HMI email account isn't set up error: fix both settings

Why does my HMI panel say the email account isn't set up?

The warning is a two-part check: the panel reports the account as not set up while either the mail-client fields or the project's address book is empty. So fixing the client fields alone leaves the same message on screen.

Each symptom you can hit points at one of the two stores:

• You filled the incoming and outgoing server fields plus your own address, downloaded the project, and the warning still appears when the alarm fires. The address book is empty. The panel has servers to talk to and nobody to send to.
• You re-enter the same server names and download again, and nothing changes. Typing the client fields a second time cannot create address-book entries, so the missing piece is still the recipient list.
• You put your own email address in the mail address field and expected alarms to arrive in your inbox. A sender address is an identity and the address book is the destination list; your address in the sender field adds no recipients.
• The warning cleared after the last download, but the inbox stays empty. That is a delivery failure with its own checks, covered at the end of this article.

Both stores live in the project on your computer, so every correction ends the same way: save, then download.

Which two settings does HMI alarm email actually need?

Two separate stores hold the setup. The mail client needs its incoming and outgoing server fields, a sender address and the port. The address book needs at least one valid recipient address before any alarm can leave the panel.

Panel makers package the same two stores under different names, and their manuals name them for you. One line's setup guide documents an Email Sender Account configuration, with the mail account and the server details for the service. Another keeps the client under the panel's network settings in an Email Client section, and keeps recipients under Database -> Address Book. A third documents an SMTP Communication service in the runtime settings, where the manual says to check the port used alongside the server details; SMTP is the protocol that carries mail between servers. Find the two stores in your own software by the job each does: connection, meaning servers, port and sender identity, and destinations, meaning the people who receive the alarm.

If the plan was to enter a webmail address and consider the job done, see why an HMI alarm email error needs an SMTP server, not a webmail address; the client fields still need a server the panel can actually reach.

Alarm email is one job among several that put screens to work on a floor; the shop-floor screens and andon-style boards category collects the rest.

Why do HMI email settings ask for IP addresses instead of names?

Panel mail clients are small, older programs. The setup guide for one common line says outright to use the IP addresses of the servers instead of the URL, and the dotted form is the one that needs nothing else on the network to work.

A server name only functions if the panel can ask a DNS server on your plant network and get an answer, and many panel networks carry no DNS at all. An IPv4 address is four numbers, each between 0 and 255, and the panel can hand it straight to the network.

Take the hostnames from the provider's mail setup page before anything else, and copy them letter by letter. The outgoing name is spelled s, m, t, p. A field that reads smpt has the last two letters swapped, and the lookup fails in a way that looks like a network problem but is a typing problem.

While the client fields are open, match the port to the provider's page. Plain SMTP is port 25 and authenticated submission is port 587. A panel that cannot sign in needs the entry point that still accepts mail from equipment, and one maker's manual says to check the port during setup for exactly this reason.

How do I find my mail server's IP address?

Resolve each provider hostname once, from a computer on the same network as the panel. Press Windows Key + R, type CMD, press Enter, then run ping with the hostname and read the address in the reply line.

Run it twice, once for the confirmed incoming hostname and once for the outgoing one: type ping, a space, and the hostname. A successful lookup prints a reply like this:

Read the address from the reply and type it into the matching client field. Then check your typing against the definition: 192.0.2.10 breaks into 192, 0, 2 and 10, and each of the four numbers sits between 0 and 255, which is what the field expects.

If ping answers that the host could not be found, stop there. Compare the spelling with the provider's page letter by letter, then ask the network administrator or the provider for the server details. A guessed address points the panel at a server that can refuse its mail, and one warning becomes two problems.

Can you show an example setup with real numbers?

In this example the provider's page confirms pop.example.net and smtp.example.net, plain SMTP on port 25. Two pings return 192.0.2.10 and 192.0.2.11, and those two addresses go into the client fields with one recipient in the address book.

Scenario: an alarm on a packaging line, one panel, one recipient who carries the plant pager.

1. The provider's setup page lists pop.example.net as the incoming server, smtp.example.net as the outgoing server, and port 25 for mail sent from a machine on its network. If the provider instead required a login on port 587, this setup would need an internal relay, because the panel cannot sign in.
2. From a computer on the plant network, ping pop.example.net returns Reply from 192.0.2.10: bytes=32 time=11ms, and ping smtp.example.net returns Reply from 192.0.2.11: bytes=32 time=12ms. The 11 ms and 12 ms figures only show the network path answers; the mail test comes at the end.
3. Check both addresses before typing them: 192.0.2.10 is the four numbers 192, 0, 2 and 10, and 192.0.2.11 is 192, 0, 2 and 11. Every number is between 0 and 255, so the client fields accept both.
4. Enter 192.0.2.10 in the incoming field, 192.0.2.11 in the outgoing field, port 25, and the line 4 mailbox address as the sender.
5. Open Database -> Address Book and add the maintenance mailbox address. One valid recipient is the minimum; a second recipient covers a shift where the first reader is off.
6. Save the project, download it to the panel, then switch the alarm condition on at the PLC. The warning no longer appears and the test message arrives in the maintenance mailbox. The setup is done, and the panel now sends on every alarm the project defines.

What do I check when the warning survives a project download?

A download carries the project you saved, so check the save first. Re-downloading cannot repair a project whose address book was empty on the computer, and it cannot add the recipient entries the panel is asking for.

Work down this list in order:

1. Save on the computer, then download. The panel runs the last saved project, and an unsaved edit exists on your screen only.
2. Mark the download by changing a text label on the start screen, so you can see the new project arrive.
3. Re-read both stores: the Email Client fields hold resolved addresses and the port, and Database -> Address Book holds at least one complete recipient address.
4. Read each recipient letter by letter. One wrong character sends the alarm to an address that belongs to nobody, and the panel's warning stays silent about it.
5. Trigger the alarm as a controlled test and watch the panel while it fires. If the warning still appears with both stores filled and the new label visible on the screen, stop editing. Take the verified settings to the panel maker's support desk through its published support channel; if you would rather ask us a related question, use the contact page.

How do I tell a setup warning from mail that never arrives?

Watch the warning when the alarm fires. If it stays away and the message still never reaches the recipient, the setup is accepted and the failure has moved to delivery, where the checks are the recipient address, the server information and whether the entered IP still answers.

Check in this order. First, the recipient's mailbox and its junk folder, where filters are hard on mail that arrives from equipment with no human behind it. Second, re-run the two pings and compare the replies with the stored addresses. Providers move mail servers without notice, and an address that worked for months can stop answering while the line itself is unchanged. Type the fresh addresses into the client fields, save, and download. Third, ask the provider whether plain SMTP on port 25 is still offered. Services that moved to authenticated submission on port 587 leave a panel that cannot sign in with nowhere to send, and the usual plant answer is an internal mail relay that accepts plain mail from machines and forwards it onward.

If the panel reports a send error instead of the account warning, name lookup on the panel's network is the first suspect: the article on an HMI panel email send error that starts at DNS walks that fault path step by step.

Recommendation: Fill both stores in one editing session, with the Email Client fields holding the resolved server addresses and the provider-confirmed port, and Database -> Address Book holding at least one valid recipient. Save, download the project, and trigger a controlled alarm before the line restarts. When the warning or a silent inbox returns months later, re-run the two pings before changing anything else.

Questions