What are the tradeoffs between GaN and LDMOS for a base station power amplifier?
GaN vs LDMOS for Base Station Power Amplifiers
The GaN vs LDMOS decision has major implications for base station cost, efficiency, size, and capability. This competition has driven rapid improvements in both technologies over the past decade.
Frequently Asked Questions
Will LDMOS eventually be completely replaced by GaN?
For new base station designs above 2 GHz, GaN has largely already won. For sub-2 GHz macro cells with very high power requirements (>100W per carrier) where cost is the primary driver, LDMOS will remain competitive for several more years. LDMOS on silicon is also evolving (LDMOS on SOI for improved performance). Long-term (5-10 year horizon), GaN cost will decrease with wafer volume and scale, likely making it dominant across all frequency bands.
What about GaN-on-SiC versus GaN-on-Si?
GaN-on-SiC provides superior thermal conductivity (SiC thermal conductivity is 3-5x higher than Si), enabling higher power density and better reliability. It is the preferred technology for high-performance base station PA. GaN-on-Si is cheaper (fabricated on standard silicon wafers in existing fabs) but has worse thermal performance, limiting maximum power density. GaN-on-Si is targeting cost-sensitive, moderate-power applications (small cells, Wi-Fi PA, consumer electronics).
How does the operating voltage affect system design?
GaN typically operates at 48-50V drain voltage, while LDMOS uses 28-32V. The higher GaN voltage presents higher impedance at the transistor output, making impedance matching to 50 ohms easier (less extreme transformation ratio needed). However, it requires 48V power supply infrastructure, which is becoming common in 5G base stations (48V DC power is the industry standard for 5G Open RAN). Legacy 4G infrastructure designed for 28V may need power supply upgrades to deploy GaN amplifiers.