Harlem: one address, four lobby doors, and a roof you can walk to

Four entry panels, yes. Four systems, no. That distinction is where the money is. Each lobby needs its own door station because there is no shared riser between the wings, but they can all report to one head-end or one cloud tenancy so the building has a single tenant directory, a single credential list and one place to revoke a fob.

The wiring topology is the design question. Depending on distances and what conduit exists between the wings, that is either a cable run through the cellar between the entrances, or independent network-connected stations per entrance with the coordination happening in software. We measure the cellar route before deciding, because a run that looks trivial on a plan often turns out to cross a boiler room, a sump and two firewalls.

Not with a chain, and not with anything that stops it opening from the inside. That door is part of the egress path, and the Fire Code requires egress hardware to be maintained in working order at all times. A padlock on it is an obvious and serious finding.

What works is knowing about it instead of preventing it. A door position switch plus a local sounder means the door still opens from the inside but announces itself and logs the event; add a camera on the bulkhead landing and you have both the alert and the record. Where a building genuinely needs to slow the door down, delayed egress hardware exists and the Building Code sets out exactly how it must behave. That is an engineered solution, not a hardware-store one.

The doors. Self-closing devices that actually close, strikes aligned so the latch throws fully, and the frame repaired where it has been kicked. That is the cheapest work in the building and it is the work that makes everything else function: a video intercom on a door that does not latch buys you a camera, not security.

After that, one entry panel at the busiest entrance, and contacts on the roof and rear doors. We write the scope in stages so the board can approve and fund them one at a time, and each stage is complete in itself rather than dependent on the next vote passing.

Only if all three conditions are met: a Class A multiple dwelling, erected or converted to residential use after 1 January 1968, and eight or more apartments. Most Harlem pre-war stock fails the date test, so §50-a does not compel an intercom there. Post-1968 buildings and post-1968 residential conversions are inside it, and there are more of those in Harlem than people assume.

Where a building is covered, the statute wants all three pieces together: a self-locking main entrance door, two-way voice with the visitor, and the ability for the occupant to release that door from the apartment. A working intercom on a door that does not self-lock does not satisfy what the law is trying to achieve. Note also that where a building is covered by the statute but the system predates it, retrofit obligations turn on written tenant consent from a majority of tenants. That is a process question worth getting right before spending money.

Yes, across Upper Manhattan. East Harlem carries a heavier concentration of large institutional developments and mid-century towers, which shifts the work toward portfolio-scale credential management and vandal-resistant hardware. Hamilton Heights and Sugar Hill run closer to the courtyard apartment house and row-house pattern described above.

Washington Heights and Inwood add another layer: long six-story pre-war buildings with multiple entrances and heavily used commercial corridors along Broadway, St. Nicholas Avenue and West 181st Street. Same trade, different geometry each time, which is why we survey before quoting anywhere above 96th Street.