Notes from the field¶
Laboratory PROTEQU, 10–17 September 2026, MHU 116 firmware 20.2, DPE-2,
display master 1, second master MHS 817 (tablo). Native TCP, one client.
Lab accepted 17 September 2026. Table:
docs/akceptacni-tabulka.md.
Watch the MHU display, the Protequ panel card (Live audit) and the element tiles. The same action from the keypad and from Protequ must produce the same two audit lines where they are two different things.
Inputs vs outputs¶
An output is a relay or sounder. The panel can switch it. Protequ has Activate and Deactivate on that node, plus bypass and test.
An input is a contact on the terminals. Protequ cannot close it. The native link has no such command. You see it close, and you can bypass it or put it in test. Closing it means the screw terminals, not a lightning icon in the tree.
If someone asks “can we activate inputs and outputs?”: outputs are the command; inputs are not — and that is correct. Activate on 556 and 558 was accepted on this firmware (17 September).
What was verified¶
Passed in the laboratory (through 17 September):
- TCP connect to DPE-2, firmware
20.2, hardware board 1. - Mid-session cable pull: receive-idle drop after 5 s, reconnect, FW 20.2, no Add device.
- Live Import configuration and XML Konfigurace: detectors 768 / 769, outputs 556 / 557 / 558, input 562, tablo master 2. Same SysAdr both ways.
- Tablo MHS 817 as a second master (address 2); its points use that master's address block.
- Spray on 769: fire on 769, output 558, confirmation window then general.
- Button 768: general immediately, no confirmation window.
- Reset and mute buzzer from Protequ (T1 and standing general). Mute buzzer is ACK-only; history 77 is keypad. Mute during T1 advances the panel to T2.
- Mute / reactivate external sounders (558).
- Bypass / cancel on modelled addresses (MHU may bounce 38 then 37).
- Test start / finish on 768 and Activate / Deactivate on 556 and 558 (the panel confirms the state; test history is on the panel card, not the detector).
- Input 562 closed on the terminals: tile active / idle, not a fire.
- Missing head on 769: trouble on the detector, then restore. An open head also raises Line 1 (SysAdr 304, event 18) until reseated.
- History after a ~4 min cable outage: 36 events in order, cursor kept.
--force-recreateloses the cursor; stop/start does not. - One event time vs the panel display: matches to the second.
Not on this project / not a pass:
- Group membership (no groups in the lab project).
- History wrap (M16) after a very long outage.
- Day / night as a Protequ command.
- Pre-alarm, conditional and technological panel-summary indications (not raised).
Do not treat these as failures¶
- A second client on DPE-2 — one session.
- Day / night from Protequ — no remote write.
- Enable / disable sirens as a separate verb (not mute acoustics).
- Line, OPPO, ZDP, slave and Ethernet as their own tiles.
- 534 / 368 / 4630 as unresolved devices — system addresses, not tree points.
- Event 30 on button 768 — MCP goes to general, not event 31.
- Cancel bypass 38 then 37 — MHU policy.
- Mute buzzer during T1 ending T1 and starting T2.
- Test history on the panel card (system address “Test”).
- Analog values from a detector.
- Per-element Audit tab on a detector.
Mute buzzer is kvitace of the panel piezo, not the external sirens. Mute sounder is the external acoustics. Falling general is not a reset.