User Manual

If the signal has frequency modulation the instrument will display the average value across the
gate time; the modulation is almost certainly not synchronous with the gate, so there will be small
random variations in the displayed value.
If the signal has amplitude modulation, its amplitude at the trough of the modulation must exceed
the sensitivity threshold of the input. Counting deeply modulated signals requires both
considerable amplitude and a sensitive adjustment of the trigger threshold.
Width, Duty Cycle and Ratio H:L Measurements
When Width mode is selected, the instrument continues to use the capture and continue method
to measure the signal period. It cannot measure the width of the active part of the signal this way
because, by definition, there are gaps between the measurements while the signal is in the
inactive state. Instead, it measures the width of a sample of individual cycles of the input signal at
a rate up to about 1000 samples per second. It accumulates up to 50 such samples spread
across the selected gate time, computes the average and displays the result. Each sample has a
resolution of 20ns, and the average is displayed with a resolution up to 1ns. The values for duty
cycle and ratio H:L (better thought of as the ratio active:inactive) are computed from the average
width and the accurately known period. The display resolution presented in these modes is a
reasonable representation of the probable measurement accuracy.
Ratio B:A
This mode is entered by a long press of the WIDTH / RATIO key when the B input is selected. It
takes as nearly simultaneous capture and continue measurements of both input signals as
possible. Because each measurement terminates on a transition of its respective signal the two
measurements are not exactly simultaneous unless the signals are synchronously related. This is
not likely to be an issue unless the signals are significantly frequency modulated.
Note that this method is completely different to the previous model (the TF830) which
implemented ratio B:A mode by counting the B input using the A signal as the reference timebase.
Timebase and Other Accuracy Considerations
The following is intended as a guide to determine the limits of measurement error.
Internal Oscillator
The instrument has an internal temperature compensated crystal oscillator (TCXO) which has
been factory set from a Rubidium reference standard such that it is within ± 0.2ppm (parts per
million) after warm-up in an ambient of 21ºC. At ambient temperatures other than 21ºC the
additional error is less than ± 1ppm over the whole operating range 5ºC to 40ºC.
The ageing rate is less than ± 1ppm in the first year and decreases exponentially with time. The
recommended calibration period is 1 year, see Maintenance section.
External Reference
If measurements are to be made which require still greater accuracy than can be obtained using
the TCXO, an external 10MHz frequency standard may be applied to the External Reference
input. The signal should be TTL, 3Vpp to 5Vpp CMOS or 1 to 2Vrms sinewave. The external
reference is used to phase lock the internal oscillator and must only be a high accuracy 10MHz
signal. It is not possible to make ratiometric measurements by applying a non-standard signal.
The presence of an external reference signal of adequate amplitude is automatically detected
and phase lock is attempted; the Ext Ref display annunciator is shown when the external
reference is detected. Note that if an improper signal is applied then the internal oscillator will be
pulled off frequency and measurement accuracy significantly impaired.
14