BIG Problem
in [Home Built PC]
|
Prev: Got it working! First post from inside Damn Small Linux! Needmore help pls
Next: ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
From: Jan Alter on 8 Jun 2010 18:51 "LSMFT" <boleyn7(a)aol.com> wrote in message news:yByPn.68090$Gx2.23495(a)newsfe20.iad... > MN wrote: >> I just completed assembling a computer. It is based upon the MSI MS-7265 >> mobo and Morex 6610 case (with 250 watt power supply). It is a mini-ITX >> computer and includes a SATA hard-drive, an IDE (PATA) CD/DVD burner and >> a >> 3.5" floppy. (The latter can come in handy, sometimes.) >> >> Needless to say, it was a tight fit, with the cables to the mobo the most >> difficult and tedious part -- especially the cable dressing. I was >> careful >> during assembly, using a grounded wrist strap-- especially when placing >> the >> memory strips, cpu, and CPU-fan onto the mobo. I had all the docs& info >> at >> my finger tips which was needed to correctly place the numerous >> connections >> to the mobo headers/connectors (e.g. SATA. IDE, USB, power button switch >> and >> LED, etc.). >> >> Well, I applied AC power to the computer with the case-cover off and NO >> connections to the monitor, mouse and keyboard. I expected to see the CPU >> fan running and some beeping since the was no mouse or keyboard >> connected. >> Much to my disappointment, NOTHING happened. So, I realized that I would >> have to diagnose what is wrong. >> >> First, I checked the AC-cord and made sure it was OK, and also connected >> the >> computer to different AC outlets -- to make sure that the AC input was >> OK. >> Then I measured voltages on the power-supply's 20-pin main ATX connector. >> I >> found the requisite standby 5 volts on pin 9 (purple wire) and pin 14 >> (green wire). This is correct with no signal from the mobo. >> >> Next, I need to JUMPER pins on the 20-pin ATX connector the simulate the >> mobo's OK signal to the power supply. I don't KNOW which pins to jumper >> to >> see if the power supply will output the other voltages on other pins. My >> pinout diagram just lists the pin #, wire color, and voltage value --so >> I >> can't determine which pins to jumper. I WANT TO BE SURE TO JUMPER THE >> CORRECT PINS! >> >> Please advise my of the pins to jumper, and other steps to perform >> in my >> diagnosis. >> >> > It's green and black. > > -- > LSMFT > > I haven't spoken to my wife in 18 months. > I don't like to interrupt her. One thing (as I'm running out the door) make sure you have the 4 pin square plug from the PS for the cpu plugged into the mb or you won't get any power to the cpu. I see you have 'cpu' listed but I'm just second guessing that's what you mean. Good luck -- Jan Alter bearpuf(a)verizon.net
From: Paul on 8 Jun 2010 21:49 MN wrote: > Paul, > Thanx for your prompt reply and VERY useful info. > > I did jumper PS_ON# to COM, to turn on the power supply. Then I measured > the voltages at the 20-pin ATX power connector. They were ALL OK! Then I > went to the 4-pin 12-volt power connector. AHA! This was NOT the12-volt > power connector -- although it looked almost identical (but without the side > latch). I was fooled since it was tie-wrapped close to the 20-pin ATX > connector. I then "dug-out" the REAL 4-pin 12-volt power connector; it's > voltages measured OK. > > Next, I removed the jumper and DVM, and I connected the 2 power connectors > to the mobo. Then, I plugged-in the computer, and pressed the power > on/power-off button. After a few seconds (that's long?), the on/off button > "glowed blue" and the CD/DVD led started blinking like normal. Also, I could > eject/retract the CD/DVD disk tray by pressing the appropriate button on the > front faceplate of the CD/DVD. That's the GOOD news. > > However, when I pressed the power on/power-off button, "power did NOT go > off' as indicated by the blue LED at the center of the the power > on/power-off button. What does this mean? > > Also, the CPU fan did NOT rotate nor were there any audible beep codes. Of > course, if the CPU was not running, the BIOS code would not execute and > generate any beeps, etc. So, now what. Can we conclude that the CPU is NOT > running as indicated by the CPU fan not rotating -- OR -- is the CPU fan not > rotating because the temperature is not "hot enough"? > > I am really perplexed. I am glad that the PS is OK and also of the minor > functionality as indicated by the CD/DVD activity, but I don't have a > rational hypothesis where to proceed next. In other words "WHAT DO I DO > NEXT"? to advance "from first base". > > Please advise. > > Regards, > Marty Oh, no... My first reaction is, I thought the pin shapes were different on the ends. The 20+4 4 pin section has a different pin shape than the 2x2 ATX12V. Something like this. S D versus D S where: S = square pin D S S D D = shaped like the letter D If you rotate one of those by 180 degrees, the D shape on one would be a "backwards D", so it shouldn't mate. I'm using the pictures on playtool to reach this conclusion, and my eyesight isn't the best. http://www.playtool.com/pages/psuconnectors/connectors.html The thing is, your 20+4 4 pin section, has +3.3, +5, +12, and Ground on it. When mated to a connector with two 12V pins and two ground pins, two pairs of rails end up shorted together. Say, for the sake of argument, that a 12V pin is shorted to a 3.3V pin, because they happen to be shoved into the two 12V motherboard holes. The 12V power output will attempt to "lift" the 3.3V output. This could damage the 3.3V circuitry on the motherboard. The power supply has several forms of protection. 1) Overcurrent. Shorting one of the rails to ground, could cause the power supply to switch off after 35 milliseconds. 2) Overvoltage. OVP detects wide excursions, like the 3.3V rail being at 4.7V. The supply can be designed to shut off when that happens. It might even take less than 35 milliseconds. But sensitive circuits can be blown in less time than that. Your CPU fan is not rotating. You could take a multimeter, and *carefully* check the 12V pin on that fan header, and see if it is getting power. To do that here, I'd make up a custom cable assembly (I have some bags of small fan connectors and pins), and then take a loose wire from a setup like that, and make a measurement. Otherwise, it would be dangerous to get the multimeter in amongst the CPU fan header pins. The CPU fan draws power from the 12V yellow wire on the main power connector. The CPU itself draws power from the two yellow wires on the ATX12V connector. If the CPU gets too hot, while sitting there in that fan-less state, the computer could shut off. Are you still jumpering PS_ON# ? Now would be a good time to stop doing that. The overheat safety feature won't work to protect the CPU, if your PS_ON# is still jumpered. I'd swap motherboards. Chances are, your CPU is not damaged. Your RAM is not damaged. They're protected by the onboard regulators. There could be damage to the motherboard. You can test your CDROM and hard drives using another computer. You've verified PSU voltages, both on the main connector and on the ATX12V, so you suspect the PSU is OK. Don't forget to check the status pin on the ATX PSU, the "power good" or whatever it is called, as the power supply may refuse to assert that, and that would prevent POST on a new motherboard. It is pretty hard to judge just what damage has been done. I would not have thought those two connectors could be mixed up (but I'm not taking my current computer apart to verify this :-) ) If you did manage to mate the connectors, one rail got shorted, and another rail got lifted. Either "insult" should have made a good power supply shut off. But maybe not fast enough to prevent damage to the motherboard. What happens to a CPU fan, is a function of its design. The CPU fan header could be a three pin or a four pin. A three pin can use voltage control on the 12V pin. That means a MOSFET may be in series in the path. Damage to the control circuit might prevent it from powering up. A four pin CPU fan header, delivers 12V to the 12V pin, and relies on the PWM pin for fan speed control. That is less likely to stop spinning entirely. I'd turn off all power, use the ohms scale on the multimeter, and check for continuity from the 12V pin on the CPU four pin fan header, to where the 12V wire would plug into the main power connector. Other than debugging, I'd swap motherboards, as long as you know for certain, that the PSU has all good voltages, and has a logic 1 on the power good output. Paul
