Adding a sensor

How to add a sensor using the Add Device wizard on your hub, and why the app's Add Sensor dialog will not make one work.

In the app: /hubs/:hubId

Adding a sensor#

Your sensors talk to the hub over Bluetooth. The hub finds them, works out what they are, reads them, and sends the readings to the app.

Sensors are added on the hub. That is the flow that scans for nearby hardware, tells you what it thinks each device is, and lets you check it works before you save it. Once your hub is linked to your account, everything you set up there appears in the app on its own, under the name you gave it.

The app also has an Add Sensor button on each hub's page. It does something narrower than it looks, and it is not a way to add hardware. See The Add Sensor button in the app before you use it.

Adding a device on the hub#

Open the hub's screen and choose Add Device in the menu. If you are using a phone or laptop on the same network as the hub, the same screen is at http://roamvitals.local/, again under Add Device. There is also a floating Add Device button on every screen except the dashboard.

The page is headed Add Device Wizard, with a Start Over button beside it. The four steps run across the top as 1. What to watch, 2. Find it, 3. Details and 4. Review.

1. What to watch#

Under "What do you want to keep an eye on?" pick the closest tile: Fridge, Freezer, Inside the RV, Outside, Propane, Fresh water, Grey or black tank, House battery, Shore power, Solar, Tyres, Water leak, or Door or window. If none of them fit, choose I'm not sure. Show me everything.

Picking a tile puts that word in the Device Name box and sets a starting check-in interval. Both can still change before you save: if the device you pick in the next step broadcasts a name of its own, that name replaces the tile's word in the box, and if the hub recognises the device it uses that device's recommended interval instead. You can type over either one.

Your answer also sorts the device list in the next step so likely matches come first. It never hides anything.

2. Find it#

Under "Choose Your Device" you get a table of everything the hub can hear right now, with the columns Name, Sensor Type, Device ID, Signal Strength, First Seen, Last Seen and Add.

Press Add on the row that is yours.

• Device ID is a short form, not the full Bluetooth address. It reads like ID 9F:2E:71, which is the last three pairs of the address. The full address is filled in for you on the next step.
• Signal Strength reads Strong, Fair, Weak or Unknown.
• First Seen and Last Seen are clock times, or Unknown.

If the list is empty you will see "No nearby devices found. Retry scan." Press Refresh Discovery, move the sensor closer to the hub, and make sure it is powered on.

If two rows look the same, compare the Device ID, the Signal Strength and the first and last seen times to work out which one is the device in front of you.

3. Details#

• Device Name is what you will see on the hub's dashboard.
• What kind of device is this? is a dropdown that opens on "- choose what this device measures -" and lists devices grouped by what they are for. Its last entry is Other - enter a device type code (advanced), which reveals a Device type code box for typing a code such as GVH5075 by hand.
• How often should we check it? has a list of recommended intervals and a box you can type a number of seconds into.
• Advanced (IDs are created for you) holds Sensor ID (generated from the device name and the end of its address, editable) and Device Address (the full Bluetooth address, filled in from the row you picked and not editable).

Press Review Device.

4. Review#

The review list repeats what you chose: Device, Name on dashboard, Type, Checked every, Device address, Sensor ID and Anonymous device report.

Press Check it works now to try a real read, then Add Sensor. Back returns you to the details.

If the address or the Sensor ID is already used by a sensor on this hub, saving is refused with "That sensor ID or MAC address is already in use. Update the details and try again." Nothing existing is overwritten.

How the hub works out what a device is#

Your hub does not need you to pair anything. It listens to what each device broadcasts and compares three things against its device list:

• the name the device announces,
• the maker's ID in the broadcast,
• the Bluetooth service IDs in the broadcast.

The Sensor Type column shows the result. When the match came from the broadcast itself, the model name is followed by a short tag saying how sure the hub is and what it went on, for example high · from data or medium · from name. A match made only from the device's name shows the model with no tag.

The hub only fills the device type in for you when the evidence is strong. That is deliberate: a maker's ID says "this is a Govee", not "this is a Govee thermometer", and picking the wrong one produces confident, made-up numbers.

