[ARC5] Drift in ARC-5s - real-world measurements. - quite long.

[Kenneth G. Gordon]

OK. I dug out the ER Magazine in which the drift of those 
transmitters was measured and discussed.

The title of the series is "Secrets of the Dead...Command Sets, 
Part 2", and written by Mike Murphy WU2D, published in Electric 
Radio Magazine number 231 of August 2008 starting on page 20.

5 different "ARC-5" variants were tested.

Power output, for all transmitters tested, was 15 watts into a 250 
watt 50 ohm dummy load, coupled by a PROPERLY 
CONFIGURED three-element "Tee" network.

Measured drift on 80 (3600 Khz) and 20 (!) (14040 Khz) averaged 
less than 1 Hz per minute.

The transmitters were the following:

1) BC-696 in complete original condition. This unit chirped badly. 
Oscillator and finals keyed during testing periods (to be described 
below). I.e., the oscillator was only "on" during the short periods 
when it was keyed during the tests. Otherwise it was off, although 
the filaments were lit and plate voltage was applied. This is the way 
they originally worked.

2) T19/ARC-5. Owned by Mark WA1QHQ, converted using the 
1962 handbook method. 12 VAC filaments, all relays removed, key 
jack added. Only the finals are keyed. Oscillator runs continuously.

3) BC-458. This unit was SEVERELY damaged, then rebuilt as an 
80 meter mobile transmitter. Only the finals are keyed. Oscillator 
runs continuously.

4) BC-696: This is unit #1 with a new 1626 and that rear-panel-
mounted triple-can capacitor replaced. Chirp is now non-existent. 
Drift is only slightly less than before.

5) BC-457A: Chassis arrived completely stripped except for the 
VFO. Rebuilt as a fully functional 20 meter CW transmitter. 1626 
replaced with a 12SK7 operating at 7 Mhz. 12A6 buffer/doubler. 
Voltage regulated oscillator screens. Only finals and buffer keyed.  
Oscillator runs continuously.

All power supplies (except in the case of # 5) were UN-regulated, 
although the power supplies were quite "robust".  I.e., were capable 
of considerable current output. Oscillator voltage is 250 VDC. Final 
voltage is 450 VDC.

The first reading of frequency was taken at 6 minutes after turn-on, 
thereafter, every 15 minutes for just long enough to get a reading.

Test equipment used was all modern.

Total time for each test was 2 hours, or 6 readings, total. 

The LAST, or 7th, reading was an "Old Buzzard" test in which the 
key was held down for 6 minutes.

The author states that the 1626 oscillators "...drifted mightly..." for 
the first 15 minutes.

Personally, I find that his description of "...mightly..." would not be 
something I would agree with. The results for the first 15 minute 
test (AFTER the 6 minute "warm-up") are as follows:

1) -273 Hz.

2) -467 Hz.

3) -930 Hz

4) -507 Hz

5) -1030 Hz (on 20 meters, doubling from 7 Mhz, or -515 Hz at 7 
Mhz.)

I am going to skip most of the rest of the "standard" tests, but will 
tell you that after the first 15 minutes, the drift slowed markedly.

For instance, between test 2 and test 3 the drift was as follows:

1) - 56 Hz

2) + 34 Hz

3) + 66 Hz

4) + 12 Hz

5) - 30 Hz (!)

Total drift for the entire 2 hour period, or 120 minutes, start to stop, 
was as follows:

1) - 542 Hz.

2) - 437 Hz.

3) - 1093 Hz

4) - 669 Hz

5) - 870 Hz

Skipping to the "Old Buzzard" test, i.e. transmitters keyed on for 6 
minutes at full output of 15 watts the total drift measured DURING 
that 6 minute test was as follows:

1) - 519 Hz

2) + 68 Hz

3) - 10 Hz

4) - 463 Hz 

5) - 340 Hz (again, on 20 meters).

So, as some of us have repeatedly stated here, the "ARC-5" 
transmitters do NOT drift unusually if run at original specified 
ratings.

In fact, at the time they were plentiful, they were FAR more stable 
than most of the VFOs sold commercially or built from handbook 
articles of the time.

Even today, 65 years and more after their first use, they will give 
many more-modern VFOs a run for their money... if run according 
to original specifications.

I rest my case.

Ken Gordon W7EKB