A numeric entry object writes the operator's value to the PLC on Enter, and a plain write needs neither notification nor handshake. Turn the two bits on only when the PLC program must react to an operator change, and then the PLC, not the panel, sets the handshake bit after it sees the notification. That division of labor is the whole exchange.
Are the notification and handshake options necessary on a numeric entry object? No. The entry writes the operator's value straight to its configured PLC address on Enter, and a plain write needs neither bit. The pair earns its place only when the PLC program must react to the fact that an operator changed a value. When you enable them, the division of labor is fixed: the panel raises the notification bit to say a value is changing, and the PLC, never the panel, raises the handshake bit once its logic has seen the notification. This page explains what each bit does, when each earns its keep, and what the exchange adds in time.
What a numeric entry does without notification and handshake
A numeric entry object has one core job. When the operator commits a value, the panel writes it to the PLC address the object names. That is the entire transaction. Nothing in the PLC learns who wrote the value, or that a person was involved rather than a recipe or a math rung, and for most setpoints nothing needs to. The optional pair exists for one documented purpose on the platforms that offer it: vendor support notes describe the notify and handshake tags as the way to confirm that the entered value has landed on the write tag before the on-screen keypad closes. If no rung in your program cares when an operator edits a value, leave both options off and let the write stand on its own.
What the notification bit does
The notification bit is a single bit in the PLC that the panel sets at the moment it writes an entry. It carries exactly one fact: an operator is changing a value. The value itself still travels in the address; the bit is a doorbell, not a message. The PLC side does the listening: a rung that examines the bit and acts on it. A notification bit that no rung reads is as silent as no bit at all, so write the bit and the rung in the same project edit. Give each entry object its own bit so the PLC knows which value changed.
What the handshake bit does, and who sets it
The handshake bit is the PLC's reply. When the PLC logic has seen the notification and done what the application requires, the PLC sets the handshake bit. The panel reads it as confirmation that the notification arrived. The panel never raises this bit; if it ends up true, a PLC rung put it there. One setup guide for pulse-style panels spells out the timing: the panel sees the handshake go true, drops its enter pulse, and waits for the handshake to drop before it accepts the next entry.
That held-until-answered request matters for scan time. A pulse the panel holds open until the handshake arrives survives any scan rate; a fixed pulse shorter than one scan can vanish between two scans. How many PLC scans an HMI momentary button bit stays on works through the scan arithmetic.
| Signal | Raised by | What it tells the other side |
|---|---|---|
| Entered value | Panel, on Enter | The operator's committed value, at the configured address |
| Notification bit | Panel, with the write | An operator is changing a value |
| Handshake bit | PLC, after it sees the notification | The PLC logic received the notification |
Vendor guides leave the reset to the application. The tidy pattern: the PLC clears the handshake after the panel's request has dropped and clears the notification in the same rung group, so both bits sit idle before the next entry.
When a plain write is enough
A plain write covers the common cases:
- The value only has to be current by the next time the PLC reads the address, and nothing in the program reacts to the edit itself.
- The panel display already shows the entry, and the PLC polls the address on its own schedule.
- One operator enters one value, with no recipe, logging or audit line attached.
In these cases the two optional bits add rungs to write, bits to initialize and one more way for an entry to stall. The simple path is the right one.
When to turn both bits on
Turn both on when the application needs the conversation:
- The PLC must flag or log that an operator changed a value, for a batch record or an audit line.
- Permissives or related setpoints must re-evaluate the moment the edit commits, inside the same scan sequence.
- The keypad must stay open until the PLC confirms the value has landed on the write tag.
- The program needs to settle who owns the value at power-up, before the first entry arrives.
The same acknowledgment pattern drives alarm confirmations, and signaling an HMI alarm acknowledgment back to the PLC walks that variant rung by rung.
Recommendation: Leave notification and handshake off for every numeric entry that only has to land a value in the PLC. Turn both on, and write the PLC side in the same project edit, when the program must react to an operator change: a rung that sees the notification and sets the handshake. Configure the panel options alone and the panel waits after every entry for a reply nothing sends.
