I have been using quarz based oscillators for most of my professional life, as well as making a first 1 transistor  Collpits osc  as a kid for RF down convertor to receive VHF on an HF reciever..  I may be able to advise ;-)

So is there anything unusual about this system ? What package is the crystal in, what is its crystal cut, I assume it is fundamental resonant not overtone, and how accurate do you need the frequency will determine the temp control you will need.
Phase noise will determine the drive level and topology.  DO you need any pulling ability ?

Mike

Hey Mike, that's great to hear!

It's in the creation of a microbalance system, it's going to be used under vacuum, in terms of package they're 1" research crystals that look like this (1 inch (25.4mm) Research Crystals (testbourne.com))

We're driving this at 9Mhz, and monitoring the amount of molecular deposits that land on it over time, and comparing this to a reference crystal which is the same but not exposed to the deposits (so you can see the difference over time). 

the way we'd been interacting with the crystal currently was using off the shelf pogo pins but they're a bit heavy handed and it meant that assembly of the system tended to result in the potential to crack the crystal if you weren't ultra careful putting everything together. 

I was looking towards potentially a spring contact this time to lessen the impact it has mechanically on it's oscillation. 

Yeah we're looking at resonant frequency, accuracy isn't ultra critical as we have the reference crystal to compare to which is in the same environment, just not open to the deposits (so it should always give a reference at 9Mhz), 

in terms of monitoring we're in the realm of days and weeks, but accuracy, my understanding is we're only going to be looking at incredibly small drops in frequency even when the crystal is fully saturated, so it's not going to drop from 9 to 5 mhz, it's going to be something like 9mhz and a drop of tens of Hz. 

ah yes.

Interesting.

I have done something similar for a measuring metal deposition in an evaporator jig we had in a clean room  - about 20 years ago,  when I worked at a place with a clean room on-site.  Actually we used to buy cheap crystals made for TV in a  metal can and then drill, well mill actually,  the case to expose the actual crystal  of about 10mm dia and  250um thick.
. As we were mainly measuring aluminium thickness we used to etch back and re-use the crystals sometimes. From memory we used to have a mixer and counter arrangement  looked at the difference in frequency between the one being plated and the other not, and with metal at the vibration antinode we got something like few  100Hz per micron on the test slide so I reckon some tens of nanograms per Hz on the crystal sweet splot of about 1cm area. My memory of the numbers may be a bit off.

Being canned the connection  to crystal problem was replaced by a connection to wire problem and we had berylium copper or phosphor bronze springs, not sure which now, that were rather like a PCB edge connector.

 Yours look rather  more up-market - are you using them with liquid contact as that datasheet suggests ? I presume that you do not clamp the crystal wafer edges, or do you?  Can you sketch or photo your arrangement? Also how far away is the oscillator electronics? If you are not careful you will end up measuring things that do not just depend on the crystal - is your oscillator usinf the crystal as a series (transmission) element or parallel (so one wire and gnd to the Xtal.)

To temp control the reference outside the vacuum we cheated and soldered a power transistor to the can of the reference rock, and had a circuit that put more  or less current into it to maintain a steady voltage on a diode also in good thermal contact used as a thermometer and tracking a similar diode in the vacuum. Nowadays I'd probably buy something.

M.

ah yes.

Interesting.

I have done something similar for a measuring metal deposition in an evaporator jig we had in a clean room  - about 20 years ago,  when I worked at a place with a clean room on-site.  Actually we used to buy cheap crystals made for TV in a  metal can and then drill, well mill actually,  the case to expose the actual crystal  of about 10mm dia and  250um thick.
. As we were mainly measuring aluminium thickness we used to etch back and re-use the crystals sometimes. From memory we used to have a mixer and counter arrangement  looked at the difference in frequency between the one being plated and the other not, and with metal at the vibration antinode we got something like few  100Hz per micron on the test slide so I reckon some tens of nanograms per Hz on the crystal sweet splot of about 1cm area. My memory of the numbers may be a bit off.

Being canned the connection  to crystal problem was replaced by a connection to wire problem and we had berylium copper or phosphor bronze springs, not sure which now, that were rather like a PCB edge connector.

 Yours look rather  more up-market - are you using them with liquid contact as that datasheet suggests ? I presume that you do not clamp the crystal wafer edges, or do you?  Can you sketch or photo your arrangement? Also how far away is the oscillator electronics? If you are not careful you will end up measuring things that do not just depend on the crystal - is your oscillator usinf the crystal as a series (transmission) element or parallel (so one wire and gnd to the Xtal.)

To temp control the reference outside the vacuum we cheated and soldered a power transistor to the can of the reference rock, and had a circuit that put more  or less current into it to maintain a steady voltage on a diode also in good thermal contact used as a thermometer and tracking a similar diode in the vacuum. Nowadays I'd probably buy something.

M.

Yeah i was thinking something like spring contacts to drive the crystal to replace the pogo pins as they were a royal pain to fit anyway. 

No liquid contact no, at present the entire crystal sits on an o-ring, but I did see other parties using clamps on the crystal edge, they almost looked like keyrings you prise apart slightly and gently put over the sides of the crystal. 

In terms of distance to the oscillator electronics, currently the design we had (which we're trying to revise), has the electronics on the other end of the pogo pins, so if you think there's the crystal, then two sets of pogo pins, and they're mounted to a pcb so it all sits quite compact, so distance wise you're looking at a centimetre or two at most. 

I believe the crystals in parallel, on X1 and X2 of a 74LVC1GX04 chip.

The reference crystal is also under vacuum. 

Yeah we've been thinking of using a TEC, or power resistor cooling / heating wise.