ups malfuntioning at 204volts

[T.M.Haja Sahib]

I think the incoming power supply cable to the UPS may not be correctly sized or there may be poor connections in the cable so that on start up the voltage available to the UPS is not sufficient so that it goes to battery mode. This may be checked by placing two voltmeters one at the source of voltage and another at the input terminals to the UPS and noting the two voltages on switching on the UPS. This may happen, when the UPS is fully loaded.

 

[kwired]

Ok met with the tech guy today he has a detailed version of events. He told me 208v but infact it is on a 208 panel the voltage is 120v. At the source where the UPS plugs in it reads 114v. coming out of the UPS to feed the strip all the servers are on it reads 108v. He has downloaded more firmware so he can adjust the voltage to 106v minimum. He refuses to believe his equipment is malfuntioning even though the battery will not charge to 100%. It only charges to 99.2%. I apologize I did not get the model of the UPS. When he does a system check it says its getting 117v. When I check it is 114v at source and 107.7 coming out of the ups feeding the servers? im proposing a sub panel in the server room as the circuit has a long way to travel to the panel its in. They have them on 30amp circuits combineing nuetral and ground at the main panel nuetral bar? My opinion is His UPS needs replaced as the brand new one next to it is functioning fine off of the same power source.

Why is UPS out put only 108 volts- is there more load connected than the rating of the UPS? I would think it should be very near input voltage, or if it is supposed to regulate the supply I would think the output would be closer to 115, especially if input is at least that high.

 

[gar]

120926-1137 EDT

highpowered:

It is quite inappropriate to start a thread with totally incorrect information. There is a total disconnect in the information in the original post with the problem of the post.

Ok met with the tech guy today he has a detailed version of events. He told me 208v but infact it is on a 208 panel the voltage is 120v. At the source where the UPS plugs in it reads 114v. coming out of the UPS to feed the strip all the servers are on it reads 108v. He has downloaded more firmware so he can adjust the voltage to 106v minimum. He refuses to believe his equipment is malfuntioning even though the battery will not charge to 100%. It only charges to 99.2%. I apologize I did not get the model of the UPS. When he does a system check it says its getting 117v. When I check it is 114v at source and 107.7 coming out of the ups feeding the servers? im proposing a sub panel in the server room as the circuit has a long way to travel to the panel its in. They have them on 30amp circuits combineing nuetral and ground at the main panel nuetral bar? My opinion is His UPS needs replaced as the brand new one next to it is functioning fine off of the same power source.

With this new information it appears that the UPS is designed for a nominal input of 120 V, and there is a long cable from the breaker panel to the UPS.

It would be useful to use two good meters to make the voltage measurements, one a true RMS type (Fluke 87) and the other an average reading type calibrated to read the RMS of a sine wave (Fluke 27).

These two meters will probably read within about 0.2 V of each other on the likely somewhat distorted sine wave at the breaker panel. Fifty years ago the waveform from the power company at the user's location was a quite good sine wave. Today it has a somewhat flattened or rounded top. This results from all the DC power supplies with capacitor input filters that are loads today.

After the main panel measurement, then use the same two meters at the input to the UPS. The voltage should be somewhat less, and the difference between the meters may be somewhat greater. Note: over short periods of time there are fluctuations in the supply voltage.

Visit the page http://beta-a2.com/EE-photos.html on my website for an idea of how voltage varies with time. This is at my home. At an industrial site the variations may be greater or less.

Photo P23 was before I had a better data collection and plotting setup. But this plot is interesting because in the voltage waveform (red curve) there is a square wave modulation at the left of about 1 V. This was on the primary lines. Either a power company voltage regulator was the cause, or a very large load. Because of the strange times of the day and week that these variations occurred my guess is a large system load. One guess for such a load is the University high powered laser which is one of the most powerful in the world.

Photo P26 is a typical day at my home over the full day. Voltage is the black curve and is resolved to 1 second, and quantized to 0.2 V changes. This provides an overall perspective.

