Supermicro motherboard beeping
I have a Supermicro PDSBM-LN2+ motherboard (yes, it is ancient, but it does the job) that just started making a short beep (about 1 second long) every few seconds. After every beep, the fans speed up momentarily and then slow down again (the BIOS is set to the Super Quiet, 4-pin fan speed option).
It is probably not a fan speed sensor, because I slowed down one of the fans with my fingers and the warning beep was higher in pitch than what I am hearing. The BIOS shows the CPU and system temperatures to be well below the alarm threshold. None of the status LED's on the motherboard are lit.
I read one discussion online where someone said that the RAM might be responsible; I removed both sticks and wiped the contacts with a soft cloth- no change. I am running memtest86+ now- too soon to know the results, but the system is beeping constantly.
Assuming that the RAM tests OK, does anyone have a clue as to what might be wrong?
Update- The first pass of memtest86+ shows no RAM errors. I will let the test run for a few more passes to see if anything turns up.
I also checked the CMOS battery- it is OK.
Crash GordonDrive It Like You Stole It
|reply to daveinpoway |
Mountain View, CA
|reply to daveinpoway |
The fact you get VGA output/etc. and the system works means that it's not failing RAM. The system would beep one long beep and not
POST if it was the RAM. So whoever gave you that advice is barking up the wrong tree.
1. Have you contacted Supermicro Technical Support yet? If not, why not?
2. Are you using a Supermicro chassis? Does the chassis have have a SATA or SAS hot-swap backplane? If so, those can also have buzzers pertaining to overheating and fan status.
3. Any orange LEDs on the motherboard?
4. There's a jumper on the 4-pin SPKR header (sometimes labelled J9), across pins 3-4 (indicating to use the on-board buzzer). When the alarm is going off, try pulling the jumper off and/or putting it on pins 1-2 -- does the noise stop? If so, then yeah, it's the on-board buzzer/speaker which is making the noise. The only thing that ties in to that, that I know of, is the HWM/SuperIO chip.
I don't think Chassis Intrusion will cause this noise, but it might. Not 100% sure.
Anyway, if it's the HWM/SuperIO chip, then it might be failing or one of the related traces or components (there's usually resistors involved) has gone bad or is faulty -- this may happen so quick that the BIOS does not show it. Documentation says it's a Winbond (now known as Nuvoton) 83627DHG. I think this is one of those board models where the silkscreening + docs say one thing, but the actual chip operation is completely different.
5. One final thing to try would be flashing the BIOS (even to the same version). If it's the HWM/SuperIO chip, that may actually reset things to a default state (much more than "Load Factory Defaults" in BIOS), or may reset the HWM/SuperIO chip.
I'd recommend talking to Supermicro above all else though. Don't fear communicating with them, they're usually pretty good, just give them all the details and try to be precise/terse.--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.
|reply to Crash Gordon |
The site specs say Phoenix. The list does not show anything relative to the description of single beeps. Maybe count exactly how many beeps and when might help.
According to the manual though the bios code for long continual beeps say "no memory detected" but there is also noted codes that could be listed in the top left hand corner of the screen.
See the following manual:
Appendix B specifically, but motherboard listed errors start at Appendix A--
The only thing necessary for the triumph of evil is for good men to do nothing - Edmund Burke
|reply to koitsu |
1) This just started happening this morning; since it is a weekend (and a close-to-holiday weekend), I doubt that I can get any Supermicro support until Monday.
2) No, the board is not in a Supermicro chassis; it was originally in a huge server-size chassis (I forget the brand); I took it out and put it into a "standard" ATX case (which I purchased from a place that dismantles electronics for recycling). There is no hot-swap backplane- only the 4 SATA ports on the motherboard itself.
3) None of the status lights turn on.
4) I can hear that it is the on-board buzzer.
5) I will try the BIOS flash if I need to (see below).
Status: A few passes of memtest86+ showed no faults. I pulled out both RAM sticks again and turned the system on- the beeping that resulted was faster and higher in pitch, so what I was hearing was not caused by RAM problems. I then put the RAM back in and turned the system on- the beeping suddenly went away. Since I have no idea why it was beeping, I have no clue as to why it stopped, but, if it is an intermittent condition, it almost certainly will happen again.
I see that Supermicro has a "Supero Doctor" monitoring program; this cannot be installed in the software (a custom firewall application) that I normally run, but I have an extra hard drive installed (which is normally disconnected); this has Windows XP Pro, SP3 installed on it, so I can install the monitoring program on this drive. The next time that the beeping starts, I will switch to the Windows XP hard drive and see if Supero Doctor tells me anything.
Thanks to everyone who responded. Since the problem has (probably only temporarily) gone away, there seems to be little point in continuing the discussion right now. I will revisit this topic when the problem comes back.
|reply to daveinpoway |
I'd guess it's an overheating warning tone.