So you will sometimes be asked to choose:

• Several devices share the same signature. The type is left blank and you pick from the candidates.
• The device's own name disagrees with the closest match. The Details step shows "We're not sure exactly what this one is." with a button offering the candidate by its code, for example Try GVH5044 and see. Pressing it sets that type and runs a test read, so you find out in a few seconds instead of guessing.
• Nothing matched at all. The Sensor Type column reads "Unknown / choose manually".

What "Check it works now" tells you#

The check connects to the device and tries a real read. There are six answers:

• "It works. Right now it says ..." with the value it just read, followed by a question about whether that looks right. If you picked a tile in step 1, the question is about that: a fridge sensor reading 22 degrees gets "That does not look like a fridge reading yet. Is the sensor already where you want it?" rather than a pass mark.
• "That doesn't look right: ..." when the number is physically impossible, such as a temperature far outside anything a thermometer could see. That means the hub is almost certainly decoding the device as the wrong model. Go back and pick a different type, or report it.
• "Found it, but it hasn't said anything yet." Some sensors only speak every few minutes or when something changes. You can still add it.
• "It answered, but it didn't send a measurement." It replied with nothing worth showing, which often means a probe is not plugged in. Check anything that plugs into it and try again. You can still add it.
• "Couldn't reach it just now." It may be out of range or asleep. Move it closer and try again. You can still add it.
• "We can see it, but we can't read it yet." This one is different, see below.

If you have not picked both a device and a type yet, the button answers "Pick a device and a type first."

When a device cannot be read#

Recognising a device and being able to get numbers out of it are two different things. A handful of devices are in the list because the hub can spot them, but nobody has worked out yet how to decode what they send.

When you select one of those, the Details step says:

We can see your <device>, but we can't read it yet.

If you try to save it anyway, the hub refuses:

We can see this device but we can't read it yet, so adding it would give you a sensor that never shows a reading. Report it below and we'll add support.

That is why these devices are also missing from the device list in the app. A sensor that can never report is indistinguishable from a broken hub, so neither screen offers you one.

Telling us you have one is the fastest way to get it working. The section for that sits at the bottom of the Details step, headed Device not working? (the notice above calls it "Report an unrecognised device", which is the same section).

1. Tick Yes, share an anonymous report about this device. Nothing is sent without it: pressing Send Report unticked answers "Consent is required before a fingerprint-only unsupported-device report can be sent."
2. Under "What went wrong?", tick what applies: It isn't in the device list (ticked for you), It wouldn't pair or connect, It connects but no readings appear, It keeps disconnecting, Something else.
3. Press Send Report. You need a device selected from step 2 for this to work.

The report contains a one-way fingerprint of the device's broadcast, a signal-strength band and the reasons you ticked. It contains nothing about you, your RV or your location. Sharing is never required to add a device.

You are shown a report ID afterwards. Keep it. Delete This Hub's Copy deletes the copy on your hub and asks us to withdraw the copy you shared. If the shared copy could not be withdrawn, it says so plainly and tells you to quote that ID to support rather than pretending it is gone.

Unknown devices#

Rows the hub could not match from the device's broadcast name get a second button next to Add: Teach. It can appear on a row that already shows a model name, because the model there was recognised from the hardware signal rather than the name.

Teach This Device watches the device's broadcasts and works out which part of the signal carries the reading. Keep the device near the hub for the whole process.

There are two ways through it, and the first needs nothing but the device itself:

• Easiest: do something to it and let us watch. Pick something you can do right now from the list of acts, do it, and press I've done it. Then press See what we found. The hub looks for the part of the signal that moved the way that act implies.
• Or type in what its own app shows (needs the maker's app). This is a collapsed section lower down. Choose which reading you are entering, type the value, press Record reading, then press Analyze.

Once a field is found, a section appears headed Last step: what does it really read? Pick a unit, type in a value from anything you trust (a fridge thermometer, a tyre gauge, the weather on your phone), and press Use this reading. It does not have to be exact and it does not have to come from the device's own app.

Finally give it a Device name for the catalog and a Category, decide whether to leave Share with the RoamVitals community (reviewed before release) ticked, and press Save Device Rule.

The Add Sensor button in the app#

On your hub's page there is an Add Sensor button in the top right, and a second one in the middle of the page while the hub has no sensors.