Photo P27 is a short time segment in the middle of the night. Note: there is no significant correlation between the voltage and my load changes. The voltage is at the main panel.

Photo P29 is of one of my freezers in the garage. This is about 100 ft from the main panel with about 70 ft of #6 copper, followed by about 40 ft of #12 copper. Here the voltage is measured at the freezer. Note: the substantial voltage variation at the destination from the wire resistance and inductance from the load current variations.

These voltage measurements are RMS averaged over 1 second.

Back to your problem.

What are the voltage readings at the input to the UPS compared to the readings at the main panel?

What are the voltage readings at the immediate output of the UPS, not at some later outlet strip?

It would be nice to know these values when the UPS was in bypass mode (normal operation) and when the UPS was running from battery. Note: I am assuming this UPS only switches the output to a synthesized sine wave with energy from the battery when the input AC voltage drops below a threshold. You have not yet defined the type of UPS that is being used.

If the UPS is of the type where the output is always synthesized, then where is this detection of low voltage occurring. Certainly, if this UPS is a nominal 120 V unit, then its output voltage should be much closer to 120 than what you have indicated. At all times the UPS's output AC voltage should be near 120 until its battery voltage drops to near a discharged value.

99.2% of full charge is certainly close enough to 100 % to qualify as 100 %. It is not easy to obtain an accurate determination of when a battery is fully charged. You need a definition of what is a full charge.

On a cheap small UPS with a 100 W incandescent load I measured 122.8 on an RMS meter and 122.7 on an average meter at the output when in bypass mode. The input voltage is essentially the same as the output because there is only a relay contact between input and output. On loss of input AC, thus in battery mode, the readings were 119.1 RMS and 112.1 average from a synthesized sine wave.

The AC output voltage when on battery supply has nothing to do with the AC source voltage other than the assumption that the battery is reasonably fully charged. Had the input AC voltage been 108 V in bypass mode the output would have been 108 V, but when input AC was lost and the UPS switched to battery mode the output would be the 119.1 V measured above.

Neither an RMS or average measurement is a good measure of a suitable voltage for a computer load. A computer input will most likely be a full wave bridge rectifier feeding an input filter capacitor. Here peak voltage is most important. At 120 V RMS the sine wave peak is 170 V. This is what the capacitor charges to. If I applied a square wave to the input of the rectifier of 120 V RMS, then the peak would be 120 V and the capacitor would only charge to 120 V. This is a big disparity.

Therefore, to use the AC RMS meter as a useful tool it is necessary that the waveform approximate a sine wave.

.

 

[highpowered]

120926-1137 EDT

highpowered:

It is quite inappropriate to start a thread with totally incorrect information. There is a total disconnect in the information in the original post with the problem of the post.



With this new information it appears that the UPS is designed for a nominal input of 120 V, and there is a long cable from the breaker panel to the UPS.

It would be useful to use two good meters to make the voltage measurements, one a true RMS type (Fluke 87) and the other an average reading type calibrated to read the RMS of a sine wave (Fluke 27).

These two meters will probably read within about 0.2 V of each other on the likely somewhat distorted sine wave at the breaker panel. Fifty years ago the waveform from the power company at the user's location was a quite good sine wave. Today it has a somewhat flattened or rounded top. This results from all the DC power supplies with capacitor input filters that are loads today.

After the main panel measurement, then use the same two meters at the input to the UPS. The voltage should be somewhat less, and the difference between the meters may be somewhat greater. Note: over short periods of time there are fluctuations in the supply voltage.

Visit the page http://beta-a2.com/EE-photos.html on my website for an idea of how voltage varies with time. This is at my home. At an industrial site the variations may be greater or less.

