Generators and Computers

[Brcobrem]

Hi,

In the aftermath of Hurricane Wilma, I am having a heck of a time with one of my customer's computer networks. I am hoping that someone here will be able to give me a little insight into a problem. If this post is completely off topic here, please let me know and I will be happy to follow your direction.

Please note that I am not a licensed electrician, but I did work as an apprentice about 25 years ago and still have my (slightly burnt) Channel Locks. Also - I am not seeking do-it-yourself electrician advice. I would have my customer's electrician do his job.

Here's the problem:

My customer has a server computer and ten workstation computers that login to that server to do the daily accounting chores. The power lines are still down outside and they are running on several portable generators (rated ~5000 watts each).

The server and computers start up fine, but they are having extreme difficulty logging in to the network (keywords: logon, login, logging in, log on, logging on ). Note that network cable runs from each computer, to a "switch". The Switch is a powered device where you plug in all the computers' network cables. A single cable then runs from the Switch to the server computer. The cable is the commonly available low voltage unshielded Cat5 variety.

We have found that if we completely shut down all the equipment, unplug it from the generators, then restart the equipment, some of the computers are able to login to the server. Each day different computers are able to make the connection. Today it took three hours of shutting down, then working up to get only five of the computers connected to the server.

My brother is a sign electrician and he suggested a couple ideas. He suggested that the different generators may be running at Hertz. In his experience, he has seen some generators that run at 58Hz and some even up to 65Hz (ie. not the normal 60Hz we see off the utility wires). He also suggested that there may be "noise" on the Neutral line (ie. AC voltage seen between neutral and the ground wires).

I am a Microsoft Certified Partner (20 years experience) and I and Microsoft Support have determined that all the software is set up properly. Additionally, I have replaced the network switch, some network cables, and the network cards in the server and some computers. We still have this computer to server communication problem.

The only thing I can think of is that possibly might help the quality of signal on these low voltage network wires would be if the generators need grounded into a grounding rod. Also - I am aware that many electronic devices like our network cards have embedded quartz clocks that may be dependent on the utility power Hertz.

Any comments, suggestions or ideas will be greatly appreciated. I will also post this message to another computer hardware technical forum and will reply back with any info I may receive from there for your use.

Regards,
Brcobrem

 

[bphgravity]

Re: Generators and Computers

I feel the problem is independent of the electrical supply. If the computers startup and run, how is the network issue a power issue? My guess is that there is an issue with the network because of the power and communication problems in the wake of Wilma. Meaning, the problem is not at your location, its somwhere else.

All the difference in hertz and grounding stuff is a bunch of bologna. I wouldn't buy into any of that.

 

[petersonra]

Re: Generators and Computers

The PC power supplies don't care much if the voltage or frequency is off a little. They will work fine or not at all, so I would not buy that.

I suspect you have some network issues that need resolving. It sounds like the PCs themselves are fine.

A lot of times people will blame "bad" grounds or "poor" power, but it is very rare that the kind of symptoms you are seeing is really about power quality.

 

[mhulbert]

Re: Generators and Computers

I'll second (or third?) the "not a power problem" vote. If the computers run OK by themselves, things are OK. Remember, the power supply puts out pure DC, from a variety of input voltages and frequencies, there is no AC past the PS unit in the computer.

Although it shouldn't matter...are all of the computers in the area grounded to each other? I mean, if you are using a seperate gennie for different computers, are you bonding the chassis of the genereators together, as well as to a ground rod? Perhaps the voltage potential between each computer's ground is causing a problem with the networking gear. I highly doubt this, because Ethernet uses a "balanced", differential communication method, where there is no "ground" between devices. However, there is a limit of the voltage difference that the unit will take before it no longer operates. This is why fiber opticcs are used between buildings and in large campus settings. How far are your networking runs? I would put a digital voltmeter between the chassis of 2 computers powered from seperate gennies and see what potential is there.

Again, this sounds more like a computer problem than a power problem, but you never know!

mike

 

[hurk27]

Re: Generators and Computers

Try running a common bonding conductor between all the generators to give the network a common signal path, Each generator is a isolated power supply that does not have any common path between them, since it is unlikely they will be in sync with each other (generators) the clocks in each computer and the network switches can't sync up. Not to sure if this is correct but I did stay at a Holiday Inn Express last night.

 

[dereckbc]

Re: Generators and Computers

Step 1. Bond al the generator chassis together, forget a ground rod, it has nothing to do with the problem.
Step 2 if step 1 does not work, use optical isolation modems.

 

[dereckbc]

Re: Generators and Computers

Are you positive there are no signal cables anywhere that use signal gound or have shields? If so step 1 or 2.

 

[marcspages]

Re: Generators and Computers



[ November 09, 2005, 05:50 AM: Message edited by: marcspages ]

 

[jimioy]

Re: Generators and Computers

Running different Generators - you are probably experiencing unbalanced frequencies between different pieces of equipment that are "connected" in a system that is being fed off different sources, each generating voltages on its own sine wave. If you took a scope view of the differnt generators and laid them on top of each other, you would see that they would not line up - leading and lagging each other. This can cause very expensive damage to inductive loads.

