none
Dead Power Supply on Mediasmart Server, HP tells me do not use with a UPS! RRS feed

  • Question

  •  

    I had my server fail yesterday and called HP Tech Support.  They asked me if I used a UPS and I said I did.  They advised me that the Mediasmart 470/5 server for WHS does not work well with a UPS and not to use it as this may be the cause of the failure.  The Tech said it had a 60 Watt Power supply that was very sensitive to failure if hooked to a UPS. Suggested I just use a surge protector and no UPS.  I was very suprised since I had never heard this.  Has anyone else experienced or heard of this?  Very bad if it is a problem since you would expect a server to be backed up by UPS.
    Friday, June 20, 2008 5:09 PM

All replies

  • Richard,

    It's been mentioned a couple of times in the past. However, as you say, it's a poor system that won't allow a UPS. I can't imagine why a power supply should be sensitive to a UPS anyway, in my mind, it should be just the opposite, as most UPS's have a smoothed output which can be better than the mains.

     

    Colin

     

    Friday, June 20, 2008 5:21 PM
  •  

    This is also why I don't call support or buy computers where the parts are proprietary, the support sucks and proprietary parts are a pain to obtain.

     

    I don't believe for a second that a UPS would cause a power supply to fail, but if it somehow did, that's a POORLY designed system.

     

     Richard Plenty wrote:

     

    I had my server fail yesterday and called HP Tech Support.  They asked me if I used a UPS and I said I did.  They advised me that the Mediasmart 470/5 server for WHS does not work well with a UPS and not to use it as this may be the cause of the failure.  The Tech said it had a 60 Watt Power supply that was very sensitive to failure if hooked to a UPS. Suggested I just use a surge protector and no UPS.  I was very suprised since I had never heard this.  Has anyone else experienced or heard of this?  Very bad if it is a problem since you would expect a server to be backed up by UPS.
    Friday, June 20, 2008 5:30 PM
  • I think that what HP-supprt was (or should be saying) that It all depends on the quality of both the UPS and the power supply. With good commercial UPS you probably will be fine, but that will cost. Most cheep UPS's for the home use are far from ideal.


    Problems can arise when switching to backup power when the main power goes down. For some UPS types there will be a small delay (typically 1-10 msecs) in switching to batery backup which will cause a small spike on you power unit (which is know to cause problems including data-corruption in some cases btw). Next there is the quality of the power during battery operation that can disrupt proper operation of a PC's power unit).

     

    For more ready look here and here...

    Friday, June 20, 2008 10:41 PM
    Moderator
  • I've been running one of these 'cheap' UPS's as you call it just fine for over two years through several power outs and brown outs with no issues. I believe this is because the power is coming from the DC batteries inside as opposed to direct AC from the outlet. Any UPS worth anything would run this way. I've yanked the power connecter out and back in several times while transferring files for testing purposes with no issues.

     

     Theo van Elsberg wrote:

    I think that what HP-supprt was (or should be saying) that It all depends on the quality of both the UPS and the power supply. With good commercial UPS you probably will be fine, but that will cost. Most cheep UPS's for the home use are far from ideal.


    Problems can arise when switching to backup power when the main power goes down. For some UPS types there will be a small delay (typically 1-10 msecs) in switching to batery backup which will cause a small spike on you power unit (which is know to cause problems including data-corruption in some cases btw). Next there is the quality of the power during battery operation that can disrupt proper operation of a PC's power unit).

     

    For more ready look here and here...

    Friday, June 20, 2008 11:02 PM
  •  Theo van Elsberg wrote:

    I think that what HP-support was (or should be saying) that It all depends on the quality of both the UPS and the power supply. With good commercial UPS you probably will be fine, but that will cost. Most cheep UPS's for the home use are far from ideal.

     

    I agree. UPS quality counts and so does the type and quality of power output in some cases.

     

    The HP EX Home Server series uses a power factor corrected power supply which I know because I measured the power factor to be 1.0, this is a good thing, less than 1 is not as good. This surprised me because every PC I have in the house (12) has a PF of around 0.6 to 0.7 which is normal for a switching power supply. All commercial servers have PF of 1.0 and it costs money to do that, so the HP Home Server really is a server! PF of 1 means the current and voltage demands are 100% in sync. This puts less of a strain on the power grid by producing less total harmonic distortion (THD). Typical PC's with PF of less than 1 produce a lot of THD and collectively can cause problems on the grid and for UPS's.

     

    There are two types of UPS used in the home, stepped square wave output, and sine wave output. All APC Back-UPS UPS's produce a stepped square wave output. This is less expensive but produces a lot of THD. That is okay though for use with a switching power supply because switchers only take power from the line at the positive and negative peaks of each AC cycle. So a switcher doesn't care if the peak is from a square wave (UPS) or a sine wave (AC line). The other type of UPS produces a sine wave output. All APC Smart-UPS UPS's produce sine wave output with low THD and they cost more. I have a Smart-UPS 750 XL powering three PC's in my 'Digital Command Center' at home. I use 'apcupsd' to coordinate shutdowns for all three PC's.