My old EVGA motherboard had a similar feature - whenever it detected that the CPU had reached a set temperature (as configured in the BIOS settings), it would automatically throttle the CPU to 50% (or as specified), max out the fan speeds and produce the audible 'beep' until the temperature dropped below the set value.
If it is a heat warning, have you checked your system/heatsink/fan for excessive dust? Otherwise, you may have to carefully inspect the the heatsink or reinstall/reapply TIM as required. In my situation, it actually turned out to be a loose CPU Heatsink bracket (used cheap plastic tabs to mount it to the motherboard) that prevented the heatsink from making a good/tight contact with the CPU.
Since I just installed a new CPU heatsink a couple of weeks ago (I wanted one with a quieter fan), dust should not be an issue. The Hardware Monitor section of the BIOS showed that the CPU and System temperatures were about 40 degrees C below the overheat threshold (the reported temperatures were about 28 degrees C and the overheat threshold is 70), so overheating would not appear to be present. If it was overheating, the fans would probably continuously run at full speed (as in your case)- here, the fans slightly sped up after each beep, then slowed back down again.
The new heatsink is securely installed with metal brackets (no cheap plastic fasteners), so it is unlikely to become detached at any of the corners.
One puzzling thing is that the single beep which occurs when power is applied and the continuous beeps that happened when the RAM was removed are somewhat high in frequency; the mysterious beeping that I heard was lower in pitch. So, this beeping may not have been a warning signal (although if it wasn't a warning, what was it?). Maybe the Supermicro people can shed some light on this (I will contact them after the holiday).
New mystery- This morning, I wanted to change something within the Windows XP installation (which is on the extra hard drive mentioned above), so I needed to go into the BIOS to change the boot order. While I was in the BIOS configuration pages, the system started beeping constantly (I used the BIOS configuration many times in the past with no beeps ever showing up). Once I left the BIOS configuration and booted up Windows XP, the beeping stopped.
After I was through using Windows, I needed to change the boot order back to where it normally is, so I had to go into the BIOS once more. Again, the motherboard beeped constantly while I was in the BIOS configuration and the beeping stopped once I exited the BIOS.
The other day, when the beeping first started, it was happening while I was not using the BIOS configuration pages.
|reply to daveinpoway |
Do you have another power supply to test with? Could be a bum PS
It could be the power supply (yes, I have a few spares sitting around), but the BIOS Hardware Monitor shows that all of the voltages are fine.
I suppose it is worth spending a little time to swap out the PS.
OK, I tried another power supply- still beeping while I am in the BIOS setup.
Unplugged the floppy drive- still beeps.
Unplugged the optical drive- still beeps.
Running out of options here.
It turns out that the beeping stops if I use a fan setting other than the 4-pin, Super Quiet one. Not sure why- both the CPU and the rear case fan show an RPM reading in the BIOS hardware monitor, so it would appear that the tachometer circuitry built into the fans is OK and the tachometer outputs can be read by the motherboard.
Unfortunately, the system is not domesticated (meaning quiet enough to use in my bedroom 24/7/365) unless I use the fan setting which causes the problem. Since I used the "problem" setting for over a year with no beeps, something (no idea what) has obviously changed.
I tried calling Supermicro support- all I got was a recorded message that they are closed for the holiday.
So, I resorted to drastic measures- I pulled the rear case fan tachometer wire out of the connector (only the Power, Ground and PWM Control wires are hooked up). This way, the motherboard does not even know that a case fan is installed. Not a permanent solution, but it should hold up for a few days until the support department opens up again.
I finally got ahold of the Supermicro support people. The man I talked with said that the beeps indicate local overheating, but he had no explanation as to why unplugging the rear case fan (which, if anything, would cause the internal case temperatures to rise even higher) will make the beeping stop. He also said that I should add a front case fan, but, again, he could not explain why the system has been able to operate for over a year with no front fan without beeping. Since the system is basically loafing (the AC power consumption is only about 40W), adding a front fan would not seem to be needed. As it is, the air blowing out from the rear of the case is barely warmer than the room temperature.
He did have a good suggestion that I should try plugging the rear case fan into a different connector on the motherboard (there are 2 unused fan connectors). If/when the beeping starts again, I will try that. It will help determine whether there is an electronic problem related to the connector I have been using.
|reply to daveinpoway |
For the new Heatsink/Fan you installed - have you confirmed the fan type is the correct one for your motherboard?
i.e. Tachometer based vs. PWM based fans
Assuming that you didn't change any of the case fans (and that they haven't failed), I'd be betting on the changes being the cause of any new issues.
Yes, the new heatsink fan is a 4-pin (PWM-controlled) type and the package indicates that it will work for a Socket 775 application. The heatsink is a massive thing with a 110 mm fan (intended for gaming PC's), so it should have no problems cooling my system. If I feel the heatsink fins, they aren't even warm.
