IC-705 Reference Stability


Some time ago, I wrote a product review for my new Icom IC-705 and one of the observations was on the setting accuracy of the radio's master reference oscillator. At the time, I noted that the "REF Adjust" value set at the factory was markedly different from the value I ended up setting on the radio in order to place known reference frequencies at the centre of the radio's spectrum display.

The phenomenon of aging in crystal controlled oscillators is well known and manufacturers often subject oscillators they either supply or use in their products to a pre-delivery burn in to deal with it. Crystal oscillator aging is a characteristic where the unit's mean frequency will slowly drift over time; the characteristic is normally monotonic and asymptotic to a "final" frequency which occurs after a very long time. Knowing the aging characteristic vs. time, manufacturers will burn in their products over the minimum time period until the aging rate slows down to one where periodic compensation, usually by adjusting some form of trim capacitor or synthesiser divider loop, will maintain the oscillator frequency within spec over a guaranteed number of years.

So, what about the IC-705, or, in this case, my IC-705? The first assessment shortly after receiving the radio was against the WWVH time signal running at 25 MHz. and found the REF needing adjustment by a substantial amount - from factory value of 47% to a new value of 52%. About a year after that in January 2024, I bought a Leo Bodnar GPS-disciplined clock oscillator (GPSDO) and did a re-check only to find a need to trim the radio's REF from 52% to 57%. - another substantial change. Another check against the GPSDO in October 2024 revealed the need for another REF adjustment from 57% to 60%.

It's all a bit concerning that the radio's master oscillator is still aging along a rather linear curve instead of slowing down, but there are only three data points available and things may get better as time goes on; it remains to be seen. Even assuming the current linear trend remains unchanged, an extrapolation suggests that it will be 10 years before the REF adjust runs out of range. I imagine by that time technology will have advanced and it will be time to replace the, by then, old faithful.

To be continued - I'll be keeping watch on the REF characteristics in nominally 6 month intervals.

Author: Hugo VE3KTN

Appears In: October 2024 Rambler Vol 67, Issue 2