RF Utilities
Frequency accuracy, oscillator tolerance, stability budgets, and clock drift in a single precision RF utility. Convert between Hz, ppm, ppb, and percent at any carrier from 32.768 kHz to 5.8 GHz.
Free to use, no sign-up needed.
Walkthrough
Top bar with the four mode selector (Offset, Tolerance, Budget, Timing) and the Inputs versus Reference panel tabs.
Overview
Parts per million is how the industry quotes frequency accuracy, and it is a fraction rather than an absolute. That single fact is the source of most of the confusion around it. A 2 ppm TCXO sounds like a fixed specification, but it is worth 0.065 Hz on a 32.768 kHz watch crystal, 20 Hz on a 10 MHz reference, 290 Hz at 145 MHz, 1.8 kHz at 900 MHz, and 11.6 kHz at 5.8 GHz. The oscillator did not change. The carrier did. Engineers who specify references every week still want a precision tool to remove the conversion as a source of error.
The noIM₃ Hz to PPM Calculator is the precision utility for that work. Enter a nominal frequency in Hz, kHz, MHz, or GHz and convert freely between a frequency error in ppm, the equivalent offset in Hz, the frequency a counter would actually read, and the same error expressed in ppb and percent. Every field is bidirectional and updates live, so the tool works equally well from a datasheet figure forwards or from a bench measurement backwards. Conversions follow the standard relationships exactly (ppm equals Δf divided by f₀ times 10⁶) rather than rounding, and offsets render from microhertz to megahertz so the answer is readable at any carrier.
Beyond raw conversion, the calculator handles the three jobs that surround it. A datasheet plus or minus ppm tolerance becomes the allowed frequency window at your carrier, with the minimum, maximum, and total peak to peak width. An oscillator stability budget totals the contributions two ways, worst case and RSS, so you can see both the bound your specification has to respect and the statistical combination, and decide which one is defensible. Finally, because a clock derived from a reference inherits its fractional error exactly, the same ppm figure is read as accumulated time error: 1 ppm is exactly 1 µs/s, 86.4 ms per day, and 31.536 s per year.
Convert freely between a frequency error in ppm, the offset in Hz, ppb, and percent at any nominal frequency from sub Hz to tens of GHz. Every field updates the others live. Changing the nominal frequency holds the ppm and recomputes the Hz, because the headline question is what a given ppm is worth at your carrier.
Enter the frequency a counter actually read and the tool works the conversion backwards to the ppm error against nominal. A counter on a 145 MHz transmitter reading 145.000312 MHz is 2.15 ppm high. This is the real bench workflow, and it is a first class input rather than something you reach by rearranging the formula yourself.
Turn a datasheet plus or minus ppm tolerance into the allowed frequency window at your carrier: the plus or minus offset in Hz, the minimum frequency, the maximum frequency, and the total peak to peak width. The tolerance is treated as symmetric so its sign is ignored. Eight presets seed the common datasheet grades from plus or minus 0.1 ppm through plus or minus 50 ppm.
Build an oscillator stability budget from addable, renameable contributions covering initial accuracy, temperature stability, ageing, and supply and load pull. Both totals are reported side by side. Worst case sums the magnitudes and bounds the specification. RSS combines independent terms statistically and is always the smaller number. The dominant contribution and its share are called out so you know which term to attack.
A clock derived from a reference inherits its fractional error exactly, so a frequency offset is also a time error. Convert ppm into accumulated drift per second, hour, day, and year, size the drift over an arbitrary holdover period, and find how long a reference holds before it passes a 1 ms or 1 s target from the last sync.
Each mode reports a table of what the current error or tolerance is worth at every common frequency from 32.768 kHz through 10 MHz, 145 MHz, and 900 MHz to 5.8 GHz, alongside a chart of offset against frequency across the whole 10 kHz to 10 GHz span on log axes. The fractional nature of ppm stops being an abstraction.
