Liebert UPS, over voltage.

[ron]

Depending on the amplitude of the overcurrent, if it lasted several minutes, the bypass overcurrent protection should have tripped, not the UPS shut off..

 

[cschmid]

I think the problem you describing could be associated with battery maintenace..If the batteries are only lets say 50%..if the load is exceeded the machine will shut down..just because the machine says they are good does not mean they are at 100%..I am just throwing that out for an idea..

 

[SmithBuilt]

Depending on the amplitude of the overcurrent, if it lasted several minutes, the bypass overcurrent protection should have tripped, not the UPS shut off..

I believe that is what happened. But to reset the bypass ocp the UPS had to be restarted. The UPS was shut down and restarted manually after the load had been dropped.

cschmid

Wouldn't/shouldn't the UPS go into static bypass instead of shutting down.

 

[cschmid]

only if the fault was cleared..

 

[catchtwentytwo]

I think the problem you describing could be associated with battery maintenace..If the batteries are only lets say 50%..if the load is exceeded the machine will shut down..just because the machine says they are good does not mean they are at 100%..I am just throwing that out for an idea..

I agree 100%, many UPS issues are caused by the poor battery maintenance. Even with good maintenance, the typical VLRA Batteries used in many UPS are prone to unpredictable failures. They are supposed to be inspected quarterly and have an annual PM per IEEE standards. With a single module/single battery system UPS, it is even more important.

And don't believe the sale literature that boast of a 10-15-20 year warranty. Typical VLRAs last 4 years at best. If the facility has a lot of power "hits" and no generator backup, less. The warranty is prorated and is basically a sales tool.

 

[cschmid]

I agree with that..the shelf life of batteries is actually really low even with the new battery technology..

 

[weressl]

My guys were doing some work at one of my customers. They have a Liebert UPS feeding their data center. It registered an over voltage and went on bypass (utility feeding the data center). They argue that we caused the shut down by shorting something out. My guys say they did nothing to cause the problem. They were not working on any circuits fed by the UPS nor anything live.

The customer rushed to the UPS and tried to get it back online. While they were doing this the UPS disconnected the power to the data center. They argue that they did not cause the shutdown. It cost them a lot of money and many unhappy customers.

My first question is what would cause an over voltage? Could a large load that is shut off cause it? A short? There was also talk of a couple of brownouts the same day.

Second, is it possible that the customer did make a mistake and shut off the power to the data center or the UPS malfunction? I'm researching the Liebert site now.

Little basic about UPS'.

An UPS should NEVER interrupt the critical load unless there is a fault in either in the connected system, or a fatal failure within the UPS system.

Normal mode of operation is to supply power through the protected controlled path, that is the rectifier/DC-bus/inverter.

In case of power failure the ups would switch onto the batteries.

When the batteries are depleted and the normal power had still not returned, the UPS would switch over to the bypass source, if such is avaliable.

Should there be an overload or a short circuit on the critical load, the UPS would attempt to switch over to the bypass to either; supply the overload until it is corrected, or to 'burn through' the short circuit. As soon as the condition disapepars tehg UPS would trasnfer back to the normal path of power supply.

The UPS will not transfer to the bypass source if it is 'outside' of the preset parameters.

The UPS will transfer to the battery source if the normal input source is outside of the acceptable parameters.

The tolerance of voltage/frequency deviation (acceptable parameters) are much stricter for the bypass source than for the normal input source. The reason for this is that the rect/DC/inv path rpovides correction for those, and the bypass is straight through with maybe some surge protection of you lucky.

 

[dbuckley]

Static bypass on a UPS is done using electronic components, and these have definite problems handling overloads, and dont like getting hot. The static switch elements are (in my experience) always sized for a greater load than the UPS can supply, to allow for fault clearing, but if the switching elements get too hot, they will get switched off to prevent damage to the switch. And that means load is dropped.

Having the UPS room aircon fail is a good way of determining how hot static switches don't like to get. First the inverters stop, and the UPS goes into static bypass. All good so far. Then the static switch reaches its maximum allowable temperature, and thats when the data centre becomes eerily quiet...

