Get two logic level MOSFETs (if there are two on the board) that have lower-or-equal Rds(on) and higher-or-equal Id than the original ones. Then replace both MOSFETs on the board so they're same. Then replace all the TAYEH caps to prevent it from blowing again.
I never buy replacement MOSFETs as I have enough dead boards with proper types

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Get two logic level MOSFETs (if there are two on the board) that have lower-or-equal Rds(on) and higher-or-equal Id than the original ones. Then replace both MOSFETs on the board so they're same. Then replace all the TAYEH caps to prevent it from blowing again.
I never buy replacement MOSFETs as I have enough dead boards with proper types

Thank you for the reply! Yes, there are two MOSFET's near each other, located near the CPU slot1. (attached is a aimage showing the relevant zone, at left of the Slot1)

But, I don't know *for sure* which are the specs of the original MOSFET's, that's is my problem!

If I don't know *for sure* those specs, then I can't search for a proper replacement(s). (According to Kc8adu, the D50N30A is 30v, 50A, but I want to know *for sure* if that info is accurate! I've searched the web and I've seen D50N03 (not A), that have different Id's ranging from 23A to 100A.

I must reinforce this thought: not knowing *for sure* the D50N03A specs I can't go any further!

QUESTION: That is, if I replace the MOSFET with a 50A parts, can I damage the CPU? I don't know how the VRM works, so I'm a bit afraid of doing this. Does the VRM adjusts itself to the MOSFET's?


Any link pointing me to the datasheet or the specs of the D50N03A (if any of those exists), would be nice.

TY,
LP

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Get two logic level MOSFETs (if there are two on the board) that have lower-or-equal Rds(on) and higher-or-equal Id than the original ones. Then replace both MOSFETs on the board so they're same. Then replace all the TAYEH caps to prevent it from blowing again.
I never buy replacement MOSFETs as I have enough dead boards with proper types

Care to point me to some of those logic levels MOSFET's, that are compatible to these D50N03A, please? At this point, I can't continue further without any enlight. I don't know how a VRM works, neither how MOSFET works on the VRM, how it figures what is the voltage / current that it has to deliver to the CPU, etc. Is there any tutorial where I can learn how does a VRM work?

TY,
Mikeluz

[gonzo0815]

Re: EP 61BXA-M sucess story, but...

Best thing for you, look for the VRM controler chip on your board. Than google a little and look if you can find a appsheet or datasheet for the controler. In most cases, you will find a reddily desinged circuit for your exact purpose, usually with all the details abouth the mosfet`s. But i am shure, that any mosfet with low enough RDS on (something under 0,03 Ohm) and logiclevel compatible will suit your needs. I doubth, that there would be more than 12 to 20A supplied to the P3. If you wana be shure, take an 30A rated ones.

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Best thing for you, look for the VRM controler chip on your board. Than google a little and look if you can find a appsheet or datasheet for the controler. In most cases, you will find a reddily desinged circuit for your exact purpose, usually with all the details abouth the mosfet`s. But i am shure, that any mosfet with low enough RDS on (something under 0,03 Ohm) and logiclevel compatible will suit your needs. I doubth, that there would be more than 12 to 20A supplied to the P3. If you wana be shure, take an 30A rated ones.

Thank you for the reply gonzo0815 The chip you are talking about, I supose it's also called a "driver chip"? And usually there's one near the MOSFET to drive them though I don't know how that driving it's done. (I'd like to know how that driving is done, as a curiosity.) The IC chip on the picture above, near the top MOSFET, is a ALPHA AS1664S (also written is 9740 --but i think this 9740 is the manufacturer year/week, so it's not relevant), so I'll google it as you advice me!

About the Philips BUK455100A, according to it's datasheet (, it's a 26A (max) part, so I suppose it should handle the 20A current drain of the P3 (I agree with you on the 12~20A max.) Thought, I'm not sure if this part ( Philips BUK455100A), is a logic level MOSFET that you guys gonzo0815 & Rainbow are talking about?


TY for the replies!
Mikeluz

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

The above link to the BUK455 100A datasheet is not working now, so here's another:

[Rainbow]

Re: EP 61BXA-M sucess story, but...

That BUK455-100A is not suitable for CPU VRM use - Id is too low, Rds too high and it's not even logic level!

[gonzo0815]

Re: EP 61BXA-M sucess story, but...

Not shure, at least it will conduct at 5v, but RDSon will probably be even higher than stated in the datasheet. Certainely, not the ideal mosfet for that purpose.

[Rainbow]

Re: EP 61BXA-M sucess story, but...