Photo P23 was before I had a better data collection and plotting setup. But this plot is interesting because in the voltage waveform (red curve) there is a square wave modulation at the left of about 1 V. This was on the primary lines. Either a power company voltage regulator was the cause, or a very large load. Because of the strange times of the day and week that these variations occurred my guess is a large system load. One guess for such a load is the University high powered laser which is one of the most powerful in the world.

Photo P26 is a typical day at my home over the full day. Voltage is the black curve and is resolved to 1 second, and quantized to 0.2 V changes. This provides an overall perspective.

Photo P27 is a short time segment in the middle of the night. Note: there is no significant correlation between the voltage and my load changes. The voltage is at the main panel.

Photo P29 is of one of my freezers in the garage. This is about 100 ft from the main panel with about 70 ft of #6 copper, followed by about 40 ft of #12 copper. Here the voltage is measured at the freezer. Note: the substantial voltage variation at the destination from the wire resistance and inductance from the load current variations.

These voltage measurements are RMS averaged over 1 second.

Back to your problem.

What are the voltage readings at the input to the UPS compared to the readings at the main panel?

What are the voltage readings at the immediate output of the UPS, not at some later outlet strip?

It would be nice to know these values when the UPS was in bypass mode (normal operation) and when the UPS was running from battery. Note: I am assuming this UPS only switches the output to a synthesized sine wave with energy from the battery when the input AC voltage drops below a threshold. You have not yet defined the type of UPS that is being used.

If the UPS is of the type where the output is always synthesized, then where is this detection of low voltage occurring. Certainly, if this UPS is a nominal 120 V unit, then its output voltage should be much closer to 120 than what you have indicated. At all times the UPS's output AC voltage should be near 120 until its battery voltage drops to near a discharged value.

99.2% of full charge is certainly close enough to 100 % to qualify as 100 %. It is not easy to obtain an accurate determination of when a battery is fully charged. You need a definition of what is a full charge.

On a cheap small UPS with a 100 W incandescent load I measured 122.8 on an RMS meter and 122.7 on an average meter at the output when in bypass mode. The input voltage is essentially the same as the output because there is only a relay contact between input and output. On loss of input AC, thus in battery mode, the readings were 119.1 RMS and 112.1 average from a synthesized sine wave.

The AC output voltage when on battery supply has nothing to do with the AC source voltage other than the assumption that the battery is reasonably fully charged. Had the input AC voltage been 108 V in bypass mode the output would have been 108 V, but when input AC was lost and the UPS switched to battery mode the output would be the 119.1 V measured above.

Neither an RMS or average measurement is a good measure of a suitable voltage for a computer load. A computer input will most likely be a full wave bridge rectifier feeding an input filter capacitor. Here peak voltage is most important. At 120 V RMS the sine wave peak is 170 V. This is what the capacitor charges to. If I applied a square wave to the input of the rectifier of 120 V RMS, then the peak would be 120 V and the capacitor would only charge to 120 V. This is a big disparity.

Therefore, to use the AC RMS meter as a useful tool it is necessary that the waveform approximate a sine wave.

.

Again I started the thread without visiting the site I do apologize as I understand time was wasted on wrong information. So to answer your question.
bypass mode meaning input voltage is off the UPS puts out 117 volts
when the input voltage of 114 volts is on the UPS puts out 108volts
this is measured directly from the outlet on back of the UPS

All uf this aside to correct the problem I have decided to add a sub-panel with proper grounding and dedicated circuits at the UPS room. I dont know how I can change the power coming in quality any suggestions? My concern is that the new UPS operating on the same panel/voltage is fine and this UPS is on the fritz? I hope this answers your questions

 

[gar]

120926-1405 EDT

If I understand what you are saying I conclude:

1. If input voltage to UPS is 117 V, then output of UPS is 117 V. In other words the input is connected to the output by a bypass relay.

2. If the input voltage drops to 114 V or someplace between 117 and 114 V (the threshold point), then the UPS switches to battery mode and outputs 108 V. This 108 V reading could be quite dependent upon the output waveform. As I said in my previous post this voltage has nothing to do with the input AC voltage when in battery mode.