From: MN on 9 Jun 2010 00:04 Paul, Thanx for your prompt reply and suggestions. I address your ideas/concerns individually below after the corresponding text. (They are prefixed with ***). ============================================================================ "Paul" <nospam(a)needed.com> wrote in message news:humrv5$fao$1(a)news.eternal-september.org... > MN wrote: >> Paul, >> Thanx for your prompt reply and VERY useful info. >> >> I did jumper PS_ON# to COM, to turn on the power supply. Then I measured >> the voltages at the 20-pin ATX power connector. They were ALL OK! Then I >> went to the 4-pin 12-volt power connector. AHA! This was NOT the12-volt >> power connector -- although it looked almost identical (but without the >> side latch). I was fooled since it was tie-wrapped close to the 20-pin >> ATX connector. I then "dug-out" the REAL 4-pin 12-volt power connector; >> it's voltages measured OK. >> >> Next, I removed the jumper and DVM, and I connected the 2 power >> connectors to the mobo. Then, I plugged-in the computer, and pressed the >> power on/power-off button. After a few seconds (that's long?), the on/off >> button "glowed blue" and the CD/DVD led started blinking like normal. >> Also, I could eject/retract the CD/DVD disk tray by pressing the >> appropriate button on the front faceplate of the CD/DVD. That's the GOOD >> news. >> >> However, when I pressed the power on/power-off button, "power did NOT go >> off' as indicated by the blue LED at the center of the the power >> on/power-off button. What does this mean? >> >> Also, the CPU fan did NOT rotate nor were there any audible beep codes. >> Of course, if the CPU was not running, the BIOS code would not execute >> and generate any beeps, etc. So, now what. Can we conclude that the CPU >> is NOT running as indicated by the CPU fan not rotating -- OR -- is the >> CPU fan not rotating because the temperature is not "hot enough"? >> >> I am really perplexed. I am glad that the PS is OK and also of the minor >> functionality as indicated by the CD/DVD activity, but I don't have a >> rational hypothesis where to proceed next. In other words "WHAT DO I DO >> NEXT"? to advance "from first base". >> >> Please advise. >> >> Regards, >> Marty > > Oh, no... > > My first reaction is, I thought the pin shapes were different on the ends. > The 20+4 4 pin section has a different pin shape than the 2x2 ATX12V. > Something like this. > > S D versus D S where: S = square pin > D S S D D = shaped like the > letter D > > If you rotate one of those by 180 degrees, the D shape on one would be a > "backwards D", so it shouldn't mate. I'm using the pictures on playtool > to reach this conclusion, and my eyesight isn't the best. > > http://www.playtool.com/pages/psuconnectors/connectors.html > > The thing is, your 20+4 4 pin section, has +3.3, +5, +12, and Ground on > it. > When mated to a connector with two 12V pins and two ground pins, two pairs > of rails end up shorted together. Say, for the sake of argument, that a > 12V pin is shorted to a 3.3V pin, because they happen to be shoved into > the two 12V motherboard holes. The 12V power output will attempt to "lift" > the 3.3V output. This could damage the 3.3V circuitry on the motherboard. > > The power supply has several forms of protection. > > 1) Overcurrent. Shorting one of the rails to ground, could cause the > power supply to switch off after 35 milliseconds. > > 2) Overvoltage. OVP detects wide excursions, like the 3.3V rail being at > 4.7V. The supply can be designed to shut off when that > happens. It might even take less than 35 milliseconds. > But sensitive circuits can be blown in less time than > that. > > Your CPU fan is not rotating. You could take a multimeter, and *carefully* > check the 12V pin on that fan header, and see if it is getting power. To > do that > here, I'd make up a custom cable assembly (I have some bags of small fan > connectors > and pins), and then take a loose wire from a setup like that, and make a > measurement. Otherwise, it would be dangerous to get the multimeter in > amongst > the CPU fan header pins. > > The CPU fan draws power from the 12V yellow wire on the main power > connector. > > The CPU itself draws power from the two yellow wires on the ATX12V > connector. > > If the CPU gets too hot, while sitting there in that fan-less state, > the computer could shut off. Are you still jumpering PS_ON# ? Now > would be a good time to stop doing that. The overheat safety > feature won't work to protect the CPU, if your PS_ON# is still > jumpered. > > I'd swap motherboards. Chances are, your CPU is not damaged. Your > RAM is not damaged. They're protected by the onboard regulators. > There could be damage to the motherboard. You can test your CDROM > and hard drives using another computer. You've verified PSU voltages, > both on the main connector and on the ATX12V, so you suspect the > PSU is OK. Don't forget to check the status pin on the ATX PSU, > the "power good" or whatever it is