KIWI OSD working at extreme temperatures

Page last updated: 22nd February 2007

Now that KIWI OSD is working in such a diverse range of climates and situations - people have wondered if it is affected by extreme temperature conditions.

Gerhard Dangl (Austria) decided to make some measurements to test his KIWI OSD in extreme temperatures. He has been very kind to share his findings with us. Gerhard has detailed his experiment in this EXCEL spreadsheet where you can see his method and tabulated results and graphs (click on tabs at bottom of spreadsheet for the 4 pages of the experiment).

This page gives a brief explanation of method and an "executive summary" of the results.

In the following image, the YELLOW dots show the 5 positions, that Gerhard placed precision temperature sensors. 
MP1 = Air temperature outside OSD case
MP2 = Air temperature inside  OSD case
MP3 = Quartz Crystal temperature
MP4 = LM1881 sync separator I.C. temperature
MP5 = Heatsink (of power regulator) temperature
 f  = Point at which system frequency was measured.

[Temperature probe positions on KIWI OSD pcb]


In the following image, we can see how Gerhard managed to get down to -25C by using a domestic freezer. The laptop is used to show if KIWI OSD is working (i.e. the quartz oscillator has started), to the right is the frequency counter, and then the multi meter used to measure the resistance (and hence the temperature) of the various temperature probes.
[KIWI OSD in the Freezer!]


Conclusions:
Using the temperature of the Quartz Crystal (MP3) as the reference. The highest temperature Gerhard measured was 55C and coldest -25C (a range of 80 degrees C). The maximum frequency variation (of the KIWI OSD oscillator) was: 
Frequency Variation  =    (2021 - 1845) Hz   = 176 Hz
Frequency Variation  = (176 / 16.00e6) x 1e6 = 11.0 ppm (parts per million)
over this temperature range. Because the only timing measurement KIWI OSD makes is the duration of the 1PPS pulse from the GPS - then for a change of 1 millisecond on the KIWI OSD display, the system timing would need to change by 1,000ppm, whereas the measured variation was 91 times LESS than that.

So KIWI OSD can "easily" maintain its 1 millisecond specification (to UTC) over this wide temperature range!

It is the GPS receiver that provides the precision timing reference (1PPS) that is phase locked to the atomic clocks on the GPS satellites.

If we examine the shape of the "temperature curve" of the KIWI OSD oscillator - that Gerhard plotted from his experiment (see above EXCEL file), - and compare it to the family of curves for AT-cut quartz crystals:


[KIWI OSD Oscillator] [Quartz Crystal Graph]
we can see the XTAL used in KIWI OSD has the same curve as the 1' (1 minute) delta to the AT reference cut! With this angle of cut of the Quartz Crystal - we have a nice flat temperature coefficient over an extended range.

From an electronic design perspective, the main issue with extreme temperatures is, will the quartz oscillator START when the device is initially turned ON? Gerhard made a number of extra "power on" tests - to check the oscillator would indeed START with extreme conditions. No start-up problems were found - which confirms it was worth adding the extra components in the design - to make sure the Quartz Oscillator would work over a wide temperature range!

At high temperatures the Heatsink is very much required - to keep the regulator from becoming a burn hazard to wandering fingers, and to ensure components keep cool and have a long life.

The Garmin 18 LVC GPS, has an operating temperature range of -30C to +80C. 
My sincere thanks to:
Gerhard Dangl (Austria) for his detailed experiment, and encouraging results.
KIWI OSD Homepage:

Regards, Kiwi Geoff