Best Practice: LeaveOn or TurnOff after Windows Shutdown?

[BillP]

With my old computer I use a Back-UPS XS1000 to power a PC and two 19-inch CRT monitors. My practice has been to shutdown Windows and turn off the UPS at the end of each day thinking that this practice gives me better protection when the PC is not being used. At work everyone seems to leave their UPS on overnight, and most shutdown their PCs and leave their LCD monitors on. I recently purchased a Smart-UPS DLA1500 (similar to the SUA1500) for a new PC I am building. It will power the new PC and two LCD monitors. I am considering the practice of leaving the UPS on and turning the LCD monitors off after I shutdown the PC. What is the best practice for leaving an UPS connected to an unused PC and monitors overnight?

[BillP]

With my old computer I use a Back-UPS XS1000 to power a PC and two 19-inch CRT monitors. My practice has been to shutdown Windows and turn off the UPS at the end of each day thinking that this practice gives me better protection when the PC is not being used. At work everyone seems to leave their UPS on overnight, and most shutdown their PCs and leave their LCD monitors on. I recently purchased a Smart-UPS DLA1500 (similar to the SUA1500) for a new PC I am building. It will power the new PC and two LCD monitors. I am considering the practice of leaving the UPS on and turning the LCD monitors off after I shutdown the PC. What is the best practice for leaving an UPS connected to an unused PC and monitors overnight?

[Anonymous user]

It's not necessary to do what you're doing unless you're trying to kill power going to the PC.

[TheNotoriousKMP_apc]

Hello,

You're going to degrade the life of the UPS over time if you power it on and off every day after use. It's designed to be on pretty much all of the time, unless the batteries are discharged during a power outage and the UPS has no choice but to turn off.

[Anonymous user]

If the UPS battery is full and power goes out and the power supply inside your PC, LCD... is also at stand-by and there is very small load on the UPS outlets, it can make the voltage on the PSU Bulk capacitors exceed it's nominal ratings!! This only happens with approximated sinewave units. I recommend everyone to measure the VDC (Direct Current Voltage) at the the capacitors terminals while there is no load on the UPS outlets, just you PC power supply being fed by the battery. You guys will be surprised!

I've done that and found out that on a 200V capacitor, voltages exceeding that are measured while on battery and the PSU of your PC is at stand-by.

There should be a recommendation not to leave your UPS on when no considerable load is drawning power from the UPS. I don't know why APC don't make such recommendation. There is a long discussion about it on badcaps forum, on jonhy guru and on a brazilian forum.

This is a fact so APC should start putting this warning on their manuals!

Message was edited by: rau

Message was edited by: rau

[Anonymous user]

The Notorious K.M.P. wrote:
Hello,

You're going to degrade the life of the UPS over time if you power it on and off every day after use. It's designed to be on pretty much all of the time, unless the batteries are discharged during a power outage and the UPS has no choice but to turn off.

It shouldn't. I really should rip one apart and research the cyclical life of the relays involved in turning the power on/off through front panel. As long as the UPS is plugged in, the floating voltage charger is live. The on-off switch simply turns off the output side. If you can, such as on the SmartUPS, you would want to turn off the "self test at each power up" though, as relays are cycled during self test and you'd effectively be using relays twice every time you turn on if the self test is enabled and this will double the wear and tear on relays.

If you're having many 2-3 minute outages and you anticipate you're going to have them over and over and cause the battery to discharge before it can catch up, the correct action (for those who does not absolutely require the system to stay on 24/7) is to shut down the system, then use the front button to turn off the UPS, so the UPS can continue to charge when power is available, but not deplete battery power during each episode of outage by powering up the load each time. On Smart-UPS that supports full smart protocol, the "E" parameter should be set to "OFF".

[TheNotoriousKMP_apc]

The point I'm trying to get at here is the premature failure of electrical components which are not designed to last forever within a UPS. Take this for example:

UPS's typical life expectancy based on component usage is 4-6 years, we'll take the low end of 4 years for this discussion.

Within that four years, let's assume 2-3 outages a year (let's take the high of 3). So in its lifetime, you're engaging the transfer switch, inverter and rectifier 12 times throughout the 4 years. After 4 years, the inverter fails over and will no longer work as designed. Time for a replacement UPS.

Now let's look at what the poster was trying to get at:

Turning the UPS on EVERY day. 365 turns on/off a year. You're now engaging the switch, inverter and rectifier 365 times in one year, or an astounding 1460 times in a 4 year span. That is over 1000 times more than the failed component in the 4-year scenario.

