The engineering decisions behind Amrit Dhara.
Amrit Dhara evolved through practical experimentation: understanding a variable natural raw material, identifying where the process failed, testing changes, and building controls that made the product more consistent.
The journey started with a different environmental question.
Moulik's work moved from studying polluted water to asking what could be prevented from entering the river in the first place.
From water testing to source prevention
Moulik's early environmental work focused on water-quality improvement, including research on indigenous plant-based coagulants for rural water-quality optimisation. The work taught him to collect evidence, test carefully, compare results and ask what the data was actually showing.
The question gradually changed: if polluted water needs treatment, what is entering the water in the first place? Around Haridwar, temple offerings became one visible source of waste entering the environment.
“A cleaner river cannot depend only on downstream treatment. It also needs upstream prevention.”
Optimizing Rural Water Quality Using Indigenous Plant-Based Coagulants: A Comparative Analysis
Moulik's earlier research on indigenous plant-based coagulants was published in Convergence Journal, Volume 1, Issue 4, under environmental engineering and environmental science.
Five problems. Five interventions.
The chapters follow the actual sequence of learning: the raw material first, then moisture, particle size, formulation and finally the way the finished product travels to the customer.
Raw material
Turning temple flower waste into a defined process input.
Moisture control
Replacing drying by feel with measured, staged drying.
Particle size
Controlling what powder is allowed to reach moulding.
Formulation
Strength, binding and a separate study of cone burn behaviour.
Packaging
Designing protection around the physical behaviour of the cup.
It started small.
The first R&D problem was not a machine. It was learning what actually entered the process—and why small pieces of unwanted material could create larger downstream problems.
Amrit Dhara began with small quantities of temple flower waste collected from temples and ghats along the Ganga. The material consisted predominantly of flowers, with some leaves and occasional threads or similar foreign material.
Manual segregation was already being carried out. The R&D learning was that segregation quality mattered much more than initially assumed. A small piece of thread or unwanted material could pass through the early stages and later interfere with pulverisation, mixing or moulding.
Approximately eight temple/Ganga-ghat locations were considered in the collection work. The intervention at this stage was deliberately simple: improve awareness and discipline before introducing more equipment.
The first engineering learning: process consistency begins before the first machine is switched on.
Evidence from the starting material
Variation in the incoming flowers exposed a more significant downstream source of inconsistency: moisture.
From drying by observation to controlled moisture.
What looked like a drying problem became a moisture-control problem spanning more than one stage of the process.
Initially, flowers were dried using sun drying, open-floor drying, sheltered drying during rainy conditions and trays. A multi-level tray-drying oven was available but was not being used consistently. There was no defined protocol, and workers largely judged dryness by touching or squeezing the material.
The observed chain was: insufficient drying → residual moisture → dough-like formation during pulverisation → re-drying/reprocessing → potential material loss.
In approximately 10 observed batches, one batch was completely lost because of the problem and two required reprocessing through open drying.
The first tumbler is a normal manual tumbler with no heating, used to homogenize each batch. The second is a separate temperature-controlled mechanical tumble dryer used for controlled moisture removal.
The basis of the decision changed: dryness was assessed through readings rather than touch alone.
What was measured
| Stage | Purpose |
|---|---|
| Homogenisation | Reduce variation within the batch before measurement. |
| Tray drying | Rapid first-stage removal of surface moisture. |
| Heated tumble drying | More consistent removal of trapped moisture. |
| Final sensor reading | Confirm the material has reached the target range. |
The 1.5–3 hour window depends on initial moisture and weather conditions; it is a controlled process window, not a universal drying time.
From pulverisation to a defined particle-size control point.
The pulveriser itself did not need a major modification. The key change was deciding what powder was allowed to continue into moulding.
Once moisture was brought under better control, another source of variation became apparent during moulding. Powder that was too coarse or too fine affected formation, surface finish and structural behaviour.
| Particle condition | Observed behaviour |
|---|---|
| Too fine | Better fragrance and easier formation, but higher burning rate and lower resistance to cracking. |
| Too coarse | More difficult burning, poorer surface finish and less consistent mould formation. |
Moulik introduced a drum-sieving stage after pulverisation. Particles above approximately 200 μm were removed from the process stream and sent for reprocessing.