Example: a setpoint entry at a 10 ms scan and a 100 ms update
Scenario: a mix-room terminal lets the operator set a new target temperature of 180 °C, and the PLC must set a recipe-edited flag and confirm the entry before the keypad closes. Example inputs: a 10 ms processor scan, mid-range of the 1 to 10 ms band small processors commonly run (your model's published cycle time is the figure to use), and a 100 ms update interval on the entry object's communications.
The exchange runs like this:
- t = 0 ms. Enter is pressed. The panel writes 180 to the setpoint address and raises the notification bit.
- t = 10 to 20 ms. The next one or two scans see the notification, move the value into the recipe register and set the handshake bit: 2 scans × 10 ms = 20 ms worst case.
- t = 100 ms worst case. The panel's next update cycle reads the handshake, drops its enter pulse and closes the keypad.
- t = 110 to 120 ms. The PLC sees the request drop, clears the handshake and the notification, and both bits idle.
End to end: 20 ms of PLC response plus one 100 ms update cycle, about 120 ms worst case from Enter to a closed keypad. At a 1 ms scan the same chain takes about 101 ms, because the update interval dominates. The operator notices neither figure. A plain write closes the keypad at once, but nothing tells the program that an operator made the change. To catch a stalled exchange, add a timer rung that faults the sequence if the handshake stays true for 1 s, which is ten update cycles of margin: 1 s ÷ 100 ms = 10.
When one side is configured and the other is not
Most trouble with this pair comes from configuring one side only. The symptoms sort cleanly:
- The panel waits after every entry, or reports a handshake error or timeout. Handshake is enabled on the panel, and the PLC rung that sets the bit is missing, or the panel and the PLC point at different bit addresses. A troubleshooting guide for numeric entry handshake timeouts covers the timeout case.
- The handshake stays dead even though the rung exists. Check that the entry's display and write references sit on different tags; some software removes the handshake capability when both reference one tag, which the SPK107 note on a numeric entry that shares a display and write tag documents.
- The value lands but the program ignores the change. That is a PLC logic question, downstream of both bits.
- Adding the bits while the write itself fails. The signaling options add reporting; a wrong address, a mismatched data type or a write the PLC refuses will still fail exactly as before.
One test separates the cases: monitor the configured PLC address and both bits while an operator makes a single test entry. The address never changes, the write path is at fault. The address changes and the notification stays false, the panel-side configuration is at fault. Both pulse and the handshake never follows, the PLC rung is at fault. Use a value that is safe for the running process and follow the site's change-control procedure; if the machine cannot take a test write safely, run the test at a maintenance stop.
Range checks belong in the PLC either way, and clamping HMI numeric entry limits in PLC ladder logic shows the clamp pattern that keeps an out-of-range entry from reaching the process.
Questions
Does one notification bit work for several numeric entries?
One shared bit tells the PLC that something changed but nothing about which object changed it. Give each entry object its own bit pair, or raise one shared pair together with a code register that names the entry. The PLC reads the code in the same scan that sees the notification and knows which value to act on. Two entries sharing one bare bit is the setup that produces edits applied to the wrong setpoint.
Can the same bit pair serve a string entry?
Yes. The pattern is identical: the panel writes the string to its configured destination and raises the notification, the PLC acts and sets the handshake, and the panel releases the entry. Only the data type of the write changes. Size the string destination in the PLC for the longest entry an operator can type, including the terminator, so a long entry cannot overwrite the neighboring tag.
What should the notification and handshake bits read at power-up?
Both idle, and the PLC should write them there on its first scan. A handshake left true from the previous run starts the new session out of sequence, because the panel can read it as a standing acknowledgment from a change nobody just made. Force notification and handshake false in the PLC start-up logic so the first operator entry raises the pair from a clean state.
Planning information, not legal or engineering advice. Rules differ by place: check codes and permits with your local authority and your licensed trades.