What it does not do: it does not tell your hub about the device. Your hub only ever sends information up to your account; nothing sent down from the app adds a sensor to it, starts it polling, or makes it read anything. A sensor created here is a record in your account with nothing behind it, and it will sit on Offline with no readings for as long as it exists.

Use the hub's Add Device screen for real hardware. Devices added there arrive in the app by themselves.

If you open the dialog anyway, this is what it asks for:

1. Sensor Name.
2. A device dropdown whose first line reads What kind of device is it? Devices are grouped by what they are for, such as "Temperature and humidity", "Power and batteries" and "Tanks and propane". The list is filtered to what your plan includes.
3. MAC Address, with AA:BB:CC:DD:EE:FF shown as a hint. Nothing checks the format.
4. Polling Interval (seconds). It starts at 300, and picking a device type changes it to the interval recommended for that device.

Add Sensor stays greyed out until the name, the device type and the address are all filled in. Press it and the dialog closes, the Sensors count at the top of the page goes up by one, and a card appears.

If the device list will not load#

While the list is loading, the dropdown is replaced by a Device type box with the note "Loading the device list…".

If the app cannot reach the device list at all, the same box appears with "We couldn't load the device list. Enter the type code if you know it."

If your plan has no devices on it, the note reads "No devices are available on your plan yet. Enter the type code if you know it."

If you are at your plan's limit#

Above the sensor cards you will see how much room you have left across your account, for example "4 of 5 saved sensors used". When it is full it reads "5 of 5 saved sensors used - to save another, stop saving one first".

Adding one more fails with a message at the top of the dialog:

Sensor limit reached for your subscription tier (5 sensors).

Each card carries a switch and a chip that reads Saved or Live only. Saved means the readings are kept in your account, with history, charts and alerts while you are away. Live only means the sensor still runs on your hub and still sounds the hub's own alarms, but nothing is stored here and no alert reaches you when you are not in the RV. Turn one sensor off to free a slot for another.

Sensors your hub reports while you are over the limit arrive as Live only rather than being thrown away, so you can see them and choose.

What you see after adding#

A sensor with no readings yet shows the Offline badge. Once your hub starts reporting, the badge changes on its own.

Under the name, the card shows the device type and the address, separated by a dot. For a sensor your hub set up, that address is a stand-in built from the hub's own name for it, such as GVH5075 · auto:govee-h5075-2e71, rather than the Bluetooth address. Signal strength appears as a second line, RSSI: -58 dBm, once the hub has reported one.

The badge on each card means:

You can also filter the list with the chips next to the Search sensors… box: All, connected, disconnected, error and unknown. A sensor that has never reported sits under unknown.

Opening a sensor card takes you to its own page, where Sensor Info lists Type, MAC, Status, Signal and Polling.

What to do if#

The device is not in the list on the hub. Press Refresh Discovery, bring the device within a few feet of the hub, and make sure it is powered on and awake. Some sensors have a button that wakes them up.

The sensor sits at Offline and never reads. The most common cause is that it was created with Add Sensor in the app, which never reaches the hub. Add the device on the hub's Add Device screen instead. If the hub did set it up, check that the device type is right for the model you own: a device saved as the wrong type can connect and still never produce a number.

It reads for a while and then goes Offline. A single missed read is normal at the edge of range and does not change the badge. After five misses in a row the hub rebuilds its connection to the device and reports it as offline, rather than leaving it looking connected while the reading ages. Moving the sensor closer to the hub is the usual fix.

A value disappears instead of updating. Some devices broadcast a fixed nonsense value in place of a real one, such as the temperature a probe thermometer sends when nothing is plugged into it. Where the device list knows the sensible range for a reading, those values are dropped rather than stored, and if the frame carries nothing else worth keeping the whole frame goes. You will see the last good reading and a gap, not an invented number.

You deleted a sensor and want it back. If you chose to keep its history when you deleted it, it is listed on the hub page under "Deleted sensors with retained data", with how many readings are held and a chip reading either purges and a date or kept forever, plus a Purge now button. The panel's own line is "Re-adding a device restores its history; purging removes it permanently." That match is made on the address, so it has to be the same address on the same hub.