The input signal is an analog 700mVp-p frequency sweep, from 0.5 MHz to 5.75 MHz. The input operating mode is analog RGB, with 625 timing.

The passband ripple measures 0.50 dB p-p to 5.4 MHz. See Figure A-1 for a typical response. The sweep signal in the figure is measured at the input to the A/D converter and is more than 2.0 V p-p in amplitude.

The input signal is an analog 700mVp-p (PAL) 2T Pulse with a HAD of 200 nsec. The input operating mode is analog RGB, with 625 timing.

The K-factor measurements ranged from 1.1-1.3% and averaged1.2% for a typical response; see Figure A-2. For an HAD of 250 nsec(NTSC), all channels would be <1.0%.

The K-factor measurements meet a <1% figure for NTSC 2T pulses.

The S/N ratio is measured at the input to the A/D converters. This includes the effects of the breakout box on the signals. Noise components can result from analog or digital crosstalk in the breakout box and on the Sirius VME board, and from thermal, excess, or popcorn noise sources within the circuitry.

First, measurements were taken with crosstalk effects eliminated as much as possible. A 50% gray flat field was applied to one channel at a time with no signals applied to the D1 inputs or to the inputs of the channels not being tested. Finally, 5.75 MHz full amplitude sweeps were applied to all analog input except for the input under test; simultaneously, D1 4:2:2:4 5.75 MHz full amplitude digital sweeps were applied to the D1 inputs. This second case is a worst-case scenario that does not represent normal operating conditions.

The S/N results summarized in Table A-2 were obtained with the VM700A and active probe modified for variable gain. A bandwidth of 100kHz-5.0 MHz was used with no weighting. The A/D input test points were the point of measurement. The specification is >=65 dB, weighted.

Typical S/N ratio with crosstalk included is 55 dB.

A horizontal rate signal from the input genlock phase-locked loop was measured for jitter using a Hewlett-Packard time domain analyzer. This instrument measures peak-to-peak jitter and standard deviation jitter; it was set to collect data on 1000 samples.

The maximum peak-to-peak jitter was 9.8 nsec, with a standard deviation of 0.65 nsec. It was estimated that 95% of the samples had a peak-to-peak jitter of less than 4.0 nsec. These are good numbers for an H-rate genlock system.

The input genlock circuit contains a mux that selects from the various possible genlock sources. Both the S/N ratio and crosstalk were measured by applying multiburst signals to the composite, R, G, B, alpha, and sync inputs. The house (genlock) input was selected as the output genlock source, but no signal was applied to this input. The input to the genlock A/D converter was then measured for S/N ratio, which includes crosstalk.

The resulting S/N ratio was 76 dB over a bandwidth of 100 kHz to 5.0 MHz, no weighting, and with averaging on. This indicates that there is a very high S/N ratio with negligible crosstalk.

Figure A-8 shows input genlock S/N ratio; crosstalk is worst-case and not comparable to normal operation. Multiburst was applied to the applied to R, G, B, A, composite, and sync inputs. On the mux, “house” was selected, with nothing connected to that input.

The quality of the composite and Y/C inputs was measured by inputting composite analog test signals and measuring the composite output signal. This technique measures the combined performance of signal decoding and encoding. The output signal quality was measured separately and is summarized later. The frequency response and S/N ratio using the Y/C inputs would be even better, due to the absence of chroma/luma separation. The encoder and decoder only process 8-bit video.

An NTSC multiburst signal was applied to the composite input (and to all other inputs) and measured on the composite output; Figure A-4. The frequency response is quite flat to beyond 2.0 MHz. Crosscolor, which is a normal effect, is visible on the last three bursts. The level of the signal, as well as the shape and amplitude of sync and burst, are correct.

Figure A-5 shows the decoded and encoded 75% color-bar signal (TSG-100). Figure A-6 shows the vector display of this color-bar signal. Except for the slightly hot subcarrier level, the vector accuracy is within all of the small boxes, with negligible offset.

The S/N ratio of the composite in-out path was measured using a 50% gray field from a TSG-100 generator, and making the measurement with a VM-700A. Figure A-7 shows the results. A S/N ratio of -49 dB was obtained with a bandwidth of 10 0 kHz-5.0 MHz and no weighting.

D1 4:2:2:4 video in, with measurements made on the G, B, R, and alpha output channels of the breakout box. The input operating mode is parallel 4:2:2:4 input with CCIR601 525 timing. The output operating mode is RGB output, also with CCIR601 525 timing. The D1 test signal is a full amplitude 5.75 MHz sweep signal. The RGB output sweep was measured on an oscilloscope with signal-to-noise ratio amplitude markers set for a 0.5 dB p-p spread.

The measured response under +-0.5 dB ripple conditions extends to 5.2 MHz with sinX/X correction, relative to 100kHz; see Figure A-8.

The measured response under 1.0 dB p-p ripple conditions extends to 5.5 MHz, relative to 100 kHz. The specification is +-0.5 dB to 5.5 MHz. This response includes SinX/X correction for the 601 sample clocks. For the square pixel 525 mode, the response will be down by an additional 0.1 dB at 5.0 MHz; for the square pixel 625 mode, the response will rise by an additional 0.08 dB at 5.0 MHz.

