Let's Take a Trip Inside A Power Strip!

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
I'd like to but I can't get half way through it. Keep getting error messages to "recover Web Page" then a message telling me that the site keeps having issues. I'd like to finish what is shaping up to be a very good article. Could use some help from the inside!
 

A few people prefer reading picture stories in "print" format and after taking a look at it myself, I can understand why - particularly for picture stories with lots of accompanying text. You might want to give that a try - the print format button should be on the floating bar to the left of the page's content.
 


Thanks Dan! I'll give that a try. I am curious to see how this thing turns out. Paul
 
It looks like human error. The board was machine wave-soldered then reworked by hand, ie the overabundance of flux resident in places. Makes sense since the cold solder joint and the poor wire feedthrus were localized in the three power leads. The cold solder joint also exhibited movement in the wire as it had cooled - the volcanic rim. I knew that NASA soldering school the USAF sent me to forty years ago would pay off some day.
 
A better fix would to bridge the gap with a wire of the correct size based on the table below. This would restore the inherent protection of the design. Note you should only use the method in cases like this and DO NOT use it to replace a Fuse or Circuit Breaker!
 

You did not link "the table below" but assuming you meant the AWG table, that would be incorrect: the fusible trace is meant to act as a slow-blow 20-or-so-amps fuse while the AWG table defines very conservative safe values. The reason it blew up was that the short in the faulty PSU I plugged into it caused current to rise much faster than breakers could interrupt and likely peaked well over 15A before the trace blew up.

A single piece of #14 wire will pass over 100A for a while before heating up enough to fuse off even though it is only rated for 15A safe operation in fiberglass-insulated walls when run in bundles of three cables. It takes a while to burn #14 wires off even under extreme fault conditions and this is what gives breakers plenty of time to react before there is any real fire hazard.

 
I have a similar product. After reading this article, I will be opening my unit up to inspect the soldering, which is the bulk of the complaint you wrote about.

The 15amp circuit breaker is to blow, not on a power line surge, but if the total demand of the appliances plugged into the unit reaches or exceeds 15amps.

What has APC responded as an explanation I am sure you must have advised them of what you found?

I have found APC to be a product pusher, and not much else. As an example, I bought a 550watt ups, and later bought a 650 ups that was rated for more power and longer minutes of protection. I opened each to examine the battery and the electronics, and amazingly, the batteries were of the same amp hours (and part number). The transformers and electronics were similar. What this discovery told me that the more expensive UPS was not better than the lower cost unit, or there is a false claim to functionality (a 650 watt giving the superior performance of a 550 watt, while having the identically sized battery to the lower rated unit.

 
APC uses the same batteries in many UPS setups.

The actual run time is also effected by the type of inverter in use. Some will last longer than others because they are more efficient.

For a normal computer 550 or 650 should still allow ample time to shut down. With so many power supplies not liking the simulated sine wave/stepped approximate wave/ect, It is best to save and shut down fast anyway. I have an APC unit with "Sinewave" and while my computers are nice and quiet on it(unlike the stepped approximation one, for some reason my monitor actually makes noise on it(it does on both types).).

Chances are I will run this thing(SmartUPS 1400) until it dies then look for something better. Honestly I would guess it is not the most efficient, but I got it for free :)
 

APC told me they would be sending me comments and one of their current models for review but I have yet to receive either.

As for your 550 vs 650 units, APC also makes 1000VA and 1500VA units using the same RBC-4 (2x12V 7.2Ah) battery pack. UPS that share the same battery pack will have similar runtimes for a given load, give or take a minute or two due to the higher-rated unit probably using slightly less lossy components, ex.: MOSFETs rated for a few extra amps and lower Rds(on), possibly more parallel MOSFETs, possibly extra rectifiers to reduce forward drop losses, a slightly bigger transformer with thicker wire gauges, lower-ESZ capacitors or extra capacitors in parallel, etc.

Looking at my BX1000's PCB (picture-story about that incoming), only the DC-HVDC transformer, its output cap and the current-sensing transformer appear to have model-specific footprints and I doubt APC has a whole different PCB for the xx1500 models that shared the BX1000's body.
 

