New Motherboard Doesn't like to Cold Boot

[Clocker]

I got sick of the fact that my Epox 8RDA+'s northbridge would get too hot with my case enclosed so I decided to try out an Abit NF7. I am really loving my NF7 (non-S) for many reasons right now. The sound is much better than on the two 8RDA+'s I had (sound was crackling on those two boards occasionally). Also, both of my Epox 8RDA+s could not run at 200mhz for more than about 1 hour while playing UT2003 with my PC's case buttoned up (even with a huge cooler on the northbridge). Running with the case open was no problem.

The NF7 kicks butt right out of the box with the 13 BIOS. I'm cruising 200mhz FSB all day long with no problems and I haven't even thought of going higher yet.....except....

THe system does not like to cold boot. If I shut it down and then go to restart it, it won't restart until I shut off the switch on the back of the PS for a few seconds and then turn it back on.

Basically, I hit the power button and absolutely NOTHING will happen. So, i have to switch the PS switch (on the PS) off and then back on and the system fans and drives will fire up after I press the PC power button again. But, that isn't the end of it..... The video still won't come up (and the machine isn't booting WinXP, either) until I press and hold the power button again. At that point the system will semi-shut down but come right back up without any of the fans or drives being able to spin down. It's kind of like a reset but it isn't because the power to drives and fans does momentarily go down and comes back up. Then the video comes up and I'm on my way into WinXP.

My first inclination is that this is a PS related problem but I'm doubting that because I have a 420W TTGI (Traditional Technology Group) (http://www.ttgitech.com/PowerSupply.htm) PS that has been serving me well, even for my DUAL Athlon XP system (Tiger MPX).

I didn't have any power related issues with the Epox nForce2 boards but they didn't have the 4-prong ATX power connector on them like the NF7 does, either. Any advice as to what is going on with my situation? Everything is fine unless I shut the system down and try to restart it. It runs like a raped ape!! (sorry, tea).

Thanks,
Clocker

 

[Groltz]

What are your actual voltages on the +12 and +5 rails as measured by a DMM?

 

[Mercutio]

Does the problem still happen if you've got your collection of drives and case fans plugged in?

One of the students in my computer building class bought an 8RDA+. You certainly are right about that northbridge heat sink... it's way too small (his felt warm after just booting to BIOS to set his 2600+ up). Asus and Gigabyte KT400 boards both have bigger heatsinks.

Still, what are you gonna do with your 8RDA+?

(Yes, I'm tempted to buy one. I want to play with the nifty sound features they have).

 

[zx]

I had similar problems with my Abit KT7A. Sometimes, when i'd cold boot, the system would start, but there was no video, no beep, no sign of life. I had to stop the system and start it again using the PSU switch. It was much less annoying than what you are exprencing. It happened about 5% of the time.

Also, I changed my PSU to get a better one and it continued doing the same thing. Later, I sold that mobo to a friend, and he had the same problem with it! So I suspect that the problem was the motherboard.

It would be interesting to try another PSU if you have one. However, I suspect that it's the motherboard.

 

[Clocker]

Steve-
Here are my voltages:

Merc-
Here are some pix if you are interested:

http://www.wideopenwest.com/~kharper/8rda_1.jpg

http://www.wideopenwest.com/~kharper/8rda_2.jpg

 

[Mercutio]

I am interested, but I accidently kinda sort bought a Sony DRU500AX and a bunch of different media today (more on that later) - it looked so lonely in that display case that I just had to bring it home - so that basically blew my upgrade budget for a few days.

That is an awesome modification, though.

Also I am setting a precedent of giving something that says nvidia a fair shake.

 

[Clocker]

That is an awesome modification, though.

Yeah, it's amazing what a little super-glue and Arctic SIlver 3 can do. I was so pissed off at that thing (I would not accept anything less than 200mhz FSB with my case bundled up) and I had accidentally left the nice skinny tie-wraps at work so I said F*CK IT IT"S JUST A MOTHERBOARD and just glued the damn thing on.....I'm not sure if it will ever come off. I'm sure it will be great at at least 166mhz (the little baby heatsink could do that) but, for some reason, this thing needs relatively cool (84F or so or less) air to run decently at 200mhz. Maybe all it will do is 199mhz all bundled up but I have never tried anything greater than 166 or less than 200. That Ttake fan is pretty quiet at 7volts BTW.