Whether the 108 V is a bad value I do not know. It may be more important what the peak voltage is. The type of meter used for this measurement and how the loads (computers or whatever) respond to the waveform is important relative to a judgement on whether the battery powered output voltage is satisfactory.

A threshold point in the 114 to 117 V range is too high in my opinion. I would expect more in the range of 105 V.

If with the same meter you make measurements on the other working UPS, then a comparison of voltages between the two UPSs can be useful. I do not think your problem UPS is working correctly based on the supplied information.

Re-doing the input wiring might reduce the observed undesired operation, but it really might just mask the real problem which is probably the UPS. This is based on your statement that the other UPS works well.

A useful experiment would be to use a Variac at the UPS input to adjust the voltage and see how the UPS responds as input voltage is varied.

.

 

[__dan]

Again I started the thread without visiting the site ...

I'm assuming you've been to the site before and were familiar enough with their layout to know the UPS was fed from something that needed an upgrade. If you truly had never been to the site before and your first step going in cold was to shotgun the board for guesses as to their problem .. I would have to say wow, you da man.

All uf this aside to correct the problem I have decided to add a sub-panel with proper grounding and dedicated circuits at the UPS room. I dont know how I can change the power coming in quality any suggestions? My concern is that the new UPS operating on the same panel/voltage is fine and this UPS is on the fritz? I hope this answers your questions

A new step down transformer with only the IT loads on it will give you some isolation from utility transients, spikes, and a lot of isolation from all of the problems on the existing transformer/panel. It does nothing for utility low voltage, outages. You can usually get a tranny with an internal electrostatic shield for a small adder. The transformer secondary common grounding and bonding busbar, the one with the GEC, SBJ attached, should be in the area right where the IT equipment is. I would bring the 480 close to the IT room and put the tranny, panel, close to the load. This gives you the common grounding point everything else bonds and grounds to. That job is a plus, an upgrade on its own merits with a long installed lifetime.

They can dispose of, change the UPS, at any time.

 

[highpowered]

Why is UPS out put only 108 volts- is there more load connected than the rating of the UPS? I would think it should be very near input voltage, or if it is supposed to regulate the supply I would think the output would be closer to 115, especially if input is at least that high.

thats the weird part? When it is drawling power from the circuit it takes in 114 volts from the outlet, but when I test the voltage on the outlets coming out tof the back of the UPS it measures 108v. When i shut the circuit off the voltage coming out of the back UPS outlets is 117v? The battery itself is producing 117v there is no power to the system?
power on 108v
power off 117v
power source 114v

 

[highpowered]

I'm assuming you've been to the site before and were familiar enough with their layout to know the UPS was fed from something that needed an upgrade. If you truly had never been to the site before and your first step going in cold was to shotgun the board for guesses as to their problem .. I would have to say wow, you da man.




A new step down transformer with only the IT loads on it will give you some isolation from utility transients, spikes, and a lot of isolation from all of the problems on the existing transformer/panel. It does nothing for utility low voltage, outages. You can usually get a tranny with an internal electrostatic shield for a small adder. The transformer secondary common grounding and bonding busbar, the one with the GEC, SBJ attached, should be in the area right where the IT equipment is. I would bring the 480 close to the IT room and put the tranny, panel, close to the load. This gives you the common grounding point everything else bonds and grounds to. That job is a plus, an upgrade on its own merits with a long installed lifetime.

They can dispose of, change the UPS, at any time.[/QUOTE

I apologize to everyone now I know what not to do here. Yes I performed an audit for this building 8 months ago and This particular panel had a hot spot concerning there server room. So I was familar somewhat with the problem. Thank you for the solution I appreciate that very much. To bring the 480v to the IT room would cost them to much they wont bite. Plus this panel is the only one backed up on there diesel generator some im stuck with it.

