-
The answer up front: these aren't the same product at different prices
-
Where those numbers come from
- What a wall outlet surge protector actually does
-
What the APC Back-UPS 650VA reviews get right and what they skip
-
The APC Back-UPS 1000 battery is the real cost center
-
The furnace filter question, and why it belongs in this article
-
Where I'd draw the line
The answer up front: these aren't the same product at different prices
A surge protector absorbs a spike and sacrifices itself. A UPS absorbs the spike, then keeps your equipment alive for the next 10 to 30 minutes so you can shut it down cleanly. Different jobs.
If an outage costs you more than roughly $150 in lost work or hardware, the $25 strip is the more expensive option. That's not a rule of thumb I pulled off a vendor page — it's the number that fell out of our own purchasing data.
In 2023 we lost a RAID controller, a QuickBooks file, and about six hours of staff time to a 40-minute outage. Damage came to roughly $4,100. The controller was on a surge strip. The machines sitting on UPS units in the same room never noticed the outage happened.
Where those numbers come from
I'm a procurement manager at a 180-person logistics software company. I've owned our facilities and IT peripherals budget ($60,000 annually) for six years, negotiated with 40+ vendors, and logged every order in our cost tracking system.
When I audited our 2023 spending I pulled every power-related purchase going back to 2019. Analyzing $214,000 in cumulative facilities and IT spend, $18,700 of it was power protection — strips, UPS units, batteries, the whole category.
The surprise wasn't the unit prices. It was the split. Twenty-two percent of that $18,700 went to replacement batteries, not devices. Nobody budgets for that line. I certainly didn't until I saw it as one number instead of eleven separate invoices.
What a wall outlet surge protector actually does
A wall outlet surge protector is a metal oxide varistor in a plastic shell. It clamps voltage above a threshold, dissipates the energy, and that's it. Typical retail units I priced in January 2025 ran 600 to 1,200 joules with a clamping voltage around 330V on 120V circuits.
Two things people miss about them:
- They're consumables. After a large hit the MOV is spent. Some units have a "protected" indicator, but that LED typically reports a fuse, not degradation. A strip that has taken a hit can look perfectly normal and be doing nothing.
- They don't cover sags, brownouts, or outages. Which, in our building, is the actual problem. We're on the same grid as a welding shop two blocks over. We get voltage dips, not lightning.
I don't have hard data on how many of our strips have quietly died — there's no failure log, and we don't test them. What I can say anecdotally is that after the 2023 incident I checked every strip in the building with a multimeter and stopped trusting the LEDs on about a third of them.
On the surge protector Home Depot aisle
I've bought a lot of surge protector Home Depot stock over the years, and I'm not going to pretend it's bad. For a desk lamp, a phone charger, a printer nobody reveres — a $20 strip is the correct purchase and I'll defend it. The joule number on the box is a capacity rating, not a guarantee, but at that price you're buying a fuse with a power cord. That's fine. Just don't put a NAS on it and call it protected.
Same applies to plug-in wall units. They cover one duplex outlet, they're easy to lose, and in a shared office they migrate. We zip-tied ours to the outlet plates. Sounds petty until you've replaced the same unit three times.
What the APC Back-UPS 650VA reviews get right and what they skip
I read a pile of APC Back-UPS 650VA reviews before our first order of fourteen units in 2021. They were accurate about the things reviewers care about: it's plug-and-play, it beeps, it gives you enough time to save and shut down. Four and a half stars, and honestly earned.
Here's what the reviews mostly didn't cover, and what changed our math:
The 650VA name is not 650 watts. It's 650 volt-amps, about 390 watts of real load. Our units run at 120–180W (a workstation plus a monitor), so we get somewhere in the 12–20 minute range. If you're sizing for a server plus a switch plus a monitor, do the watt math first. I've watched two colleagues buy 650s for loads that really wanted a 1000.
Half the outlets are surge-only. The 650VA has eight outlets and only four are on battery. Nobody puts that in a review because it's in the spec sheet, and none of us read the spec sheet.