Come on, you know you want it and that I'll give you a good deal.... 8)

I may have to wait though...if I need to RMA this Abit board for the starting problem, I'll need something to use in the meantime....

C

 

[Clocker]

Well....unplugging my other non-necessary devices does not change the problem. I had a cheapie PowerMax PSU that I also tried. It is supposidy 400W but the whole system would just freeze during post with that thing. However, it would at least wake the system up.

I've got a BestBuy gift card that has been burning a hole in my pocket and a 10% off coupon. Maybe I'll give an Antec TruePower 430 a try to see what happens. I guess I can always return it if it doesn't help out.

Anybody have any other ideas...?

C

 

[Mercutio]

Do you get a normal reading from a VMM when the machine is plugged in but "cold" (check on the ATX power header)?

 

[Groltz]

Also I am setting a precedent of giving something that says nvidia a fair shake.

I know you are contemplating a Nforce2, but fair's fair, Mercutio...My next video card upgrade will likely be a R350 (Maybe R400 depending if my patience holds out) based ATI. 8)

 

[Clocker]

OK now...let's not get all lovey-dovy in my tech support thread...

 

[Groltz]

Steve-
Here are my voltages:

C, I meant with a Digital MultiMeter. The voltage reported by software utilities can be far from accurate. At best, they can show a rise or fall from a known benchmark for reference.

Try taking a reading off of a PSU power connector at boot and see if you can find an abnormal reading. This is easiest done with a DMM that can record maximum/minimum with fast refresh.

For instance, the 400W Sparkle I have was low on both the +5 and +12 rails out-of-the-box. I took the measurements with both CD drives spun-up and the CPU running at 100% to try and obtain a worst-case-scenario. MBM5 reported a >.25 volt difference (higher) than my multimeter on both the +5 and +12 volt rails. I aligned the PSU using the multimeter's readings.

Readings can also be taken off a PSU that isn't installed yet. If pins 13 and 14 of an PSU ATX connector are shorted together, the effect is the same as hitting a computer's "ON" button....you'll have power to the harness connectors. Readings gathered in this way may not be that good, however, since there is zero load on the PSU. This method is best for getting the PSU roughly dialed-in before installing in the case. (If necessary, that is.)

All that aside...I am wondering if the PSU is really the problem. At this point I can't think of anything else. :dunno:

OK now...let's not get all lovey-dovy in my tech support thread...

Settle down, toughguy.

 

[Clocker]

Bummer...I don't have a DMM. Guess I'll just have to try another PS...

C

 

[blakerwry]

you know clocker, your problems remind me of my Abit ka7 .. you know, the one with the bad capacitors... The 1st problem was that it would sometimes not cold boot....this was maybe 5% of the time... then it became ever time... I could warm boot.. but otherwise I'd have to turn it off (i think at the PSU) to get it going again...

after 1 year of closet time the situation reversed.. i could cold boot (only every other time)but couldn't warm boot.... just an odd little problem...

anyway... since your PSU has worked with similar boards and *should* work with this one, I think that your mobo is the problem... and that it might be a capacitor problem... best just return the board to the retailer while you can.

 

[Mercutio]

And you call yourself an engineer! Honestly. Kids these days. *I* was going to suggest you break out an oscilliscope and look to see if you were getting a nice, flat output from your PSU (something you can't always tell from a voltmeter), but... there goes that idea.

Honestly, Groltz, my only reasons for contemplating Clocker's epox board is the giant frickin' heatsink he glued to it and the happy thoughts of always-dolby-digital output to my Logitech Z640s (or, if it's really, truly good, to my HTPC setup).

Satan must be ice-staking tonight.

 

[Groltz]

Honestly, Groltz, my only reasons for contemplating Clocker's epox board is the giant frickin' heatsink he glued to it and the happy thoughts of always-dolby-digital output to my Logitech Z640s (or, if it's really, truly good, to my HTPC setup).

I know, but the irony still struck me as kind of funny.

Satan must be ice-staking tonight.