called, as the power supply > may refuse to assert that, and that would prevent POST on a > new motherboard. *** I really can't swap mobos, because I don't have another spare functioning mobo of similar architecture. > It is pretty hard to judge just what damage has been done. I would > not have thought those two connectors could be mixed up (but I'm not > taking my current computer apart to verify this :-) ) If you did > manage to mate the connectors, one rail got shorted, and another > rail got lifted. Either "insult" should have made a good power > supply shut off. But maybe not fast enough to prevent damage to > the motherboard. *** Regarding the "similar" 4-pin connector: I measured +5 volts on it when I had both ATX power connectors NOT connected to the mobo, and the PS_ON# to COM jumper in place. This +5 volts was measured AT THE SAME TIME that I measured (and found correct) the voltages on both ATX power connectors. Comment: this +5 volts on the "similar" 4-pin connector is LESS than the +12 volts that the correct 4-pin connector supplies. This is in "my favor" since the +12 volts that the mobo expected is greater than the +5 volts I possibly erroneously supplied. Also, (1) after I connected the correct 4-pin ATX connector AND the 20-pin ATX connector (no jumper, of course), and (2) applied AC power, and (3) then pressed the case power button, I got the BLUE LED to illuminate and was able to manually operate the CD/DVD from the front faceplate. Thus, I believe that my original "application of +5 volts instead of +12 volts" on the 4-pin connector did NOT damage the mobo. TO DO: I need to check into the voltages applied to the fan connector when the mobo has power applied. QUESTION: SUPPOSE, I did not have the 2 sticks of DDR2 memory in their sockets, what would be the symptoms? If, for any reason the read/write/refresh RAM memory cycles were not correct (incorrect default BIOS memory timing, SPD,etc), it would be as if no RAM memory was present. Thus my question. In other word, isn't the CPU "dead in the water" without correctly functioning DDR2 RAM? > What happens to a CPU fan, is a function of its design. The CPU fan > header could be a three pin or a four pin. A three pin can use voltage > control on the 12V pin. That means a MOSFET may be in series in the path. > Damage to the control circuit might prevent it from powering up. *** TO DO: I need to check into the voltages applied to the fan connector when the mobo has power applied. > A four pin CPU fan header, delivers 12V to the 12V pin, and relies > on the PWM pin for fan speed control. That is less likely to stop > spinning entirely. I'd turn off all power, use the ohms scale on the > multimeter, and check for continuity from the 12V pin on the CPU four > pin fan header, to where the 12V wire would plug into the main power > connector. *** The fan header is 3-pin with pins for +12v, sensor, and ground. My previous TO DO will verify the +12/ground. As for the sensor, ??? > Other than debugging, I'd swap motherboards, as long as you know for > certain, that the PSU has all good voltages, and has a logic 1 on the > power good output. > > Paul
From: Paul on 9 Jun 2010 00:31 MN wrote: > > QUESTION: SUPPOSE, I did not have the 2 sticks of DDR2 memory in their > sockets, what would be the symptoms? If, for any reason the > read/write/refresh RAM memory cycles were not correct (incorrect default > BIOS memory timing, SPD,etc), it would be as if no RAM memory was present. > Thus my question. In other word, isn't the CPU "dead in the water" without > correctly functioning DDR2 RAM? > If you turn off the power, then remove both RAM sticks, the CPU can still function at the next power up. What should happen, is the CPU should be able to beep the computer case speaker. (You need a computer case speaker, connected to the SPKR pins on the PANEL header on the motherboard.) It might be two beeps, then silence, two beeps, then silence and so on. If you hear beeps, then it means the CPU is working a little bit. If it still won't beep, then the CPU might not be executing any code at all. Another way to prove the CPU is executing code, is to insert a PCI Port 80 POST card, into the PCI slot #1 (that is the PCI slot nearest to the processor). The codes don't mean a whole lot, that flash on the card, but just getting the code value to change, from the initial 0x00 or 0xFF value is an achievement. You'd either have to borrow or buy one of those. They can be real cheap when ordered from Hong Kong via Ebay, while if you buy one from a local computer store, they can be $100.00. (This one from Startech is $40 or so. One reviewer says it only works in 5V PCI slots, not 3.3V ones. 5V slots are relatively common. I think all my motherboards here have 5V slots. Some products like this, come with a second display on the end of a ribbon cable, which makes it easier to read the display.) http://www.startech.com/Share/Gallery/Large/PCIPOST.Alarge.jpg This one plugs into PCI and ISA slots. I only have one computer with ISA slots here (11 years old). This one has a second display on the ribbon cable, to aid visibility. http://www.elstonsystems.com/prod/pc_post_diagnostic_card.html Some limited number of motherboards, have four status LEDs. Some even have the port 80 display soldered right onto the motherboard, for easy reference. But to keep motherboard costs down, they don't bother with that most of the time. Asus used to provide a Winbond voice chip at one time, and it would "speak" an error code. The solution probably cost a buck or two per motherboard, but did add some value. But they've stopped using those, and no other motherboard manufacturer seemed to be interested in doing the same thing. I have a couple motherboards here, with that error information voice device on it. The sound comes out of the external computer speakers, via the green audio jack on the motherboard. Good luck, Paul
From: larry moe 'n curly on 9 Jun 2010 04:07
MN wrote: > > *** Regarding the "similar" 4-pin connector: I measured +5 volts on it when > I had both ATX power connectors NOT connected to the mobo, and the PS_ON# to > COM jumper in place. This +5 volts was measured AT THE SAME TIME that I > measured (and found correct) the voltages on both ATX power connectors. > Comment: this +5 volts on the "similar" 4-pin connector is LESS than the +12 > volts that the correct 4-pin connector supplies. This is in "my favor" since > the +12 volts that the mobo expected is greater than the +5 volts I possibly > erroneously supplied. Also, (1) after I connected the correct 4-pin ATX > connector AND the 20-pin ATX connector (no jumper, of course), and (2) > applied AC power, and (3) then pressed the case power button, I got the BLUE > LED to illuminate and was able to manually operate the CD/DVD from the front > faceplate. Thus, I believe that my original "application of +5 volts instead > of +12 volts" on the 4-pin connector did NOT damage the mobo. TO DO: I need > to check into the voltages applied to the fan connector when the mobo has > power applied. > > QUESTION: SUPPOSE, I did not have the 2 sticks of DDR2 memory in their > sockets, what would be the symptoms? If, for any reason the > read/write/refresh RAM memory cycles were not correct (incorrect default > BIOS memory timing, SPD,etc), it would be as if no RAM memory was present. > Thus my question. In other word, isn't the CPU "dead in the water" without > correctly functioning DDR2 RAM? First of all, I wouldn't test a new computer with the disk drives installed but the keyboard and monitor not connected because the video and keyboard can indicate some signs of life in the computer. At least you had the mobo speaker installed. Another thing I do with new mobos is first remove the CMOS battery because when the CMOS clock memory contents is gone, most mobos will turn on instantly when power is applied, even before the power switch pins on the mobo aren't shorted together, plus the CPU fan will probably run at full speed, rather use any quiet mode that may not turn on the fan until the CPU warms up enough. I have a feeling Paul is right about damage occurring because the two different square 4-pin power connectors were mixed up. Connectors are supposed to be molded to make that impossible, but that wasn't the situation with my Ultra V-series PSU and a cheap ECS mobo. It's possible that damage was limited to a protective fuse (flat rectangular surface mount thing, often black or green on top and marked "F" or "X" on top or on the mobo) and diode (diode may have blown the fuse and shorted in the process). These devices are often near the power connectors. Try measuring the resistance across each, in both directions. I think you should test the mobo outside of the case, to rule out any shorts between the mobo and case. Be sure to set the mobo on top of a non-metallic anti-static surface, such as the foam sheet packed with the mobo or some pink transparent bubble wrap. Put that on top of 3/4" brown cardboard, brown paper, or non-shiney newspaper. |