Appearance
Surveys and pickups
Getting shots off an instrument and turning them into ground.
The two registers, and why there are two
Mining → Pickups and Mining → Surveys look similar and are not the same list.
| Register | What it holds |
|---|---|
| Pickups | The shots as they came off the instrument. The measurement. |
| Surveys | Surfaces built from those shots. One reading of the measurement. |
One day's shots can carry more than one surface, and a pickup can sit in the register with nothing built on it — which is a normal state, not a loose end.
A survey is never edited into shape
Beyond its name, nothing about a filed survey changes. A surface records what the ground was on a day, and moving its geometry would silently move every volume ever reported against it. A corrected survey is a new survey.
Importing a file
Use Import on either register.
1. Say which instrument wrote the file
This is the first question and the one that matters most.
The wrong instrument does not fail — it mirrors
There is no single survey export format. Total stations, GPS rovers and office packages all emit comma-separated points and then disagree about which of the first two numbers is the northing. Read the wrong way round, a file produces a surface that builds perfectly, looks entirely plausible, and reports a volume that is wrong.
The reader shows you what it understood — how many shots, how many feature lines, the extent, the level range. Check it before you file.
The layouts offered are:
| Choice | What it expects |
|---|---|
| Work it out from the file | The reader decides, and says how confident it is |
| Name, Northing, Easting, Level, Code | A named point, northing first |
| Name, Easting, Northing, Level, Code | A named point, easting first |
| Number, Name, Northing, Easting, Level, Code | A record number in front of the name, northing first |
| Number, Name, Easting, Northing, Level, Code | The same, easting first |
| Northing, Easting, Level | Bare coordinates, northing first |
| Easting, Northing, Level | Bare coordinates, easting first |
| Name, Easting, Northing, Level, String name | The fifth column names a feature line rather than a code |
Work it out from the file reads a few hundred rows and settles two things: which columns hold the coordinates, and which of the two is the northing. The first is why a Hi-Target rover's pickup reads — it numbers every record before naming the point, so its coordinates sit a column further along than a total station's. The second is a genuine answer on a national grid, where a northing in the millions and an easting in the hundreds of thousands are an order of magnitude apart; on a small local grid the two are indistinguishable, and the reader says so rather than guessing confidently. Pick the layout by hand when it does.
The columns it settled on are listed with everything else it understood — Columns read as: 2 name, 3 northing, 4 easting, 5 level, 6 code. Count along one row of the file and check them.
2. The other formats need no telling
A file's extension decides how it is read, and three of them carry their own column order:
| Format | Extension | Order in the file |
|---|---|---|
| Comma-separated points | .csv, .txt | You state it, above |
| Surpac string file | .str | Northing, easting, level |
| LandXML | .xml | Northing, easting, level |
| DXF drawing | .dxf | Easting, northing, level |
DXF is the odd one because it is a drawing format: its X runs across the page and its Y up it, which is easting then northing.
Surpac often writes a string file with a .csv extension
It is recognised for what it is by its contents rather than by its name, so you do not have to rename it first.
3. Check what the reader separated out
Three things are pulled aside rather than silently mixed into the ground:
Control, check and closing shots. Reported, never treated as ground. A control peg six hundred metres from the dump would stretch the surveyed extent across the gap and lay a triangulated sheet over land nobody visited.
Strays. Detail shots sitting well away from the body of the survey. Excluded, and listed, so you can put them back if they are real.
Separate clusters. Where the file holds two runs of ground far enough apart to be different features, they are offered as separate surveys rather than bridged by a triangle across the gap.
Shots metres out in height. Every ground shot is compared with the ground immediately around it, and one that disagrees by metres is set aside and listed. This finds the blunder — a rod left long, a level typed with two digits transposed — and nothing finer: the dialog says in metres what it could and could not have seen in your file, because a clear result here is not a check on your levels.
The two directions are not the same claim, and they are not worded alike:
- Above its neighbours. Little real ground stands metres proud of a bench on its own, so this is almost always the blunder it looks like.
- Below its neighbours. A sump, a drain, a void or an over-dug toe reads exactly the same way, and all four are real ground you are meant to have picked up. Look at these on the plan before leaving them out — you can tell a sump from a blunder at a glance, and this check cannot.
Nothing is dropped either way: tick Include them anyway and they go into the surface. A piece of ground carrying fewer than twelve shots is not judged on height at all, and the dialog says so rather than leaving silence to read as a pass.
