Annex A (informative) — Verification measurements

Annex A gives the verification methods the register names, then bench procedures for selected measurements. A product's own documented procedure, or an equivalent analysis, serves as well.

Methods

The register (4.4) names, for every requirement, how it is verified and the figure checked, from the verify field of data/requirements.json. A may row is a permission, and what is verified is the condition it sets. The methods:

Method Meaning
inspection Inspection of the product, its drawings, or its documentation
analysis Documented analysis or calculation from the product's design figures
measurement A bench measurement against the figure the requirement states

Supply ripple and noise

EIS · Measuring ripple and noise supply one rail +12 V 0 V load rated current 300 mm (12 in) twisted pair 0.1 µF ceramic 47 µF at the pair's far end oscilloscope, 20 MHz bandwidth peak-to-peak Every rail at its rated current; the pair runs from the output terminals, or from test points the product publishes, and the probe, with a short ground spring, sits across the capacitors at its far end, as common supply datasheets measure. The supply publishes the peak-to-peak figure for each rail (SUP-001); every rail stays below 150 mV (SUP-007). A linear or post-regulated supply, typically under 5 mV, suits a row of modules with VIDEO ports.

  1. Measure the ripple and noise of each rail, peak-to-peak, by the method of 6.2.
  2. Compare it with the published figure (SUP-001) and with the limit of the rail's power interface, 150 mV on the bus (SUP-007).

Module header current

  1. Run the module under its heaviest load.
  2. Measure each rail's current at the module's power header. Confirm it is within the published figure (PWR-001) and no more than 1.5 A (BUS-008), once the decoupling has charged.
  3. With the module running from a bench supply, confirm that no current flows out of any rail pin into the supply (PWR-004).

DC entry

  1. Confirm the polarity: center positive on a barrel (DC-004), the tabulated +V and −V pins on a multi-pin variant (DC-005).
  2. Confirm that the product runs across its published input voltage range (DC-002) at its published current (PWR-001).

Patch audio and CV

  1. Check each AUDIO and CV port against its claimed range (SIG-005).
  2. Hold ±12 V from a source of 1 kΩ or less on each patch input, and confirm no damage (SIG-002).
  3. Short each patch output to 0 V and to another output, and confirm no damage (SIG-003).

VIDEO levels

DC-couple the probe at the patch tip. Confirm that the output's full scale is 0 to +1 V and that it stays within the U1 ceiling, −0.1 V to +1.2 V (SIG-005).

VIDEO impedances

  1. For each VIDEO output, measure the series resistor from the buffer output node to the patch tip: 75 Ω ±1 % (SIG-010).
  2. For each VIDEO input, with the module powered, measure the resistance from the tip to 0 V: 100 kΩ ±1 % (SIG-009).

The input figure is the resistance at the tip: the resistor to 0 V and the input stage behind it read together. Where the stage's own input resistance is a few megohms, choose the resistor so that the parallel value lands inside the window.

EIS recommends checking that each patch jack barrel is insulated from the panel (7.1).

VIDEO paths

These procedures measure the path figures of 7.3: the bandwidth and delay of a linear or variable path (SIG-016), the rise time of a shaped path, and the input capacitance of a VIDEO input. The fixture below serves the first three; input capacitance has its own set-up. Methods A, B, and C give the same bandwidth for the smooth roll-off of a video buffer; where they disagree, Method A is the reference. A port list names the method by its letter (measured_by).

EIS · Measuring a VIDEO path: the fixture generator 50 Ω, 0.5 V p-p 25 Ω 75 Ω in total IN jack the path controls set for the widest response OUT jack 100 kΩ 30 pF probe counted 0 V scope at the jack jack to jack: the path's figure A, the reference: a stepped sine from 100 kHz; bandwidth where the amplitude is 0.707 of the 100 kHz reading. B: a step from +0.25 to +0.75 V; bandwidth in MHz = 350 / rise time in ns (10 % to 90 %, source and probe taken out in quadrature). C: a multiburst; the bandwidth lies between the last packet at 0.707 of the bar or above and the first below it. Delay: the B step, between the 50 % points at the two jacks. Input capacitance: drive the input through 10 kΩ with no 75 Ω; C = 1 / (2 × 3.14 × 10 kΩ × f) at the −3 dB frequency f, less the probe's.

