Quadrant

FI-116 · four sensing stations

Overall vibration

--
mm/s RMS · band limited
No capture yet
Sample rate
--
Samples
--
Bin width
--
Peak accel
--

Capture

Idle. Strap the phone to the machine with its long edge along the shaft axis before capturing.

Machine geometry

Enter geometry to compute forcing frequencies.

Spectrum

Peaks and tags

No spectrum yet. Capture or load the demo trace.

Trend

One capture tells you where the machine is. A row of captures at the same point tells you where it is going, which is the only version of this that lets you schedule instead of react.

Severity

Waterfall

No captures saved yet. Analyse a trace, then save it here.

Airborne channel

This listens through the microphone, not through the case. It is airborne sound, not structural vibration, and it hears the whole room. Use it to ask whether a bearing tone exists, never to put a number on how severe it is.

The motion sensor stops at about 30 Hz, which is well below where bearing defects live. Envelope analysis gets there another way: band-pass the high frequency region where a defect impact rings, rectify it, low-pass the result, and the repetition rate of the impacts appears at the bottom of the envelope spectrum, right where BPFO and BPFI sit.

StateMicrophone stopped
Nothing captured yet.

Flash rate

1750
flashes per minute
29.167 Hz
Display refresh
--
Achievable
--
Timing error
--
Source
Screen

Controls

Slow motion off. A non-zero offset makes the pattern drift at the offset rate, which is the way to inspect a belt seam or a spindle face while it runs.

Finding the true speed

A frozen image at rate f may also freeze at f/2, f/3 and every other integer division, so the first still image you find is often a submultiple. Work down: halve the rate. If the pattern stays frozen and single, keep halving. The true speed is the highest rate at which exactly one image appears. Two or three evenly spaced images mean the rate is a multiple of the true speed.

Never use a strobe as proof that a machine has stopped. Frozen and running look identical. Lock out the drive before you touch anything.

Readings

No readings logged.

Camera tachometer

A strobe tells you the truth only if you know the harmonic rule. This does not care. Put a strip of reflective tape on the shaft, point the camera at it, and the brightness of a small patch of the frame rises once per revolution. Count the rate and the speed is unambiguous.

Detected speed--RPM
Pulse rate--Hz
Confidence--
Frame rate--fps
Start the camera to see the speed ceiling for your frame rate.

Illuminance

--
lux · uncalibrated units until set
Pick a space type to check against a target range.
Mean luma
--
Exposure
--
ISO
--
Calibration
Not set

Camera

Point the lens at the surface you are lighting, from the surface, facing the fixture. The phone camera has no cosine corrector and no photopic filter, so treat this as a relative meter that becomes useful once calibrated against a known reading.

Calibration and target

Flicker

Percent flicker
--
Flicker index
--
Frequency
--
Verdict
--

Flicker is read from rolling shutter banding: the sensor scans rows over time, so a modulating source prints stripes across the frame. Frequency is only absolute once the row readout rate is known, which is what the calibration step measures using a plain mains-driven lamp as the reference.

Colour quality

CCT
--
Duv
--
Tint
--
Mean RGB
--

Correlated colour temperature is estimated with McCamy's cubic on the camera's own sRGB response, and Duv tells you how far off the blackbody locus the source sits. This is a hint, not a spectrometer reading: white balance is forced off where the browser allows it, but any residual camera processing lands directly in this number. Hand the swatch to Colorimeter when you want to build a palette from it.

Survey points

No points logged. Log a point at each task location to build a room survey.

Room grid

One reading at the bench tells you almost nothing about a room. A grid tells you uniformity, which is what people actually feel: the average can be perfect while the corner you work in is half of it.

Build a grid, walk the room, and capture a reading at each cell. The photometer camera must be running.
Average--lux
Minimum--lux
Maximum--lux
Min to average--
Min to max--
Load density--W/sq ft
Delivered efficacy--lux per W/sq ft

Sound level

--
dB(A) slow
Microphone stopped
Elapsed
00:00:00
Leq
--
Lmax
--
Lpeak
--

Exposure

Dose
0.0 %
Projected 8 h dose
--
TWA (8 h)
--
Time left at this level
--

Calibrating the microphone

A phone microphone is not a class 2 meter. It compresses, it rolls off below about 100 Hz, and it clips well under the levels a shop can produce. Set the offset by standing next to a real sound level meter, or an acoustic calibrator if you have one, and adding the difference here. Without that step every number on this station is relative, and the honest use of it is comparing today against last Tuesday rather than declaring compliance.

L10
--
L50
--
L90 (background)
--
Clipping
none

Session history

Octave bands

Dose tells you how much. It does not tell you what frequency it was, and that is the number you need to choose hearing protection. A muff rated for 25 dB may give you eight in the band that is actually hurting you.

Turn band capture on, then start the dosimeter. Bands accumulate as equivalent levels over the session.

Hearing protection, long method

Enter the attenuation your protector claims in each band, from its data sheet. The estimate applies the usual seven decibel safety adjustment to the A-weighted result.

Unprotected--dB(A)
Under protection--dB(A)
Effective reduction--dB