Why does my machine not boot cold?

Post Reply New Topic

Well, the drives still did not register the first time up with the new power supply, and neither was my CPU fan running ... but then everything worked correctly after a simple "warm reset". So, maybe I have something set wrong in my BIOS or maybe (as I have long suspected anyway) I just have a buggy BIOS. Either way, I now have a new power supply that is 430W rather than 300 and with more taps available so I no longer have to piggyback-feed my case fans.

Some BIOS won't boot if the CPU fan isn't running. Have you figured out why the fan wasn't running the first time?

No, and please forgive me for missing your post a few days ago.

I am glad to have a new power supply with SATA support, but that has not solved my problem ... and here is what is going on:

When started cold, my machine goes through POST and everything looks okay, but the CPU fan does not start and I get an error message saying "No active partition found." At that point, I use the reset button to come around again and go into BIOS and detect drives and "Save Settings" ... and then the POST report shows my PATA drives with a "%" in place of a "-" in their IDs and I get that same "No active partition" error. However, doing just a reset at that point brings things around a third time and all is well. So, my fan and drives are not seen the first time, then the fan shows up and starts the second time, then my BIOS finally gets everything right the third time around.

I have just downloaded the latest Biostar BIOS utility and plan to try that to see whether that might fix things, but I have also placed a bid on a replacement mobo.

Is the mobo my problem? When I detect drives in BIOS, the PATAs get detected quickly and easily, but the SATAs take longer and the second one takes just a little longer than the first, and then sometimes the "Auto" option for one of the SATAs somehow gets changed to "Large".

Well, that was a new experience for me: The BIOS update ran to 51% and stopped ... and that is the end of that! So, it looks like I will now be pulling that chip and sending it out to be re-flashed or replaced. I have never had that problem before on any machine.

Question: Will replacing the mobo with a slightly newer version of the same Biostar model send Windows into having fits about the change? Windows 7 once had to be re-activiated simply because I had added a PCI card for more USB ports, and now I suspect possibly everything from 2K on up is not going to run if I replace the mobo.

I have never seen it either. Some mb has a second bios that you can switch to in such situation.
Do you know if the fan was working when you are flashing the BIOS? If it was, then the problem may be the bios chip. It'll also explain why the settings aren't stored properly.


I haven't had the chance to test this myself. According to M$, Win7 is is able to handle mb upgrade without issue (except the need of re-activation). 2000/XP should be fine if it's the same model. BIOS version shouldn't matter to Windows, but it may to the drivers you already installed. For the SATA controllers, make sure it's running under the same mode (AHCI or ATA).

You shouldn't have an issue reactivating Windows if you change motherboards with a legitimate non-VLK OEM/retail/MSDN key. The GUID MS sees will change but they keep track.

And Biostar mobos should have a "dead flash" feature - I know Asus boards will try to reflash with the CD provided with the board or burned onto a CD. Gigabyte boards have a 2nd BIOS stored on a ROM chip separate from the main system flash chip.

I replaced my mobo with a (used) newer version of the same model, but I switched the processors to keep that ID number from being an issue ... and all went well. And, I can already tell this newer board is somehow a bit faster or at least smoother-running than the other one.

Well, I spoke too soon. The mobo swap went well, but then this morning's cold start revealed the same deal: My drives do not get properly identified or recognized until after I go back into BIOS and re-detect them. So, that means something like maybe Mint's bug for disk checks at startup has done something that now confuses something else ...

Any suggestions?

Oh, and the "Morse Code" sounds I had previously mentioned are not from this computer at all. I still had them even while this machine was completely down and unplugged.

Is the CPU fan still down from time to time or is that solved?

As for the disk problem. Is there any thing that makes you suspect Mint?

It's a long shot considering it has worked before but check the generation of the SATA controller and HDs. SATA2 supports lower voltage signals and some early SATA1 controllers has trouble with it. Some harddisk have a jumper to force the drive to always use the higher voltage.

Also, if you mb is old enough and the boot drive is SATA, you may need to add it to the boot order.

You mean the "Morse Code" still sound on your other machine? You need to jot down the beep sequence and the BIOS company to decipher the meaning. If your mb has a debug LED, also jot down the POST code.

