The Hobbs is an electric clock. They will run at a constant speed when fed with anywhere between 10 and 80 volts dc, so unless your electrical system has real problems (and dropping below 10 volts), the Hobbs will continue to count fairly accurately. The Hobbs knows nothing other than someone applied power to it. I've seen Hobbs connected to the master, an oil pressure switch, the gear squat switch, the power to the gas-fired heater, etc...
Mechanical recording tachs are just revolution counters. They assume some nominal cruise setting and divide that to get "hours."
Finally. Give that man a cookie.
The tach hours are proportional to the engine RPM. There's no magic "start" RPM, it's a straight mechanical ratio.
If the tach is calibrated to show 1.0 tach hour for 1.0 clock hour when the engine is operating at 2500 rpm (a typical 65% power value for a fixed pitch O-320), then what it's really doing is moving the drums 1 "hour" for every 150,000 engine revolutions (2500x60=150,000). At a constant 2500 rpm it'll show 1.0 tach hour for every 1.0 Hobbs hour.
If you are instead cruising at a constant 2100 rpm, then your engine is only turning over 126,000 revolutions every Hobbs hour and the tach will only move the drums .84 tach hours (which will show as .8 hours on the tach). 126,000/150,000 = .84.
If your tach is calibrated for the above mentioned 2500 rpm/65% cruise rpm, but you instead cruise at 2600 rpm (74% power in my aircraft), then the engine will turn 156,000 revolutions in 1 Hobbs hour and the tach time will be 1.04 tach hours, which still shows as "1.0" on the tach. You won't see a difference until you are 3 hours into the flight when the tach will reflect the 468,000 engine revolutions and 3.12 tach hours (3.1 hours on the tach) compared to 3.0 hours on the Hobbs.
If you cruise at 2700 rpm and 81% power, the tach time will be 1.08 hours, so you'll see 1.1 on the tach shortly after the Hobbs shows 1.0.
----
In my Citabria, doing a mix of stop and goes, full power climbs to 1000 ft-3500 ft, and airwork at 2100-2300 rpm, my tach time averages about .73 hours for every hour on the Hobbs.
On a cross country with a full power climb to 3500-4500 ft and a 2200 rpm cruise, it averages .92 tach hours for every hobbs hour on a 2.5 hour flight. (As opposed to a straight .88 tach hours for every Hobbs hour at a constant 2200 rpm).
That tendency for the tach time to be less than the actual clock time, nearly (but not always) all the time is why Hobbs meters became a thing. Renting aircraft where a pilot could fly for 1 hour and only show 45 minutes on the tach cost FBOs revenue. To be fair, the engine sees less wear at low rpm, engine TBOs are based on tach hours, and fuel burn is lower at the lower rpm settings, so the actual costs are at the the 45 minute figure (not counting instructor time), but it's still lost revenue.
Once Hobbs meters became a thing, it didn't take renters long to figure out they could turn off the master switch for awhile and reduce the bill. Some FBOs started tracking both Hobbs and Tach and could quickly figure out who was cheating. But it also became common to tie the Hobbs to an oil pressure sensor, so that the Hobbs ran anytime the oil pressure was over a certain minimum value. I suppose if you were renting something like a DA 20 with an 11-1 glide ratio, you could climb to 10,000 ft, shut down the engine, slow enough to stop the prop, glide 19-20 miles or so, restart the engine, and then repeat the process.