The output frequency response typically meets the specifications.

The input signal is a D1 4:2:2:4 PAL 2T Pulse with a HAD of 200 nsec. The output operating mode is analog RGB, with 625 timing.

The PAL K-factor measurements uniformly measured 0.8% for a typical response; see Figure A-9. For an HAD of 250 nsec (NTSC), all channels would be <0.8%.

The output K-factor measurements meet a <1% figure for both NTSC and PAL 2T Pulses.

Disregarding crosstalk, the inherent S/N ratio of any of the R, G, B, or alpha channels is 78dB, measured 100kHz to 5.0MHz, with no weighting. The effect of output crosstalk on S/N ratio was measured by outputting a 5.75MHz sweep signal on the RGB and Alpha outputs, but disabling the output lookup table on one channel at a time, and measuring the S/N ratio of that channel. A bandwidth of 100 kHz to 5.0 MHz and no weighting was used. The results are summarized in Table A-3.

The output sync signal is available in either 300mVolt or TTL (4V) level, as controlled by software.

The 300mV version has a 75 Ohm output impedance and is measured into a 75 Ohm load. The level is very precise, with an accuracy of <3%, is Gaussian-shaped, and has a rise-fall time of 140 nsec (+-10%). Figure A-10, Figure A-11, and Figure A-12 show typical waveforms.

There is a small 25 mV 30 nsec glitch at the beginning of each transition; but this is unlikely to cause a problem with any external sync detector. The glitch is due to crosstalk around the TTL buffer that produces the 4V sync.

The 4 Volt sync output was designed to be terminated in a 75-Ohm load; with a typical amplitude of 4.5 V p-p +-.25V under these conditions. It has a rise-fall time of 18 nsec, and a very low output impedance. The DC level of sync tip is ground, the polarity is negative-going.

The output jitter was measured with a Tektronix VM700A video measurement set by measuring the video and sync waveforms on the output of the breakout box, and by a time domain analyzer (TDA) that measures jitter on TTL logic waveforms. The VM700A has a minimum jitter of 3 to 4 nsec, when measuring a high-quality generator output signal. Jitter was measured with no additional filtering on the VM700A, and with a 1.5 MHz Gaussian filter inserted in the path.

The output genlock jitter is very low, at approximately 1nsec measured with the TDA. The results shown in Table A-4 were obtained with the VM700A, and are for the CCIR 601 525/625 modes, with essentially the same results obtained for the square pixel 525/625 modes.

The SYNC OUT and GBRA/YUVA COMPONENT SYNC outputs on the breakout box have very low jitter when high frequency noise is filtered out. The jitter approaches the residual of 3-4 nsec. When correlated with the results of the TDA, the conclusion is that the non-filtered jitter is the result of high-frequency noise contaminating the video appearing at the breakout box outputs.

These conclusions also apply to NTSC composite output, which exhibits similar measurements as the Sync and component outputs. However, the PAL composite output has much higher jitter. This increased jitter is due to noise on the chroma output that occurs during the sync edge. The Y/C Luminance output signal does not have the noise, and does not exhibit the increased jitter. The noise on the chroma channel may be produced within the encoder chip.

By using the “square root of the sum of squares” formula, with a residual jitter measurement of 3.5 nsec and a measured jitter of 5 nsec, the output genlock jitter is calculated at 3.5 nsec. This is an acceptable figure for a line-locked system

Output genlock crosstalk and S/N ratio were measured by applying full amplitude 5.75 MHz sweep signals to all inputs except for either sync or house (genlock) input. The S/N ratio was measured at the analog input to the genlock circuit, with a bandwidth of 100 kHz to 5.0 MHz, no weighting, and averaging on. The S/N ratio was measured both on line 17, where there could be no crosstalk because it is not an active line, and on line 50, where there would be crosstalk. The output genlock mux was used to select either the sync or house input. Table A-5 summarizes these results.

These results are worst-case conditions not comparable to normal operation. The 59 dB figure, when the sync input is selected, is due primarily to equal crosstalk from the green and composite inputs. This figure indicates that the house input is the preferable one to use to minimize crosstalk.

The quality of the composite and Y/C outputs was measured by inputting digital 601 75% color bars and outputting composite color bars.

Frequency response was measured by inputting a digital 601 5.75 MHz sweep signal and viewing the composite output. The response is flat to beyond 5.0 MHz, as seen in Figure A-16.

Digital 75% color bars were input to the system. In the composite output, the levels and appearance of the bars is correct, and the levels, shape, and timing of sync and burst are correct.

Figure A-19 shows the vector display of color bars, and shows a slightly hot subcarrier amplitude with excellent vector accuracy and negligible offset. A flat 50% gray field was applied to the digital 601 input and output on composite. Figure A-20 shows that the S/N ratio is 50 dB.

Figure A-19 and Figure A-20 show the output timing of green versus red, and green versus blue; and indicate that the channels are coincident. The input to the breakout box was composite color bars from a TSG-100 generator, and the output was RGB from the breakout box.