Exhibitor Info: (214) 521-9430 Registration into: (214)423-3498

Date 1 April 1987

Spacecraft NOAA-9 NOAA-10

Orbit Number 11844 2778

Eq, Crossing Time (UTC) 0054 19 0102.14

Nodal Period (Hin.) 102.0638 101.2979

Frequency (MHz) 137.62 137.50

These orbital parameters arc-projected two months ^n advance due to deadline considerations Accumulated errors due to uncompensated orbital decay and other anomalies result in expectation of errors up to iwo minutes and possibly as many degrees in terms of the crossing data and possible small changes in the indicated period. Users requiring precision tracking data snould rely on more current sources.

Table 2 TIROS/NOAA orbital predict data chain can be used for display timing as well! Such tricks are one of the primary reasons thai the entire WSH scan converter circuit, including all satellite video, timebase. computer interface, and TV display circuits require only 13 ICs with a total component cost ot about $60'

Timebase Adjustment

If a timebase is to do its job. it must be operating at just the nghi frequency As you shall see in a moment, close is not good enough Theprincipal virtueol the direct subcarner lock approach, aside from simplicity, is that the only adjustment to be made is adjusting the PLL for Jock on the subcarrier All of the other approaches require that a crystal oscillator be irimmed to a precise frequency. There are three ways to do this.

The first and most obvious approach is to use a frequency counter on the osciHaior in question and adjust the relevant trimmer capacitor for the proper frequen* cy. Don't bother with this one unless you are sure about the absolute calibration of your counter, If you are a few hundred Hz off in the 2-5-MHz range, you will get unac* ceptable results!

Your run-of-the-mill counter can be used, however, if you first examine the results of a live printout or display. The geometry of the display can give you a(l the information you need to get the oscillator on frequency regardless of the calibration of your counter.

Photo A is an example of a visible-light NOAA FAX print I re*

ceived in the mail with an under^ standabie "What s wrong? query. What is wrong is that the ctock frequency is off but the question is, by how much?

Note that the readout is progressively offset to the left from its initial starling posiron at ihe top This indicates thai the clock frequency is LOW, If the tiTt had been to the right, the frequency would have been HIGH Had the image been precisely vertical, the frequency would be right on (see the third option below), but then I never would have gotten the fetter! OK, it's low in frequency, Now let's find out how Jow!

Using the trailing edge of ihe sync pulse, together with the minute markers (white horizontal lines in the pre-earth space scan), you si art by physically measuring the offset over a specific time interval. In this case, the offset totals 14 mm (measured from the original print)

over a period ot 4 rmnuies.

Since the print width is 142 mm, the total offset error accumulated over 4 minutes is 9 86%—(14; 142) x 100. Each ¡fine is 250 ms long, making ihe accumulated time error 250 x 0 0986r or 24 65 ms, over 4 minutes Since 4 minutes is 240 seconds, you have an error of 24 65 ms/240 seconds, or 0.102 ms each second. This may not seem like much of an error, but it represents 0 01%—[0.102i 1,000) x 100.

This particular recorder was using the circuit in Fig 1(c) as the timebase. so the 0.01% frequency error represents 368 Hz— 0 0001 X 3,679,545. You already knew that the frequency was low, so in this case you want to raise the clock frequency by 368 Hz,

Now comes the reason for ail this calculator punching No matter how well or poorly your counter is calibrated, simply hook ii to the TUNE point in the circuit, note whatever reading you get, and then adjust the trimmer until it is higher by 368 Hz! This technique can be used with any display as long as you can measure the offset error over a known time period. All this requires is counter resolution, not precision calibration, since you are making a relative frequency adjustment

The finai approach, to be used if you have no counter but plenty of pat:ence. is to r-ake very small adjustments in your trimmer while looking at the results of lave printouts or displays with each change. The goal <s to get the printout precisely vertical which means that everything is on frequency!

You must use live transmissions for each run since a recording will always preserve any error that was present when the recording was made and what you really want to do is check each adjustment against a live reference Signal, Once properly adjusted, your recordings will also come out cor recity since the reference frequency is now on the money1

Well, 1 have thoroughly run out of space this month, but at least t have made good on an introduction to timebases. Next month, I will look at the many aspects ol image resolution, an often misunderstood subject'


References to WSH refer to ihe third edition of the Weather Satellite Handbook, available from yours truly for$12.50 plus $1 shipping in the U S, and £2 elsewhere. ■

Was this article helpful?

0 0

Post a comment