Testing crystal oscillators in high pressure conditions
01 April 2020
Figure 1. Diagram showing arrangement of test samples within pressure chamber
Following questions from customers with demanding applications, frequency control specialist, Euroquartz recently undertook a series of tests to understand what would happen when its crystal oscillators were subjected to up to 500 bar of pressure.
This case study was originally featured in the April 2020 issue of EPDT magazine [read the digital issue]. Sign up to receive your own copy each month.
Applications where high pressures might be encountered include subsea systems, downhole installations and other hostile industrial environments. While the company was able to speculate with reasoned argument, it had no firm evidence to hand. So, as Andy Treble, Sales & Marketing Director at Euroquartz tells us in this case study, testing its crystal oscillators under high atmospheric pressure was necessary to confirm the outcome.
The industrial grade crystal oscillators used for testing were produced in ceramic packages, with metal lids. Under increasing pressure, it was expected that deformation of these external packages and lids will, at some point, become so extreme as to cause the oscillator to fail. As well as determining the point of pressure increase where the first failure occurs, knowledge of the different mechanisms of failure also need to be gained and understood to help improve product design and quality.
The method used to test the samples was to hire a third-party pressure test facility (Avalon Sciences research centre in Somerset).
Due to limitations of time and resources available, an efficient test method had to be devised. Because of the extreme pressures the facility can generate, physical windows or real-time viewing media was not possible. As such, the pressure facility was not able to support real-time visual monitoring of test samples.
Each IUT (item-under-test) had to be placed within a sealed steel chamber and immersed in a solution of water mixed with anti-corrosion agents for the duration of test. The chamber offers the ability to make limited connections to the IUTs via a set of pressure sealed through-studs and crimp connections (figure 1). This would allow for a minimal level of real-time monitoring of oscillator output.
IUTs can only be visually inspected after pressure has been returned to atmospheric, thus allowing them to be removed from the test chamber.
With the resources available, the most practical way to examine function and physical resilience was to split the testing into two separate methodologies:
1. Function only test
Figure 2. Showing flat plane of metal lid pushed down due to increase in differential pressure
• Prior to the commencement of testing, integrity of the samples was confirmed by visual magnified inspection and a helium fine leak test.
• A methodology was developed to test the function of the oscillators while at pressure, and a stepped change in pressure was arranged. Once each pressure stage was reached, the pressure was stabilised to allow for a reading to be taken.
• Each oscillator shared a common supply, with a common ground and common signal output line.
2. Physical structure test
• This was a purely mechanical test to assess the increased pressure differential between the outside of the package and the inside, no electrical performance was considered.
• The following pressure targets were agreed: 25; 50; 100; 200; 300; 400; and finally, 500 bar.
• Five sample parts of each oscillator type were placed in a polyethylene bag inside the sealed chamber and pressure increased to each target value.
• Seven independent pressure tests were carried out. Each subjected the 15-piece batch to one complete pressure cycle, holding each pressure for five minutes before returning to atmospheric pressure.
• Each sample was visually examined after each test, and a fresh set was used for each of the seven pressure levels.
Stage 1: Metal lid distortion
The flat plane of the metal lid is pushed down on its upper surface by the increasing pressure differential, such that the flexible metal distorts downwards in a concave fashion (figure 2). This distortion then applies inwards tension all around the sides of the welded lid seal to accommodate the change of shape, thus the tops of the brittle ceramic walls have increased inwards forces applied. These forces are greater at the top of the walls, near the weld, than at the bottom, as the hard brittle (non-flexible) ceramic base has resisted distortion and, unlike the lid, its dimensions remain unchanged. The tensile forces have the effect of pulling the top of the side walls inwards, rotational around a pivot point where the walls join the ceramic base.
Stage 2: Ceramic case fracture (side wall)
When the forces become large enough, the brittle ceramic will fail at a weak point. In some cases, the walls fracture away from the base at the pivot point, and in others, a triangular (or wedge-shape) section of the walls appear to crumble around a midpoint along the length of the oscillator (figure 3). This appears to allow the lid to transfer greater tensile force to the two shorter ends of the package
Stage 3: Ceramic case fracture (base)
Eventually, the forces are so great that the ceramic base shatters to accommodate whichever of the forces prevails in the two variations seen in stage 1 above (figure 4). It is not clear whether there is significant time between stages 2 and 3. However, it is more likely that the failure of the base in stage 3 occurs as an instantaneous chain reaction following the wall failure of stage 2 as, once the ceramic is breeched at any point, water will rush in. After that, all distorting forces are immediately removed by the fact that the internal and external pressures have been equalised.
The testing provided a valuable understanding into the mechanisms of failure and areas of weakness in this design of ceramic case with metal lid. All three package options tested withstood pressures up to 25 bar, and a number managed beyond. However, slight deformities were recorded at indeterminable points between 25 to 50 bar. Although functionality was maintained in all samples up to 98 bar, any deformation of package, no matter how small should be considered unacceptable. Inward buckling of a flexible lid will place strain on the brittle ceramic side walls.
Movement of the metal lid as it distorts towards the crystal will also influence the oscillator output frequency, by altering the parallel/parasitic capacitance of the circuit.
Repeated cycling, even at lower pressures, would also most likely result in premature catastrophic failure of the component through fatigue. Moreover, this testing does not measure the effects of pressure over extended time periods.
Therefore, while all samples appeared to withstand at least 25 bar, both with no visible damage and passing a Helium fine leak test afterwards, the current design of standard Euroquartz XO oscillator packaging is not recommended for operation up to that level without further evaluation.
The effect of repeated exposure is also unknown. At this stage, customers needing such a component operating up to 25 bar should only consider this after satisfying themselves by conducting their own extensive trials in application prototyping. As such, customers would need to accept full liability for any such failure, either during trials, application production or in full in-field use of the application.
Contact Details and Archive...