Tear-Down: Let's Take a Trip Inside A UPS

[Guest]

In response to our power bar tear-down, many of you asked to see a UPS receive the same treatment. So, today, we go on a tour inside a vintage APC BX1000.

Tear-Down: Let's Take a Trip Inside A UPS : Read more

 

[blackmagnum]

Thanks Daniel for the interesting write up of this old beauty. I have always trusted APC for my electricity supply backups.

 

[arunphilip]

Great article, and particularly informative narrative.

 

[olsaltydog]

Good article, I find it funny you just put this up. I just turned in a 8 page research paper on UPS systems and the potential technology that could replace the traditional lead-acid UPS/ Diesel Generator combination. Could have used you in part of my paper for a citation.

 

[Daniel Sauvageau]

[quotemsg=14138873,0,1353241]Good article, I find it funny you just put this up. I just turned in a 8 page research paper on UPS systems and the potential technology that could replace the traditional lead-acid UPS/ Diesel Generator combination. Could have used you in part of my paper for a citation.[/quotemsg]
I actually sent in this article two weeks ago. It just happened to finish bubbling up the editorial pipeline last weekend and get a publication slot today.

As for replacements for the good old UPS + Generator combo, I presume you mean things like the Bloom Box - basically a fuel-cell-powered UPS - removes the lead, the acid, periodic battery replacement, periodic maintenance on the mechanical parts, etc. from the equation.

 

[olsaltydog]

Yeah one portion of the paper dealt with the Bloom Box being implemented and used by Microsoft, E-bay, and Apple. Another part of the paper was about just replacing the lead acid batteries themselves with the Durathon battery design created by GE. Only thing is GE's battery will be used in larger applications just like the Bloom Box. The last portion I researched was how Hybrid topologies where essentially a combination of of the Star and Bus/Ring topologies you could essentially create a new entire category of UPS' by integrating the UPS into the generator itself.

Im just glad the paper is done and the UPS's we have at work dont need their batteries replaced hopefully for another few years. The ones we are using have a tray of about 20-26 batteries linked together and can make a great sparkler if the terminals make contacts to the wrong parts.

 

[f-14]

Having trouble unlocking your "slightly used" UPS to take a look inside? At the top of the BX1000, there is a snap about halfway in at the top holding the shell's two halves together. Its location is circled in red on this picture. You need a long, flat screwdriver to easily reach and disengage it.[quote/]

or go to your utensil draw and grab those chop sticks you've been saving!

 

[rahulkmr391]

i have a APC Back UPS RS600 which i purchased in june 2011 and its battery has been replaced on april of 2014 it worked fine till august of 2014 suddenly it it stated to give me problems it was refusing to start up but battery power was working but it refuses to supply power to my equipment when i asked tech support they said the main Board has to be changed the reason they said to me is fluctuation killed my ups. what an utter piece of junk, my older ups served me for about 9 years with regular battery changes but the apc one died in less than 3yrs is this what apc makes quality ups..?

 

[olsaltydog]

So you had one failure with how many long lasting APC UPS' out there. UPS' have many parts that can wear or fail at some point and should not be looked at as the ultimate safeguard. They can fail and most people find that out when they need it not to. I know it sucks but this one instance you had with APC probably should not deter you from their product. Check online if the model has been having issues then yeah avoid it, no different then what we do with power supplies which is a good reason there is a list here on site. Other then that I have seen many types of there products hold strong and true for years, hope your next experience even if with another company goes better.

 

[nukemaster]

Guess I no longer need to look inside of my old XS 1200

 

[TheDraac]

Just a note: those rivets you refer to are actually screws.

 

[lalitkr]

Thanks for posting nice words..

 

[Daniel Sauvageau]

[quotemsg=14143906,0,164009]Just a note: those rivets you refer to are actually screws. [/quotemsg]
Hollow screws? Guess I'll have to take a second look.

Even on my full-resolution originals I cannot really see inside their bore but they do look like Allen-key heads. I do not remember seeing threads in those holes but do not have clear-enough back-side shots to verify that without going back in - the unpopulated holes are clearly non-threaded. Part of the reason I thought those were rivets is head shape variations - as if they had been pressed into place. I suppose this could just as easily have been part of the hollow Allen screw forming process. Now I'm going to have to remove one of these to take a look.

 

[stoatwblr]

The last portion I researched was how Hybrid topologies where essentially a combination of of the Star and Bus/Ring topologies you could essentially create a new entire category of UPS' by integrating the UPS into the generator itself.

