Do you need a pure sine wave inverter?

The wall socket in your house delivers a sine wave. An inverter’s whole job is to build one out of a battery, and there are two ways to go about it: draw the curve properly, or fake it with steps and hope the appliance doesn’t notice.

For a long time the second way was how you saved money. It still is, at the component level — but the two companies below who sell both kinds tell you, in their own published material, to buy the first.

What is a modified sine wave, actually?

A stepped square wave that spends part of every cycle sitting at zero volts. It is not a rougher sine wave. It’s a different shape.

Samlex America — which builds both a pure sine PST line and a modified sine SAM line — puts it plainly: “In a modified sine wave, the voltage rises and falls abruptly, the phase angle also changes abruptly and it sits at 0 Volts for some time before changing its polarity.” Renogy describes the same thing as “a stair-step, square pattern, where the polarity is flipped back and forth”.

That shape isn’t one frequency. Samlex’s engineering note calls it “multiple sine waves of odd harmonics (multiples) of the fundamental frequency”, and says the consequence is “enhanced radio interference, higher heating effect in motors / microwaves and produces overloading.”

Those extra harmonics are the entire story. Everything that goes wrong downstream goes wrong because of them.

Which devices actually care about the waveform?

Anything that reads the waveform’s timing or shape to decide what to do — plus anything with a motor. A resistive load doesn’t read anything. It just gets hot, and it gets hot on either waveform.

Rather than compose a list, here’s Samlex’s own, from its comparison page: laser printers, photocopiers and magneto-optical hard drives; built-in clocks in clock radios, alarm clocks, coffee makers, bread-makers and microwave ovens; output voltage control devices such as dimmers and ceiling fan speed controls; sewing machines with microprocessor speed control; transformer-less capacitive input devices including razors, night-lights, smoke detectors and battery rechargers; devices that carry radio frequency signals on the AC wiring; furnaces with microprocessor control; high intensity discharge (metal halide) lamps; and some fluorescent lamps with power factor correction capacitors.

Samlex’s SAM-1500C-12 manual — the manual for a modified sine product — repeats the warning against zero-voltage-crossing timing, RF signal modulation, Triac-based phase control, temperature controllers and high-capacitance voltage multipliers, naming clocks, X-10 home automation, light dimmers, electric blankets, strobe lights and laser printers.

Motors are the separate case, because they run — they just run badly. Samlex says an inductive load on a modified sine wave “may be running hotter than normal, which equates to less power efficiency and more power consumption”. Renogy quantifies it: motors “will use about 20% more power,” and “will run hotter, which also means it will make them not last as long.”

Note what that 20 percent is measuring. It’s the appliance drawing more, not the inverter converting worse. Two different efficiencies get muddled together in almost every article on this subject, including some of the makers’ own.

Audio is the one you’ll hear before you measure it. Samlex says “you could encounter a buzzing or humming noise when using modified inverters to power your AC devices”; Renogy says some fluorescent lights “may buzz or make humming noises.”

On medical devices, we’ll quote and stop. Renogy names medical equipment, and CPAP machines with humidifiers specifically, among loads it treats as incompatible. Samlex writes that “if you are aiming to power delicate medical equipment, a laptop, or other devices that require clean AC power to function properly, the safest option is to go with a pure sine wave inverter.” Those are the manufacturers’ words. We’re not adding to them.

Do the makers who sell both waveforms take a side?

Yes — and they take the side that costs them the cheaper sale. Samlex still lists a full modified sine range from 100 W to 3000 W, and still writes that “if your budget allows for it, Samlex suggests to always go with the safe option of purchasing a pure sine wave power inverter.” Renogy’s guide ends the same way: “we recommend purchasing a pure sine wave inverter.”

The portable power station category has already settled it by omission. Goal Zero’s Yeti 1500X specification reads “2x AC: 2000W, 3500W Surge (120VAC 60Hz, 16.5A), Pure Sine Wave”. EcoFlow’s DELTA 2 manual lists “AC (x6) Pure Sine Wave, 1800W total (surge 2700W), 120V~ (50/60Hz)”. Nobody in that category sells a modified sine box.

Where the “settled” story breaks down is price, and it’s worth being honest about that rather than repeating the comfortable version. Samlex’s own figure is that modified sine inverters “can range from 40-75% cheaper in comparison to pure sine wave power inverters of the same power output.” That gap has not closed. What changed is that the cheap end got good enough that paying it stops being a real decision for most people — not that the two now cost the same.

