APC's Current-Gen SurgeArrest: A Modern Tear-Down
APC's Current-Gen SurgeArrest: A Modern Tear-Down
By
It Finally Arrived
It took several weeks after we published Let's Take a Trip Inside a Power Strip! to line up a modern equivalent, which we'd use for a follow-up tear-down. But it's here now. Hidden somewhere under this ocean of bubble-wrap is the star of today's show. While I dig it out, you may want to revisit my exploration of APC's old SurgeArrest power strip, the company's thoughtful response to my coverage, and a deconstruction of its BX1000 uninterruptible power supply in Tear-Down: Let's Take a Trip Inside A UPS; I'll be referring back to both features throughout today's piece.
See more
See less
Actually, not surprising because it's not valid to compare the ratings between the different generations. The old surge protector was rated under UL 1449 2nd Edition, whereas the new one was rated under UL 1449 3rd Edition. The 3rd Edition test applies a much higher current, so the higher voltage is expected.
Much more interesting would be a teardown of a unit obtained independently through normal retail channels.
I wanted to take a look at that but could not find a free copy of it online; only various companies' notes highlighting differences relevant to their respective business. Without access to the original document, I have no idea how comprehensive and accurate those whitepapers's change lists are.
For example, GE's version does describe the new test conditions (6kV at up to 500A for 2nd vs 6kV input at up to 3kA for 3rd) but it is preceded by a mention that only type-1 and type-2 SPDs (meter/panel-mount) will be further discussed, which seemed to imply that the above was specific to those types.
They most likely did at least check it out for any obvious issues but then again, the new design has fewer wires attaching to the PCB and none of them directly to the traces from the copper side, which should considerably reduce the potential for handling mistakes.
I recommend this unit in particular for its value/$:
http://www.brickwall.com/collections/surge-protectors-home-theater-hdtv/products/eight-outlet-audio-surge-protector
Made in USA, not China like almost every other surge suppressor you can buy today. You get what you pay for, and I learned the hard way with cheap suppressors.
Agreed .... same here .... and replacing the batteries is not done because it's cheaper to buy a new unit. Since The Schneider take over, it's not the same company and certainly not the same product we used to see.
Do you happen to have an appnote or something that details how Brickwall/SurgeX's "non-sacrificial" surge protection works? Both companies have the exact same diagram that does not give any useful details about it.
Everything I can find about anything resembling that from other vendors effectively boils down to slapping an LC filter in front of the MOVs to block or dampen most of the high-frequency, high-energy stuff so the MOVs do not need to deal with (as much of) it.
Much more interesting would be a teardown of a unit obtained independently through normal retail channels.
I don't see what wrong with that, most of review got their material from the company..
It's the same with GPU, CPU, mobo, phone, etc..
most (all) of them is supplied by the manufacturer,
the cost will be too much to handle, if all review need to buy the product and then sells it with lower price later..
But I'm agree, real market sample is much better and objective for review...
I don't believe their claim "clamps 2V above peak" as inductors are not sharp (the proverbial "brick wall") filters, Ask anyone who has worked with hobby electronics or electric motors knows that back-EMF can easily be higher than the supply voltage. Though I don't think this configuration would generate back EMF, other than perhaps at power off.
Most surge protection (the so called "sacrificial" variety) are in fact shunt based designs that shunt the extra energy from an over-voltage event (at or above the clamping voltage). This includes Metal-Oxide Varistors (MOV), Transient Voltage Suppressor (TVS) diodes, Zener (Avalanche) diodes, potentially combined with a slower shunting device such as the gas discharge tubes, and most surge protection shunt devices do have a non-trivial failure rate.
Please find an EE to write about electronics. I'm not a professional, but even as a hobbyist I cringed at some of the lack of knowledge. See: EDN, EE Times, Clive Maxfield http://www.clivemaxfield.com/ the Amp Hour podcast, and even manufacturers such as Texas Instruments or OnSemi, are likely willing to offer their own in-house engineering staff to assist in such writing such articles.
Personally I do prefer brands that are better known in industrial / enterprise markets including TrippLite, and PowerWave (also owned by Eaton).
I don't see what wrong with that, most of review got their material from the company..
It's the same with GPU, CPU, mobo, phone, etc..
The problem, which is the current norm in web based review sites, is that manufacturers have been repeatedly accused, rightly and wrongly, of cherry-picking the review samples. In fact in recent CPU reviews here, the reviewers have disclosed their opinions when samples are not as a "prime" as they expected that perhaps they didn't received hand-picked CPU samples for review purposes.
The US consumer magazine Consumer Report, published by their own not-for-profit organization has been considered a leader in unbiased reviewing because it has strict policies on only reviewing products that they buy themselves through normal retail channels.
This technique also works against fraudsters, who may provide a real (but expensive) review sample as a means of generating free marketing, but plan on selling low-manufacturing-cost (defective or fake) products in the market.
The ethics, and integrity of product reviews has a long and complicated history.