When you go upscale, that's exactly what you get wit stuff like Caterpilar Power Systems 500kVA-1MVA units.

The UPS part is a flywheel motor/generator spinning in a vacuum. If the power is off longer than 30 seconds, a large diesel kicks in.

The whole thing fits in a standard 40 foot shipping container.

 

[Daniel Sauvageau]

Well, just checked one of those screws (the one over the unused position on the diode heatsink steel clip) and they are indeed 2mm Allen-key heads.

However, the screw's threads are heavily "beaten down" and that tells me they were used as self-tappers. Removing them pretty much ruins both whatever is left of the screw's threads along with most of the threads they formed inside the hole. If you do not want to ruin the heatsink's threads, you have to drill the head off and run the rest of the screw through. Guess that might explain why they picked hollow screws.

At that point, they might just as well be rivets.

 

[nukemaster]

I think some makers recommend against rivets(at least pop rivets) because of the sudden shock when they snap.

I have seen transistors pop riveted in place many times however.

 

[Daniel Sauvageau]

[quotemsg=14148154,0,35532]I think some makers recommend against rivets(at least pop rivets) because of the sudden shock when they snap.[/quotemsg]
I was thinking more along the lines of press-fitted rivets: load the rivet on a hydraulic press, line it up with holes, press it in, done.

Another possibility I thought of was ram-expansion: insert a hollow rivet and ram a die through the center to lock it in place. Considering how abnormal those threads felt when I removed this "screw," I would be tempted to believe this might be how these things were set: insert the "screw" and ram a 2mm hex die through the middle to expand its threads into place instead of screwing it in.

Maybe I am just over-thinking why they used hollow screws... but I cannot help thinking there must be a reason.

 

[rahulkmr391]

So you had one failure with how many long lasting APC UPS' out there. UPS' have many parts that can wear or fail at some point and should not be looked at as the ultimate safeguard. They can fail and most people find that out when they need it not to. I know it sucks but this one instance you had with APC probably should not deter you from their product. Check online if the model has been having issues then yeah avoid it, no different then what we do with power supplies which is a good reason there is a list here on site. Other then that I have seen many types of there products hold strong and true for years, hope your next experience even if with another company goes better.

today i got my ups back which service centers had replaced PCB of UPS after i inspected i was shocked they just gave me a old working unit, they just placed my unit's serial number on it and then they gave it to me. i'm seriously disappointed and i feel cheated.. i don't get it, what's wrong with these guys..

 

[olsaltydog]

[quotemsg=14146319,0,564680]

The last portion I researched was how Hybrid topologies where essentially a combination of of the Star and Bus/Ring topologies you could essentially create a new entire category of UPS' by integrating the UPS into the generator itself.

When you go upscale, that's exactly what you get wit stuff like Caterpilar Power Systems 500kVA-1MVA units.

The UPS part is a flywheel motor/generator spinning in a vacuum. If the power is off longer than 30 seconds, a large diesel kicks in.

The whole thing fits in a standard 40 foot shipping container.[/quotemsg]

The flywheel systems are nice and do handle a heavy load, they just will not replace the traditional lead-acid battery design. At this time the only system that can attempt to change that is those working with the sodium based batteries like GE. They can store the batteries outside without concern of environmental factors unlike lead acid, and they last alot longer. Office space inside a building cost a premium, storing everything outside in a utility shed costs almost nothing.

Next UPS I take apart, I will try to remember to take pictures of everything to add to this.

 

[TheDraac]

[quotemsg=14146485,0,1736083]Well, just checked one of those screws (the one over the unused position on the diode heatsink steel clip) and they are indeed 2mm Allen-key heads.

However, the screw's threads are heavily "beaten down" and that tells me they were used as self-tappers. Removing them pretty much ruins both whatever is left of the screw's threads along with most of the threads they formed inside the hole. If you do not want to ruin the heatsink's threads, you have to drill the head off and run the rest of the screw through. Guess that might explain why they picked hollow screws.

At that point, they might just as well be rivets.[/quotemsg]

I had an RS 1500 that just stopped working one night. After some dis-assembly and some research, it was at least one of those components that went bad. I was able to remove the screws so I could desolder all the components, think it was 6 of them. I think 2 tested out bad, but I replaced all of them, fresh thermal paste where needed and was able to screw em back onto the heat sinks.

It's still working now, 2 years later, a new set of batteries and its gotta be over 10 years old by now, I bought it used on ebay as it was.