4. Say what each code means
Every code in the file is listed with the role the reader gave it — ground, breakline or control — and you can change any of them. Codes you have classified before on this account are remembered and offered back.
A breakline is a feature the surface must respect rather than smooth over: a crest, a toe, a road edge, a drain. Getting these right is the difference between a surface that follows the bench and one that rounds it off.
Numbered codes. Surveyors number their control as they go — CP1, CP2, TBM3, BASE1 — and the reader takes a numbered code as the code it is numbering, so CP7 is read as control without you saying anything. It only does this where the name it strips back is a code it already knows: an invented code that happens to end in a digit is still just that code, and is still read as ground.
Where you have told the reader what a code means, that answer covers its numbered forms too — but only when it moves them towards ground or towards a breakline. Say CP is ground on your site and CP1..CP8 come in as ground with it. An answer the other way, towards control, stays on the exact code you gave it, so CP1 is left in the ground until you name it. Ground read as control disappears from the surface with nothing on screen to say so; control left in the ground shows up as a cone somebody queries.
This does not apply to the layout whose fifth column holds the name of a feature rather than a code. There, numbering is how two features are told apart — BM1 and BM2 on a mine are berm 1 and berm 2 far more often than they are two benchmarks — and a name read wrongly there costs the whole line, not one shot.
A name that could be a mark. On that same layout, a crest the surveyor genuinely called BM cannot be told from a benchmark by its name. The reader does not guess. It asks you about the name — how many shots carry it, and whether BM is a control point here or the name of a feature — and the file cannot be brought in until every such name has been answered. Your answer is remembered for this site, so the question is asked once and not again.
Every one of them blocks, with no attempt to sort the dangerous from the harmless, because sorting them would be the same guess the question exists to avoid. One click, and the rule is one you can hold in your head.
It is worth the click. A site's control network is routinely booked as a dozen marks scattered around the workings under one string name — PEG, CP, STN, TRIG. Read as ground, those become a single line threaded across the whole site. On a 64-shot floor with a ring of twelve marks 25 m clear of it, the fill against a datum comes out 159% over; at 60 m it is 648% over. Nothing else in the import catches that: the separate-pieces-of-ground check cannot see it on this layout, and the longest-triangle clip only helps when the network is far enough away to be obvious. A crest read as control is quieter still — it is simply absent from the surface. Both mistakes are silent, which is why the file waits for you rather than proceeding on a guess.
Two cases are never asked about, because nothing is at stake in them. A name carried by a single shot has no line to lose, so a benchmark shot once is read as control. And a run of shots that never moves is repeat observations of one point whatever it is called, so a real BM peg booked at setup stays control even when a crest in the same file shares its name — and it cannot stretch a triangle anywhere, which is the damage the rule exists to prevent.
Write the description
A code on its own tells the next surveyor nothing. The description is optional and worth the ten seconds.
5. Name it and file it
| Field | What it is |
|---|---|
| Name | What this survey is called in the register |
| Date | The day the ground looked like this — not the day it was filed |
| Mine | Who may see it. Only members whitelisted on the mine's Tracker Group can open it |
| What was surveyed | The piece of ground this is a survey of — a pit, a dump or a stockpile |
| Notes | Anything that changes how the figures should be read |
What was surveyed ties a survey to the ones before it
Cut and fill are only measured between surveys of the same pit, dump or stockpile. An import filed against the wrong one is not comparable to anything, and the fix is to import it again.
Filing opens what the import just created, rather than returning you to the register it grew.
A pickup walked with no signal still files
An import with no network is queued on the device — the pickup and the surface it joins or starts, together — and filed the moment the device next sees a connection, whether or not the app is still open. The register reloads instead of opening it, because there is nothing to open yet. See Working offline.
Correcting a feature line
Open a pickup from Mining → Pickups.
The points cannot be changed — they are the measurement. What you can correct is the strings: which shots the codes joined into a crest, a toe or a road edge, and what each of those runs is called.
Saving a correction corrects every surface built from that pickup. This is the one place in the module where an edit reaches back into finished work, and it is why it lives on the pickup rather than on the surface.
A stray point is dragged back onto the feature it belongs to, and catches on the geometry already drawn like any other placement. Where two feature lines meet at a point, moving that point moves both — the join is found to the millimetre and kept, so a triangulation cannot slip a sliver through the gap a torn join would leave.
Reading feature lines off the ground
On an open survey, the header's line-reading action reads the feature lines off the ground — the crests, toes and ramp edges the surface implies. They can be drawn and exported there. Holding them permanently means correcting the pickup the survey reads from, as above, so every surface built from it gets them too.
