The AM Forum

Ola,

I have a solid state 6M transverter that has some drift in the 22 Mhz crystal oscillator. The manufacturer sent me a thermistor to correct this problem.  

As per their directions, one side of the thermistor body is soldered to the metal outside shell of the xtal and the other lead goes to +12Vdc.   The metal shield of the xtal is floating.  I measured infinite resistance from the xtal shell to ground, so there is no DC path through the thermistor to ground.  

So what is it actually doing here?  All I can figure is that the thermistor is like a variable resistor from the xtal shell to B+.  What does that do to stabilize the xtal circuit???

I actually had my own homebrew heat oven in there, but figured this was a better solution when I was told about it...

T

Tom, they saw your get up and decided to see if you would actually try it.
actually you hace the thermister in series with the case C

It must work - cuz they wouldn't screw with Elvis!

So, what is actually happening in the circuit to null out drift?  As the thermistor gets warmer it shunts the case so ground and pulls the osc lower in freq, or something like that?

The thermistor is a 60 degree part according to the sheet.

I've been watching the osc freq and so far I'm not impressed. It starts out cold at 19.99989  and works its way up to 20.00013 within 5 minutes.  Then hovers around 20.00016.  Does not seem much better than before and worse than when I used an oven.  I can adj the trimmer to put it on 20.00000, but in the warmup mode, 200hz drift is noticable on ssb.

This is for long term stability AFTER warmup, right?

T

Sounds like Jersey Black Magic to me. But now you have a do-dad hanging on your crystal shell. Thermistor is floating one lead(case) in air. Can't see how anything would change the stability of the crystal. I think you need more input from them.

Tom,
I've seen a lot of compensation tricks but that is a new one on me.
I have seen how the case termination effects frequency.
I would use the thermistor and a varactor diode to compensate the circuit.
Every crystal has a temperature where it oscillates at the highest frequency. above and below that temperature the frequency drops.
This temperature is usually 25 to 35 degrees C. First you need to find out which way you need to go after the circuit has warmed up. Changing the bias current or voltage to the case sounds like it is missing something.
fc

A-HA!

Looks like you were both correct in being suspicious.

I just called them and they said  indeed, the Xtal case must go to ground. They omitted it on the instructions.

What happens is that the thermistor actually heats up the xtal case and turns on and off like a thermostat - at least that's what I understood from them.

I asked if maybe I shud supply a constant 12V there all the time when the gear is off and they agreed.

I'll add this and make some measurements on drift and post here.

Tnx for the heads up, OM's.

T

Tom,
Sounds like the thermister needs to be clamped to the case so it is seeing the case temperature and actually heating it up. Epoxy, super glue
or dental cement will work. Be carefule soldering the case of a crystal so the seam doesn't pop apart. A good crystal has nitrogen gas inside and any crap that falls on on the crystal like flux will change the frequency and
temperature effects.

May be their new version of the gear has modification, grounded can?

Hey T,
That avatar splits my gut every time I look at it.
Bob

Yep, Frank, one side of the flat side of the thermistor gets soldered to the xtal case and the xtal case gets grounded.

I am running it this way and see a good amount of current is flowing thru the case and it stays warm. I leave it on all the time and see that the drift is much better from a cold start.

I see the voltage drop on the 12V supply slowly creep around showing that the thermistor is working it's thing.

Appears to be a good mod. No damage to the xtal after soldering.

Yes, Bob, that Elvis shot is about my favorite of the last 5 years. It never gets old...    There's lots of new heads to add...:lol:  

T

gads, Thats an old gimick.  Used it before, about 1971ish for a UHF radio that would wander with temp as a quck fix.

The thermister is a positive temerature coeffiecient resistor, IE it;s resistance goes up as it heats up.

When you put one of those across the power they tend to heat up to a point and self regulate at that temperature, attanching it to the crystal then heats the rock to stabilize it.   Call it a poor mans oven.

My .02$, any 22mhz crystal osc that drifts that much really suckith!   I've built 5 6M radios (my design) and with microprocesor crystals and a decent osc I expect and get less than 50hz warmup drift.

A better way is to use the thermister to react to the temperature and apply a bias to a varicap to pull the rock to nominal.  This is a cheap and dirty TCXO approach.

Allison
KB1GMX