 

[petersonra]

Re: Generators and Computers

Originally posted by jimioy:
If you took a scope view of the differnt generators and laid them on top of each other, you would see that they would not line up - leading and lagging each other. This can cause very expensive damage to inductive loads.

since each load only sees the power from one generator I don't see how this statement can be true.

and all the load in a PC is DC. The only thing that uses AC is the power supply. Even the internal fans are DC.

 

[rattus]

Re: Generators and Computers

Furthermore, the 32KHz quartz crystals in the computers are independent of the AC frequency.

Seems to me that someone needs to check out all the signal levels with a scope or test set. All sorts of strange things may be going on here, and I would suspect first the grounding system.

 

[catchtwentytwo]

Re: Generators and Computers

The power lines are still down outside and they are running on several portable generators (rated ~5000 watts each).

How is the generator power being distributed? Are there dedicated temporary lines from the genset(s) to the equipment (point-to-point)?

Or were they tapped into to existing site wiring?

[ November 06, 2005, 07:59 AM: Message edited by: catchtwentytwo ]

 

[hbiss]

Re: Generators and Computers

Link the generators together so that you have one speed of supply (it does not matter if it is less, more, or exactly 60Hz, but as loing as it is one speed).

Ain't gonna happen with those little generators you have all over the place. They are not made for it. As far as this and your other theories, sorry but you don't know what you are talking about. I too would suspect the grounding or lack thereof.


While we are on the subject (kind of) the only portable generators I know of that can be paralleled (I think only 2 together) are made by Honda and they incorporate an electronic inverter after the generator. These give you a regulated output voltage and a nice clean sine wave which is great for electronic equipment.

-Hal

[ November 07, 2005, 08:01 PM: Message edited by: hbiss ]

 

[Leitmotif]

Re: Generators and Computers

Tying all these little generators together would cause them all to run at the same frequncy - IF you could get them to parallel and do it without damaging them they MUST be connected (paralleled) when in phase - that is why there are synchroscopes or "synch" lites.

ON a simple standalone unit the governor (throttle on prime mover) controls the speed. On a large when in standalone the governor controls speed. When paralleled the governor controls speed and more important LOAD. YOu can "take' load from other unit by opening throttle and on other unit closing throttle. The governors MUST be designed to operate together or it just is NOT gonna work. This is one reason they put reverse power trips on alternators you can get one alternator driving the other as a motor that is all OK but it can tear up the prime mover.

The other tricky thing with alternators paralleled is the voltage regulators ALSO have to be designed to operate in parallel.

Power factor has nothing to do with the alternator it is the result of resistive and reactive load (capacitive and or inductive).

Dan Bentler

 

[charlie tuna]

Re: Generators and Computers

energizing a computerized device with a standard generator is asking for trouble.......... total impeadance of an a.c. circuit is directly determined by frequency.......... as loads change on such a small generator(5000w) so does the frequency. now if you have a mini-ups unit ahead of each comuter and they are charged up they may solve your problem. also there are generators that put out pure power -- they have a pulsating d.c. output ---- the es series manufactured by honda. and then you could have all your computers properly powered and have dirty power supplying your phone network and cause communication problems within your system. i get a kick out of my freinds recently without power for near two weeks -- and operating their $3000.00 plasma t.v. off their $500.00 briggs and straton generator!

 

[Brcobrem]

Hi,

I know it's been almost a year but, I did finally determine why the workstations would not log onto the server *consistently". It took about about two weeks after the original 11/05 posts to get this figured out. Remember that there was no utility power for 3 weeks after the hurricane at that location.

First I had them move four 5000 watt generators to a common location, drive in a 6 foot grounding rod, and ground them all to that rod. All the PCs's, servers and newtworking gear were connected to UPS's (as usual). The UPS's were then connected to these generators. The PC's that could log into the domian server before the grounding rod, now would stay logged on for a longer period of time without dropping their connection (sometimes all day), but they too would occasionally drop the connection. Some PC's still could not connect to the domain at all.

Without getting too geeky on you, it turns out that because the server had lost it's connection to another network, a security function would not allow local logons. Once I made some temporary changes, all PC's could then log into the domain and the local network ran flawlessly for the next 7 days on these cheapo generators (until the utility poles were put back up).

The reason why certain PCs could log on to the domain before the "fix" turned out to be a real technical puzzle. That too was finally solved and will be the subject of an upcoming article for one of the technical journals that I write for occasionally.

I just want to thank you all for your input and suggestions during that very difficult time down here. You can only imagine working in dark offices with a hiking LED headband for light, it was 90-95 degrees inside, humid and tempers were short. I really enjoyed the technical aspect that you presented. In retrospect, grounding these generators together and using UPS's on the equipment may help some other poor soles the next time the wind blows.

Regards,
Brcobrem