What a surface build tells you
Opening a survey triangulates it — honouring its breaklines — and reports:
- Plan area actually covered, in square metres.
- Level range, lowest to highest.
- How long the triangulation took, so a slow survey is explicable rather than mysterious.
- Feature lines that were released.
Released lines
Where two feature lines contradict each other — they cross, at different levels, which cannot both be true — the file is not refused. One of them is read as ordinary ground instead, and the fact is reported.
Why it is reported rather than swallowed
Releasing a line changes the shape of the surface slightly. Anybody comparing this figure against the same ground measured in another package needs to know it happened.
Taking a survey out again
Everything that reads in also writes back out. From an open survey, the header offers a download in three formats, all on the mine grid with no re-projection:
- LandXML — the format every survey package reads.
- Surpac string file — shots and feature lines.
- DXF drawing — shots and feature lines, with levels.
And, for a package that wants this triangulation rather than its own, the surface as triangles: LandXML with its faces, or a DXF of one 3D face per triangle. Where the account has named its mine grid under Integrations, the LandXML says which grid it is on.
The writers are proved by a round trip through Trackapt's own readers
Which is the test that catches a swapped northing and easting — the fault that produces a mirrored plan and a plausible, wrong volume. A file Trackapt writes comes back in as the same ground.
So a consultant, a client's engineer or a colleague who works in another package can be handed the survey in a form they can open, without either of you re-keying anything. See Working with other packages.
See also Volumes and plans for the printed plan sheet, and Taking data away for the export actions common to every screen.
A surface that keeps up to date
An imported survey is a picture of one date. A running surface is the ground as it stands, kept up to date by every survey imported into it: each pickup replaces the ground it covers and extends the ground it reaches past. Volumes are still taken between two dates; the running surface is read as it stood on each.
Start one from the layers icon on Surveys (or Start a running surface on an empty register): the Tracker Group, the item, a name. It begins empty. Importing a survey against the same item then offers, under This survey, Add to the running surface or File as its own survey.
Order comes from the capture date
Which pickup wins a piece of ground is decided by its Surveyed on date, not by the order of import. The newest pickup that covers a spot is the ground there, so a backdated pickup slots into its place.
Opening a running surface shows Pickups in this surface, newest first, each row counting its shots as used, superseded, set aside and lines released. In plan view a scrub bar plays the ground forward one pickup at a time.
Refills, and setting ground aside
A pit should come down; a dump should build up. When a pickup moves the ground the other way, a ramp rebuilt inside a pit out of cleared waste and shot by mistake, the client convention is usually that the original floor wins. Check for refills on a pickup row finds those places.
- The first time, say what this ground does between surveys: Coming down, Building up or Both. It is remembered for the item.
- Each flagged region shows its volume, plan area and worst height difference.
- Set aside a region, and say why. The reason is shown beside every volume the adjustment changes.
Setting a region aside suppresses that pickup's shots there and its claim on the ground, so the earlier survey wins there again and no hole is left. An adjustment is listed under its pickup with a Lift button.
When something looks wrong
| What you see | Why, and what to do |
|---|---|
| The surface is mirrored, or the volume is wildly wrong | The wrong layout was chosen at import. Import the file again with the right one — a filed survey is not edited into shape. |
| The extent stretches away to one corner | A control peg was read as ground. Re-import and mark that code as control. |
| The bench has been rounded off | The crest and toe codes were read as ground. Re-import and mark them as breakline. |
| A spike or a cone stands out of the surface | A shot metres out in height that the check could not see — its file was too rough, or too small to judge. The dialog states that floor in metres at import. |
| The sump is missing from the floor | Its shots were set aside as metres below the ground around them. Re-import and tick Include them anyway. |
| A crest or berm is missing, on a file whose fifth column holds names | Its name is also a control mark — BM, CP, TRIG. The import asks you about names like these; answer The name of a feature and the line goes into the surface. |
| The import asks about a name that really is a benchmark | Answer A control point. Every shot with that name is then kept out of the surface, and the answer is remembered for the next file. |
| File survey is greyed and the footer names a code | That name has still to be answered — a control point, or the name of a feature. Answer it and the button comes back. |
| A survey will not compare against last month's | The two are filed against different items. Only surveys of the same item can be compared. |
| The register does not show a survey somebody else filed | You are not whitelisted on its tracker group. Ask the group owner or the account holder. |
| A pickup has no survey against it | Normal. A pickup is the measurement; building a surface from it is a separate act. |