It outputs a stepped sine wave, not pure sine. Fine for most workstations. Not fine for some active-PFC power supplies and definitely not for certain motors. If a device has a motor or a fussy PSU, that's a Smart-UPS question, not a 650 question.
We standardized on APC UPS units (Schneider Electric) partly because of that spec clarity and partly because our electrician already knew the product line. We compared three vendors before the first order and all three were competent — this wasn't a case of one obvious winner, it was a case of one option our contractor could service without a learning curve.
The APC Back-UPS 1000 battery is the real cost center
If you take one line off this page, take this one. When I modeled five-year ownership on our nine 1000VA units, the battery replacement ran $65–90 per unit, and we've replaced them at roughly the three-to-four-year mark. That's not an APC problem, it's lead-acid chemistry. Every UPS in this class does it.
Which means the honest comparison isn't $210 for a UPS versus $25 for a strip. It's:
- Strip over five years: $25, assume one replacement after a hit — call it $50. No runtime. No data protection.
- Back-UPS 1000 over five years: $210 unit + roughly $75 battery = $285. About $57 a year, per protected workstation.
$57 a year. That's the number I bring to budget meetings now instead of the sticker price, because the sticker price is the number that gets the purchase order rejected and the annual cost is the number that gets it approved.
We tried third-party batteries once to save about $25 a unit. The upside was $225 across the order. The risk was a swollen cell cooking a $210 unit or, worse, failing to hold load during an actual outage. I kept asking myself whether $225 was worth a repeat of the 2023 mess. We went OEM. I'm not saying third-party is wrong for everyone — I'm saying I couldn't defend the savings to myself if the next outage took a controller with it.
The furnace filter question, and why it belongs in this article
Somebody searching "can you run a furnace without an air filter" and somebody running a UPS on four-year-old batteries are asking the same question with different nouns.
Yes, you can run a furnace without an air filter. It will fire, the blower will push warm air, and nothing will appear broken for a while. What happens instead is that the blower wheel and the heat exchanger get coated, static pressure climbs, the system runs longer to hit the same setpoint, and a $15 consumable turns into a $300 service call plus a few hundred dollars of wasted gas. The filter was never the expensive part. Skipping it was.
A UPS works the same way. The box is not the system. The battery is the system, and it's a consumable with a three-to-five-year clock. Our 2019 units didn't fail — they just quietly dropped from 18 minutes of runtime to about 90 seconds, with no error, no alarm, nothing. We found out during an outage. (Should mention: we now load-test every unit in January and log the runtime in a spreadsheet. Wish I'd started that in year one.)
In both cases the line item that gets cut is the cheap one that makes the expensive one work. That's the pattern I keep running into, across categories that have nothing to do with each other.
Where I'd draw the line
I'm not going to tell you everything needs a UPS. That'd be a sales pitch, not a procurement decision.
A surge protector is probably enough if the device has no state to lose — phone chargers, desk lamps, a monitor you'd replace without thinking, anything with a wall brick feeding a device that shuts off and comes back on cleanly. If your only real risk is a lightning surge and you never lose power, spend the money elsewhere.
Step up to a Smart-UPS when you've got a spinning disk array with a write cache, a network closet that serves a floor, anything lab or medical, or a device with active-PFC and a picky power supply. The step up isn't about runtime so much as voltage regulation and pure sine output.
And none of this survives a long outage. A UPS is measured in minutes, not hours. If your grid is genuinely unreliable — like ours, the welding shop notwithstanding — the conversation moves to generators and transfer switches, and that's a different budget entirely.
Per UL 1449 (3rd edition), surge protective devices sold in the US carry standardized ratings, so the joule and clamping numbers on the box are at least made comparable across brands. Verify current specs and pricing at the manufacturer's site before you buy — the figures above reflect what we paid as of January 2025 and battery prices move.
The one thing I'd tell my 2019 self: count the consumable, not the box. The battery and the filter are where the actual decision lives. Everything upstream of that is just packaging.