Not sure what you are cringing at since everything you said, I had already written and drawn - if you look at the simplified diagrams, you see the MOVs in shunt configuration, the inductors in series with live and neutral (ignore the MOVs with both terminals on the mains-side) and the capacitor across those L/N inductors to complete the LC filter.
This is a picture story for the typical THG reader; not a reference paper for EMI/SPD engineers and advanced hobbyists. With a target budget of about 750 characters per image (that's the editorial guideline for picture stories - if I go too far above that, I get asked to axe stuff because the text overrun starts breaking the picture-story page layout,) I am already well over-budget on character count by merely scratching the surface.
As for the retail vs review sample thing, I agree that reviewing from retail would be better but that cost would come right out of my own pocket and I cannot afford that unless I can get 5+ picture-story slides or 2+ review-style pages per $10 spent - unless it is something I already had (the old SurgeArrests and BX1000) or something I would have bought anyway.
Companies like ZeroSurge, BrickWall, and Torus Power license their technologies from SurgeX, as SurgeX holds the patent for its unique surge elimination design. Note how I say surge "elimination" and not "suppression". We're talking about a whole different beast here. Cheaper knockoff products from companies that do not use the SurgeX design as a base may use MOVs and will most likely not offer the same level of protection as SurgeX's "series mode" design. SurgeX has since improved its design and is now marketing it under the term "Advanced Series Mode". SurgeX has not yet licensed this newer design afaik so all derivatives currently use the older SurgeX design from the 90s, which is not necessarily a bad thing as the design is solid, but it may not be as efficient or robust.
The following videos may help further illustrate the SurgeX design and differences with other typical surge suppressors:
SurgeX Technology Overview:
https://www.youtube.com/watch?v=79p3ysUnx_Y
SurgeX vs MOV Suppressors:
https://www.youtube.com/watch?v=RixUrc-FRcM
As shown in the second video, typical surge suppressors pollute the ground and/or neutral connections which is a huge no-no for audio equipment and other delicate electronics and only SurgeX has a viable solution for those who want real protection for their equipment.
Many people have been confused about SurgeX and similar products, as can be seen here:
http://www.avsforum.com/forum/40-oled-technology-flat-panels-general/1146963-surge-protectors-brickwall-zero-surge-furman-surgex-etc.html
Most of the other technical questions people have can be answered by contacting SurgeX, as I am not a certified technician.
Not that different when you read SurgeX's patents including US 7 068 487 from 2004 - the biggest difference is the use of a fancy RCD snubber (yes, I noticed the SCRs but all they do is switch in extra capacitance during surges to bring snubber voltage down further) network with three large electrolytic capacitors in place of MOVs to eat the surge energy and those tend to have finite lifespan when you charge them using 100-200A pulses. The use of series inductors to limit instantaneous peak current is every bit as applicable to RCDs as it is to MOVs - SurgeX conveniently omitted that fact in their '89 patents to make MOVs look as bad as possible.
If you put a 120µH inductor in the path of a 6kV, 8µs surge, you are effectively current-limiting it to a much more manageable 400A peak and you need something to shunt that current or otherwise you will not get any voltage drop across the inductor. Unless SurgeX has found a way to break Kirchoff's current law, that 400A coming out of the source, through the inductor and RCD network still has to return to the source through the neutral wire - the sum of all currents coming in or out of a node must be zero. There is still surge current getting shunted to neutral; the inductor is simply making it about an order of magnitude more manageable than having none - enough to make the resistor demonstration work.
10yrs of 100 surges of the what IEEE considers some of the harshest environments out there, [B3 strikes] isn't bad
"You do not have to take our word for it. Consider the following:
IEEE (The Institute of Electrical and Electronic Engineers) states that 6000V is the largest transient that the interior of a building would experience.
IEEE defines its harshest interior surge environment as one that could experience 100 surges of 6000V, 3000A in a years time (category B3).
A new federal guideline recommends that a surge protector utilized in a harsh environment should be capable of withstanding 1000 surges of 6000V, 3000A or ten years worth of IEEE's category B3.
UL (Underwriters Laboratories) now provides a new adjunct testing service (in addition to the 1449 safety classification) that will test surge protectors to the 1000 surge, 6000V, 3000A federal protocol."
""Early in 1996, 1000 surges (at 60 second intervals) of 6000V and 3000A were applied to an off-the-shelf Series Mode Surge Protector.
There were no failures. There was no performance degradation. Let through voltage did not exceed 400V.
We still perform this testing on our equipment today."
These things can take this and live. How long would your MOV last?
http://surgex.com/aboutus.html
"100% Fail Safe Protection: For the past twelve years, SurgeX technology has never failed to protect connected equipment, due to an AC surge-related event. That’s a fact. "
So are you saying they're lying? ZERO failures is impressive if true (and they are all claiming it). You are essentially saying they'll fail just like MOV's here right?