One thing which doesn't make sense to me- the Supermicro rep indicated that the BIOS may not report the temperatures accurately. The question is: If the BIOS does not accurately know how hot the CPU is, how does it correctly know when it should activate an overheat alarm? I suspect some sort of BIOS bug(s), since the BIOS version I am using is very old. Updating to a later version doesn't work for me, since someone apparently rewrote the fan-speed code, causing the rear case fan to turn significantly faster in the later BIOS versions. The faster speed causes the noise level to be unacceptable for the location of the system.
I suppose that something could be wrong with the rear case fan tachometer circuit, but the BIOS shows a speed reading that makes sense, so everything seems to be fine. As I indicated earlier, I will try plugging the rear case fan into a different motherboard connector if the beeping returns. I don't presently have another 120mm 4-pin fan to try.
If I was really brave, I could take the later BIOS code and edit it to lower the rear case fan speed, but one mistake and the board will be ruined. I will leave BIOS editing to the experts (I have never done it before).
Just pulled up the manual for your motherboard - PWM fans should be fine.
If it is an overheating warning, there's also an overheating LED on the motherboard (LE2 beside the CPU) that should light when the warning sounds. The manual only mentions beep codes for video, memory or overheating - other error codes would attempted to be displayed at the top left corner of the display.
There should be a BIOS setting for 'CPU Overheat LED and Control' where you should be able to set the CPU temperature threshold.
You could try setting the temperature threshold lower in order to intentionally generate the overheat warning tone at a lower/safer CPU temperature and determine whether or not it's the same tone that you're hearing now. If it's not, then your problem must be something else. If it's the same, than I'd suggest re-installing the Heatsink/Fan/TIM (or even going back and installing the original HSF to see if there's any change).
If it is an overheating warning but the Heatsink doesn't feel warm, then its likely a HSF mounting issue - i.e. the CPU IS hot and the heat isn't transferring to the heatsink.
Trying the Supero Doctor III software to monitor voltages/temps would provide more details as well.
|reply to daveinpoway |
An easier and safer alternative to BIOS editing would be to use a potentiometer to slow down the case fan. Here's
one that Zalman includes with a lot of their heatsinks. I've used it a few times and it works pretty well.--
|reply to Mister_E |
The motherboard overheat LED never comes on. I will have to check to see how low I can set the overheat threshold- it is not convenient to shut the PC down to hook up a keyboard and monitor right now.
One confusing thing- I reconnected the rear case fan tachometer wire and the beeping has not returned (so far). If the CPU was overheating, I would expect it to happen all of the time (the CPU workload is pretty constant and the daily report I receive from the software I am running tells me that the CPU usage never goes over 20%).
The software I use on this PC does not allow any other programs (such as Supero Doctor) to be installed, so this monitoring program cannot be easily used.
|reply to n_w95482 |
Yes, I was thinking about one of those speed control potentiometers, but I haven't been able to find any locally. One disadvantage of this approach, however- if the system should overheat, the PWM fan will automatically go to full speed. With the potentiometer, unless someone is around to change the setting, the fan will not be able to attain full speed. This system runs 24/7/365 (even when I am not home), so it needs to be fail-safe.
OK- I'm learning something. Yesterday, I went to Best Buy and bought a Rocketfish 120mm PWM fan (the reviews on this have been mixed, but I can always take it back if the quality is junk). I just shut down the PC (there was no beeping) and replaced the Cooler Master Excaliber rear case fan with the Rocketfish. While things were apart, there was time for the system to cool down (not that it was very warm to begin with).
Immediately upon turning the system back on (before anything could heat up), there was beeping (sometimes the low-pitch beep, sometimes a low pitch which quickly changed to a high pitch). Changing the rear case fan to another motherboard connector had no effect. Looking in the BIOS Hardware Monitor, the Rocketfish seems to run about 100 RPM slower than the Excaliber.
So, it would seem that the beeping is somehow related to the motherboard detecting that the fan speed is slower than it would like to see- heat appears to have nothing to do with it. This would explain why disconnecting the case fan tachometer wire stopped the beeping- if the motherboard did not know that a fan was there, it wouldn't be checking to see if the speed is satisfactory.
As to why this condition has suddenly shown up, I have no clue. There is no place I can see in the BIOS to set a fan speed alarm threshold, so I guess that I will have to leave the case fan tachometer wire disconnected.
Yesterday, when the system was running (no beeps at the time), I used my finger to stop the rear case fan for a few seconds. Sure enough, the alternating high-pitch/low-pitch beeping started immediately. So, even though the Supermicro support rep swore up and down that this beep condition meant local overheating, it is definitely a fan-speed warning.
Perhaps on other Supermicro motherboards (or even on this one with a different version BIOS) the 2-tone beep signifies overheating, but not for my situation.