 

[dave__asdf]

Having the UPS room aircon fail is a good way of determining how hot static switches don't like to get. First the inverters stop, and the UPS goes into static bypass. All good so far. Then the static switch reaches its maximum allowable temperature, and thats when the data centre becomes eerily quiet...

i'm a ups tech and i've worked on just about every type of UPS including a lot of lieberts. it sounds like you have either a ap366 model or ap600. lieberts in general have a static switch that is only a partial duty, it shuts just long enough for the bypass contactor to shut, then opens. MGE UPS's have a full duty static switch that is on when in bypass until the UPS bypass breaker is manually shut (Q4N or Q4S i think...). Most UPS's built today are microprocessor controlled. lieberts however (with the exception of n-power) are still fully analog. if you look inside the unit there's about 25 little blue pots that control how the UPS acts. lieberts are notorious for having bad connectors. it was mentioned earlier how opening up a door caused the UPS to EPO. this is not uncommon. a lot of times since these machines are analog a bad connector can cause a false reading to do similar things to the control circuitry. the worst part about an analog controlled UPS is there is no fault history. they will store alarms, such as over current but it is not often easy to tell why it went into that condition. was there an alarm "hardware shutdown" in the history? this is one of the only alarms that liebert has that tells you there is something wrong with the UPS logic.

if the customer is still looking for resolution let me know the model number, i'll let you know the most common failures for that model.

Liebert techs are generally good, but sometimes get used to doing PM's only and passing problems on to the local mr. fixit for the company.

also, do you know how much load was on the phase during normal operation before it tripped and what the rating of it is? i would guess that a 20a breaker supplied the motor or piece of equipment that is being blamed for the outage, but generally 20a of added load is not enough to put the liebert up to 150% (it is rated for 10 minutes @150%)

 

[zbang]

I hate to disagree, but pretty much all of the Lieberts I've seen in the last 10+ years have at least one microprocessor buried somewhere inside. For one thing, it's mighty hard to do serial-port-based monitoring/control without it. They may use analog controls for parts of the converters, but definitely have microprocessor monitoring. (This doesn't imply that they keep an internal log, but any data center operator with an ounce of sense has their UPS connected to a monitoring system of some sort.)

I think you'd have to go back to the 1980's to find a UPS that didn't have an embedded controller. The Deltec's and Emersons of the early '80s come to mind.

 

[SmithBuilt]

also, do you know how much load was on the phase during normal operation before it tripped and what the rating of it is? i would guess that a 20a breaker supplied the motor or piece of equipment that is being blamed for the outage, but generally 20a of added load is not enough to put the liebert up to 150% (it is rated for 10 minutes @150%)

The UPS had plenty of capacity. It was running at 71%. I agree that a 20a load should not matter.

The Liebert tech stated that the UPS would sense a motor load as a short. That there have been several cases of someone plugging in a vacuum cleaner and causing the UPS to go into bypass.

 

[rdelauter]

I have heard many stories of this, buffers and vacuum cleaners faulting off UPS systems as well but normally these are smaller systems. If you said I missed how large the system was so 71% of what could matter.

I have seen a 20 amp circuit shorted to ground on the output of a PDU cause a relatively large Liebert 600 series UPS, think it was 150 KVA, transfer to bypass, sometimes they auto transfer back and sometimes they don't, depends on many things.

When I started working on UPS systems, one of the golden rules were, "the system is designed to protect the load as long as it can without destroying itself." I have used this many times when trying to determine why a system failed the way it did and caused a load to drop.

 

[SmithBuilt]

It is 100kva Liebert.

The UPS went into bypass due to overvoltage at 1:20:28PM and then tried to go back online at 1:20:33 but did not. Then about 5 minutes later the static bypass dropped the load. The Liebert Tech called Liebert tech support who said when the UPS sensed the circuit was ok then the UPS would attempt to transfer back online. That only happened once 5 seconds after the initial problem. The tech stated that the UPS would not record the static bypass dropping the load and there was not record on the log. Tech support stated that the overcurrent was happening from the time it recorded the problem until the static bypass dropped the load.