Sounds like something Greg would say. :eekers:

 

[Clocker]

you know clocker, your problems remind me of my Abit ka7 .. you know, the one with the bad capacitors... The 1st problem was that it would sometimes not cold boot....this was maybe 5% of the time... then it became ever time... I could warm boot.. but otherwise I'd have to turn it off (i think at the PSU) to get it going again...

after 1 year of closet time the situation reversed.. i could cold boot (only every other time)but couldn't warm boot.... just an odd little problem...

anyway... since your PSU has worked with similar boards and *should* work with this one, I think that your mobo is the problem... and that it might be a capacitor problem... best just return the board to the retailer while you can.

Blake-
Right now I kind of doubt it is a capacitor problem. You might find THIS STORY interesting...at least Abit acknowledged the problem and I believe that have switched to using Japanese Cap. manufacturers (at least that's what I gleaned from the article). My dad's KA7 also had the cap. problem...too bad it was out of warranty. It was cheaper to buy a new board than ship it around and wait, anyway.

Merc-
I'm a ME Can't you tell by my big friggin' heatsink!

 

[Mercutio]

Eh. Never stopped my father (PhD EE). That man is dangerous with a hammer.

 

[Clocker]

Well, I am now pretty convinced this is a motherboard problem.

I connected an Antec TruePower 430W PS to the board and it did the same thing as when I tried my cheapie 400W PowerMax....the video signal would initialize but the system would just freeze after POSTing with the message on the screen stating the memory speed (ie. DDR266 or whatever).

Anyway, I pulled the Antec PS and reinstalled my nice TTGI PS that the system will at least boot with (after I fiddle around with the power switch a couple times or so).

I pulled the Radeon 9500 128MB card I was using and, to my suprise, the AGP slot won't accept my old spare AGP cards (Diamond Riva128 4MB or a Trident Blade3D 9880 8MB). THe slot is keyed differently so I could not try a different AGP card.

I had a REALLY old Cirrus Logic 512K PCI video card. With that in place of the Radeon 9500, the system would never just not turn on at all like it has been doing with the 9500 installed but it's only about a 50/50 chance that I will actually get a video signal on the monitor. THe video card is pretty old but I did manage to get into Windows once. Damn it looked like sheot!

There is another guy with a similar problem in the Abit forums here:

http://forum.abit-usa.com/showthread.php?s=&postid=24904

so I guess I'll see if the RMA he did solves his problem....

C

 

[Mercutio]

At least now you know what the "Init Display first" BIOS setting does.

I know I've mentioned this already this week, but a $5 Matrox Millenium is supported by everything under the sun and distinctly does not look like sheot.
They're awesome test cards.

 

[Clocker]

At least now you know what the "Init Display first" BIOS setting does.

I know I've mentioned this already this week, but a $5 Matrox Millenium is supported by everything under the sun and distinctly does not look like sheot.
They're awesome test cards.

Will it fit in this new fangled AGP 8x slot?

C

 

[Mercutio]

Nope. But it makes nice for PCI.

I run into the same thing with my KT400 boards... the G200 I'd been using for a test card doesn't work in those.

 

[The JoJo]

Merc might not Nvidia, and I DEFFINITELY do not like Abit.

The reason? Too many problems like the one Clocker is experiencing. Just RMA the mobo and hope you'll get a better one.

 

[The JoJo]

Merc might not like....

 

[time]

I agree with The Jojo ...

 

[Mercutio]

What's this "might" stuff? I definately don't like nvidia. I definately don't like Abit, either.

Abit and leaky caps might as well be synonymous, but in the last couple of years I've run into dead AGP slots, motherboards with no working DIMM (or RIMM, in one case) slots and a motherboard that won't detect a hard disk, but work fine when you enter CHS data (that was my favorite).

I don't see Abit boards that often since I stopped selling them, fortunately.

 

[Clocker]

Merc-
Regarding leaky caps, you might find this interesting:

http://www.spectrum.ieee.org/WEBONLY/resource/feb03/ncap.html

C

 

[Cliptin]

Maybe the 200MHz problem is endemic!

Conclusion

Needless to say, the Athlon XP is one awesome overclocking CPU! If you are wondering where the 200MHz FSB benchmarks are, there aren't any. I've had some serious issues with the NF7-S v1.0 in getting it over 180/190MHz bus, so 166MHz bus is as far as I'm taking the 2400+. Sorry, maybe next time... Please visit our Forums for more on the 2700+ & NF7-S issue.

http://www.cluboc.net/reviews/cpu/amd/xp2400/p5.htm

 

[Clocker]

Clip-

I'm having hte boot issue overclock or not. 200MHZ FSB is really no problem on this board it seems though with Vdd set to 1.7V....