Built in reference ranges for each oscillator technology class (XO crystal, VCXO, TCXO, OCXO, rubidium, and GPS disciplined) place a quoted figure against the technology it implies, and report what each grade is worth in Hz at your carrier and in clock error per day. These are ballpark ranges for orientation only, not a specification, and no substitute for the datasheet of the actual part.
Runs entirely in your browser. No frequencies, tolerances, or stability budgets are submitted to a server. Useful for commercially confidential work, classified projects, or any environment where information security policy prohibits sending engineering data to third party services.
Inputs and outputs
Inputs 8
Outputs 9
Use cases
Divide the frequency offset by the nominal frequency and multiply by a million. ppm equals Δf divided by f₀ times 10⁶. A 290 Hz offset on a 145 MHz carrier is 290 divided by 145,000,000 times 10⁶, which is 2 ppm. Going the other way, Δf equals ppm times f₀ divided by 10⁶. The calculator does both directions live and exactly, so the output preserves enough significant figures for engineering documentation.
Because ppm is a fraction, not an absolute. It expresses the error as a proportion of the carrier, so the Hz it represents scales directly with the carrier. A 2 ppm oscillator is 0.065 Hz on a 32.768 kHz watch crystal and 11.6 kHz at 5.8 GHz. Same specification, same part, a difference of more than five orders of magnitude in Hz. This is why a tolerance that is comfortable at VHF can be unworkable at microwave, and it is the single most common source of surprise when a design moves up in frequency.
They are the same fractional error at different scales. ppb equals ppm times 1000, so 0.001 ppm is 1 ppb. Percent equals ppm divided by 10,000, so 10,000 ppm is 1 percent. ppm suits crystal and TCXO work, ppb suits OCXO and rubidium work where ppm figures get uncomfortably small, and percent almost never appears in frequency work but does turn up when a specification has been written by someone outside the discipline. The calculator surfaces all three at once so the conversion is a read across.
It depends on what the number is for. Worst case sums the magnitudes of every contribution and assumes each lands at its specification limit, in the same direction, at the same time. It is pessimistic, and it is the bound a datasheet total stability figure has to respect. RSS takes the root sum square, which assumes the contributions are independent and treats each as a statistical spread rather than a hard limit, so it is always the smaller number. RSS is only valid when the terms genuinely are independent. An ageing trend and a temperature ramp both pulling the same way are correlated and do not qualify. The calculator reports both and does not pick for you.
A clock derived from a reference inherits its fractional error exactly, so the two are the same number wearing different units. 1 ppm is exactly 1 µs per second by definition, which works out to 86.4 ms per day and 31.536 s per year on an ordinary 365 day year. A 50 ppm uncompensated crystal drifts about 4.3 s per day, which is why a cheap digital clock needs resetting. A 0.05 ppm OCXO drifts about 4.3 ms per day. The Timing mode makes this conversion directly and also answers the operational question, which is how long you hold before you pass a target.
Cumulative ageing is the quoted per year rate multiplied by the number of years, which is a linear extrapolation. Real crystal ageing is steepest in the first months after manufacture and flattens towards a logarithmic trend, so a linear projection over a long service life is a conservative bound rather than a prediction. It will overstate the drift, not understate it, which is the right direction for a specification. The default budget carries ageing as an explicit named row so the service life assumption stays visible rather than disappearing into a single total.
They are indicative ranges for each oscillator technology class, offered so a quoted figure can be placed against the technology it implies. They are not a specification, they are not drawn from any regulatory instrument, and they are not a substitute for the datasheet of the part you are actually specifying. Real parts vary widely inside each class, and vendors quote stability over different conditions. Use the grades for orientation and the datasheet for decisions.
No. The calculator runs entirely in your browser. No frequencies, tolerances, or stability budgets are submitted to a server. Useful for commercially confidential work, classified projects, or environments where information security policy prohibits sending engineering data to third party services.