You are right a 14 AWG wire is Fuse Rated at 166A which would be totally ridiculous

The reason the table is not here is I have not found a way to post a table or image to this forum
so I'll add a link to the Table http://www.gxk.org.uk/info/wire.htm this is a British version but a USA Table is the same a number 26 AWG is rated for 21A which would be more than enough That should restore the design intent. If you choose to continue to uses the method of using wire braid to bridge the gape it will work just fine until another PSU failure when the damage will be so bad that you have to scrap it. The design was careful though out that link was there expressly for the event you experienced, altering the design puts the device and home at risk of fire. I can see the law suits Piling up on you desk now!
 

Not really: the solder braid I uses (with the stiffening solder on top) probably has a gauge and specific heat equivalent to #12-14 wire like all other power traces and inductors on that PCB so it is not going to blow up even under 100A fault current. Protecting against "normal" electrical faults is supposed to be the breakers' job.

Even though I may have eliminated a PCB fuse on the live side as a possible failure point, I still have quite a few protections still working:
1- the mains 20A breaker
2- the power bar's 15A breaker
3- the power bar's neutral fusible trace
4- the power bar's thermal fuse - these things will heat up and trip themselves like a normal slow-blow fuse would under sustained overload conditions

As I said in the story, the PCB fuse is most likely part of the high-energy surge-protection design (blow the trace and let the arc switch to ground) and it just happens to make the power bar vulnerable to faulty equipment that would otherwise get caught by breakers - as it did in my post-repair short.
 
My one interaction with a surge protector "protecting" was interesting. I was on a trip to Europe, and took a power strip, as several of my colleagues were attending, and we would all have US plugs. The idea would be that I would use a plug converter for the power strip, and we would plug our laptops into the power strips. The laptops could take 240V input, but as found, that tripped the surge protector, as well as making it a short circuit to the mains, which tripped the breaker, which powered the WiFi in the room. I had forgotten that the power strip even had a surge protector, and didn't realize what the problem was until I got home and it tripped the breaker at home too. While the surge protector did it's job, it was quite surprising that the protection mode was to make the outlet strip a short circuit.
 

I would not be so surprised since MOVs are shorting all that current across L-N and all that heat is going to melt their semiconductor material at some point, turning them into a short. This is why stuff with MOVs usually has a fuse or breaker along with it - so that something will break the circuit after the MOVs have failed.

In the case of APC's power bar, you have two thermal fuses, a breaker and two fusible traces (still waiting for a response from APC regarding whether this is an "accidental feature" or intentional) as fail-safes.
 
I have long held that, as consumers, we are not able to purchase decent equipment, probably due to the pervasive "Its not my job" attitude rampant throughout much of the world.
We can only obtain a kit to build something from. However, many things cannot be repaired and must be exchanged or replaced.To wit, 3 tries to get a working Nexus7, 5 units to get a properly-working monitor, dress shirts that shrink by 2 sizes when washed.
 
That's why I tend to buy things not for the average consumer, at least when it comes to electrical equipment. I had an enterprise laptop made by Dell in 1999, still working despite heavy abuse in 2014 when I sold it. Yet a Dell desktop my parents had years ago only lasted about 2 years before it died.
 
What happened to pride in one's products?
Click to expand...

If there are several similar looking products on offer your typical shopper is going to buy the cheapest one. Let the race to the bottom commence! One corner after another will be cut, materials will get cheaper and crappier, production will move farther and farther afield. Everything is fine as long as the shareholders see good number on the quarterly report!
 
One thing that most people don't seem to catch is that this unit is just poorly designed. For a well designed used unit, it is very difficult to assemble it badly. This is the opposite of that.

Of course, then, visual inspection fails, as well.

I'd also give this unit low chances of survival in a high vibration environment; now, you might think this is not really the intended application, and you'd be right, but I'm betting there isn't a warning not to do this, so some people will.

And, I've got awful suspicions about its EMC performance; it won't really generate any noise in normal operation, but, with design like this, you wouldn't want to bet on noise on one connection not coupling to another. You'd really have to do some proper analysis and testing to be sure, but APC haven't been bothered to do it, so I don't know why I should.

You refer to this as a 'vintage' unit; any idea how old it is? It would be a bit less worrying if it were 20 years old, and a bit more if it were 2.

All in all, this wouldn't be a surprise coming from a no-name supplier, but APC? Very disappointing.
 

I call it a "vintage" unit because I have probably owned it for 10+ years - don't remember how long ago that was exactly. APC retired that rounded design years ago and replaced it by a more angular one..

BTW, I received APC's modern equivalent today so I will probably put something together about it soon... might do a side-by-side (re-)teardown..
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom