One Ground Station, Three Modems: The Choice That Delayed a Satellite Launch
Archi4 min read
Contents
The modems on the table#
In late 2023, a small satellite operator we will call Helios Data faced a common but painful choice. Its first two spacecraft were scheduled for a rideshare launch in seven months. The ground segment was not yet built. The company needed a modem to talk to its X-band downlink, and three options were on the table: a proven commercial unit from a large vendor, a cheaper software-defined radio (SDR) from a mid-size supplier, or an in-house build using off-the-shelf components.
The decision had a deadline. The launch provider required a ground station compatibility test four months before liftoff. Any modem that could not pass that test would push the mission to the next rideshare slot, nine months later.
Option A: The proven commercial modem#
The first option was a rack-mounted unit from a vendor that had been shipping satellite modems for two decades. The price was $85,000 per unit, plus a $12,000 annual support contract. Lead time was 14 weeks. The unit supported the required DVB-S2X standard and came with a tested interface for the company's existing antenna control software.
The upside was predictability. The vendor had hundreds of installations. The downside was flexibility. Changing modulation schemes or adding a new frequency band meant buying a new unit or an expensive firmware upgrade. For a startup planning to test new waveforms in orbit, that felt like a cage.
Option B: The software-defined radio#
The second option was an SDR platform from a smaller supplier. The hardware cost $28,000, and the software license was $9,000 per year. Lead time was 22 weeks. The SDR could be reprogrammed to handle different modulation schemes and frequencies, which appealed to the engineering team. But the vendor had fewer than 30 space deployments, and the interface to the antenna control software would need custom work.
The hidden cost was integration time. The team estimated six to eight weeks of engineering effort to get the SDR working with their existing ground station. That put the compatibility test at risk.
Option C: The in-house build#
The third option was to build a modem using commercial off-the-shelf digitizer cards and open-source signal processing libraries. The bill of materials was roughly $15,000. The team already had two engineers with SDR experience. But those engineers were also responsible for the satellite's payload software, which was behind schedule.
Building in-house would save money on paper. It would also create a maintenance burden and a single point of failure if the engineers left. The CTO called it "the option that looks cheap until you count the salaries."
The decision framework#
The team scored each option on four criteria: total cost over three years, probability of passing the compatibility test on time, flexibility for future missions, and operational risk. The three-year cost estimates were $121,000 for the commercial modem, $74,000 for the SDR (including integration labor), and $95,000 for the in-house build (mostly engineering time).
The compatibility test probability was the decisive factor. The commercial modem had a 95 percent chance of passing on the first attempt, based on the vendor's track record and the existing interface. The SDR had a 70 percent chance, and the in-house build had a 50 percent chance. A failed test would cost the company an estimated $400,000 in delayed revenue and rebooking fees.
What the numbers said#
The expected cost of the commercial modem, including the risk of delay, was $121,000 plus 5 percent of $400,000, or $141,000. The SDR was $74,000 plus 30 percent of $400,000, or $194,000. The in-house build was $95,000 plus 50 percent of $400,000, or $295,000. On pure expected cost, the commercial modem won.
But the flexibility score told a different story. The company planned to add a second frequency band in 2025 and to test a new modulation scheme for a government contract. The commercial modem would require a hardware upgrade for the new band, estimated at $40,000 and 12 weeks. The SDR could handle it with a software update. The in-house build could do anything, if the team had time.
The choice and the fallout#
Helios chose the commercial modem. The unit arrived in week 12, passed the compatibility test in week 15, and supported the launch on schedule. The company saved the $400,000 delay cost and kept its launch slot.
The flexibility problem surfaced 18 months later. When the government contract required a new modulation scheme, the commercial modem could not be upgraded in time. Helios had to buy a second modem from the same vendor, this time at a negotiated price of $70,000, and run two units in parallel. The SDR option would have handled the change with a software patch, but the company would have faced a higher risk of missing the first launch.
What the team would do differently#
The CTO's postmortem was blunt: the decision was correct for the first mission but wrong for the second. The company had optimized for the launch date and underweighted the cost of being locked into a vendor. A hybrid approach, buying the commercial modem for the first satellite and funding an SDR development track in parallel, would have cost more upfront but preserved optionality.
That is the trade-off that satellite operators face repeatedly. The ground segment is not glamorous, but it shapes what the spacecraft can do and how fast the company can respond to new business. The modem choice is rarely just a hardware purchase. It is a bet on which risks the company can afford to carry.
More articles