I don't understand why the UPS cannot clear a 20 amp fault or motor load. It looks like they should be built to protect the load better. Maybe design it to use static bypass until maintenance bypass can take over the load.

I appreciate the help I would like to understand what the problem was. The main thing is my company is in the clear as far as the power outage.



Edit: I'm not sure what model UPS it does have the wheel for start up ect.

 

[dave__asdf]

i know the two chips you're talking about. i troubleshoot to the component on lieberts. all those chips are used for is logic comparison. they're microchips not microprocessors. with microprocessor control you can store settings and adjust voltage using the controlling microprocessor. if you look at the front of your liebert there will be a voltage adjust knob with a fine and coarse adjust on it. if you open up the door you'll see all the blue potentiometers that adjust the operating parameters of the machine.

the reason the stuck with them for so long (or so i've been told) is when liebert first came out they sold a whole bunch of machines. when people wanted to upgrade, they wanted something that was compatible with their old machine so liebert just kept "updating" the old boards and ups's but kept most of it the same.

 

[dave__asdf]

that previous post is a response to the last post on page 3.... i missed page 4 all together. i'm not too forum savvy as you can tell by my upcoming "quotes"


quote: smithbuilt
Edit: I'm not sure what model UPS it does have the wheel for start up ect.

that's an AP366.

i haven't seen as many problems with the 366 as the 600 as far as logic errors but i have replaced a lot of capacitors in them. the output filter might have had something to do with the over current.
there's a couple banks that you can easily inspect with the unit online by opening up the door. with a flashlight you can look at the ac cap banks in the bottom. you'll be looking for oil on the top of it. it might be possible that a bad filter bank is putting too much load on the inverter but dont think the bypass uses the filter...


was it an input or output over voltage? or current?

the best thing to tell the customer is buy a new mitsubishi ups.

 

[SmithBuilt]

was it an input or output over voltage? or current?

the best thing to tell the customer is buy a new mitsubishi ups.

It was over current on the output.

I have wired several mitsubishi air condition units for an hvac contractor I work with a lot. If the rest of their(mitsubishi) products are as good I'm sold.

 

[dave__asdf]

the ups's are bullet proof, the new big toshibas are the same thing. my company does maintenance for a guy that's had a mistu ups for 20 some years and never had a failure. we've tried to sell him a new one but he'd rather not mess with something that's never been down.

 

[catchtwentytwo]

Am I correct that this UPS is a 300 series?

If yes, Liebert lists it as a discontinued product but there seems to be plenty of documentation available:

discontinued Series 300 Dual Input IOM SL-24537_R07-97.pdf

discontinued Series 300 IOM SL_24535_R07-97.pdf

discontinued Series 300 Load Bus Synch operation manual sl-24536_R05-00.pdf

In my experience, UPS manufacturers seem to have a "don't ask, don't tell" philosophy with "known" issues. Unless the customer/owner pushes VERY hard, the truth often doesn't come out. The owner usually has to go above the technician level to get the truth.

 

[jstrick2]

I will add that the facility I worked at had tons of issues with Liebert UPS. Batteries commonly failed about 3-4 years out, multiple capacitor failures, one leading to a fire alarm going off due to smoke.


Solution?


Installed a make before break bypass switch for the maintenance/bypass switching. We assumed they would fail, just wanted to be able to pull the UPS out and do a complete replace if needed while running on "dirty" power.

UPS is a "fun" game to play

 

[elmorseyma]

Common problem in Liebert

Common problem in Liebert

Leibert UPSs have these common problems; this is done many times in data centers and core sites in the major GSM companies in world.
in a famous case Leibert (high pulse 160KVA) transferred to bypass with instant of interruption occurred in the UPS output that caused complete outage in data center.

These cases are repetitive in all GSM companies, some GMS operators decided that "Liebert in black list"