# Test station development
Story: “I build and rework test rigs while the hardware is still changing. I want to see my data right away, and decide how to organize it once I know what it is.”
This page is for engineers who develop [test stations](https://dev.nominal.io/platform/reference/glossary.md#test-station "Another name for a test stand: the equipment that runs tests on a unit under test.") while the hardware on them is still in development. Boards get swapped, instruments get added, the [channel](https://dev.nominal.io/platform/reference/glossary.md#channel "A named signal: a series of measurements or computed values, such as voltage, pressure, or system state.") list and specifications change from one week to the next, and the right structure isn’t known yet. If you’re developing the board itself, on a handful of dev boards, see [Controller development](https://dev.nominal.io/platform/data-model/paradigms/stories/control-loop.md) instead.
The [quick visualization](https://dev.nominal.io/platform/data-model/paradigms/stories/data-fast.md) approach is good for one-off testing, but a test station engineer needs the same flexibility while keeping the history of their work for analysis and feedback cycles.
Start by implementing the [tagged approach](https://dev.nominal.io/platform/data-model/organization/tagged-data.md). Once your tests settle into a routine, [give each unit an asset of its own](https://dev.nominal.io/platform/data-model/paradigms/stories/test-engineer-stabilizing.md) on the station’s [dataset](https://dev.nominal.io/platform/reference/glossary.md#dataset "A store in Nominal for time-series and log data, which grows as you upload or stream more into it."). *Consider* also [giving each unit a dataset of its own](https://dev.nominal.io/platform/data-model/paradigms/stories/test-engineer-stabilizing.md#giving-each-unit-a-dataset-of-its-own) only if the units themselves are long-lived, tested over many campaigns, or have data of their own, such as an onboard recorder. That [combines the default and tagged approaches](https://dev.nominal.io/platform/data-model/organization/combined.md). See also the [Repeatable test execution](https://dev.nominal.io/platform/data-model/paradigms/stories/repeatable-runs.md) story.
## While the station keeps changing
Create one [asset](https://dev.nominal.io/platform/reference/glossary.md#asset "A physical or simulated device or system you track in Nominal, such as a vehicle, a component, a unit under test, or a test stand.") for the station and one [dataset](https://dev.nominal.io/platform/reference/glossary.md#dataset "A store in Nominal for time-series and log data, which grows as you upload or stream more into it."). Everything the station records goes into the dataset, which is linked to the asset under the [ref name](https://dev.nominal.io/platform/reference/glossary.md#ref-name "The name an asset or run gives each data source attached to it, such as default. Workbooks and checklists find channels by ref name and channel name.") `default`.
To tell boards and setups apart, put [tags](https://dev.nominal.io/platform/reference/glossary.md#tag "A key-value pair on your data points, such as motor_id: 1, that lets you filter a subset of points from a single dataset.") on your data, such as `board=A` or `config=2`, or any other key and value you would find valuable.
Because the station’s asset is linked to the whole dataset with no tag filter, a [workbook](https://dev.nominal.io/platform/reference/glossary.md#workbook "Where you analyze time-series data from assets and runs, in tabs and panels, after a test or live.") on the station’s asset also lets you [monitor the station itself](https://dev.nominal.io/platform/data-model/paradigms/stories/test-engineer.md#station-health).
```mermaid
---
config:
layout: elk
---
flowchart BT
classDef purple fill:#e0e7ff,stroke:#4338ca,color:#1f2937
classDef darkgreen fill:#bbf7d0,stroke:#15803d,color:#1f2937
subgraph dataset1["Dataset: Test station"]
row2["Tuesday
board=B
config=1
10:30:00 · 11.8
10:30:01 · 11.9
..."]:::purple
row3["Thursday
board=B
config=2
14:15:00 · 13.0
14:15:01 · 12.8
..."]:::purple
row1["Monday
board=A
config=1
09:00:00 · 12.1
09:00:01 · 12.3
..."]:::purple
end
style dataset1 fill:#f5f7ff,stroke:#4338ca,color:#1f2937
station["Asset: Test station"]:::darkgreen
dataset1 -->|"ref name=default"| station
click station href "../../../#asset" "Asset"
click dataset1,row1,row2,row3 href "../../../#time-series-data" "Time-series data"
```
By tagging the data you can look at just a portion of it in your workbooks, filtering the data for `board=A` or `config=2` or both.
### Walkthrough: upload with tags
1. In the sidebar, click **+** (**New**), then **Upload data**, and add your files.
2. Click **SELECT A DESTINATION**:
- The first time, choose **Create an asset**, name it after the station, and click **CREATE ASSET**.
- After that, choose **Link to an existing asset** and pick the station.
3. In **Organize your data**:
- The first time, click **ADD DATA**, then **New dataset**, name it, and click **CREATE DATASET**. Keep the **Refname** `default`, then click **CONTINUE**.
- After that, click the station’s existing [data scope](https://dev.nominal.io/platform/reference/glossary.md#data-scope "The link between an asset or run and a data source, named by its ref name. It can filter the data source by tags and shift its timestamps.").
4. In the **Settings** panel, check the timestamp, then under **Tags** click **ADD TAG** to tag everything in the [upload](https://dev.nominal.io/platform/reference/glossary.md#upload "To send a finished file to Nominal, from the app or a client library, so it can be ingested into a dataset."), for example `board` = `A`. If a column in the file already says which unit each row came from, use **ADD TAG COLUMN** under **Tag columns** instead.
5. Turn on **Create run**, name the [run](https://dev.nominal.io/platform/reference/glossary.md#run "A named, time-bounded period of interest on one or more assets, such as a test, a flight, or a race session."), and click **CREATE RUN**, so the workbook you open next shows just this upload’s time range.