Isn't 1000 strikes of 6000v or 3000amps every 60 secs higher than your 100-200a pulses? They hit these with 1000 of them and they lived. Doesn't this kind of blow a hole in your example? Can you show me an MOV based surge that can take this much testing an live? They don't EAT the voltage, they slow it down.
http://www.brickwall.com/pages/how-it-works
How brickwall says their units work.
http://surgex.com/library/10001_WhatisTrueSeriesMode.html
Their claims of what is different, for anyone who cares
http://www.apc.com/products/resource/include/techspec_index.cfm?base_sku=P11VNT3&tab=features
APC says it can take A 6000v spike, but no claim of taking A 1000 of them.
http://espei.com/catalog/pdfs/product-sheets/ESP-SurgeX-Product-Line-STANDALONE.pdf
Let through of 0v and can take 1000+ spikes.
In this article of yours, aren't you saying they can take maybe one or two big hits and then fail? I'll take 1000 vs. 1 or 2. I used to sell tripplite/apc and I've seen a tripplite I sold, fail with one night of strikes (and they didn't cover the customers stuff either). So you'll have to forgive me if I go with the one that says it can withstand 10yrs in the the harshest environments of 100 strikes each year.
https://en.wikipedia.org/wiki/Surge_protector
More reading...There seems to be no argument MOV's don't hold a candle to these SM (series mode) units.
"These devices are not rated in joules because they operate differently from the earlier suppressors, and they do not depend on materials that inherently wear out during repeated surges."
"Because SM work on both the current rise and the voltage rise, they can safely operate in the worst surge environments."
Ok then...Not that wikipedia is the best source, just an example. You might say these have a finite lifespan, but at 1000+ B3 strikes I'd say in my area I'll be dead LONG before my surge will be. But go ahead and get the $50 APC if you'd like
"But since the SM devices do not wear out and are not required to be replaced every few years, the overall cost of ownership is much lower."
Either they are saying they are BETTER, or I'm just not picking up what this guy is putting down.
"Surges are not diverted but actually suppressed. The inductors slow down the energy. Since the inductor in series with the circuit path slows the current spike, the peak surge energy is spread out in the time domain and harmlessly absorbed and slowly released from a capacitor bank.[31]"
They sure sound DIFFERENT to me. Slowing the current down and EATING it are two completely different things. EATING=quick death. Slowing=longer life. I could go on, but I'm thinking you should get the point by now. The OP was correct, these are a different BEAST (thus ~5-10x more expensive up front).
BTW my CyberPower UPS has been going strong since 2010. Pretty happy with it, but the battery age will probably be creeping up soon.
An RCD network is obviously not exactly the same thing as a MOV and the SurgeX snubber uses SCRs to trigger its secondary snubbers that give the SurgeX its "voltage dip" characteristic in mid-surge. But they are similar enough in function for SurgeX to choose to use the MOV symbol to represent their snubbers in their simplified schematics for prior-art-vs-invention comparisons.
SurgeX's patent does not state what capacitor model they use; only that they are 100-200µF 450V. I do not think there are any electrolytic caps rated for 200-400A charge current. I do not think I have seen electrolytics rated over 10A RMS. It may take thousands of cycles at such a charge current but I would expect them to fail eventually.
3000A is the current limit on the source but the actual output will not go that high unless the load has low enough impedance to allow it, which is exactly what the inductor is there to prevent: if you put a a 120µH inductor in series, current rise from a 6kV voltage difference gets limited to 50A/µs, which translates into about 400A peak at the end of a 8µs pulse. Split this between two MOVs, that becomes about 200A each, split four-ways and it becomes about 100A each. 20DnnnK and most equivalent MOVs are rated for 10 000 shots at 100A with less than 10% parameter deviation. The potential for a surge suppressor manufacturer to build high-endurance MOV-based designs is certainly there. Also keep in mind that 6kV is what UL deems to be the worst-case surge voltage that should be allowed to get indoors so the typical everyday surge would be nowhere near that bad and the MOVs' useful life would end up that much longer.
As noted in my previous responses and alluded to in the article itself ("taking the edge off surges"), "slowing the current" by using inductors is every bit as applicable to MOV-based designs as it is to SurgeX's snubbers. Since you get no voltage drop across an inductor unless there is a current change through it, something has to eat that 50A/µs ramp for any surge suppression, elimination or whatever marketing wants to call it to occur. Without the shunt components sinking that current, whatever they might be, the small inductors become useless.
The SurgeX still needs to eat (snub) all the excess energy coming through that inductor like the MOV does. The difference is that a simple MOV cannot do anything more than follow its I-V curve as the inductor current rises while the SurgeX can switch in its discharged caps to soak up that energy when it detects a fast edge and momentarily drag the voltage down. Both will need to eat the same 400A ramp and both will need to dissipate about the same amount of energy.
Is SurgeX's design better? In most circumstances, yes. Is it worth paying 5-10X as much for? Not when everything you are going to plug into it already has tolerance to 500-600V peak due to universal input SMPS which can easily accommodate a good MOV-based suppressor's 340-400V peak clamping voltage.