MaxLinear announced the extension of its Telluride 100G per wavelength PAM4 DSP family with the availability of the MxL93515 and MxL93516, optimized for 100G applications.
MaxLinear said its second-generation DSPs offer significantly lower power and smaller footprint, comfortably enabling sub-3.5W QSFP28 designs while also meeting the form factor needs of the smaller SFP-DD/DSFP optical modules. Like its first-generation DSPs, the MxL93516 includes a monolithically integrated EA-EML laser driver.
The new MxL93515 and MxL93516 also feature a comprehensive digital pre-distortion (DPD) engine in the transmit direction to compensate for laser non-linearity and to cancel packaging limitations that cause reflections and bandwidth degradation at these extremely high signal frequencies. This DPD engine combined with the integrated EA-EML laser driver make these DSPs ideally suited for uncooled EA-EML lasers offering a substantially lower cost model to compete against existing 100G PSM4/CWDM4/LR4.
All Telluride DSPs include an auto-adaptive signal enhancement engine that enables robust SNR and BER performance. The engine integrates a continuous time linear equalizer (CTLE), automatic gain control (AGC), a feed forward equalizer (FFE), and a decision feedback equalizer (DFE).
The MxL93515 and MxL93516 are available in a tiny 7mm x 8.5mm package that supports both 4x25Gbps NRZ (QSFP28) and 2x50Gbps (SFP-DD/DSFP) PAM4 signals on the electrical interface. The package includes either an integrated 1.8Vpp EA-EML laser driver or a differential output that supports an external silicon photonics driver.
“Following the spectacular success of our first-generation Telluride products, we optimized the power and cost of our second-generation devices to enable a new generation of 100G single lambda QSFP and SFP modules,” said Will Torgerson, Vice President and General Manager of MaxLinear’s High-Speed Interconnect Group. “Our second-generation Telluride DSPs combined with the need for fewer optical components will drive the cost per bit down across the industry and we anticipate single lambda 100G solutions to become a majority of the data center and front-haul deployments over the next 2-3 years.”