Oops! I just edited out my previous post ...

Solved, and that might have been related to the fact I had disconnected the case fan from the board before getting the new power supply.


Just that Mint's disk-check bug might have done something to the drive IDs or something.


My board has SATA 1.0, and my WD drives do not have that jumper.


The PATAs are listed first and I boot from one of those.


I actually think that might be coming from a communications tower about three blocks from my house.

I just talked with one of the techs at WD, and he said the cold-start problem is with my board since a drive either works or it does not ... and I agree. Something is different/wrong only when the mobo is completely cold in the morning, and it is fine after that unless turned off and left to sit without power for at least a hour or two ... so maybe I just need to treat it like an old Ford and place a lightbulb underneath to keep it warm overnight?!

How do the capacitors on your motherboard look?

I do not know enough to even specifically locate them, but both boards look fine to me in the overall sense. However, I do have another completely new board on the way and I will likely try that board when it arrives. What I am dealing with here is the Biostar P4M800-M7A. My original board was Version 1.0, my current replacement is a used Version 7.0, and the *new* one still on its way is a "Pro" model (but I do presently know its Version number). In any case, I suspect your concern about capacitors is correct. This morning I started this machine and it did not work (as usual), and then I just let it sit for about five minutes ... and then a simple reset brought it around again and all was well. So, my problem is not about a bad BIOS or my drives but about something taking place and having a positive affect/effect on BIOS function after the board warms up even just a little bit.

That is weird. The VIA VT8237 south bridge is supposed to be incompatible w/ SATA2 device. I don't understand how it worked for you before. You could try setting your harddisk to SATA1 mode. I think WD still have the jumper. My Caviar Black got it, albeit without any jumper cap. You can find which jumper to set on this document 2579-001037.pdf.

You may want to check what south bridge the Pro model uses. VIA has an updated VT8237 PLUS that fixed the issue. If it's still using the plain VT8237, you may just face the same problem again.

btw google "capacitor" and "blown capacitor" to see what they looks like and what sign to look for. While it's a possible cause, blown capacitors usually cause system wide instability instead of something so specific.

VIA chipsets have assorted compatiblity issues with all kinds of devices. Intel, Nvidia, and AMD/ULI core logic and south bridges have less issues. However, Intel holds near-monopoly status on core logic for their processors - Apple is using Nvidia's nForce GeForce 3x0M/9400M which also include their nForce core logic on the same chip. Intel shut out AMD and NVidia from creating a onboard graphics system that interacts with their Core i3/i5/i7 in the same matter as onchip graphics but the 9th circuit court allowed Nvidia to use those patents Intel has.

Whew. I would never try this, but I can imagine somebody somewhere just makes his or her own custom board, eh?!


My SATAs are not having a problem, but maybe their presence is somehow *causing* a problem. It is the PATAs that do not get identified properly when the board is completely cold. I had meant to disconnect the SATAs before turning things on this morning to see whether that made a difference, but I forgot. So, and like usual now, a simple reset was all it took to get things going after a two-minute wait following the first failed startup attempt.


That chip on my original (V1.0) board and the one running right now is a VT8237R PLUS.


I have done that, and even with a magnifying glass I cannot see anything bulged, leaking or corroded.

I will try to remember to disconnect the SATAs tomorrow morning to see whether the PATAs then show up correctly.

Just thinking out loud a bit ...

When I first got this machine about a year ago, it did fine with just one PATA drive. Then I added a second PATA during all my multi-boot trials (and I report that even though I believe the matter of a second drive is irrelevant) ... and then the problem I have now began showing up somewhere along in there after my main drive (MBR and boot sector) had been altered and re-altered and re-formatted a few times. So, I had then actually pulled that drive and replaced it with a new one (still in place now) and the new one did just fine until recently. So ...

I think in the morning I will temporarily disconnect *all* drives and temporarily connect another new (raw) WD1600AAJS and see what the BIOS reports. I have no idea what the problem might actually be, but I cannot keep from thinking it has something to do either with my BIOS needing to get the sleepers out of its eyes for a second look or with my drives needing to clear their throats before they report.