planning to mod Tyan Tiger MP mobo, please review my plans

[Jake the Dog]

as some of you might know, i've purchased James' Tyan Tiger MP mobo, it should arrive any day. anyhoo, in doing a some research into these boards is seems they can have one major issue. the ATX power connector that Tyan chose to use on this board apparently isn't up to the job. this board draws a lot of current through the 5V rail and has in the many cases i found of this issue the result was a melted the power connecter. in almost all cases the PSU died and more often than not the mobo as well. i asked James what PSU he was running and thought i might get the same but he must be busy as i've not heard from him for a while. in any case, i've decided to make sure this issue doesn't happen to me. i'm planning to use a powerful PSU and mod the power connector on the mobo.

tyan has apparently stated that the 5V rail needs a 30A supply which is a huge amount across 5V (i'm somewhat doubtful) but in any case even if it's 2/3's of that i can see how this could cause problems. i've purchased a Codegen 450W PSU which is rated at 40A on its 5V rail. i haven't heard of Codegen before but i'm assured they are a decent PSU and they are listed on AMD's site as a maker of approved PSU's.

the mod involves a molex power splitter. i plan to strip one side of the splitter and solder it to the back side of the power connector on the mobo. the red 5V wire should go to to pins 19 & 20, the yellow 12V wire to pin 10 and both black ground wires across pins 16, 16 & 17.

below is a pin diagramme of an ATX power connector.

seems like a safe solution to me but i'd like to get some opinons first. any comments?

cheers,

tim

 

[blakerwry]

what about the 5v pins 4 and 6?

Personally, I think you're probably better off leaving it alone., but if you ahve the knowledge and expertise you can try.

I'm kind of wondering if the problem is related more towards the Tyan power connector or the power connector of the PSU. It's quite possible that the Tyan uses alot of 5v voltage and the cheaper made PSU's can't hack it. I would try to look for some kind of correlation in the cases you've seen.

You couls always try to reduce the 5v draw. I assume that your PCI/AGP devices are going to be the biggest leechers of your 5v power...

 

[Cliptin]

May the Force be with you.

 

[Jake the Dog]

what about the 5v pins 4 and 6?

being that they share a common bus, i don't think it should matter which pins i connect to. if you know this to be incorrect please tell me! since
most times, pins 19 & 20 were the ones that were melted 1st i thought they would be the best candidates to solder to.

I'm kind of wondering if the problem is related more towards the Tyan power connector or the power connector of the PSU. It's quite possible that the Tyan uses alot of 5v voltage and the cheaper made PSU's can't hack it. I would try to look for some kind of correlation in the cases you've seen.

that's exactly it blakerwry. if you re-read my post you'll see that with the investigation i’ve already done, i've already established that as the problem is that the power connecter is to small for the current draw requirements of the board. the contact area between the 5V power plug and socket ito small to carry a large current and this heats the pin at the point of contact. the resultant heat generated melts the socket and causes all sorts of nasty problems. the solution, as i see it then, must be to increase the connector size which can be done a few ways, the most obvious two being a) replace the power connector for a larger, up rated item or b) add another power connector. i'm going with b) since it's the easiest solution although certainly less elegant.

You couls always try to reduce the 5v draw. I assume that your PCI/AGP devices are going to be the biggest leechers of your 5v power...

you're right, add on cards are going to draw lot of 5V but to reduce that i have to reduce the amount of cards or change my video card from a GF4/4200 to a lesser cars - neither are an option unfortunately.

 

[James]

I'm just using an Enermax Whisper 350W.

Before that it worked fine on an AOpen/Sparkle/Powerman 300W one.

Tim, all being well (my girlfriend being organised, basically) the board went today. Later than I thought, but last week was absolutely flat out and I never got home before about 10pm. Sorry.

 

[Jake the Dog]

no worries James. i was never bothered

 

[Cliptin]

Will the MB fit in the case with stuff soldered to the backside?

 

[Tea]

Ahhha! You have not seen Jake's skill with tools, Cliptin. In particular, his deft touch with soldering iron, pliers, Dremel, tinsnips, spray gun, rivet gun, hammer, axe, and tactical nuclear weapon. If it don't fit, don't worry: Jake will make it fit. :wink:

 

[Pradeep]

I wouldn't worry about the mod Jake. Just plug it in and turn it on. Just avoid those dreadful PowerMax/Leadman POS PSU's, had a 500W blow up on me with my TigerMP. The AOpen 300W worked fine.

 

[Jake the Dog]

too late mate! the mod seems to have worked well.

CPU temps are a bit of a concern... with both folding, CPU #1 sits at 51C and CPU #2 sits between 59-63C...

 

[James]

I might not have done the most brilliant thermal paste-ing job, Tim. You might want to check.

 

[Jake the Dog]

thanks for the heads-up James. I ended changing cases and reseating the HSF using Artic Silver thermal paste. temps are now at 48-50C & 52-54C

 

[James]

Glad to hear it!

How's the dual CPU goodness?

 

[Jake the Dog]

i'm very happy it except for one issue. it doesn't like restarting... did you have this prob James? it stalls at the post screen. i have to physically turn it off, and keep it off for a few minutes before it will boot into XP.

 

[James]

Nope, never had that problem. Make sure the board is properly grounded and your motherboard standoffs are properly fitted.

Also make sure it's not hanging because the CPU/chassis fans aren't fast enough or whatever.

You can also try going with the latest released BIOS as opposed to the beta I had on there, although I never had any problems with it and it fixes a cosmetic issue in Win2K/XP.

I did find it was slow to POST when I first bought it - it used to hang for (literally) a minute while detecting the secondary slave IDE. That turned out to be because the secondary IDE cable wasn't the greatest (it was a rounded one) and once I replaced it all was well.

 

[Jake the Dog]

hmm IDE cables... i think you're onto something there. the point at which it stalls is just as should start reading off the drive to begin booting. initially i couldn't get a drive to be recognized(*) i tried a high-quality shielded rounded cable and it worked. perhaps these boards are very sensitive less than perfect IDE connectors. i'll reseat them and try again.





* z=s: is that better Tea?

 

[James]

If you leave it for 60 seconds you'll probably find it boots anyway - at least it did with me.