     

    I replaced an old CyberPower UPS. It has a warning on the bottom that says it has THD of 39% and not to use it on anything but a PC. The Smart-UPS has THD of about 3% max. It can run any electrical device within its VA rating.

     

    Now, to the question at hand. Did your UPS blow your EX power supply? Maybe if your UPS is a cheap unit that has way high THD or large transfer spikes. The fact that the EX475 has a PF corrected power supply tells me it is not a cheapo P/S and in fact must be pretty descent. It should not have any trouble running on a good quality UPS, square wave or sine wave. I recommend APC as I have never had trouble with any of the several models I have used over the years. I have had trouble with small IBM, Tripp Lite and CyberPower UPS's and have since replaced them with APC.

     

    I had an old APC Back-UPS Pro 650 on my HP EX475 for over a year. No problems. I just replaced it with an APC Back-UPS XS 1300VA (BX1300LCD) to get more run time and to get the cool LCD display. I am using the GridJuction Add-In to manage the UPS.

     

    In addition to power factor, crest factor comes into play when using a UPS. Below I include an excerpt from an APC white paper discussing crest factor as it applies to PC's. (I am not an APC employee, but that is the UPS I like to buy and use).

     

    "In addition to a low power factor, some computer loads are also unusual in that they exhibit a very high crest factor. Crest factor is the ratio between the instantaneous peak current required by the load and the RMS current (RMS stands for Root Mean Square, which is a type of average). Most common electrical appliances exhibit a crest factor of 1.4 (1.4 is the ratio of the peak value of a sine wave to its RMS value). Computers and IT equipment with Power Factor Corrected power supplies exhibit a crest factor of 1.4. Personal computers and stackable hubs exhibit a crest factor of 2 to 3.

    When a load exhibits a crest factor of more than 1.4, the source (UPS) must supply the peak current desired by the load. If the source does not supply the current, then the source voltage will become deformed (distorted) by the excess peak current. Therefore, if a UPS is not sized to supply the crest factor desired by the load, the output voltage waveform of the UPS will be distorted.

     

    The crest factor requirement of a computer load will vary depending on the source which it is supplied from. The crest factor may even vary when the computer load is moved from one AC receptacle to another in the same room. It is widely believed that the crest factor is an inherent characteristic of a computer load, when in fact crest factor results from an interaction between the load and the AC source. The crest factor required by a computer load depends on the AC source waveform. For a sine wave source, a non-power factor corrected personal computer will typically exhibit a crest factor of 2 to 3. For a source waveform which is a stepped approximation to a sine wave, as used in most UPS below 1kW, a computer will exhibit a crest factor of 1.4 to 1.9. It is widely but mistakenly believed that it is desirable to operate a computer at as high a crest factor as possible. In fact, computer manufacturers go to great lengths to reduce the crest factor of the computer because high crest factor causes overheating of power supply components.

     

    The reduction in crest factor which occurs when a computer load is operated from a UPS, surge suppressor, or power conditioner is a positive side benefit, except if the reduction is accompanied by excessive distortion of the input voltage waveform to the computer load. Such distortion may result in a significantly reduced peak voltage being supplied to the load, which is equivalent to a brownout condition. The UPS or line conditioner must be designed to maintain the proper peak voltage.

     

    Typical sine wave UPS systems such as the APC Symmetra or Smart-UPS models have a crest factor capability of approximately 3 when operated at full load, 4 when operated at 1/2 load, and 8 when operated at 1/4 load. Typical smaller stepped wave models such as the APC Back-UPS have a crest factor capability of 1.6 at full load and 2 at 1/2 load. UPS systems with this performance will maintain the proper peak voltage into the computer load for computers with any input crest factor specification. Lower quality UPS systems on the market have limited peak output current capability and consequently low crest factor performance and will distort the output voltage under personal computer loads. Typically this does not cause a malfunction with smaller PC installations. However in large multi-PC installations such as call centers can cause significant degradation of the UPS output waveform. In these cases it is not uncommon to see a UPS output waveform that is nearly a square-wave. This situation can cause various types of office machines to malfunction.

     

    The fact that virtually all equipment installed in data centers today such as servers, routers, and storage devices are power factor corrected has virtually eliminated crest factor as a problem in the data center. Crest factor problems today are limited to call center or trading floor installations with a high density of Personal Computers."

     

    Saturday, June 21, 2008 1:46 PM
  • There are issues that may result from using a very inexpensive UPS with a computer. Very inexpensive UPSes almost always output a square wave, or a "modified square wave" or "modified sine wave" (which are the same thing, in UPS-speak). PC power supplies are built to expect a sine wave as input, so they tend to run hotter and be less efficient. A small PC power supply has smaller components, and less ability to deal with dirty power. I can conceive of a PC power supply failing eventually as a result, but I think it would take more than a couple of hours...
    Saturday, June 21, 2008 1:58 PM
    Moderator