Fixture.

  1. Drive the input jack from a generator through 75 Ω in total: a 75 Ω generator, or a 50 Ω generator with 25 Ω in series.
  2. Load the output jack with 100 kΩ ±1 % in parallel with 30 pF ±5 pF, the probe's own capacitance counted: a 10:1 passive probe of 10 to 15 pF beside a 100 kΩ resistor and a 15 to 20 pF capacitor across the jack, or an active probe with 30 pF across it.
  3. Measure at the jacks with the shortest leads the probe allows. A 300 mm patch cable left in the fixture adds 30 pF and halves the figure of a wide path.
  4. Set every control in the path for its widest response, unity gain where there is a gain control. Record any setting that is not the default.
  5. Drive 0.5 V peak-to-peak about +0.5 V, which keeps the output inside U1 at unity gain; drive a path with gain proportionally lower.

Method A, stepped sine (the reference method).

  1. Apply a 100 kHz sine and read the output amplitude.
  2. Step the frequency up by a ratio of 1.4 or finer, reading the amplitude at each step.
  3. The bandwidth is the frequency, interpolated between steps, at which the amplitude is 0.707 of the 100 kHz reading.

A network analyzer, or an oscilloscope with a tracking generator, does the same in one sweep. A path whose response rises before it falls is reported at its −3 dB point all the same, with the peak noted.

Method B, step.

  1. Apply a 0.5 V step, from +0.25 V to +0.75 V, from a source whose own rise time is under a fifth of the path's.
  2. Read the 10 % to 90 % rise time at the output jack.
  3. Take out the source's and the probe's rise times in quadrature: the path's rise time is the square root of the read time squared less the source's squared and the probe's squared.
  4. The bandwidth in MHz is 350 divided by the path's rise time in nanoseconds.

The result is within about 10 % of Method A for a path with a smooth roll-off, which the buffers of the trade have; measure a path that peaks or rings by Method A. The same step, read at the input and the output jack on two channels, gives the delay between the 50 % points. The rise time of a shaped path, the figure 7.3 recommends for it, is read the same way with a step through the path's threshold.

Method C, multiburst.

  1. Apply a multiburst test signal through the path, from a video test generator or a module that produces one: a reference bar, then packets of sine at fixed frequencies, each at one amplitude.
  2. Read each packet's amplitude at the output against the bar.
  3. The bandwidth lies between the last packet at or above 0.707 and the first below it; Method A places it within that interval.

A standard-definition generator's packets run to 4.2 or 5.8 MHz and a high-definition generator's to 30 MHz, so Method C places a figure only within the generator's reach.

Input capacitance.

  1. With the module powered and the input otherwise unpatched, drive the input jack from a generator through 10 kΩ, with no 75 Ω.
  2. Read the −3 dB frequency f at the jack itself.
  3. The capacitance is 1 / (2 × 3.14 × 10 kΩ × f), less the probe's own.

An LCR meter at 1 MHz across the jack reads it directly.

Worked figures. One pole of 75 Ω into 30 pF is 70.7 MHz, with a 10 % to 90 % rise time of 4.9 ns. 75 Ω into 130 pF, a meter of cable and the input, is 16 MHz: the 1000 mm row of the link table read the other way. A 35.7 MHz path has a rise time near 9.8 ns and is written BW35.

VID-75 termination and drive

With the product powered:

  1. At each COAX-IN, measure 75 Ω ±5 % from the center pin to the shell (V75-010).
  2. At each COAX-OUT and buffered COAX-THRU, measure twice the line level open-circuit and the line level into a 75 Ω load (V75-010), and a −3 dB bandwidth of at least 30 MHz into 75 Ω (V75-012).
  3. At a looped THRU, measure 75 Ω across the line with TERM on and no load with TERM off, and confirm that the product locks either way (V75-009).