The change moved the process from discovering powder problems after moulding to screening the material before moulding.
Defined control point
Particle size became a process variable to control, not merely an output of the pulveriser.
From a customer complaint to a formulation study.
This chapter contains two distinct investigations: a customer-driven study of sambrani-cup strength and binding, and a separate self-initiated study of incense-cone burn behaviour.
Customer feedback showed that sambrani cups could reach customers cracked or damaged. A simple in-house transport simulation was created by placing cups in a box and shaking it to reproduce repeated movement and turbulence during handling.
The investigation pointed to the material going into the mould: particle size was already an important variable, and binding also needed stronger control. Moulik standardized the binder and improved mixing control. Guar gum, Jigat powder and wood powder were incorporated into formulation trials to improve binding while also influencing burning behaviour.
The objective was not simply to make the cup harder. It was to achieve sufficient structural strength while retaining the desired burning and fragrance characteristics.
What was learned
Burn rate was not the original customer complaint. Once formulation changes were being explored, Moulik initiated a separate study to understand how formulation affected the burning behaviour of incense cones.
The measurements were performed manually using simple, low-cost experimental methods. No high-end laboratory instruments were used; the work was conducted within the project's practical budget constraints.
| Scenario | Formulation by weight | Linear burn rate | Observed behaviour |
|---|---|---|---|
| The Sweet Spot | 60% Flower + 25% Wood Powder + 15% Jigat Powder | 1.2–1.5 mm/min | Clean, steady burn; strong flower aroma; good airflow. |
| Fast-Burning Base | 40% Flower + 45% Wood Powder + 15% Jigat Powder | 1.8–2.2 mm/min | Faster, heavier burn; wood smoke can mask delicate flower scent. |
| Too Dense | 80% Flower + 15% Jigat Powder + 5% Guar Gum | <0.5 mm/min | Smoulders and may go out; dense structure restricts internal airflow. |
| Gummy / Heavy Resin | 50% Flower + 20% Wood Powder + 25% Jigat + 5% Loban Resin | 0.8–1.0 mm/min | Slow, steady burn; suitable where traditional resin is desired. |
The trials identified a formulation that provided the most balanced results under the test conditions. The documented preferred formulation for the cone study was 60% Flower + 25% Wood Powder + 15% Jigat Powder, with an observed linear burn rate of approximately 1.2–1.5 mm/min.
These are manual comparative measurements from the project's R&D work, not laboratory-standard certification values.
From a common box to product-specific protection.
The final problem was discovered in the real world: the package used for both cones and cups did not account for how a sambrani cup behaves during transportation.
Customer complaints about damaged sambrani cups triggered the investigation. The existing packaging used a common box format for both sambrani cups and incense cones. That approach did not individually locate or restrain the cups.
During transportation, cups could move inside the box. The cup also contains filling material, which could come out during movement and contribute to stress on the cup wall. The key realization was simple: packaging designed around the product category was not necessarily suitable for the physical behaviour of each product.
The packaging change was implemented alongside the formulation and binding improvements. The two interventions addressed different sides of the same transport problem: the product needed to withstand handling, while the package needed to limit unnecessary movement.
Before → after
The post-improvement outcome is reported from customer feedback rather than as a statistically measured zero-damage rate.
Redesigned product-specific packaging: 15 sambrani cups individually positioned.
The R&D was not one big invention.
It was a sequence of small engineering decisions: identify where variation enters, make the variable visible, test a practical intervention, and move the control point closer to the cause of the problem.
Start with what actually happens in the process, including failures that appear downstream.
Replace assumptions such as “feels dry” with measurable process conditions where possible.
Use practical, budget-conscious trials to compare formulations, materials and process choices.
Turn the learning into a repeatable control, not merely a one-time fix.