V
VK2RK
Guest
In the development of an accurate ovenized crystal oscillator, many lessons have been learned along the way, such as cause due temperature drift of components with effects of aging in the short and long term with further environmental effects that needed to be understood and considered.
The first effect I noticed was that altering the position of the oven from a horizontal to a vertical position caused a frequency shift, not a large change but never the less a change took place. I searched the web for any other that noticed this repeatable effect, I came across a NIST article also sponsored by NASA that spoke about the effect of magnetic field upon a quartz crystal, it also spoke about atmospheric pressure along with temperature shock etc. all effects that effect the stability of the oscillator. This was very revealing read providing me with answer and better understanding.
In the process of development I noted that the second hand modules seemed to require an aging (Heat soak) cycle before the design criteria was met. Again I went looking on the web and I found this document that explains what I have observed.
The lesson is don't assume an initial calibration to hold for very long, the unit needs to age before the required stability is achieved.
The problem is to ascertain what is causing the drift as the observed change is the sum of all drift factors, once the components used that provide the controlled voltage is as stable as possible against environmental temperature change are met, what is left is the aging along with physical effects placed upon the crystal by the environment.
QUOTE (Power cycling of OCXO)
" What does all this mean in practice? First, an ovenized oscillator should be continuously powered if at all possible. If power interruptions are unavoidable, be aware that the oscillator will take some time beyond normal warm-up to return to the prior aging rate and, because of aging and hysteresis, is unlikely to return to exactly the same frequency. Hysteresis for AT-cut resonators is unlikely to be much better than a few parts in 10-8. Frequency adjustment during the re-stabilization period is not a good idea. With its extensive experience designing precision ovenized oscillators, Vectron is able to assist in determining an appropriate re-stabilization period, especially if the power-offperiod exceeds a few days and a longer re-stabilization period is necessary."
The first effect I noticed was that altering the position of the oven from a horizontal to a vertical position caused a frequency shift, not a large change but never the less a change took place. I searched the web for any other that noticed this repeatable effect, I came across a NIST article also sponsored by NASA that spoke about the effect of magnetic field upon a quartz crystal, it also spoke about atmospheric pressure along with temperature shock etc. all effects that effect the stability of the oscillator. This was very revealing read providing me with answer and better understanding.
In the process of development I noted that the second hand modules seemed to require an aging (Heat soak) cycle before the design criteria was met. Again I went looking on the web and I found this document that explains what I have observed.
The lesson is don't assume an initial calibration to hold for very long, the unit needs to age before the required stability is achieved.
The problem is to ascertain what is causing the drift as the observed change is the sum of all drift factors, once the components used that provide the controlled voltage is as stable as possible against environmental temperature change are met, what is left is the aging along with physical effects placed upon the crystal by the environment.
QUOTE (Power cycling of OCXO)
" What does all this mean in practice? First, an ovenized oscillator should be continuously powered if at all possible. If power interruptions are unavoidable, be aware that the oscillator will take some time beyond normal warm-up to return to the prior aging rate and, because of aging and hysteresis, is unlikely to return to exactly the same frequency. Hysteresis for AT-cut resonators is unlikely to be much better than a few parts in 10-8. Frequency adjustment during the re-stabilization period is not a good idea. With its extensive experience designing precision ovenized oscillators, Vectron is able to assist in determining an appropriate re-stabilization period, especially if the power-offperiod exceeds a few days and a longer re-stabilization period is necessary."
Last edited by a moderator:
