The Cheapest Insurance in Mining
A metallurgical test programme on a representative sample typically costs well under one percent of the plant it informs. It is the only thing standing between a flowsheet designed around your ore and a flowsheet designed around a brochure.
Almost every underperforming small plant in the region can be traced to one of four things that test work would have found: the ore is harder than assumed, the gold is finer than assumed, the gold is not free-milling, or the grind size needed for recovery is finer than the mill can deliver. All four are cheap to discover in a laboratory and ruinous to discover after commissioning.
Start With the Sample, Not the Test
Every result below is worthless if the sample is not representative. This is where most programmes fail, and no laboratory can correct for it.
- Cover the ore types you will actually mine. Oxide, transition and fresh sulphide behave completely differently. A sample of easy near-surface oxide will flatter a plant that must later treat sulphide
- Cover the grade range, not just the best material. Nobody sends the laboratory their marginal ore, and that is exactly the ore that determines cut-off
- Take enough. 20 to 50 kg supports a basic programme. Comminution testing on coarse material and any pilot work needs considerably more, sometimes hundreds of kilograms
- Document where it came from. Location, depth, ore type and how it was collected. An undocumented sample cannot be defended later
- Use drill core or channel samples where possible. Grab samples from a stockpile are biased towards whatever is easy to pick up
The Tests That Matter, and What Each Changes
1. Head assay and mineralogy
Where the gold actually sits. Is it free, locked in sulphides, or associated with carbonaceous material. This single result determines whether gravity alone can work, whether you need leaching, and whether the ore is refractory and needs pre-treatment. Everything downstream depends on it.
Changes: the entire flowsheet category.
2. Bond Ball Mill Work Index
How hard the ore is to grind, expressed as kilowatt hours per tonne. Fed into the Bond equation with your target tonnage and grind size, it gives the mill power you actually need.
Changes: mill diameter, length and motor rating, which are the largest single equipment cost in the plant and the one that cannot be fixed later without replacing the mill. Skipping this test is the most expensive omission available.
3. Abrasion index
How fast the ore will destroy liners, jaw plates and pump wet ends. Ore across the Lake Victoria Goldfields tends to be competent and abrasive.
Changes: liner and wear part material selection, and the wear part line in your operating cost model. Operations routinely underestimate this by a factor of two.
4. Gravity Recoverable Gold
How much of the gold a gravity circuit can recover, and at what grind size. Run in stages, it also shows how recovery improves as you grind finer, which tells you where the economic grind sits.
Changes: whether you need a leach circuit at all. On many small Tanzanian operations this test alone saves the cost of an entire CIL plant, because gravity turns out to be sufficient.
5. Bottle roll cyanide leach
Recovery achievable by leaching, cyanide consumption, lime demand and leach kinetics over time. Run at several grind sizes.
Changes: residence time, so tank count and tank size, plus your reagent consumption budget. Cyanide and lime consumption vary hugely between orebodies and are a major operating cost.
6. Diagnostic leach
A staged dissolution that identifies exactly what is preventing recovery: gold locked in sulphides, gold in silicates, or preg-robbing carbonaceous material stealing dissolved gold back out of solution.
Changes: whether CIL is mandatory rather than optional, and whether the ore needs flotation, roasting or oxidation ahead of leaching. Discovering preg-robbing after commissioning is among the worst outcomes in gold processing.
7. Settling and rheology
How the slurry thickens and how it behaves when pumped.
Changes: thickener sizing, flocculant selection and water recovery, which in water-constrained districts can decide whether the site is viable.
What a Programme Costs and How Long It Takes
A basic programme of head assay, Bond Work Index, gravity recoverable gold and bottle roll leach commonly returns in four to eight weeks from sample receipt, with cost measured in low thousands of dollars rather than tens of thousands. Adding diagnostic leach, abrasion index and settling work extends both.
Accredited commercial laboratories in Dar es Salaam handle the routine work, with more specialised comminution and diagnostic testing sometimes sent to South Africa or further afield. Build the turnaround into your project schedule from the start, because test work on the critical path is the most common cause of a delayed plant order.
Reading the Results Honestly
- Recovery is reported against a grind size. A recovery figure without the grind size attached is meaningless
- Laboratory recovery flatters plant recovery. Expect the plant to fall a few percentage points below bench results, and design with that margin rather than against the laboratory number
- Kinetics matter as much as the endpoint. Ore reaching 92 percent in 12 hours needs a very different tank train from ore reaching 92 percent in 36
- Check the variability. If two samples of the same ore type disagree, you have a variability problem and you need more samples, not an average
- Reagent consumption is an operating cost, not a footnote. High cyanide or lime demand can move project economics more than a few points of recovery
The Sequence
- Collect representative samples across ore type, grade and depth, and document them
- Head assay and mineralogy first, because it determines which tests are worth running
- Comminution testing: Bond Work Index and abrasion index
- Recovery testing: gravity recoverable gold, then bottle roll leach at several grind sizes
- Diagnostic leach if recovery is disappointing or preg-robbing is suspected
- Design the flowsheet from the results, and only then request equipment quotations
Regions We Serve
Frequently Asked Questions
How much ore do I need to send for test work?
20 to 50 kg supports a basic programme of head assay, gravity recoverable gold and bottle roll leach. Comminution testing on coarse material needs more, and pilot-scale work can need several hundred kilograms. Confirm the requirement with the laboratory before collecting, because a second sampling trip is expensive.
Can I skip test work if the ore looks like my neighbour's?
No. Adjacent orebodies in the same belt routinely differ in hardness, liberation and leach behaviour, and the differences are invisible in hand specimen. A neighbouring operation's flowsheet is a useful starting hypothesis and nothing more.
What is the single most important test?
Head assay with mineralogy, because it determines whether the ore is free-milling or refractory and therefore which flowsheet category you are in. Bond Work Index is a close second, because it sets the mill, which is the largest and least reversible equipment decision in the plant.
Where can I get test work done in Tanzania?
Accredited commercial laboratories in Dar es Salaam handle fire assay, multi-element ICP and routine metallurgical work. Specialised comminution and diagnostic testing is sometimes sent to South Africa or further afield. We can scope the programme and manage the submission if that is easier than arranging it yourself.