Oscillators are a critical component of any cellular basestation. TCXO technology first appeared in early femtocells to meet the low price point required. The latest TCXO products can now match the performance of older and more expensive OCXOs, making them suitable for any cellsite. We spoke with the key engineers at Pletronics to learn more about the technology behind them.
Quick recap on OCXO and TCXO oscillators progress to date
Just as LTE and LTE Advanced are pushing right up against the limits of Shannon’s Law and semiconductor fabrication below 7nm is almost designed with individual atoms, so TCXO technology is also approaching what’s physically possible with the materials used. TCXOs are the lower cost oscillators used extensively in residential femtocells and arguably are now suitable for almost all cellular basestation designs.
Older OCXO (oven controlled) technology heats the crystal oscillator and maintains it at a known temperature for which the performance has been calibrated. The benefit is highly predictable and stable frequency performance. Downsides include the time taken on start-up to reach and stabilise to operating temperature, higher power consumption and associated heat generated.
Newer TCXO (temperature compensated) designs compensate for the ambient temperature by adjusting the voltage used to drive it. A detailed profile of the characteristics of the crystal has to be known and accurately balanced out to ensure consistent performance throughout the frequency range.
How do these compare with other oscillator applications?
Cellular networks have much tighter tolerances with maximum permitted deviation from frequency measured in parts per million:
|Smartphones||1 to 2 ppm|
|GPS receivers||typically 0.5 ppm|
|Residential femtocells||0.25ppm (although often specified 0.1ppm)|
While some mass market consumer goods may use crystals that have been sorted and screened, it needs much more rigorous engineering to achieve the consistency, long lifetime and predictability. Adding a poorly performing device into a congested network can seriously impact the total system throughput.
We go to great lengths to design inherent performance into our products from the outset, rather than just sorting and picking the best from the manufacturing line. This gives our customers greater confidence that quality will be met and ensures a longer lifetime for the end product.
The latest TCXO products satisfy both small and macrocell specifications
Pletronics latest OeM8 oscillator is a TXCO that has pushed the boundaries, comfortably fitting into the 0.05ppm envelope through the full industrial temperature range.
Where TXCOs were originally used indoors, with temperatures of 0-70C, these latest devices can cope with anything from -40C to +85C. This allows them to be designed in to any application – indoor/outdoor, urban/rural/remote. They are equally suitable for backhaul products.
Pletronics also have a companion OeM4 product that for indoor use (0-70C) at lower cost.
Do you expect those frequency tolerance requirements to become even stricter for LTE-Advanced?
Actually no. We expect the 0.05ppm frequency tolerance to remain the gold standard for both 3G and LTE for the foreseeable future. The cost versus performance tradeoffs makes this an optimal choice for system designers.
For the highest spectral efficiency in dense urban areas, phase synchronisation is used to achieve very tight co-ordination between small cells and macrocells sharing the same spectrum. That’s achieved using external timing electronics with some clever software algorithms.
With so many different LTE frequencies being introduced, do these require different oscillator variants?
Also no. There is an optimal frequency range related to the specific crystal materials, and we use one of those. Commonly 10.0MHz, 19.2MHz and 26.0MHz are used across the entire product range. The master clock output is then divided down and/or multiplied up to obtain the required frequency.
For example, this would be no different when supporting 5GHz LTE-U compared to LTE at 2.6GHz.
What about Phase Noise?
A critical aspect for the performance of any small (or macro) cell is how stable the frequency source is. Phase noise is the short term variation in that signal in the frequency domain. Jitter is a similar measure but in the time domain. You have to be careful when comparing these figures. For the same phase noise, jitter would be lower at a higher frequency because it’s frequency dependent.
The quality of the clock source sets the bar for jitter throughout the whole system. A high quality oscillator gives designers fewer constraints and gains them some margin to play with affecting other component choices.
Overall, low phase noise increases system performance, reduces Bit Error Rate and improves throughput.
Does this mean the end for OCXOs?
If you wanted better than 0.05ppm, then you’d be back in OCXO territory and need some additional circuitry. The demarcation line between OCXO and TXCO was about 0.25ppm some 10 years ago. That’s moved forward to 0.05ppm today, more than enough to meet the needs of 3G and LTE. If anything, it’s a little bit ahead of the technology requirements.
TCXOs have been widely used in a variety of Stratum 3 applications over the past decade which mandates 0.05pm in 0-70C. The technology has also been successfully deployed in millions of femtocells and small cells to date, and has evolved and matured substantially over that time.
We think that eventually TCXOs will consume most of the market volume. These latest products provide a very good baseline for the next few years. The advantages in the technology are much more appropriate for field use, avoiding issues such as heat dissipation and warm up time. It will just take one or two design cycles for these components to become widely embedded.
Any thoughts for the longer term including 5G?
For those researching and designing 5G, it’s important to investigate and appreciate the available clocking technology and incorporate it in the design from the outset.
Although 5G may still be somewhat undetermined at this stage, we’d expect it to involve tight co-ordination and synchronisation as a fundamental element. Making best use of available and emerging timing technology will be inherent to achieving the highest performance.
Our thanks to Dave Kenny, VP R&D and Rob Henry, VP Engineering, at Pletronics for their assistance in developing this article.