Whether it degrades the life of the battery, we'll leave that for KVAr's expertise to state. However, there's no denying that ultimately, using the UPS as a master on/off, 365 days a year is going to significantly degrade the life expectancy of the UPS.

[Anonymous user]

The Notorious K.M.P. wrote:
The point I'm trying to get at here is the premature failure of electrical components which are not designed to last forever within a UPS. Take this for example:

UPS's typical life expectancy based on component usage is 4-6 years, we'll take the low end of 4 years for this discussion.

Within that four years, let's assume 2-3 outages a year (let's take the high of 3). So in its lifetime, you're engaging the transfer switch, inverter and rectifier 12 times throughout the 4 years. After 4 years, the inverter fails over and will no longer work as designed. Time for a replacement UPS.

Now let's look at what the poster was trying to get at:

Turning the UPS on EVERY day. 365 turns on/off a year. You're now engaging the switch, inverter and rectifier 365 times in one year, or an astounding 1460 times in a 4 year span. That is over 1000 times more than the failed component in the 4-year scenario.

Whether it degrades the life of the battery, we'll leave that for KVAr's expertise to state. However, there's no denying that ultimately, using the UPS as a master on/off, 365 days a year is going to significantly degrade the life expectancy of the UPS.

Kevin,
I know you're a tech support, but are you involved in the actual operation of the UPS beyond what is visible to the user? I rip 'em down to components.

Under default setting, the transfer switch operates a MINIMUM of 26 times a year. APC Smart-Protocol UPSs are programmed to run a self-test every 336 hours, unless you set parameter E to "OFF", however the "E" parameter can be set to every 168 hours too, so if the engineers knew what they're doing the UPS was designed to accommodate a minimum of 52 engagements a year.

So perhaps the MTBF of the UPS will be reduced to 8 years as opposed to 16 years.

All 120v plug-in Smart-UPSs I've ripped apart uses the same transformer for charging/discharging and the MOSFETs are utilized for both inverter and charging functions.

What's actually harmful to the relays is when the AVR function hunts hundreds of times a day due to a presence of laser printer and such on the same circuit and the sensitivity is set too high or the AVR low voltage engagement threshold is too high. This situation is like when your automatic transmission can't decide a gear and goes in and out of over-drive. With cars, you resolve it by manually disabling over-drive and with APC UPSs, you resolve this by isolating the problem load and putting it elsewhere, or "turn a blind eye" by changing the sensitivity.


The MOSFETs and transformer stays on 24/7/365 without a regard to the front switch status. The only way to deactivate these components is to turn off the power, then unplug the UPS from the outlet. So the inverter components don't sit "idle" like an emergency generator.

[TheNotoriousKMP_apc]

You're right. I need to re-do my math on the UPS on the four year span. I'll work on that and come up with something a little more defined tomorrow afternoon.

My apologies for forgetting about the 2-week self-test (which is even more remarkable since you and I just discussed the effects of a Self-Test on another thread).

Anyways, no I am not directly involved in the actual operation. Sure we can take apart units as desired and learn the components. But since the components are mostly manufactured overseas, we learn what can cause what to fail along the lines of power flow, and leave it at that. I don't have an EE degree so it doesn't stimulate me as much as it probably should. I have an IS degree with End User focusing. So I try to emphasize my troubleshooting side.

[Anonymous user · Accepted solution]

The Notorious K.M.P. wrote:
You're right. I need to re-do my math on the UPS on the four year span. I'll work on that and come up with something a little more defined tomorrow afternoon.

My apologies for forgetting about the 2-week self-test (which is even more remarkable since you and I just discussed the effects of a Self-Test on another thread).

Anyways, no I am not directly involved in the actual operation. Sure we can take apart units as desired and learn the components. But since the components are mostly manufactured overseas, we learn what can cause what to fail along the lines of power flow, and leave it at that. I don't have an EE degree so it doesn't stimulate me as much as it probably should. I have an IS degree with End User focusing. So I try to emphasize my troubleshooting side.

Care to explain where you got a lifespan of "4-6 years" ? APC publish an expected battery life of 3-6 years. It sort of doesn't make sense to offer "user replaceable battery" system when the useful product life is the same as the useful battery life, does it?

I can't say this for the modern outsourced plastic chassis cost reduction model units, however the US made, metal chassis commercial grade APC UPSs are very well made. I just pulled the SN on the UPS this PC is on via Telnet and it's a 1998.5 vintage. It did get a new battery last year and it's still working perfectly and the runtime tests within original specs.

This unit can no longer be managed on today's consumer PCs as most of them don't have a serial port and this UPS can only talk through serial interface. Unlike the rest of IT equipment, UPSs haven't really changed in its core components in the last 15 years or so.