6. Click **CONFIRM & INGEST**, then **VIEW RUN**, then **NEW WORKBOOK**.
To [stream](https://dev.nominal.io/platform/reference/glossary.md#stream "To send points to Nominal as they're produced, instead of uploading a finished file.") instead, attach the tags to the data you write. In Python:
```python
from nominal.core import NominalClient
client = NominalClient.from_profile("default")
dataset = client.get_dataset("") # the station's dataset
with dataset.get_write_stream() as stream:
stream.enqueue("voltage", "2026-01-01T12:00:00Z", 12.1, tags={"board": "A", "config": "1"})
```
If [instro](https://instro.nominal.io/) is used to drive the station’s instruments, create the instruments for each unit and pass that unit’s tags to each constructor. Everything an instrument reads or sends then carries those tags:
```python
from instro.psu import InstroPSU
from instro.psu.drivers import BK9115
def run_test(board: str) -> None:
psu = InstroPSU(
name="psu",
driver=BK9115(""),
num_channels=1,
dataset_rid="", # the station's dataset
board=board, # other keyword arguments become tags on every point
)
with psu: # opens the instrument, and closes it when the test ends
... # the test sequence
run_test("A") # once board B is in the socket, run_test("B")
```
instro’s [test rack example](https://instro.nominal.io/examples/test_rack_example/test_rack/) drives a power supply, an electronic load, and a DAQ into one dataset.
To look at one board at a time in a workbook on the station, filter by the `board` tag. To compare boards side by side, split by it instead.
### Filter by a tag
Show only the points with certain tag values:
1. In the workbook’s left sidebar, click the **Data** icon to open the **DATA** panel.
2. In its **TAG FILTERS** section, click **+** (**Add tag filter**). Choose the **Tag key**, pick the values under **Tag values**, and click **ADD TAG FILTER**.
3. On the filter’s row, after **Apply to**, click **TAB** to filter every [variable](https://dev.nominal.io/platform/reference/glossary.md#variable "A reference to a channel, with any tag filters, splits, and transforms applied, or a formula built from other variables. Workbooks plot variables, and checklists check them.") in the current tab, or **WORKBOOK** for every variable in the workbook.
### Split by a tag
Plot each tag value as its own series by splitting the variable that plots the channel:
1. In the left sidebar’s **Channels & variables** view, find the **VARIABLES** section, hover over the variable, and click its **Transform** icon.
2. In the **Transform** window, click the settings icon on the **Series** row.
3. In the **Series** window, choose the tag under **Split by tags**, and click **APPLY**.
## Monitor the station itself
Comparing units tells you how each one performed, but as a test station engineer you may also need to validate the test station itself is working. Is an instrument drifting? A voltage source slowly sagging? The station’s asset already lets you watch for problems like these.
With the tagged approach set up above, every measurement the station has taken is in its one dataset. A single workbook on the station’s asset can plot all of the station’s signals over time, where problems such as supply voltage drift or rising signal noise show up. When you find one, split by the `board` tag to tell whether it comes from a unit or from the station.
Instro also writes each command it sends to a channel ending in `.cmd`. E.g. a measurement of `psu.ch1.voltage` will be recorded alongside `psu.ch1.voltage.cmd`. This allows a test engineer to look for failures where a setpoint is not reached.
```mermaid
---
config:
layout: elk
---
flowchart BT
classDef purple fill:#e0e7ff,stroke:#4338ca,color:#1f2937
classDef blue fill:#e0f2fe,stroke:#0369a1,color:#1f2937
classDef darkgreen fill:#bbf7d0,stroke:#15803d,color:#1f2937
subgraph dataset1["Dataset: Test station"]
unitA["Monday
board=A
09:00:00 · 12.1
..."]:::purple
unitB["Tuesday
board=B
10:30:00 · 11.8
..."]:::purple
end
style dataset1 fill:#f5f7ff,stroke:#4338ca,color:#1f2937
station["Asset: Test station"]:::darkgreen
dataset1 -->|"ref name=default
no filter"| station
workbook1["Workbook for
Station health
Every point, every board"]:::blue
station --> workbook1
click station href "../../../#asset" "Asset"
click dataset1,unitA,unitB href "../../../#time-series-data" "Time-series data"
click workbook1 href "../../../../workbooks/" "Workbooks"
```
### Walkthrough: check the station’s health
1. In the sidebar, click **Assets** and open the station’s asset, the one you created in [your first upload](https://dev.nominal.io/platform/data-model/paradigms/stories/test-engineer.md#walkthrough-upload-with-tags), then click **NEW WORKBOOK**.
2. Plot the station’s own signals, such as its supply voltage, and widen the time range to cover several weeks of tests.
3. To tell the station from the units, split the signal’s variable by `board`: in the **VARIABLES** section, click the variable’s **Transform** icon, then the settings icon on the **Series** row, choose `board` under **Split by tags**, and click **APPLY**.
## Next steps
As the design stabilizes, [give each unit an asset of its own](https://dev.nominal.io/platform/data-model/paradigms/stories/test-engineer-stabilizing.md).
Detailed guides:
- [Using a single dataset for many identical assets](https://dev.nominal.io/platform/data-model/organization/tagged-data.md): the tagged approach in full.
- [The default approach](https://dev.nominal.io/platform/data-model/organization/per-asset.md): one asset and dataset per unit, and production [tester](https://dev.nominal.io/platform/reference/glossary.md#tester "Another name for a test stand: the equipment that runs tests on a unit under test.") monitoring.
- [Combining the default and tagged approaches](https://dev.nominal.io/platform/data-model/organization/combined.md): units with datasets of their own alongside the station’s tagged dataset.