 

[BobSmith809]

Thanks for the great article! I had to comment here as I have this exact model of APC 1000 still running strong after over a decade. And believe it or not, the ORIGINAL battery is still working although it was offline for 3-4 years while I was using a Belkin (which was quieter but charging circuit died .....lost its ability to charge a new battery after about 6 years) and when I re-connected the Belkin to my system (Core i7 loaded to the gills +accessories) I was assuming I'd have to replace the battery but so far (going on 3 years) it's still holding a good charge and, utilizing Powerchute monitor, it passes self-tests 5x5.

 

[gawkowsk]

the poor performance of the replacement (not original) batteries can have quite different origin than their quality. I've experienced that in APC UPS 1500 and Smart 1000 as well - after replacement the UPS was calibrated with the PowerShute, however, it was keeping up only for 8 minutes. Then, after some googling I've found the solution - the UPS has to be calibrated quite different way: after charging up you have to bias it with constant current (like a 100W bulb) while totally unplugged (from the power source, computer etc.) - it will be discharged till it shut down by itself. After this my UPS started to keep my system for 40 minutes. It seems like the PowerShute's calibration is not sufficient in case of battery replacement.

 

[Daniel Sauvageau]

[quotemsg=14181210,0,560548]It seems like the PowerShute's calibration is not sufficient in case of battery replacement.[/quotemsg]
I doubt it is a calibration issue in my case since the UPS starts beeping the low-battery alarm within seconds from switching to battery power, meaning battery voltage is dropping close to the UPS' minimum operating voltage as soon as load is applied. No amount of calibration is going to magically change the UPS's under-voltage lock-out and low-battery alarm thresholds. Once battery voltage starts collapsing (around 1.8V per cell for lead-acid), the battery is pretty much empty... this is not something you can calibrate your way around of.

When my cheap replacement batteries were fresh, it took 2-3 minutes for the low-battery alarm to go off. On the original batteries, it only started after ~10 minutes for the few times I had to dare stretch my luck that far. That was my first clue about how my new batteries were nowhere near the same performance class as the originals - but I already suspected that much from past shopping for a battery to put in my BK650.

The last time I had a power outage, it took me about three minutes to put my computer in hibernation and the UPS shut down from under-voltage lock-out just as I was about to press the UPS' button to turn it off. That's where my 3-4 minutes run-time came from; not PowerChute's estimation.

 

[PaulR08]

Great detail on the components inside a UPS, how I have a better understanding of how they work.

 

[gawkowsk]

[quotemsg=14185846,0,1736083]
I doubt it is a calibration issue in my case since the UPS starts beeping the low-battery alarm within seconds from switching to battery power, meaning battery voltage is dropping close to the UPS' minimum operating voltage as soon as load is applied. No amount of calibration is going to magically change the UPS's under-voltage lock-out and low-battery alarm thresholds. Once battery voltage starts collapsing (around 1.8V per cell for lead-acid), the battery is pretty much empty... this is not something you can calibrate your way around of.

[...]That was my first clue about how my new batteries were nowhere near the same performance class as the originals - but I already suspected that much from past shopping for a battery to put in my BK650.

[...] That's where my 3-4 minutes run-time came from; not PowerChute's estimation.[/quotemsg]

I agree that the original batteries are more suitable for the UPS - they discharge characteristic is "fitted" in the UPS control logic. That is why you observed very short uptime after the replacement - UPS simply thinks that the battery is discharged - no doubt about it. I had the same experience. However, during the calibration (made my way - not a powershute one) the UPS also started to signal that it is discharged after a few minutes - however, the power at the UPS output was still provided - the bulb was still lighting for almost an hour till the full UPS shutdown. It is important that during this process there is only UPS and the discharging bulb connected - no power at UPS input, no connection to the computer - just a constant current discharging device (bulb in my case). After this process the UPS learned this new battery characteristic and knows that even the voltage is lower (comparing to the original battery) the battery is not discharged yet. That is the idea - no matter the battery is (different discharging characteristic) it has to keep the declared parameters (power capacity in this case) - it can behave differently but the total power should be the same - it only a matter if the UPS can recognize and work with this new characteristic. It worked for me for two different APC UPS models. The powershute calibration is very short so it seems like it is just a "characteristic perfection" as the battery life goes (here, the battery characteristic also changes), but apparently (in my case) UPS needs different type of calibration to learn the characteristic of a new, not original battery. Just give it a try and let me know the results - I'm really interested. Regards. Peter