When Vikram, the lander in Isro's Chandrayaan-3 mission, safely touched down on the Moon at 6.03pm on August 23, a 500-strong team of scientists looked back upon four years of work with pride.

They were years when the team "breathed Chandrayaan-3," says Isro's associate project director K Kalpana.

And she means it.

Hundreds of tests were done, thousands of simulations run.In the words of project director P Veeramuthuvel, "Failure was not an option." TOI spoke to some scientists in Chandrayaan-3's 34-member core team to find out how they ensured everything worked this time.

Isro chairmanS Somanath, and M Sankaran, director of UR Rao Satellite Centre - the lead centre for Chandrayaan-3 - give all the credit to the relentless toil and unwavering commitment of their scientists and engineers.

Chandrayaan-3 succeeded because nothing was left to chance, says Veeramuthuvel.

"The lander was meticulously crafted to adapt to any descent path it encountered.

No room for compromise was left." For this, the team developed multiple mission plans and designed tests to fix the shortcomings identified in Chandrayaan-2.

Putting things to the testIt was crucial to anticipate every possible fault, and build systems capable of overcoming them.

That's why, after hundreds of lab and field tests, the team was fully confident of success on landing day.

They had done more than 80 integrated cold tests (no engines), integrated hot tests (with engines) and drop tests.

The cold tests alone had involved 25 hours of flight time and 23 sorties in an IAF helicopter.

As issues arose, they were addressed to guarantee reliability.

The preparation differed from Chandrayaan-2 in one crucial way, says Veeramuthuvel: "In Chandrayaan-2, we used an aeroplane so we couldn't hover or bring it to low altitudes, like 10 metres." As the equipment was tested on a flat bed at a 6km altitude, system-level sensors could not be evaluated thoroughly for Chandrayaan-2.

However, the use of a helicopter for Chandrayaan-3's testing allowed Isro to test sensors "at different levels of the power descent phase for five months.

We even did tests while it hovered at 800m and 150m, just as Vikram did while landing on the Moon," adds Veeramuthuvel.

While the cold testing focused on sensors and navigation, the hot testing at Sriharikota in Andhra Pradesh checked for engine firing.

It included dry runs, static tests, closed-loop evaluations, and truncated de-boost tests under conditions similar to those on the Moon.

Sure-footed VikramThey also had to ensure that Vikram landed on its feet, so they built seven models of the lander, three of which were scaled-down pieces, and did extensive lander leg drop tests at Chitradurga in Karnataka where craters and boulders served as a test bed for helicopter experiments.

These tests improved the team's know-how of powered descent and landing.

"We closely assessed integrated sensors and navigation performance, ensuring every system was in sync," says Veeramuthuvel.

The lander leg tests were done for conditions ranging from steep slopes to flat surfaces, hard and soft terrain, and various combinations of horizontal and vertical velocity.

Practice in virtual worldChandrayaan-3 had something Chandrayaan-2 did not - a dedicated simulation group.

The 2019 mission's simulations were done by its control system teams and were less elaborate.

Aditya Rallapalli, project manager (simulations), says, "We have 25TB simulation data from more than one lakh tests." Bharath Kumar GV, deputy project director for navigation, guidance and control (NGC) simulations, says, "Testing for nominal conditions wasn't enough.

We predicted various parameters that could go wrong and built models.

Then corrections were carried out at each level." Four different simulation test beds - six degrees of freedom (6-DOF), onboard in-loop simulation (OILS), software in-loop simulation (SILS), and hardware simulation - were used with one clear goal: making Vikram soft-land on the Moon.

A team led by Prashant Kulshreshtha built and exhaustively tested the onboard software, a key component of the redesigned lander.

Guiding it homeAnd finally, someone had to ensure that the systems behaved exactly like the simulations showed they would.

Madhavraj, project manager (trajectory), Kuldeep Negi and their team were on the job.

"Our role was to see that all the planning on the ground worked.

It involved a lot of mathematics," says Negi.

The five Earth-bound manoeuvres, the trans-lunar injection, lunar orbit insertion, five lunar-bound manoeuvres, and the two de-boosts to get into the pre-landing orbit had to happen exactly as planned.

"We had a Plan B for each of them, but our Plan A worked every single time," says Madhavraj.

"If I've to sum everything up, I'd say, guidance couldn't fail," says Rijesh MP, also a deputy project director with NGC controls and dynamics.

One of their key challenges was to ensure the guidance systems were in sync with the engines.

If the engines were slow to respond to a command, the guidance system shouldn't have read it as an error.

It didn't.