C

 

[Handruin]

Clip-

I'm having hte boot issue overclock or not. 200MHZ FSB is really no problem on this board it seems though with Vdd set to 1.7V....

C

Do you remember if the problem existed before you began to over clock the system. Point being, do you think this could be the result of over clocking?

 

[Clocker]

No, I'm not really sure. It could have been the result of a BIOS flash for all I can remember....

C

 

[Cliptin]

Yeah, it was more for general info.

 

[Newtun]

Maybe the 200MHz problem is endemic!

Interesting Sudhian article, Opening the nForce Floodgates--Modding the AthlonXP for a 400 MHz FSB.: ". . . XtremeSystems . . . forum readers . . . have found a solution to this problem. . . . The problem, apparently, is that the nForce2 (and possibly the KT400 as well) destabilizes when the front-side bus is overclocked using a 333 MHz FSB CPU. The solution (which we've tested) is to cut one of the L12 bridges on the CPU -- the exact one is pictured below. Cutting this particular bridge forces a motherboard to recognize a CPU as a 266 MHz part, rather than a 333. . . . Although the procedure had the potential to kill a processor (and thereby kill a $300 investment) we went ahead and cut the bridge. . . . We've gone back and tested multiple boards and, in every case, seen dramatic stability jumps at high FSB's. Our Chaintech 7NJS (which formerly would die around the 190 mark) now runs at 200 MHz rock-solid with very low latencies."

 

[Bartender]

Thanks for the helpful information Newtun.

 

[blakerwry]

I saw that too.. but damn if I'm going to cut apart my 333mhz FSB CPU!

 

[Newtun]

I saw that too.. but damn if I'm going to cut apart my 333mhz FSB CPU!

Dayem, they just did it to a Barton!

 

[Groltz]

.

Severing bridges on Athlon XP CPU's is easy.

Here are two posts I made on the topic in a different forum.

Originally posted by Nico67
Just wondering how you cut the bridges. I'm hanging out for a NF7-S to get NF2 with 5 bit multiplier select, but I'd like to see what my 8K3A+ would do also.

I use a Xacto knife. Not held to cut as a knife would, but with the blade turned 90° so tip is more like a scraper. Very slowly scratch down a tiny bit at a time until the copper trace is revealed, then scratch through it. Make sure the trace is removed in a wide enough section as to make sure there will be no electrical continuity. As a bonus, the pit that has been dug with the Xacto won't be deep enough to hit that conductive floor that you see in the factory laser pits. If I want to rebridge the trace, nonconductive filler isn't necessary.

I might mention that I do this work under a 20x/40x stereoscopic dissection microscope. There is no way I would try it with any less magnification than that.

Originally posted by Lazgoat
One slip of the Xacto blade........dead CPU?

No...A scratch in the CPU substrate.
You don't bear down hard enough on the blade to slip anyway. All that is required is to make small, light scratches over the trace until it is severed. Opening the traces by that method is much easier than closing them by means of pit-filler and conductive grease/paint. Assuming you have enough magnification to see what you're doing, anyway.

 

[Clocker]

Well I finially broke down and bought an Aopen AK79D-400VN to replace my POS NF7 that won't boot or shut down correctly. I tried the RMA with Abit for the original bad NF7 and the RMA board I got from them had the exact same problems (but only seemed to get worse over time). What's really errie is that the serial numbers were only about 100-150 apart!

I replaced NF7 with the Aopen and everything is perfect. While I was in there I replaced my Vantec AeroFlow/TMD fan heatsink with a Falcon Rock and threw in an extra 256MB of Corsair for a grand total of 768MB.

For $71 after $10 rebate at Newegg you really can't beat this NF2 Ultra 400 board. It is just a great board with excellent support/BIOS updates from Aopen. It's the 2nd one I have installed and I just love it.

I will be RMAing the NF7 AGAIN with Abit so I can sell the replacement as new on eBay. Sigh..what a friggin waste of time.

C