Exactly - if the VRM controller requires logic level MOSFETs, the Rds(on) will be awfully high and it will smoke and die.

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Hi Rainbow and gonzo0815!

Well, I've measured temperatures of the BU455during P95 and the results are:

T _P2 350@434 (FSB=124MHz) _ = 95ºC (max)
T _P3 450@554 (FSB=124MHz) _ = 105ºC (and raising)

so I've decided to limit the CPU to the P2350@ 434, but I continue t think that 95ºC @ room temperature of about 18ºC is way too much! On Summer ambient temeratures go up to 38ºC so BUK455 will probably not survive! Anyway on the computer I'm writing now, a P4 1800A, when Oc'ed to 2.4 (FSB=133MHz) and Priming the temperatures of one of the two "MOSFET drivers also raise to about 95ºC), so I think this is a normal temperature under heavy load, like "Priming95"?

Returning to the main subject, the BUK455 MOSFET. What do you guys think that would be a proper replacement for the D50N30A MOSFET that had been replaced for this BUK455? For instance what Phillips type MOSFET would be the one to look for?

TY for your repplies guys!
Mikeluz

P.S. I have an school exam during the next 2 days, so I'll probably won't feedback during these 2 days. Sorry!

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Hi Rainbow and gonzo0815!

Well, I've measured temperatures of the BUK455-100Aduring Prime95 and the results are:

T _P2 350@434 (FSB=124MHz) _ = 95ºC (max)
T _P3 450@554 (FSB=124MHz) _ = 105ºC (and raising!!!)

so I've decided to limit the CPU to the P2350@ 434, but I continue to think that 95ºC @ room temperature of about 18ºC is way too much! On Summer ambient temperatures here at Portugal go as high as 38ºC so BUK455 will probably fry! (As a side note, on the computer I'm writing now, a P4 1800A, when OC'ed to 2.4MHz (FSB=133MHz) and Priming95 the temperatures of one of the two "MOSFET drivers also raise to about 95ºC, so I think this is a normal temperature under heavy load, like "Priming95"? Maybe the mobo's manufacturer's trend is similar to the trend seen on low end (pricewise) PSU's... to use components above its rated specs...leading to a shorter life of the components and as a consequence of overheat follows that the PSU dies...selling more and more often to increase profit...at expense of quality and longevity of products...)

Returning to the main subject, the BUK455 MOSFET. What do you guys think that would be a proper replacement for the D50N30A MOSFET that had been replaced for this BUK455? For instance what Phillips type MOSFET would be the one to look for?

I recall that the original MOSFET (that is, the good one that has left on the motherboard) has voltages VGS~=5V and VDS=~4V, but the BUK455-100A has voltages VGS~=3.8V and VDS~=3V.

TY for your repplies guys!
Mikeluz

P.S. I have an school exam during the next 2 days, so I'll probably won't feedback during these 2 days. Sorry!

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

That BUK455-100A is not suitable for CPU VRM use - Id is too low, Rds too high and it's not even logic level!

What Id value should I look for? 30A? 40A? It's a Slot 1, P3 600 (Katmai core) that I'll run on it.

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Not shure, at least it will conduct at 5v, but RDSon will probably be even higher than stated in the datasheet. Certainely, not the ideal mosfet for that purpose.

What RDSon value should I look for?

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Exactly - if the VRM controller requires logic level MOSFETs, the Rds(on) will be awfully high and it will smoke and die.

How do I determine if I can use logic level MOSFET's? (I can't find the VRM controller datasheet anywhere in the www). OR, in another words, can I use, by default, logic level MOSFET with VRM's? IF the answer is Yes, then I'll buy the couple of MOSFET needed, based on the Id & RDSon that you advice me to. If the answer is NO, what MOSFET's do you advice me to buy (any example)?

TIA,
Mikeluz

[Rainbow]

Re: EP 61BXA-M sucess story, but...

Look, for example, at CET CEB6030L - these MOSFETs were used on many Slot 1 boards - you need something like that:
logic level, Rds(on) = 20mohm@Vgs=4.5V, Id = 52A continuous, 156A pulsed


Yes, you can use logic level mosfets for all motherboard VRMs.

[Mikeluz]

Re: EP 61BXA-M sucess story, but...

Look, for example, at CET CEB6030L - these MOSFETs were used on many Slot 1 boards - you need something like that:
logic level, Rds(on) = 20mohm@Vgs=4.5V, Id = 52A continuous, 156A pulsed


Yes, you can use logic level mosfets for all motherboard VRMs.

Thank you Rainbow, for the input!

I'll post the results here after I replace both MOSFET's, after school exams

Best luck for you!
Mikeluz