[NLRS] Freq stability improvement

[Dr. Gerald N. Johnson]

I claim that dissipating drive power to the transverter case is the 
prime cause of drifting while transmitting. Unless the transverter is a 
high power version (and I think most are not) there is very little 
difference in power dissipated in the transverter circuits between 
receiving and transmitting. The mixer needs only about 0.1 mw so all the 
drive over that is turned into heat. And the 50 ohm load is bolted to 
the case and the case is not far from the crystal on my 10 10G 
transverter. It already had the thermistor for crude frequency control 
and I added all the dow blue board foam bits I could around the crystal 
checking for stability with an independent microwave receiver I know I 
improved it considerably with the foam. I turned down the drive as far 
as I could without changing TC jumpers. I should change those (like you 
did) to reduce the drive requirement further and so the transmitter 
drift which isn't one I heard any complaints about this year. I also 
took care to keep the transverter powered for my drive to the first (or 
all day) operating site and powered all day. I'm sure the solar heating 
did affect the frequency some. I'm sure my frequency stability isn't 
perfect but the drift is slower than the tuning rate on receivers or I'd 
hear complaints or at least comments.

Controlling the temperature environment of any crystal will improve its 
stability no matter what the temperature held. There are optimum 
temperatures for a given crystal where the temperature coefficient is 
nearly zero and on a rounded peak or valley which makes temperature 
control less critical.

About a week ago I built up a one transistor Driscoll oscillator circuit 
as a candidate for driving an external reference brick for a 24 GHz 
transverter. Once I got it over my biases of choke and coil sizes, it 
works great and even without breeze protection of the crystal says 
within 100 Hz at 62 MHz. With a little environmental protection like a 
cover on the box and the box inside an insulate box I expect it will do 
better. And if that doesn't work I'm off on a wild track of using GPS to 
discipline a 96 MHz VCXO (Paul Wade scheme and boards) which inherently 
in the division down to 10 kHz for comparison to the GPS 10 kHz output 
(a few do, most don't) has stabile signals available at 12, 6, 3, 1, .5, 
.1, .02, and .1 MHz which might be handy in making a direct frequency 
synthesis for other handy frequencies. Like 3 divided by 2 is 1.5 and 12 
- 1.5 is 10.5 with added to 96 makes 106.5, a common crystal frequency 
for 10 GHz transverters. With a different prescaler (setting or chip) 
the frequencies could be 16, 8, 4, 2, 1 and down or 10, 5, 2.5, 1 and 
down. Without any modifications to Paul's $6 circuit board. The there's 
the 96 Mhz output that could be divided on a separate board with the 
handy MC120xx family of programmable prescalers.

Things are quite interesting these days in frequency control at moderate 
to low cost.

73, Jerry, K0CQ

On 11/10/2013 1:35 PM, Bill Davis wrote:
>
>
>     Thanks to DEMI and a lucky draw at MUD, I was a "lucky dog" at the prize table and came home with a DEM L144-28INTCK kit. This is Steve's 2m transverter that is designed for use with higher band/microwave transverters with 144MHz IF. Part of the incentive for selecting the transverter was  limited "good" choices at the time my number was read AND I had not built any DEM gear in a very long time. The new designed transverter would be a fun project and with some input level mods for 902/1296 and 10GHz, would be something new in the shack. One of the "new features" of this design, is an improved Xtal heater and insulated enclosure for the osc. (Important for what follows)
>
>     I built up the new transverter this past week, a very nice kit and modified the transverters to allow lower drive levels (prior 1-2W on all). All bands seem to be playing well. Yesterday I started to wonder, I wonder how much of my 902/1296/10GHz freq stability issues while attempting WSJT in the past on those bands, were associated with the older designed 25W DEM 2m tansverter that I was using, as the 144Mhz IF? Answer, MOST if not All! This morning Gary and I ran WSJT on 10GHz and 902 with VERY good freq stability. Gary is GPS locked on both bands. We didn't try 1296, the older DEM designed 1296 transverter may be stability challenged, we will see. But 902 and 10GHz was doing well. Solid copy on 902 and one good decode of Gary on 10GHz. What appears to be multipath coupled with the very weak sig on 10GHz didn't allow much more in out brief quick test. We were using WSJT JT65B, the same mode as we use with 2m WSJT EME.
>
>    More will follow as we make other tests. A FUN few min added to our regular twice daily sked. Thanks Gary!
>
>
>    Thanks also to Steve for the nice items on the prize table AND the osc stability.
>
> 73  Bill   K0AWU
>