 

[__dan]

So, if I'm reading this correctly, I still have no idea.

The UPS is single phase, 208 volt line side with neutral. The line current is 17 amps, 17 amps, with 17 amps on the neutral (?).

Line side: L to neutral is 114 v but the load side: L to neutral is 108 v on utility.

Really need a UPS model number. The cheap disposable UPS units use an auto tap changing transformer at the output to regulate incoming voltage sags/swells. What has been described so far is the output voltage regulator is toast and the input power factor and harmonic correcting filter section is toast.

If the replacement cost on the UPS is less than $1000., they need a new one, cited previously a few times.

A new circuit from the same panel that is a long way away, the grounding does not improve. The common grounding and bonding point for the IT equipment is at the origin point of that separately derived system, not at a new subpanel from that system.

The general range of response from the beginning of this thread is the UPS is toasty, maybe by cheap design.

In addition to looking closely at the UPS to confirm the source of the problem, check the ground to neutral voltage at the UPS output. Also check the voltage from the output neutral to the nearby building steel. Anything more than 2.5 volts and they have a problem they need to spend money on.

Added: with specific UPS types, the UPS output is separately derived, putting the common grounding busbar, SBJ, GEC at the same point, in the same area as the IT equipment, which should be a cleaner design.

 

[brian john]

the company has a UPS tech on payroll and he says its the voltage? I only bring the power to these racks I have no idea how to troubleshoot once it passes through the cleaner. I used infrared on this circuit and everything is running hot. The nuetral is returning almost the same amount of power and heat the conductors are. its pulling 17 amps on all lines (hot and nuetral) on a 20 amp 208 volt 2 pole circuit. Any thoughts?

Typically UPS techs know as much about voltage as I do plumbing. They are electronic techs without a clue.

Not unusual for the neutral to be carrying high amperage due to the type of loads being served.

Without being on site and doing some basic test I have not real idea and can only guess.

1. It is not low voltage.
2. The UPS voltage settings are not correct for the input voltage.
3. There is an issue with the UPS
4. POSSIBLY the voltage sine wave on the input is distorted.

In the past issues like you describe it was the UPS and/or the UPS tech not setting the UPS operational setting properly.

 

[dave__asdf]

<--- ups tech

<--- ups tech

thanks brian....

what kind of UPS is it? how old is it? it sounds like the input filter is crap. is this a modular or APC ups? if you can gain access to the inside there are a few things we can take a look at but i need more information.

on a side note. it is VERY unlikely that the waveform on the input of the ups is affecting the operation. it's more likely that it's setup as a 240 input ups than that. also, since it is happening at the same time at night, take a look outside the box. these might be two different problems. i've had people complain about ups's tripping because teh cleaning personnel would hit the EPO thinking it was the "push to exit" button every wednesday night when the left the data center. i've also found a guy that would plug in a refrigerator every night so he could have cold drinks in the morning. these two things could be unrelated.

i think the biggest deal is that you have two, one works, one does not. what's teh difference?

 

[iceworm]

that is what the tech is saying theres a volt problem. My problem is if they spend $15,000 to upgrade there transformer and panel and the UPS is still malfunctioning they are going to be a lil pissed off. there is alot of stuff connected to the panel. There is another UPS connected to the panel as well and I believe it is functioning properly.

high -
As well they should be. Makes me never want to guess.

First thing I usually do is to look at the IOM and the specs. Yes, read the book first.

Second is to put recording monitors on the input voltage and current, and output voltage and current. That way I'm not guessing. I use a PX5. There are plenty of other good ones out there.

You said there are two units, one working. Identical units? Identical loading? Identical feeds? If you are going to compare, these would be good to know.

Be embarassing to find out the malfunctioning unit is failing because it is operating outside of it's specs.

I really would start by reading the book. If nothing else, you will know the specs for the new feeders you are having installed. Really be good if the new feeder met the UPS specs.

ice