Jackery’s Explorer 1000 v2 page is the odd one out: it publishes “120V ~ 60Hz AC Total Output 1500W Rated, 3000W Surge peak” and states no waveform at all.

What do the published numbers look like side by side?

Sparse, and inconsistent about what’s being measured. Only Victron Energy states the temperature its continuous rating assumes, and only Samlex states a surge duration.

Inverter	Waveform	Continuous W	Surge W	Published efficiency	THD
Samlex SAM-1500-12	Modified sine	1500	3000	not published	not published
Samlex SAM-1500C-12	Modified sine	1500	3000 (<4 ms)	87% at 100% load	not published
Samlex PST-2000-12 / -24 / -48	Pure sine	2000	3500	>85% / >88% / >90% peak, load not stated	<3%
Samlex PSR-1200-24	Pure sine	1200	not published	not published	<2%
Victron Phoenix 12/1200	Pure sine	1000 at 25°C, 900 at 40°C	2400	92% max at 12V, load not stated	not published
Victron Phoenix 24/3000	Pure sine	2400 at 25°C, 2200 at 40°C	6000	94% max at 24V, load not stated	not published
Renogy 2000W 12V	Pure sine	2000	4000	“over 90%”, load not stated	<3% linear, <5% non-linear
Renogy 3000W 12V	Pure sine	3000	6000	“>90%” nominal, load not stated	<3% linear, <5% non-linear
Goal Zero Yeti 1500X	Pure sine	2000	3500	not published	not published
EcoFlow DELTA 2	Pure sine	1800	2700	not published	not published
Jackery Explorer 1000 v2	not published	1500	3000	not published	not published

Read the efficiency column carefully: Samlex’s modified sine SAM-1500C-12 is the only unit here that names its load point, and its 87 percent at full load is numerically higher than the “>85%” peak Samlex publishes for the 12 V pure sine PST-2000. The two numbers are not comparable, which is exactly the problem.

What does a surge rating actually promise?

A wattage, and almost never a duration. Only one line in that table carries a time: Samlex’s SAM-1500C-12 manual gives “MAXIMUM SURGE POWER 3000W (<4 millisec; Power Factor = 1).”

Four milliseconds is a quarter of one 60 Hz cycle. That is enough to survive the instantaneous inrush of a switch-mode supply and nothing like enough to spin up a stalled compressor, which wants its starting current for a good fraction of a second. Every other maker in the table publishes the wattage and stops.

“Continuous” hides a second assumption. Victron is unusually direct about it — “In our datasheets inverters, and the inverter function of Multis and Quattros, are rated at 25ºC (75ºF)” — and publishes the derating that follows: 82 percent of continuous output at 50 °C, 75 percent at 60 °C. A 3000 W inverter in a hot engine bay is a 2250 W inverter. Sizing against the badge is covered in more detail in our guide to what size portable power station you actually need.

Why is one efficiency number not enough?

Because inverter efficiency is a curve, and the single number every maker prints is its highest point. Victron publishes the curve for a Phoenix 24/3000, and it moves a long way: 83.0 percent delivering 100 W, 90.1 percent at 200 W, 93.5 percent at 400 W, 93.9 percent at 800 W, 91.5 percent at 1600 W and 86.5 percent at 3000 W.

The reason the left end collapses is idle draw. That inverter burns 20 W doing nothing, and 20 W is a rounding error against 800 W and a fifth of a 100 W load. Samlex publishes the same cost differently, as no-load current: “< 1.0 A” for the PST-2000-12 against “0.65A to 0.85A” for the modified sine SAM-1500C-12. Renogy’s 2000 W pure sine unit draws “< 2A” with nothing plugged in — roughly 24 W burned to keep an empty inverter awake.

If your loads are small and constant, idle draw matters more than peak efficiency, and a smaller inverter beats a bigger one. The battery chemistry feeding it matters too, and we’ve written separately about LiFePO4 versus other lithium chemistries.

The short version

Modified sine is a stepped approximation with harmonics that resistive loads ignore and motors, clocks, dimmers, speed controls and switch-mode supplies do not. Samlex sells both waveforms and tells you to buy pure sine; Renogy says the same; nobody builds a modified sine power station any more. The price gap is still real — Samlex puts it at 40 to 75 percent — so the honest answer is that the technical question is settled and the budget question isn’t. When you’re comparing spec sheets, treat the surge figure as meaningless without a duration and the efficiency figure as a peak rather than a promise. Only Victron publishes enough to check either.